\_sh v3.0 560 Linguistic Notes \_DateStampHasFourDigitYear \key AAAA -- Introduction to Manual \shd A Field Manual for Descriptive Linguistics \ftx by Thomas E. Payne \txt This volume is intended to serve as a Field Manual for Descriptive Linguistics for use by field linguists. It attempts to step people through the process of analyzing the morphosyntax of any language by the process of asking pertinent questions about the structure of the language at various levels. A series of helps are provided to assist people in dealing with the concepts and theory involved. I have broken the original manuscript down into chunks of a size more convenient for Shoebox and for viewing on a computer screen. This has meant adding a great many more section headings and question fields than were in the original document. Any shortcomings arising out of these additions and changes are entirely our responsibility, and should not be blamed on Tom Payne. \txt There are a number of places where more examples are needed to illustrate particular points, and in some instances the examples need to be checked for accuracy. There are also some bibliographic references which are not complete. It is hoped that these discrepancies will be corrected in later revisions. A few of the Help cards are empty at this stage. Eventually these should also have some appropriate material added to them. This introductory section is essentially background material, and does not have a separate help file as the rest of the chapters do, but is a help file in itself. \txt Some of the special characters occurring in the examples will not be accurate, since we are limited to ANSI's upper 128 character set. So don't worry if you sometimes see something strange, or not all the contrasts are preserved. As far as we have been able to tell this does not at any time take away from the grammtical significance of the example. \txt It is hoped that these files will prove useful to field linguists analysing specific languages. It is intended that these questions and helps, or something very like them, will eventually appear as part of the Linguistic Field Manual in the LinguaLinks Library. Any feedback relating to this manual, or suggestions for improvements, will be welcome, and should be sent to us in Dallas. We need to know if this material is helpful, and what needs to be done to it to make it more helpful, in order, eventually, to make the LinguaLinks implementation itself more valuable. \ftx Linguistic Field Manual Development Team Intl. \ftx Linguistics Department \ftx Dallas \fln __________________________________________________ \txt To continue this introduction, use Database Search (ALT-D,S) or highlight and jump to: \cf Hintro_dling \txt To see an overview of the contents of the Field Manual use Database Search (ALT-D,S) or highlight and jump to: \cf Htopics_outline \fln __________________________________________________ \dt 27/Jul/1998 \key Hcplx \shd Complex constructions \txt This topic contains all of the following subsections. Highlight and jump to the topic of interest: \cf Hcplx_intro Complex constructions--Introduction \shd2 Serial verbs \cf Hcplx_sv Serial Verbs \cf Hcplx_sv_tam Serial verbs--Tense/aspect/mode \cf Hcplx_sv_extr Serial verbs--Extraction \cf Hcplx_sv_subj Serial verbs--Subjects \cf Hcplx_sv_vschn Serial verbs vs Clause chaining \cf Hcplx_sv_vsaux Serial verbs vs Auxillaries \cf Hcplx_sv_ex Examples \shd2 Complement clauses \cf Hcplx_cplcl Complement Clauses \cf Hcplx_cplcl_cplx The Complexity Continuum \cf Hcplx_cplcl_prop Propositional complements \cf Hcplx_cplcl_mrgcpl Merged complements \cf Hcplx_cplcl_indq Indirect questions \shd2 Adverbial clauses \cf Hcplx_advcl Adverbial Clauses \cf Hcplx_advcl_time Time \cf Hcplx_advcl_loc Location \cf Hcplx_advcl_man Manner \cf Hcplx_advcl_purp Purpose \cf Hcplx_advcl_rsn Reason \cf Hcplx_advcl_circ Circumstantial \cf Hcplx_advcl_sim Simultaneous \cf Hcplx_advcl_cnd Conditional \cf Hcplx_advcl_cncs Concessive \cf Hcplx_advcl_sbst Substitutive \cf Hcplx_advcl_add Additive \cf Hcplx_advcl_abs Absolutive \shd2 Medial clauses \cf Hcplx_med Medial Clauses \cf Hcplx_med_chn Clause Chaining \cf Hcplx_med_medcl Medial Clauses \cf Hcplx_med_sw Switch Reference Systems \cf Hcplx_med_sw2 Switch reference systems II \cf Hcplx_med_sw3 Switch reference systems III \shd2 Relative clauses \cf Hcplx_relcl Relative Clauses \cf Hcplx_relcl_typ Typological parameters \cf Hcplx_relcl_pre-n Pre-nominal examples \cf Hcplx_relcl_post-n Post-nominal examples [INCOMPLETE] \cf Hcplx_relcl_inthd Internally headed examples \cf Hcplx_relcl_hdless Headless Examples \cf Hcplx_relcl_vscplcl Headless Relative clauses vs complement clauses \cf Hcplx_relcl_np NP Encoding--verb medial languages \cf Hcplx_relcl_np_2 NP Encoding--verb initial languages II \cf Hcplx_relcl_np_3 NP Encoding--Pronoun retention \cf Hcplx_relcl_np_4 NP Encoding--Relative pronouns \cf Hcplx_relcl_elem Clause Elements Capable of being Relativized \shd2 Coordination \cf Hcplx_coord Coordination \cf Hcplx_coord_np Parallel to Noun Phrases \cf Hcplx_coord_zero Zero \cf Hcplx_coord_cnj_1 Coordinating Conjunction I \cf Hcplx_coord_cnj_2 Coordinating Conjunction II \cf Hcplx_coord_other Other Coordination Strategies \cf Hcplx_coord_log Logical Relations \cf Hcplx_other **Other Complex Constructions \ftx \dt 29/May/1998 \key Hcplx_advcl \shd Adverbial clauses \txt ADVERBIAL CLAUSES are clauses that serve an 'adverbial' function. They modify a verb phrase or a whole clause. They are not an argument of the clause. Sometimes adverbial clauses are termed 'adjuncts' (as opposed to complements). This is a good term since the term 'complement' implies completion, and a predicate is not a proposition until all its argument positions are filled, i.e. completed. On other hand, adverbials attach to constructions that are already complete propositions. The adverbial simply adds some information to the proposition. \ftx \fln Sometimes adverbial clauses look like complements: \ftx \fln (47) a. He ran to get help. (purpose) \ftx b. We're sorry that you feel that way. (reason) \ftx c. She went out, locking the door behind her. (sequence) \ftx \txt The adverbial clauses in these examples all have the same morphosyntax as certain complement types of English. Nevertheless, they are not complements because they do no constitute logical arguments of the main verb. Rather they simply add 'adverbial' information. \ftx \txt The kinds of information embodied in adverbial clauses are the same kinds of information encoded by adverbs, e.g. time, place, manner, purpose, reason, condition, etc. Examples of each of these will be provided in the following paragraphs. \ftx \dt 27/Apr/1998 \key Hcplx_advcl_abs \shd Adverbial clauses--Absolutive \fln 'Absolutive': \ftx a. Having told a few bad jokes, Harvey proceeded to introduce \ftx the speaker. \ftx b. Seeing me, Harvey hid behind his mother's skirt. \ftx \txt The distinguishing characteristic of 'absolutive' adverbial clauses is that there is no explicit marking of the semantic relationship between the two clauses. The adverbial clause simply presents the general background for the situation expressed in the main clause. \ftx \txt If the language employs this sort of 'absolutive' clause extensively, you should consider the possibility of calling them medial clauses (see section Hcplx_med_medcl). In English these gerundive clauses really fall in between adverbial and medial clauses. However, because there are no canonical medial clauses in English, and because these gerundive clauses are not extremely common, it makes more sense for the purposes of this typology to think of them as a type of adverbial clause. \ftx \dt 07/Jul/1999 \key Hcplx_advcl_add \shd Adverbial clauses--Additive \txt Additive: In addition to having your hand stamped, you also have to have your ticket stub. \ftx \dt 27/Apr/1998 \key Hcplx_advcl_circ \shd Adverbial clauses--Circumstantial \fln Circumstantial: He got into the army by lying about his age. \ftx (Typologically rare.) \ftx \key Hcplx_advcl_cncs \shd Adverbial clauses--Concessive \fln Concessive clause: \ftx a. Although she hates Bartok, she agreed to go to the \ftx concert. \ftx b. Even though it's still early, we'd better find our \ftx seats. \ftx \key Hcplx_advcl_cnd \shd Adverbial clauses--Conditional \fln Conditional: \ftx \fln Simple: a. If it's raining outside, then my car is getting wet. \ftx b. If you step on the brake, the car slows down. \ftx c. If you were at the party, then you know about Sue and \ftx Fred. \ftx \fln Hypothetical: If I (were to see/saw) David, I would speak Quechua \ftx with him. \ftx \fln Counterfactual: If you had been at the concert, you would have \ftx seen Ravi Shankar. \ftx \fln Concessive conditional: Even if it rains, we'll have our picnic. \ftx \txt Most languages use a subordinating morpheme like if in concessive conditionals, but some languages use a different morpheme than that which occurs in other types of conditionals: \ftx \fln (55) Mandarin: Jiushi ta song gei wo, wo dou bu yao. \ftx Even:if he give to I I still NEG want \ftx 'Even if he gave it to me I wouldn't take it.' \ftx \fln 'Speech act': If you're thirsty there's coke in the refrigerator. \ftx \dt 27/Apr/1998 \key Hcplx_advcl_loc \shd Adverbial clauses--Location \fln Location: I'll meet you where the statue used to be. \ftx \fln (49) Turkish: Sen Erol-un otur-dug-u yer-e otur. \ftx you Erol-GEN sit-OBJ-POSS place-DAT sit \ftx 'You sit where Erol was sitting.' \ftx (requires the word for 'place') \ftx \key Hcplx_advcl_man \shd Adverbial clauses--Manner \fln Manner: a. She talks like she has a cold. \ftx b. Carry this as I told you. \ftx \fln (50) Quechua: Alista-pan kuura ni-shan-naw-qa. \ftx prepare-BEN3 priest say-REL-MAN \ftx 'They prepared it for him like the priest said.' \ftx \txt Manner clauses in Quechua take the marker for relative clauses in addition to the suffix that indicates manner. A literal translation of this sentence might be '... the way that the priest said.' \ftx \dt 27/Apr/1998 \key Hcplx_advcl_purp \shd Adverbial clauses--Purpose \fln Purpose: He stood on his tiptoes in order to see better. \ftx \fln (51) Panare: Tyenche' e'ñapa tu'ñen i'yatatópe \ftx t-yen-se' e'ñapa tu'ñen i'ya-ta-tópe \ftx IRR-take-GNO people medicine chief-INCHO-PURP \ftx 'People take medicine in order to become a chief.' (MCW.78.103a) \ftx \txt In Panare the inflectional suffix '-tópe' marks a clause as being a purpose adverbial. \ftx \dt 27/Apr/1998 \key Hcplx_advcl_rsn \shd Adverbial clauses--Reason \fln Reason: He got here early because he wanted to get a good seat. \ftx \fln Most languages treat purpose and reason alike, e.g. Yoruba: \ftx \fln (52) Vêru gàadà dà shi sêma \ftx go:out:PERF PURP IRR drink beer \ftx 'He went out to drink beer.' (purpose) \ftx \fln (53) A-ta abên gàadà aci ngaa \ftx eat-PERF food REASON he well \ftx 'He ate because he was well.' (reason) \ftx \txt The only formal difference between purpose and reason clauses in Yoruba is that the purpose clause contains the irrealis marker 'dà'. \ftx \dt 27/Apr/1998 \key Hcplx_advcl_sbst \shd Adverbial clauses--Substitutive \fln Substitutive: We barbecued chicken instead of going out to \ftx eat. \ftx \key Hcplx_advcl_sim \shd Adverbial clauses--Simultaneous \fln Simultaneous: \ftx a. While (we were) eating, we heard a noise outside the window. \ftx b. He woke up crying. \ftx \fln (54) Yavapai: \ftx Kwawa '-chkyat-a-k vak '-unuu-t-m swach'skyap-ch vqaov-k yuny \ftx hair 1-CUT-IRR-SS here 1-INCOMPL-SIM-DS scissors-SUBJ break-SS TNS \ftx 'While I was cutting my hair the scissors broke.' \ftx \key Hcplx_advcl_time \shd Adverbial clauses--Time \fln Time: We'll go when Sandy gets here. (also 'before', 'after' etc.) \ftx \fln (48) Barai (PNG): Bae-mo-gana e ije bu-ne ke. \ftx ripe-PAST:SEQ-DS people these 3PL-FOC take \ftx 'When it was ripe, these people took it.' \ftx (takes one of several 'sequence' markers) \ftx \key Hcplx_coord \shd Coordination \txt Languages often have morphosyntactic means of linking two clauses of equal grammatical status. Such linkage is termed COORDINATION. It is distinct from SUBORDINATION in that in subordination, one clause is grammatically dependent on the other. Coordination is sometimes difficult to distinguish from mere juxtaposition of clauses in discourse. In fact, in spoken discourse some kind of morphosyntactic clause linkage, either coordination or subordination, may be evident at nearly all clause junctures. We are all familiar with the English colloquial narrative style that inserts 'and ...' or 'and then ...' after each clause. \ftx \txt In general the fact that two clauses are grammatically coordinated simply conveys that 1) the two clauses have more or less the same function in terms of the event structure of the text (e.g. they both code events, they both code non-events, they both code foregrounded information or they both code background information, etc.) and 2) they are presented as being conceptually linked in some way. \ftx \txt Interpropositional (logical) relations that often obtain between coordinate clauses include conjunction, disjunction and exclusion. Coordinating devices used to distinguish these relations will be discussed in this section. It should be kept in mind, however, that just about any semantic relation between clauses in discourse can obtain in a coordinate structure. Section #12.1.2 describes one framework for analyzing the interpropositional relations in a text. \ftx \txt To continue reading on this topic, highlight the key below: \cf Hcplx_coord_np \ftx \dt 27/Apr/1998 \key Hcplx_coord_cnj_1 \shd Coordination strategies--Coordinating conjunction I \txt The most common means of indicating conjunction is by the use of a COORDINATING CONJUNCTION such as 'and' in English. For VP languages this conjunction normally occurs in between the two conjoined clauses: \ftx \ftx English: \fln (88) Robespierre fell out of favor and the revolutionaries killed \ftx him. \ftx \fln (88A) Central Buang, (PNG): \ftx Avëh sis anöö lob maluhnö lev lok kele degwa \ftx woman killed dog and brother buried in tree base \ftx `The woman killed the dog and her brother buried it at the foot \ftx of the tree.' \ftx \txt However, sometimes in VP languages the coordinating conjunction follows the first element of the second clause: Yoruba: \ftx \fln (89) mo mú ìwé; mo sì wá ilé 'I took a book and I came home.' \ftx I take book I and come house \ftx \txt For PV languages, the coordinating conjunction comes either between the two conjoined elements (90a, b, c), or after the last element (91): \ftx \fln (90) Farsi (From J. Payne 1985:28): \ftx a. Jân [ xandid va dast tekân dâd ]vp \ftx John smiled and hand sign gave \ftx 'John smiled and waved.' (V-Phrase coordination) \ftx \ftx b. Jân [ puldar va mashur ]ap bud \ftx John rich and famous was \ftx 'John was rich and famous.' (Adj. coordination) \ftx \ftx c. [ Jân raft va meri dast tekân dâd ]Cl \ftx John left and Mary hand sign gave \ftx 'John left and Mary waved.' (Clausal coordination) \ftx \txt To continue reading on this topic, highlight the key below: \cf Hcplx_coord_cnj_2 \ftx \dt 27/Apr/1998 \key Hcplx_coord_cnj_2 \shd Coordination strategies--Coordinating conjunction II \txt The form that conjoins two elements is often the same as the operator that encodes the comitative sense of with. In Walapai (Yuman), both instrumental and comitative elements are signalled with the enclitic '-m'. Example 92 illustrates this '-m' in its common role as an instrumental case marker (Redden 1966:160-61 as cited in J. Payne 1985:30): \ftx \fln (92) na-c nikwáì-c-a avon-a-m taT-k-wíl \ftx 1SG-NOM clothes-PL-DEF soap-DEF-with wash-1SG-CONT \ftx 'I washed the clothes with soap'. \ftx \txt This '-m' operator also functions as a phrasal and clausal coordinator: \ftx \fln (93) Wàlpáìkwáùk háìkùkwáùk-m íce \ftx Walapai:speech white:man:speech-with we:speak \ftx 'We speak Walapai and English.' \ftx \txt Such isomorphism among the instrumental, comitative and coordinating operators is extremely common in the world's languages. \ftx \txt To continue reading on this topic, highlight the key below: \cf Hcplx_coord_log \ftx \dt 27/Apr/1998 \key Hcplx_coord_log \shd Coordination--Logical relations \txt CONJUNCTION is primarily a logical relationship between propositions. If the conjunction of two propositions is true then each of the component propositions must be true. According to this definition, nearly any two propositions in discourse could be conjoined. \ftx \txt DISJUNCTION, like conjunction, is a logical relationship between propositions. If the logical disjunction of two propositions is true, then one and only one of the component propositions can be true. There is no particular reason why a language would gramaticalize exactly this notion of disjunction, and in fact it seems to be quite rare for languages to possess a device specifically for this purpose. English 'or' can signal a logical disjunctive relationship if both component clauses are affirmative: \ftx \fln (95) He came in through the window or the door. \ftx \txt This sentence can be thought of as the disjunction of the two propositions 'he came in through the window' and 'he came in through the door'. That is, 95 would normally assert that one of these propositions holds true but the other one does not. The word 'either' reinforces this interpretation. In fact, in discourse this use of 'or' without 'either' is quite rare. Instead 'or' without 'either' is used almost exclusively when one or both of the component propositions is negated. In this case 'or' ceases to convey logical disjunction: \ftx \fln (96) I didn't break the window or the door. \ftx \txt Under any natural circumstances this sentence expresses the conjunction of 'I didn't break the window' and 'I didn't break the door'. Hence it is simply inaccurate to characterize 'or' as a disjunctive particle in English. Logical disjunction is but one, relatively rare, function of this particle. \ftx \txt In other languages it is similarly rare for logical disjunction to have its own unique morphosyntax. If disjunction is conveyed at all it will usually be via some periphrastic device such as 'I might have broken the window and I might have broken the door.' In Yagua one interesting case of disjunction is conveyed via the use of contrastive pronouns: \ftx \fln (97) a. Ra-dyéétya-rúú-kyey, \ftx 1SG-know-POT-EVID \ftx 'I want to know \ftx \ftx b. níí-numáá-tiy vátan-tán-dyé-ryéy, munuñú-niy, \ftx 3SG:PRO-now-COND curse-cause-DAY-1SG savage-NIY \ftx if HE cursed me, the savage, \ftx \ftx c. ráñiy vátan-tán-dye-ryéy. \ftx 1SG:PRO curse-cause-DAY-1SG \ftx or I cursed myself.' \ftx \txt In this case the speaker is not claiming that either the savages or the speaker himself did the cursing, but only that one did and the other didn't. There is no explicit coding of the disjunctive relation between 97b and 97c other than the use of contrastive pronouns. \ftx \dt 27/Apr/1998 \key Hcplx_coord_np \shd Coordination strategies--Parallel to noun phrases \txt Often some strategies for conjoining clauses are identical to strategies for conjoining noun phrases. For example, English uses the conjunction 'and' for both phrasal and clausal conjunction: \ftx \fln (85) John and Mary NP + NP \ftx John cried and Mary laughed Cl + Cl \ftx \txt However, it is also common for there to be special strategies for conjoining clauses that are not used for conjoining phrases. For example, the English 'but' does not easily function as a noun phrase conjunction: \ftx \fln (86) *John but Mary NP + NP \ftx John cried but Mary laughed Cl + Cl \ftx \txt To continue reading on this topic, highlight the key below: \cf Hcplx_coord_zero \ftx \dt 27/Apr/1998 \key Hcplx_coord_other \shd Other Coordination strategies \txt Latin possesses a 'negative conjunctive' particle 'nec', in addition to the affirmative conjunction 'et'. The meaning of the negative conjunctive particle can be characterized as `and not' in English (Kühner and Stegmann 1955:48 as cited in J. Payne 1985:37): \ftx \fln (94) eques Romanus [ nec infacetus et satis litteratus ]ap \ftx knight Roman and:not dull and moderately literate \ftx 'a not dull and moderately literate Roman knight.' \ftx \txt Unlike the English translation 'not . . . and', the negative conjunctive particle in Latin does not have scope over the entire conjoined phrase. In other words, only dullness is negated in the Latin example, whereas the English translation could be taken as ambiguous as to whether 'moderately literate' should be taken as being negated as well. \ftx \dt 27/Apr/1998 \key Hcplx_coord_zero \shd Coordination strategies--Zero \txt The simplest means of conjoining two clauses is what J. Payne (1985) describes as the ZERO STRATEGY. This is where two phrases or clauses are simply juxtaposed. According to J. Payne, most languages probably allow the zero strategy at least as a stylistic variation. Some languages, however, use it more extensively than do others. Vietnamese is a language that uses the zero strategy extensively in both phrasal and clausal coordination (examples from Watson 1966:170 as quoted in J. Payne, J. 1985:26): \ftx \fln (87) a. Nháng tiráp [ tilêt, callóh, acôq ]np \ftx we prepare basket spear knife \ftx 'We prepare baskets, spears and knives.' (NP coordination) \ftx \ftx b. Do chô [ tôq cayâq, tôq apây ]pp \ftx she return to husband to grandmother \ftx 'She returns to (her) husband and to (her) grandmother.' \ftx (PP coordination) \ftx \ftx c. Do chô [ tôq cayâq, chô tôq apây ]pp \ftx she return to husband return to grandmother \ftx 'She returns to (her) husband and returns to (her) grandmother.' \ftx (Clausal coordination) \ftx \txt To continue reading on this topic, highlight the key below: \cf Hcplx_coord_cnj_1 \ftx \dt 27/Apr/1998 \key Hcplx_cplcl \shd Complement clauses \txt A prototypical COMPLEMENT CLAUSE is a clause which functions as an argument (subject or object) of some other clause. A MAIN (or MATRIX) clause is a clause that has another clause as one of its core arguments. However, a much wider range of clauses have been called 'complements'. For our purposes we can consider a complement clause to be any clause that is EMBEDDED within another clause (Foley and Van Valin 1984:##). \ftx \txt To continue this discussion, highlight and jump to the topic below: \cf Hcplx_cplcl_cplx \ftx \dt 27/Apr/1998 \key Hcplx_cplcl_cplx \shd Complement clauses--the Complexity Continuum \txt Complements can be described as falling somewhere in a continuum defined in terms of its extremes as follows: \ftx \ftx 'Merged 'Propositional \ftx complements' complements' \ftx |---------------------------------------------------------| \ftx \txt We will refer to this continuum as the 'complexity continuum'. It will guide our presentation of various kinds of complements in the following paragraphs. As we saw with causatives (Hval_incr_caus), the closer the structural integration between complement and main verb, the closer the conceptual integration is likely to be. \ftx \txt NOTE: The terminology and the concept of a 'complexity continuum' stems from class lectures presented by Sandra Thompson in 1979. The notion of a typology of complement types is represented in many works, including notably Givón 1984 and Haiman 19##). \ftx \txt To continue this discussion, highlight and jump to the topic below: \cf Hcplx_cplcl_prop \ftx \dt 27/Apr/1998 \key Hcplx_cplcl_indq \shd Complement clauses--Indirect questions \txt INDIRECT QUESTIONS are a sub-type of complement clauses. English uses wh- type complementizers in indirect questions: \ftx \fln English indirect questions -- subject complements: \ftx \fln (43) a. Whether they're here is not known. \ftx b. It is a mystery to me who saw you. \ftx \txt Whether corresponds to yes/no type questions, while what, when, where, who/whom, how, why, which, etc. correspond to question-word questions. \ftx \fln English indirect questions -- object complements: \ftx \fln (44) a. I wonder whether they're here. \ftx b. I wonder who saw you. \ftx \txt Indirect questions may share formal properties with interrogative clauses or relative clauses. For example, in Yoruba, 'ti' is the complementizer used in relative clauses, and 'woni' is a question word. In indirect questions, however, 'ti' rather than 'woni' is the complementizer: \ftx \fln (45) Tale mo okunrin ti obinrin na lu \ftx Tale know man that woman the hit \ftx 'Tale knows which man the woman hit.' \ftx \fln English can go either way: \ftx \fln (46) a. I know the year that Mary was born. \ftx b. I know which year Mary was born. \ftx \txt In 46a the emphasized portion resembles a relative clause. It means the same thing as 46b in which the complement resembles a question word question. \ftx \txt Malagasy is the same as Yoruba. 'izay' is the relative clause complementizer, and ''inona' is the question word. However, indirect questions use 'izay'. \ftx \dt 27/Apr/1998 \key Hcplx_cplcl_mrgcpl \shd Complement clauses--Merged complements \txt MERGED COMPLEMENTS are more tightly knit, less independent, less like a separate sentence from the matrix clause than are propositional complements. Merged complements tend to have the following properties: \ftx \ftx 1. The reference of the subject is highly constrained. It \ftx often must be identical to the subject of the matrix verb. \ftx \ftx 2. Tense, aspect and mode is highly constrained or not \ftx specified at all. The complement verb is usually non-finite. \ftx \fln Some examples of merged complements follow: \ftx \fln English merged subject complements: \ftx \fln (38) a. To cook a meal like that requires a lot of patience. \ftx b. It isn't so easy to do linguistics. \ftx \fln English merged object complements: \ftx \fln (39) a. I enjoy washing my car. \ftx b. She likes to do linguistics. \ftx \ftx Mandarin: \fln (40) wo yao nian (*-le) shu 'I want to read a book.' \ftx 1SG want read (-PERF) book \ftx \txt In Mandarin it is ungrammatical to attach the perfective aspect marker to verbs that are complements of certain matrix verbs. \ftx \ftx Wappo: \fln (41) ¿ah ce k'ew ew tum-uhk hak'se¿ 'I want that man to buy fish.' \ftx 1SG that man fish buy-INF want \ftx \txt In this Wappo sentence, the complement verb does not take an independent tense/aspect marker. Instead the infinitive suffix marks it as a non-finite verb. \ftx \txt A high degree of merger often results in the matrix and complement verbs becoming a single unit. This is sometimes referred to as CLAUSE UNION. For example, DESIDERATIVE affixes typically derive from older matrix verbs: \ftx \ftx Yagua: \fln (42) a. Ra-yá. 'I go.' \ftx 1SG-go \ftx \ftx b. Ra-ya-rúúy. 'I want to go.' \ftx 1SG-go-DESID \ftx \txt Diachronically the desiderative suffix '-rúúy' arose from the matrix verb 'nirúúy'-- 'want' in the following manner: \ftx \fln Stage I: A COMP V \ftx I [I go] want \ftx \fln Stage II: I [0 go] want Merged complement \ftx \fln Stage III: I-go-want Clause union \ftx \txt To continue this discussion, highlight and jump to the topic below: \cf Hcplx_cplcl_indq \ftx \dt 27/Apr/1998 \key Hcplx_cplcl_prop \shd Complement clauses--Propositional complements \txt Prototypical PROPOSITIONAL COMPLEMENTS are like independent clauses as evidenced by the following characteristics: \ftx \ftx 1. They carry their own tense and aspect -- the complement \ftx verb is a finite verb form. \ftx \ftx 2. They code their subjects directly; subject reference is \ftx not restricted to that of the matrix clause. \ftx \txt Typical matrix verbs for propositional complements are verbs of utterance and cognition. For example: \ftx \fln English propositional object complements: \ftx \fln (34) a. I know that it's raining. \ftx b. I emphasized that she knows Swahili. \ftx \fln English propositional subject complements: \ftx \fln (35) a. That it had rained surprised me. \ftx b. It is well known that she is terribly rude. \ftx \txt Note that, except for the complementizer, each of the emphasized complement clauses in 34 and 35 could stand alone as a complete and understandable utterance in English. Each one is independently marked for tense and subject reference. As might be expected, in VO languages object complements tend to follow the matrix verb: \ftx \ftx Mandarin: \fln (36) wo zhidao neige ren chi-le san wan fan \ftx I know that person eat-PERF three bowl rice \ftx 'I know that that person ate three bowls of rice.' \ftx \txt The perfective marker in the complement verb shows that this is a propositional complement. \ftx \txt In OV languages, object complements tend to precede the matrix verb: \ftx \ftx Wappo (Charles Li and Sandra Thompson, p.c.): \fln (37) ¿ah ce k'ew ew tum-tah hatiskhi¿ \ftx 1SG that man fish buy-PAST know \ftx 'I know that man bought fish.' \ftx \txt Again, the complement clause in 37 is a propositional complement because it contains all of the inflectional information necessary to an independent clause of the language. \ftx \txt To continue this discussion, highlight and jump to the topic below: \cf Hcplx_cplcl_mrgcpl \ftx \dt 27/Apr/1998 \key Hcplx_cplcl_s/o \shd Subject & Object Complement clauses \txt The kinds of complement clauses that we will discuss and illustrate in this section can be either subjects or objects of the matrix clause. For example: \ftx \fln (30) SUBJECT COMPLEMENT: \ftx A V P \ftx [[That Lady Lucretia trod on his toe] stunned the Duke of Wimple]. \ftx <---------- Complement -------------> \ftx <------------------------ Main (matrix clause -------------------> \ftx \fln (31) OBJECT COMPLEMENT: \ftx A V P \ftx [Lady Lucretia wants [to tread on the Duke of Wimple's toe]]. \ftx <------------ Complement ------------> \ftx <---------------- Main (matrix) clause --------------------> \ftx \txt In English we usually place subject complements after the verb and replace the subject with the neuter pronoun it. This is called post-posing of subject complements: \ftx \fln (32) It stunned the Duke of Wimple that Lady Lucretia trod on his toe. \ftx \txt A clause can be both a complement and a matrix clause, i.e. it can be an argument of one clause and at the same time have a third clause as one of its own core arguments. For example: \ftx \fln (33) [Lucretia wants [to believe [that that oaf is the Duke of Wimple]]]. \ftx <------ Complement of believe ------> \ftx <--------------- Complement of want -------------> \ftx Matrix clause for that oaf is ... \ftx <----------------------- Main clause -----------------------------> \ftx \dt 27/Apr/1998 \key Hcplx_intro \shd Complex constructions--Introduction \txt In previous chapters we have discussed several means of altering the form of basic level lexical verbs to shape the semantic force of the concepts they embody. In every language there exist as well different ways of combining basic lexical items, such as verbs, to form more complex expressions. In this chapter we will discuss several constructions that involve combinations of verbs. \ftx \txt Most of the multi-verb constructions described in this chapter involve one INDEPENDENT clause and one or more DEPENDENT clauses. An independent clause is one that is fully inflected and capable of being integrated into discourse on its own (see Hmor_infl_gen -- inflectional morphology). A dependent clause is one that depends on some other clause for at least part of its inflectional information. For example, in the following sentence, clause b is dependent on clause a because the subject and tense of clause b are only understood via the subject and tense of clause a: \ftx \fln (1) a. He came in, b. locking the door behind him. \ftx \txt Clause b. by itself does not qualify as a fully inflected clause, able to be integrated into discourse on its own. Sometimes fully inflected verbs are called finite verbs, whereas dependent verbs are termed NON-FINITE. However, this distinction must be understood as a continuum, as some verbs are dependent in one respect, but independent in another. Thus we may talk about one verb being more finite or less finite than another. \ftx \txt The present chapter will be organized according to six general types of multi-clause construction: 1) coordination, 2) serial verbs, 3) complement clauses, 4) medial clauses, 5) adverbial clauses and 6) relative clauses. \ftx \dt 27/Apr/1998 \key Hcplx_med \shd Medial clauses, clause chaining and switch reference \fln For help on these questions, see the following sections. \ftx \key Hcplx_med_chn \shd Clause chaining \txt Since the mid-sixties there have been many studies dealing with CLAUSE-CHAINING languages (McCarthy 1965, Healy 1966, Hetzron 1969, 1977, Longacre 1972, Olson 1973, Thurman 1975, Gerdel and Slocum 1976 inter alia). The paradigm examples of clause chaining languages occur in the highlands of New Guinea, both Irian Jaya and Papua New Guinea (Elson 1964), though clause chaining is a well-recognized phenomenon in Australia (Austin 1980) and the Americas (Longacre 1985). In descriptions of such languages there is normally a distinction drawn between 'final' and 'non-final' clauses. These terms are based on the fact that in clause chaining languages as identified in these previous studies, there are significant morphosyntactic differences between the sequentially final clause in a clause chain and those that precede it. Longacre (1985:264, and footnote 6) hypothesizes that, though it is a logical possibility for languages to have clause chains in which the functionally 'final' clause actually occurs initially, no clear examples of such languages have been documented. \ftx \dt 27/Apr/1998 \key Hcplx_med_medcl \shd Medial clauses \txt More recently the term MEDIAL clause has begun to replace the term non-final clause in descriptions of clause chaining structures (Haiman 198 7). This term reflects the fact that this clause type occurs sentence-internally, i.e. in the 'middle' of a clause chain. As defined by Longacre 1985 :263, a medial clause is one which a) has a reduced range of tense-aspect possibilities in comparison to final clauses, b) usually specifies 'subject' reference in terms of (i.e. as same as or different from) the subject of the final clause, and c) usually directly expresses temporal relations such as 'overlap' and 'succession' with respect to other clauses in the sequence. A clause chaining language, then, is a language that employs sequences of medial clauses completed by a final clause as a major discourse structuring device. Longacre (op cit) describes such languages as 'radically different' in surface structure from more familiar 'co-ranking' languages. \ftx \txt NOTE: Haiman (1987) uses the term 'medial verb' to describe the morphologically distinct verb type that heads non-final clauses. Longacre (1985), however, seems to use the terms medial clause and non-final clause interchangeably. \ftx \dt 27/Apr/1998 \key Hcplx_med_sw \shd Switch reference systems \txt A prototypical SWITCH-REFERENCE system is verbal inflection that indicates whether the subject of the verb is coreferential with (i.e. the same as) the subject of some other verb. For example in Yuman languages, like Maricopa, the verbal suffix '-k' indicates that the subject of the verb is the same as the subject of the next verb in a sequence. The suffix '-m' indicates that the subject is different from the subject of the next verb: \ftx \fln (56) a. Nyaa '-ashvar-k '-iima-k \ftx I 1-sing-SS 1-dance-ASPECT \ftx 'I sang and I danced.' \ftx \ftx b. Bonnie-sh 0-ashvar-m '-iima-k. \ftx Bonnie-SUBJ 3-sing-DS 1-dance-ASPECT \ftx 'Bonnie sang and I danced.' \ftx \txt In Maricopa switch-reference markers are distinct from verb agreement, i.e they are a different inflectional category (note that both verbs 'agree' with their subjects by way of prefixes). Hence there is a lot of redundancy. Sometimes, however, coreference markers are incorporated right into the system of verb agreement. In this case the category of 'third person' is subdivided into two, one for same reference and another for switch reference. Various terminology is used to refer to such systems, e.g. reflexive, fourth person, recurrent, etc. Yup'ik Eskimo exhibits such a system: \ftx \fln (57) a. Dena-q quya-u-q Toni-aq cinga-llra-0-ku. \ftx -ABS happy-INTRNS-3 -ABS kiss-because-3/3-DEP \ftx 'Denai is happy because shej kissed Tony.' \ftx \ftx b. Dena-q quya-u-q Toni-aq cinga-llra-mi-ku. \ftx -SS \ftx 'Denai is happy because shei kissed Tony.' \ftx \txt To continue reading on this topic, highlight the key below: \cf Hcplx_med_sw2 \ftx \dt 27/Apr/1998 \key Hcplx_med_sw2 \shd Switch reference systems II \txt Perhaps the most complex systems of switch reference occur in the languages of highland Papua New Guinea. For example Kâte has a switch-reference system consisting of four markers: \ftx \ftx +---------------------------------------+ \ftx |Overlap ('while') | Succession ('then')| \ftx +--------+------------------+--------------------| \ftx | SS | -huk | -ra | \ftx |--------+------------------+--------------------| \ftx | DS | -ha | -0 | \ftx +------------------------------------------------+ \ftx \fln (58) a. Fisi-huk na-wek. 'As hei arrived, hei was eating.' \ftx arrive-SS ate-3SG \ftx \ftx b. Fisi-ra na-wek. 'Hei arrived, then hei ate.' \ftx arrive-SS ate-3SG \ftx \ftx c. Mu-ha-pie kio-wek 'As they spoke, he was weeping.' \ftx speak-DS-3PL weep-3SG \ftx \ftx d. Mu-0-pie kio-wek 'After they spoke, he wept.' \ftx speak-DS-3PL weep-3SG \ftx \fln Some systems are even more complex than this. \ftx \ftx Kanite: \fln (59) a. his-u'a-ke-'ka, 'If we do this,' \ftx do-we-DS-you \ftx \ftx b. naki a'nemo-ka hoya ali-'ka, 'you women work the garden,' \ftx so women-you garden work-you \ftx \ftx c. naki ali ha'anoma hu-ne'atale-'ka, 'when the work is finished,' \ftx so work finish do-COMPL-you \ftx \ftx d. popo hu-'ka, (e.) inuna kae-'ka, 'hoe and burn the weeds' \ftx hoe do-you weeds burn-you \ftx \ftx f. naki ha'no hu-talete-ke-ta'a 'when that is finished,' \ftx so finish do-COMPL-DS-we \ftx \ftx g. naki viemoka-ta'a keki'yamo'ma ha'noma ne-his-i-ana \ftx so men-we fence finish FUT-do-it-CONJ \ftx 'we men will finish building the fence.' \ftx \ftx '-ana' in (g) marks the end of the chain. \ftx \txt (a) is marked with 'u'e' indicating the subject. Then '-ke' indicates that the next clause will have a different subject, and finally '-'ka' indicates that the next subject will be second person. Each of the following 4 clauses is marked with '-'ka' indicating that the following clause will have a second person subject. Same subject is indicated with zero. In clause (f) we have the '-ke' marker again indicating that the final clause will have a different subject. After the '-ke' comes the form '-ta'a' indicating that the final clause will have 'we' as its subject. (g) is the only final verb in this series. All others are medial, i.e. they cannot stand alone as full propositions. This is probably because of the '-'ka' markers. Notice (g) has a tense marker, whereas none of the other verbs do. \ftx \txt To continue reading on this topic, highlight the key below: \cf Hcplx_med_sw3 \ftx \dt 27/Apr/1998 \key Hcplx_med_sw3 \shd Switch reference systems III \txt In Panare, operators that indicate same or switch reference relations between clauses also indicate temporal or logical relations. The following table illustrates these operators and the various relations that they convey (T. Payne 1991): \ftx \ftx Table Table 1: Medial clause suffixes in Panare (HC = high continuity, \ftx LC = low continuity) (Most prototypical medial clauses are listed first) \ftx \fln Operator Temporal relation Reference Other relations conveyed \ftx -------- ------------------- ---------------------------------- \fln -séjpe Succession (HC) Actor=Actor purpose \ftx -sé'ñape Succession (HC) Absolutive=Patient result \ftx -ñépe Succession (LC) Actor=/Actor movement/purpose \ftx \fln -npan Overlap Actor=Actor none \ftx -tááñe Overlap Actor=/Actor none \ftx \fln -pómën Anteriority Actor=Actor reason \ftx \fln References: Longacre and Thompson (1985), Munro and Haiman (1982). \ftx \dt 27/Apr/1998 \key Hcplx_other \shd Other Complex Constructions \ftx \ftx \key Hcplx_relcl \shd Relative clauses--Introduction \txt A relative clause is a clause that functions as a nominal modifier, e.g. \ftx \fln (60) The oaf that [ 0 trod on Lady Lucretia's toe ] . . . \ftx \fln The relevant parts of a relative clause are the following: \ftx \ftx 1. The HEAD of a relative clause is the noun phrase that is \ftx modified by the clause. In 60 the head is 'the oaf'. \ftx \ftx 2. The RESTRICTING CLAUSE is the clause itself. In 60 the \ftx restricting clause is indicated in brackets. \ftx \ftx 3. The RELATIVIZED NOUN PHRASE (NPrel) is the element within \ftx the restricting clause that is coreferential with the head \ftx noun. In 60 the NPrel is represented as 0 (a gap). \ftx \ftx 4. The RELATIVIZER is the morpheme or particle that sets off \ftx the restricting clause as a relative clause. In 60 the \ftx relativizer is 'that'. If the relativizer reflects some \ftx properties of the NPrel within the restricting clause \ftx (e.g. humanness, grammatical relation in the restricting \ftx clause, etc.), then it can be termed a RELATIVE PRONOUN. \ftx \txt To continue this discussion, highlight and jump to one of the topics below: \cf Hcplx_relcl_typ Typological parameters \cf Hcplx_relcl_elem Clause Elements Capable of being Relativized \ftx \dt 27/Apr/1998 \key Hcplx_relcl_elem \shd Clause elements capable of being relativized \txt Finally, in outlining the typology of relative clauses in a language, it is important to specify for each type of relative clause encountered which elements can be relativized. Keenan and Comrie (1977) observe that any given relative clause strategy will allow relativization on a continuous segment of the following hierarchy: \ftx \ftx Subject > Direct object > Indirect object > Oblique > Possessor \ftx \txt For example, according to Keenan and Comrie, no language allows relativization on subjects and indirect objects but not direct objects using a single strategy. Also, Keenan and Comrie assert that in any given language, if one position on this hierarchy is relativizable, all positions to the left will also be relativizable, though not necessarily with the same strategy. For example, some languages allow relativization of subjects, but no other clausal arguments. No language, however, allows relativization of direct objects but not subjects. Different case recoverability strategies (e.g. gap, pronoun retention, etc.) may be employed for different positions, but there will never be a language which allows relativization on one position while not allowing it on a position to the left. \ftx \txt For example, standard written English allows relativization on all positions on this hierarchy, except possessor: \ftx \fln (84) a. I hate the alligator that 0 ate Mildred. SUBJECT \ftx b. I hate the alligator that Mildred saw 0. D.O. \ftx c. I hate the alligator that Mildred threw the ball to 0. I.D.O. \ftx d. I hate the alligator that Mildred rode on 0. OBLIQUE \ftx e.*I hate the alligator that 0 teeth are huge. POSSESSOR \ftx \txt For every relative clause strategy noted in the grammar sketch, be sure to clarify which 'positions' (i.e. grammatical relations) can be relativized with that strategy. Chances are the more explicit strategies (e.g. pronoun retention, internal head) will be used to relativize arguments further down (to the right) the hierarchy than the less explicit strategies. \ftx \dt 27/Apr/1998 \key Hcplx_relcl_hdless \shd Relative clauses--Headless examples \txt Headless relative clauses are those clauses which themselves refer to the noun that they modify. In general, languages in which nominal modifiers are themselves nouns are more likely to employ headless relative clauses as a major RC strategy than languages for which there is a distinct and large class of adjectives. English, and many other languages, can use headless relative clauses when the head noun is non-specific: \ftx \fln (66) a. [ Whenever I'm afraid ], I call her. \ftx (c.f. 'Any time that I am afraid . . .') \ftx \ftx b. [ Whoever goes to the store ] should get some water balloons. \ftx (c.f. 'The person who goes to the store . . .') \ftx \txt Some languages use headless relative clauses whenever specific reference to the head is clear. Often the relative pronoun specifies as clearly as necessary, e.g. 'who went to the store' = 'the person who went to the store', 'where I live' = 'the place where I live.' etc. Ndjuká (Suriname Creole) apparently employs headless relative clauses for both non-specific and specific referents (examples courtesy of George Huttar): \ftx \fln (67) a. [ Di o doo fosi ] o wini. SUBJECT, NON-SPECIFIC \ftx REL FUT arrive first FUT win \ftx 'Whoever arrives first will win.' \ftx \ftx b. A mainsi ya a [ di e tan a ini se ] \ftx the eel here COP REL CONT stay LOC inside sea \ftx 'This eel is what (the one that) lives in the sea.' \ftx SUBJECT, SPECIFIC \ftx \ftx c. A daai go anga [ di a be puu ] OBJECT, SPECIFIC \ftx 3SG turn go with REL 3SG ANT remove \ftx 'He turned and returned with what (the ones) he had removed.' \ftx \fln To continue this discussion, highlight and jump to the topic below: \cf Hcplx_relcl_vscplcl \ftx \dt 27/Apr/1998 \key Hcplx_relcl_inthd \shd Relative clauses--Internally headed examples \txt Internally headed relative clauses are those for which the head is within the relative clause. Many OV languages, including Bambara, a Niger-Congo language of West Africa, exhibit internally headed relative clauses: \ftx \fln (65 ) a. ne ye so ye 'I saw a horse.' \ftx 1SG PAST horse see \ftx \ftx b. ce ye [ ne ye so min ye ] san \ftx man PAST 1SG PAST horse REL see buy \ftx 'The man bought the horse that I saw.' \ftx \txt Note that the relativizer 'min' is the only thing that marks the clause in brackets as a relative clause. The head noun remains intact within the relative clause and is not repeated external to the relative clause, as in the other examples above. Internally headed relative clauses can be thought of as another means of avoiding having a phonologically large and semantically complex modifier precede the head noun. It obviates the hearer having to 'wait' until the relative clause has been uttered to learn what noun the clause modifies. \ftx \fln To continue this discussion, highlight and jump to the topic below: \cf Hcplx_relcl_hdless \ftx \dt 27/Apr/1998 \key Hcplx_relcl_np \shd Noun phrase encoding in relative clauses, verb medial languages \txt The second major parameter by which relative clauses can vary is how the NPrel is encoded. This parameter is sometimes stated as a 'case recoverability' problem (e.g. Keenan 1985). That is, in any relative clause there must be some way of determining what the role of the head noun is within the relative clause. Since the head noun itself functions in another clause (the 'main') clause, it directly codes its role in that clause, e.g. by its position in the main clause, case marking etc. However, the head noun always has a coreferent within the relative clause (the NPrel in our terms). The role of that NP can be different from the role of the head noun within the main clause. For example, in 71a the head noun is the subject of the main clause verb 'ate'. It is also the subject of the relative clause verb 'saw'. In 71b, however, 'the alligator' is still the subject of 'ate', but it is now the object of the relative clause verb: \ftx \fln (71) a. The alligator that saw me ate Alice. \ftx b. The alligator that I saw ate Alice. \ftx \txt These sentences can be considered to be reductions of the following two abstract structures: \ftx \fln (72) a. The alligator that [ the alligator saw me ] ate Alice. \ftx HEAD NOUN NPrel \ftx \ftx b. The alligator that [ I saw the alligator ] ate Alice. \ftx HEAD NOUN NPrel \ftx \txt Since the NPrel is 'gapped' (i.e. left out) in the surface structure of these sentences (e.g 72), a problem arises as to how the hearer is to identify the grammatical relation of this invisible noun phrase. English solves this problem by simply leaving a conspicuous 'gap' in the position where the NPrel would be if it were overtly expressed. This is called the GAP STRATEGY. This strategy works for languages that have a fairly fixed constituent order, i.e. those for which grammatical relations are conveyed via the position of the core nominals in a clause. In such languages a missing argument is very obvious. However, if the language allows many constituent orders, and/or if grammatical relations are specified via some device other than constituent order, the gap strategy may leave the relative clause ambiguous. \ftx \txt To continue reading on this topic, highlight the key below: \cf Hcplx_relcl_np_2 \ftx \dt 27/Apr/1998 \key Hcplx_relcl_np_2 \shd Noun phrase encoding in relative clauses, verb initial languages II \txt The gap strategy is only effective in recovering the grammatical relation of the NPrel in verb medial languages. For example, Isthmus Zapotec is a VAP language that allows an NPrel to be coded with a gap, but for which the gap is useless as a case recovery strategy: \ftx \fln (73) a. najii Juan junaa 'John loves a woman.' \ftx loves John woman \ftx \ftx b. najii junaa Juan 'A woman loves John.' \ftx \ftx c. junaa ni najii Juan 0 'A woman that John loves.' \ftx woman REL loves John \ftx \ftx d. junaa ni najii 0 Juan 'A woman that loves John.' \ftx \txt Examples 73a and b show that constituent order is in fact the means of determining the grammatical relations of A and P arguments in transitive clauses. Example 73c shows that when the NPrel is the object of the relative clause, the gap appears in the object position, i.e. after the subject. Example 73d illustrate that when the NPrel is the subject of the relative clause, the gap appears in the subject position. Notice, however, that 73c and d are identical. The gap is an abstract symbol that doesn't actually have any phonetic realization (though sometimes there may be intonational cues as to where such gaps appear). In fact the grammatical relation of the NPrel is simply unrecoverable in these sentences out of context. In context, of course, the pragmatics of the situation would normally disambiguate. These examples show that though case recoverability is an important property of relative clauses, at times a certain amount of ambiguity is tolerated. However, the fact that the gap strategy is potentially ambiguous in VAP languages that don't have overt marking of grammatical relations, such as Zapotec, explains why this strategy is uncommon in such languages. The same holds true for APV languages. \ftx \txt To continue reading on this topic, highlight the key below: \cf Hcplx_relcl_np_3 \ftx \dt 27/Apr/1998 \key Hcplx_relcl_np_3 \shd Noun phrase encoding in relative clauses-- Pronoun retention \txt If the gap strategy is insufficient, the language is likely to use a more explicit device to express the grammatical relation of the NPrel. The next device we will discuss is termed PRONOUN RETENTION. In this strategy a pronoun that explicitly references the grammatical relation of the NPrel, either by its position, its form or both, is retained within the relative clause. Pronoun retention is used in many types of relative clauses in spoken English: \ftx \fln (74) That's the guy who [ I can never remember his name ]. \ftx \fln In this sentence the NPrel is coded by the pronoun 'his'. \ftx \fln (75) We've got 16 drums that we don't even know what's in them. \ftx (Heard on a television news interview). \ftx \txt Here is an example of the pronoun retention strategy in modern Israeli Hebrew (Keenan 1985:146): \ftx \fln (76) ha-sarim she [ ha-nasi shalax otam la-mitsraim. ] \ftx DEF-ministers REL DEF-president sent them to-Egypt \ftx 'The ministers that the President sent to Egypt.' \ftx \txt It is rare for pronoun retention to be used to relativize the subject of the relative clause. For example, the following RC is ungrammatical in Hebrew: \ftx \fln (77) *ha-ish she [ hu makir oti ] \ftx DEF-man REL he knows me \ftx `*The man who he know me.' \ftx \txt Keenan (1985) claims that Urhobo (Kwa, Niger-Kordofanian, Nigeria) and Yiddish are the only languages which clearly employ the pronoun retention strategy to relativize the subject position. \ftx \txt To continue reading on this topic, highlight the key below: \cf Hcplx_relcl_np_4 \ftx \dt 27/Apr/1998 \key Hcplx_relcl_np_4 \shd Noun phrase encoding in relative clauses-- Relative pronouns \txt Sometimes a special pronoun functions both to identify a structure as a relative clause, and to identify the relativized NP. These special pronouns are called RELATIVE PRONOUNS. Relative pronouns are typically similar to other special pronouns in the language, either the question words or pronouns used to refer to non-specific, indefinite items (see Hprag_ndecl_q_qwd on the parallel between relative pronouns and question words in English). Relative pronouns can be thought of as combining the functions of a plain relativizer and a clause internal pronoun that refers to the relativized NP. English allows the relative pronoun strategy, a relativizer plus gap strategy, and an unmarked 'no relativizer' plus gap strategy. Sometimes all three are allowed in the same environment, and it is difficult to determine what semantic nuances are conveyed, if any, by the various allowable structures: \ftx \fln (78) a. Rel Pro: The man who saw me \ftx b. Rel + gap: The man that saw me \ftx c. No Rel: *The man [ 0 saw me ] \ftx \fln (79) a. Rel Pro: The man whom [ I saw] \ftx b. Rel + gap : The man that [ I saw 0 ] \ftx c. No Rel: The man [ I saw 0 ] \ftx \fln (80 ) a. Rel pro: The place where I live \ftx b. Rel + gap: *The place that I live \ftx c. No Rel : The place I live \ftx \fln (81) a. Rel pro: The reason why I came \ftx b. Rel + gap: The reason that I came \ftx c. No Rel: The reason I came \ftx \fln (82) a. Rel pro: ?The way how he did it (acceptable to some speakers) \ftx b. Rel + gap: The way that he did it \ftx c. No Rel: The way he did it \ftx \fln (83) a. Rel pro: The table which he put it on \ftx b. Rel + gap: The table that he put it on \ftx c. No Rel : The table he put it on \ftx \dt 27/Apr/1998 \key Hcplx_relcl_post-n \shd Relative clauses--Post-nominal examples \ftx \fln To continue this discussion, highlight and jump to the topic below: \cf Hcplx_relcl_inthd \ftx \dt 20/Aug/1997 \key Hcplx_relcl_pre-n \shd Relative clauses--Pre-nominal examples \fln Prenominal relative clauses occur in some OV languages. \ftx \ftx Japanese: \fln (62) a. Yamada-san ga sa'ru o ka't-te i-ru \ftx Yamada-Mr. NOM monkey ACC keep-PTCP be-PRES \ftx 'Mr. Yamada is keeping a monkey.' \ftx \ftx b. [ Yamada-san ga ka'tte iru ] sa'ru \ftx 'The monkey that Mr. Yamada is keeping' \ftx \ftx c. [ sa'ru o ka'tte iru ] Yamada-san \ftx 'the Mr. Yamada who is keeping a monkey' \ftx \txt Examples 62b and c illustrate two relative clauses based on the independent clause in 62a. In both of the relative clauses, the restricting clause comes before the head. \ftx \txt Example 63 illustrates an entire Turkish sentence in which a relative clause modifies one of the nominals: \ftx \fln (63) Eser [ uyuy-na ] kadïn-ï tanyor \ftx Eser sleep-PTCP woman-ACC knows \ftx 'Eser knows the woman who is sleeping.' \ftx \txt Turkish is an OV language and, true to its type, it employs prenominal relative clauses. The head of the relative clause in 63 is 'kadïn'-- 'woman'. This noun is preceded by the relative clause in brackets. Note also that the verb within the relative clause is marked as a participle. This is a very common feature of relative clauses, especially in languages that have a lot of verbal morphology (polysynthetic languages). Even English has a marginal participial relative clause strategy: \ftx \fln (64) a. Eser knows the [ sleep-ing ] woman. \ftx b. Eser sat on a [ fall-en ] log. \ftx c. Eser ripped up her [ reject-ed ] novel. \ftx \txt All of the highlighted morphemes in these examples are markers of adjectives derived from verbs (participial verbs) of one type or another. Though traditional English grammar would not call such verb forms 'clauses' at all, nevertheless they fulfill our definition of relative clause. For many languages (e.g., Turkish) constructions analogous to these are the only means of modyfying a noun phrase using anything like a clause, i.e., they function just like relative clauses even though they may not be very clause-like formally. \ftx \txt To continue this discussion, highlight and jump to the topic below: \cf Hcplx_relcl_post-n \ftx \dt 27/Apr/1998 \key Hcplx_relcl_typ \shd Relative clauses--typological parameters \txt There are several typological parameters by which relative clauses can be grouped. The parameters to be discussed and exemplified in this section are 1) the position of the clause with respect to the head noun, 2) the mode of expression of the relativized NP (sometimes called the 'case recovery strategy'), and 3) which grammatical relations can be relativized. \ftx \txt The first typological parameter by which relative clauses can vary is the position of the clause with respect to the head. Relative clauses can be either PRENOMINAL (the clause occurs before the head), POST-NOMINAL (the clause occurs after the head), INTERNALLY HEADED (the head occurs within the relative clause) or they may be HEADLESS (self-explanatory). Since relative clauses are noun modifiers, one might expect that they would occur in the same position as other noun modifiers, e.g. descriptive adjectives, numerals, etc. Though it is generally true that the position of the relative clause with respect to the head noun often is the same as the position of descriptive modifiers, there is a distinct tendency for relative clauses to be post-nominal, even in languages for which descriptive modifiers are prenominal. This tendency is probably due to a universal pragmatic principle that shifts 'heavy', i.e. long, phonologically complex, information late in the clause. This is the same principle that motivates post-posing of subject complements in English (see Hcplx_cplcl) \ftx \txt For examples of these types, jump to: \cf Hcplx_relcl_pre-n Pre-nominal examples \cf Hcplx_relcl_post-n Post-nominal examples [INCOMPLETE] \cf Hcplx_relcl_inthd Internally headed examples \cf Hcplx_relcl_hdless Headless Examples \ftx \dt 27/Apr/1998 \key Hcplx_relcl_vscplcl \shd Headless Relative clauses versus complement clauses \txt Sometimes headless relative clauses are hard to distinguish from complement clauses. They are, however, distinct at least at the semantic level. The following examples of headless relative clauses (HRCs and corresponding complement clauses (CCs) in English: \ftx \fln (68) a. HRC: That which John said annoyed her. \ftx b. That John said it/something/anything annoyed her. \ftx \txt Note that the semantic representation of 68a is ANNOY(x,y), i.e. the content of what John said annoyed her. The representation of 68b, on the other hand is better represented as ANNOY(P,y). That is, it wasn't what he said that annoyed her, but the fact that he said anything at all. His act of saying is annoying rather than the specific thing he said. Here are some further examples: \ftx \fln (69) a. HRC: I hate where I live. (marginally acceptable to some?) \ftx \ftx b. CC: I know where I live. \ftx \txt The semantic representation of 69a would be HATE(x,y), where y is a thing, i.e., the particular place in the world where I live. The semantic representation of 69b, on the other hand, would be KNOW(x,P) where P is a proposition. i.e. 'I live there.' It is not the same use of the verb 'know' as in 'I know your brother.' \ftx \fln (70) a. HRC: Whoever goes to the store has to buy me some rice. \ftx \ftx b. CC: I don't know whether he'll go to the store. \ftx \txt Example 70a is very similar in function to an adverbial (conditional) clause (see Hcplx_advcl): 'If someone goes to the store, he has to buy me some rice'. In some languages this type of conditional clause is not distinct morphosyntactically from relative clauses. Example 70b is an embedded question. It can be paraphrased as 'I don't know the answer to the question "will he go to the store?"'. \ftx \dt 27/Apr/1998 \key Hcplx_sv \shd Serial verbs \txt A serial verb construction contains two or more verb roots that are neither compounded (see Hv_cmp_v) nor are members of separate clauses. Serial verbs are most common in languages that have little or no verbal morphology (analytic languages, see Hmor_typ). English marginally employs serial verbs: \ftx \fln (2) ?Run go get me a newspaper. \ftx \txt In many other languages, serial verbs are a much more well-installed characteristic of the grammar. Typically, verbs in a series will express various facets of one complex event. For example, the concept expressed by the English verb 'bring' is divisible into at least two components, the picking up or taking of an object and the movement towards a deictic center. In many languages, this complex concept is embodied in a serial verb construction ( 3a): Yoruba (Bamgbose 1974): \ftx \fln (3) a. mo mú ìwé wá ilé 'I brought a book home.' \ftx I take book come house \ftx \ftx b. mo mú ìwé; mo sì wá ilé 'I took a book and came home.' \ftx I take book I and come house \ftx \txt Example 3b illustrates a pair of coordinate clauses that employ the same two verb roots as the serial construction in 3a. The formal factors that distinguish 3a as a serial construction are the following: \ftx \ftx 1. There is no independent marking of the subject of the \ftx second verb. \ftx \ftx 2. There is no independent tense/aspect marking of the \ftx second verb. \ftx \ftx 3. The intonation is characteristic of a single clause. \ftx \txt To continue this discussion, highlight and jump to the topics below: \cf Hcplx_sv_tam Serial verbs--Tense/aspect/mode \cf Hcplx_sv_extr Serial verbs--Extraction \cf Hcplx_sv_subj Serial verbs--Subjects \cf Hcplx_sv_vschn Serial verbs versus Clause chaining \cf Hcplx_sv_vsaux Serial verbs vs Auxillaries \cf Hcplx_sv_ex Examples \ftx \dt 27/Apr/1998 \key Hcplx_sv_ex \shd Serial verbs--Examples \txt The following examples illustrate various serial verb constructions: \ftx \ftx Nupe (Hyman 1971): \fln (16) u lá dùkù là 'He broke the pot.' \ftx he took pot break \ftx \ftx Supyire: \fln (17) U-a lì tàha-a nùnke pwò 'She tied her hair with it.' \ftx she-PERF it use-NF head.DEF tie \ftx \ftx Yoruba (Stahlke 1970): \fln (18) mo sô fún ô ... 'I said to you ...' \ftx I say give you \ftx \ftx Efik (Welmers 1973): \fln (19) nám útom emì nê mì 'Do this work for me!' \ftx do work this give me \ftx \ftx Tok Pisin (Givón 1988): \fln (20) ... em i-wokim paya pinis ... 'she got the fire started' \ftx she PRED-make fire finish \ftx \fln (21) ... em i-brukim i-stap 'he keeps breaking (it)' \ftx he PRED-break PRED-be \ftx \ftx Supyire (again): \fln (22) Zànhe sí dùfugé keege 'The rain will spoil the maize.' \ftx rain go maize.DEF spoil \ftx \fln (23) mu gú n`-jà m`-pà wíí 'You can come see me.' \ftx you POT FUT-be.able FUT-come look \ftx \ftx Tibetan (DeLancey 1990): \fln (24) qhó phoo (cee) cî-pêréè 'He escaped away.' \ftx he:ABS escape NF went-PERF.DISJUNCT \ftx \ftx Supyire: \fln (25) ... fó kà pi-í m-pá lye '...till they grow up.' \ftx till and they-SEQ CN-come be.old \ftx \fln (26) U-a sòla-a kàrè 'She left early (in the a.m.)' \ftx she-PERF be.early-NF go \ftx \fln (27) ka-a tònna-a tòrò 'It's too/very long.' \ftx it-PERF be.long-NF pass \ftx \fln (28) u-a wyera-a kàrè 'He left me in a hurry/too soon' \ftx he-PERF be.hot-NF go \ftx \fln (29) u-a cì cyán-á màhà 'She dropped them all over the \ftx she-PERF them drop-NF do.DIST place' \ftx \dt 27/Apr/1998 \key Hcplx_sv_extr \shd Serial verbs--Extraction \txt Another interesting formal characteristic of prototypical serial verb constructions is that when a constituent of the second verb is extracted , i.e. moved for pragmatic purposes, it extracts to the front of the entire serial construction. Example 6a illustrates the same Yoruba sentence with the constituent 'to the house' fronted. Example 6b illustrates that such fronting cannot occur in a coordinate construction: \ftx \fln (6) a. ilé ni mo mú ìwé wá 'It was to the house \ftx house is I take book come that I brought a book.' \ftx \ftx b *ilé ni mo mú ìwé mo sì wá \ftx I and come \ftx \txt To continue this discussion, highlight and jump to the topic below: \cf Hcplx_sv_subj \ftx \dt 27/Apr/1998 \key Hcplx_sv_subj \shd Serial verbs--Subjects \txt Some serial verb constructions are less than prototypical in that some inflectional information may be carried by both verbs. For example, in Akan both verbs in a serial construction must have the same subject, but the subject is redundantly specified on both: \ftx \ftx Akan (Schachter 1974): \fln (7) mede aburow migu msum 'I pour corn into the water.' \ftx I.take corn I.flow water.in \ftx \txt Supyìrè and Minyanka are closely related Senufo languages of Mali, West Africa. In Supyìrè both verbs in a serial construction may contain a reference to the subject (8), while in Minyanka, the subject reference in the second clause is omitted (9): \ftx \ftx Supyire: \fln (8) pi-a yì yàha pí-á kàrè fó Bàmàko e \ftx they-PERF them leave they.SUB-PERF go till Bamako to \ftx 'They let them go to Bamako.' \ftx \ftx Minyanka (Dan Brubaker, p.c.): \fln (9) pá yì yáhá kárì fó Bàmàkò nì. \ftx they.ASP them leave go till Bamako to \ftx 'They sent them to Bamako.' (Lit: 'let them go.') \ftx \txt One might say that the Minyanka serial verb construction is 'further along' in the diachronic path from fully independent clauses to compound verbs. \ftx \txt To continue this discussion, highlight and jump to the topic below: \cf Hcplx_sv_vschn \ftx \dt 27/Apr/1998 \key Hcplx_sv_tam \shd Serial verbs--Tense/aspect/mode \txt The following examples illustrate that in the Yoruba serial verb construction, tense/aspect/mode information is carried by the first verb: \ftx \fln (4) mò n mú ìwé bô / *wá 'I am bringing a book.' \ftx I PROG take book come.PROG \ftx \txt In example 4 the auxiliary that signals progressive aspect occurs before the first verb. It is not repeated before the second verb. Nevertheless, the form of the verb meaning 'come' must be consistent with progressive rather than perfective aspect. \ftx \txt Example 5a illustrates that the negative particle is associated with the first verb. Nevertheless, negation has scope over the entire clause (see Hprag_neg_n-cl_sco -- Negative Scope). Example 5b illustrates that the negative cannot be associated with the second verb: \ftx \fln (5) a. èmi kò mú ìwé wá 'I did not bring a book.' \ftx I.NEG not take book come \ftx \ftx b. *èmi mú ìwé kò wá \ftx \txt In coordinate clauses, each clause can have its own tense, aspect and mode. \ftx \txt To continue this discussion, highlight and jump to the topic below: \cf Hcplx_sv_extr \ftx \dt 27/Apr/1998 \key Hcplx_sv_vsaux \shd Serial verbs versus Auxillaries \txt Semantically, serial verb constructions often mean something slightly different from what the same series of verbs would mean if they were cast in separate clauses. However, if the semantics have changed very much, it is possible that the one of the verbs in the series has been re-analyzed as an auxiliary. Serial verbs are one major diachronic source for auxiliaries. In Lahu, some verb pairs are ambiguous, out of context, as to whether they are to be construed as a series of co-equal serial verbs, or as an auxiliary plus a main verb. Only the semantics reveal any difference whatsoever: \ftx \ftx Lahu: \fln (13) lò chê a. beg to be there (Verb series) \ftx beg be.there b. is begging (verb + aux) \ftx \fln (14) ga kì a. is busy getting (Verb series) \ftx get be.busy b. must be busy (aux + verb) \ftx \fln (15) ta sa a. easy to begin (Verb ser.) \ftx begin be.easy b. begin to be easy (aux + verb) \ftx \txt To continue this discussion, highlight and jump to the topic below: \cf Hcplx_sv_ex \ftx \dt 27/Apr/1998 \key Hcplx_sv_vschn \shd Serial verbs versus Clause chaining \txt In Lahu, the difference between a serial construction and a clause chain (see Hcplx_med_chn) is that in a clause chain non-final verbs may take a special particle. This particle may not appear on the final verb: \ftx \ftx Lahu (Matisoff 1973): \fln (10) lâ pô¿(-le) chè¿(-le) câ(*-le) pê . . . \ftx tiger jump(-NF) bite(-NF) eat(*-NF) finish \ftx 'The tiger jumped (on them), bit into (them), and ate (them) up.' \ftx \txt In serial constructions, on the other hand, none of the verbs may take the non-final particle: \ftx \fln (11) súqhu nîqhu kê lò¿ chï ve \ftx pipes three put.into be.enough roll ? \ftx 'He rolls enough to put into three pipes.' \ftx \txt The actual meaning of a serial verb construction as a whole can often be ambiguous out of context. The following example from Thai is provided by Foley and Olson 1985: \ftx \ftx Thai: \fln (12) John khàp rót chon khwaay taay \ftx John drive car collide buffalo die \ftx \ftx a. John drove the car into a buffalo and it (buffalo) died. \ftx b. John drove the car into a buffalo and it (car) stalled. \ftx c. John drove the car into a buffalo and he (John) died. \ftx \txt Out of context this sentence is ambiguous in the three ways illustrated above. In discourse, only the pragmatics of the situation can disambiguate. \ftx \txt To continue this discussion, highlight and jump to the topic below: \cf Hcplx_sv_vsaux \ftx \dt 27/Apr/1998 \key Hdisc \shd The language as a vehicle for discourse \txt This topic contains all of the following subsections. Highlight and jump to the topic of interest: \shd2 Inroduction \cf Hdisc_intro_ling Discourse analysis and linguistic analysis \cf Hdisc_intro_intrp Discourse analysis vs interpretation \shd2 Continuity \cf Hdisc_cnt Continuity (cohesion) and Discontinuity \cf Hdisc_cnt_ref Topic (referential) Continuity \cf Hdisc_cnt_ref_refd Referential devices \cf Hdisc_cnt_ref_other Other referential devices \cf Hdisc_cnt_ref_inher Inherent vs context imparted factors \cf Hdisc_cnt_ref_ind Indices of Topic continuity \cf Hdisc_cnt_thm Thematic Continuity \cf Hdisc_cnt_thm_infr Inference \cf Hdisc_cnt_thm_ex Examples \cf Hdisc_cnt_thm_strc Structural frameworks \cf Hdisc_cnt_act Action Continuity \cf Hdisc_cnt_act_proc Procedural Text \cf Hdisc_cnt_prom Episodic Prominence \cf Hdisc_cnt_prom_intns Intensification \cf Hdisc_cnt_prom_clim Climax/Peak \shd2 Genres \cf Hdisc_genr Genres \cf Hdisc_genr_cnvs Conversation \cf Hdisc_genr_nar Narrative \cf Hdisc_genr_nar_pers Personal Experience \cf Hdisc_genr_nar_hist Historical \cf Hdisc_genr_nar_flk Folk Stories \cf Hdisc_genr_nar_myth Mythology \cf Hdisc_genr_hort Hortatory \cf Hdisc_genr_proc Procedural \cf Hdisc_genr_exp Expository \cf Hdisc_genr_desc Descriptive \cf Hdisc_genr_rit Ritual Speech \shd2 Other topics \cf Hdisc_idm Idiomatic Expressions/Proverbs \cf Hdisc_symb **Sound Symbolism \cf Hdisc_typ **Typological Findings \cf Hdisc_misc **Miscellaneous and Conclusions \cf Hdisc_bib Bibliography or References Cited \dt 29/Jun/1998 \key Hdisc_bib \shd Bibliography or references cited \txt To qualify as a BIBLIOGRAPHY, a list of works should contain substantially everything that has been written on a certain well-defined topic. If your list of works contains only full references to works cited in the text of the sketch, call this list `references cited'. The following are suggested items to include in a true bibliography for a given language: \ftx \ftx 1. References to all works cited in the sketch. \ftx \ftx 2. References to all purely linguistic works on the language. \ftx \ftx 3. References to all published linguistic works in the language \ftx family. \ftx \ftx 4. References to all major published classifications which \ftx include the language family. \ftx \ftx 5. References to all published anthropological works on the \ftx speakers of the language. \ftx \ftx 6. References to all major published anthropological works on \ftx related peoples (i.e. speakers of languages in the same \ftx language family). \ftx \txt A `major' work is defined roughly as a published book-length work, or an influential article in a refereed journal. \ftx \dt 27/Apr/1998 \key Hdisc_cnt \shd Continuity (cohesion) and discontinuity \txt DISCOURSE consists of strings of propositions. A text is a linguistic artifact, a record of language used during a segment of discourse. Text, then, normally consists of strings of clauses, i.e. linguistic instantiations of propositions. However, not every series of clauses is a text. In order to be a text, a series of clauses must hang together in certain definable ways. This is because discourse hangs together. If a text does not reflect the cohesive character of discourse, then it is disfunctional, just as if it did not reflect the referential or temporal character of discourse. Thus text exhibits COHESION or CONTINUITY. \ftx \txt There are three kinds of continuity that will concern us here: TOPIC CONTINUITY, ACTION CONTINUITY and THEMATIC CONTINUITY. This division is somewhat arbitrary since the categories clearly overlap and interact with each other to a high degree. Also, there are probably kinds of continuity that aren't captured under these headings. Nevertheless, this tripartite division is reasonably well-defined, and will serve as a convenient framework within which a field linguist might organize observations regarding the discourse structuring devices of a language. \ftx \shd2 Topic Continuity \ftx \txt Topic continuity refers to the fact that discourse tends to evoke the same referents over and over again. Pronouns and other referential devices are morphosyntactic means of expressing this kind of continuity, as well as its converse, topic discontinuity, i.e., the introduction of new, unexpected referents. \ftx \shd2 Action Continuity \ftx \txt Action (or event, or situation) continuity refers to the fact that discourse tends to develop along predictable parameters, e.g., location (X happened here, Y happened there), time (first X happened then Y happened, then Z happened), or logic/causation (X happened because of Y, Y happened so that Z). Different kinds of discourse, or GENRES (see Hdisc_genr), rely on different organizational parameters. Foregrounding and backgrounding are defined in relation to the particular parameter employed, e.g., foregrounded clauses denote progress along the major organizational parameter, whereas backgrounded clauses provide ancillary, supportive information (Longacre 1976, Hopper and Thompson 1980). Tense/aspect marking and clause connectors are morphosyntactic structures that aid speakers and hearers in expressing and recovering this kind of continuity. \ftx \shd2 Thematic Continuity \ftx \txt Thematic continuity refers to the fact that discourse tends to revolve around recurring 'themes', e.g., 'how to make a blowgun,' or 'latest styles.' Inference is probably the major process whereby thematic continuity is recovered. However, devices that normally code other kinds of continuity also may be used, by extension, to code or reinforce thematic continuity. Thematic continuity is probably the most difficult kind of discourse continuity for linguists to deal with precisely because there is so little in the way of empirical evidence that identifies it. The most promising results so far are drawn from experimental studies (e.g., Tomlin 1991). Unfortunately, the methodologies developed in such studies are not particularly amenable to implementation in a field situation. Nevertheless, some important general principles may be forthcoming from this research. \ftx \txt It should be noted that the term 'theme', and related expressions, such as 'thematic' or 'thematicity', has been used in a variety of senses by linguists. We will use this term in the sense of Jones (1977) to mean the `main idea' of a text or portion of text. Themes are expressed via propositions akin to titles, rather than via referring expressions. This use departs rather significantly from the way the term is used by the Prague school, Halliday, Grimes and others. For these linguists, though theme is supposed to be a semantic notion, it is identified as the 'left-most constituent of a sentence' (Brown and Yule 1983:126). For the field linguist, it is extremely important to keep definitions of formal and functional categories independent of one another, even though that often makes functions more difficult to identify. \ftx \ftx This section also contains the following subsections: \cf Hdisc_cnt_ref Topic (referential)Continuity \cf Hdisc_cnt_thm Thematic Continuity \cf Hdisc_cnt_act Action Continuity \cf Hdisc_cnt_prom Episodic Prominence \ftx \dt 27/Apr/1998 \key Hdisc_cnt_act \shd Action continuity \txt Different kinds of discourse are organized according to different principles. For example, narrative discourse (see Hdisc_genr_nar) is normally organized according to time: first X happened, then Y happened, etc. The events in a narrative are those propositions that are related to one another sequentially, i.e., those that clearly end before the next one begins. Sometimes the series of propositions that express the events of a narrative is referred to as the 'time line', the 'main event line' or the 'backbone' of the text. In a good narrative, however, there is always a great deal of material that isn't on the main event line. For example, the following narrative excerpt has been divided up into events and non-events. Non-events have been tagged as to whether they express descriptive, evaluative or non-sequential information (based on Grimes' 1975 analysis of a section of text from p.1 of Out of the Silent Planet by C. S. Lewis): \ftx \fln Descriptive: The last drops of the thundershower had hardly \ftx ceased falling \ftx \fln EVENT: when the Pedestrian stuffed his map into his pocket, \ftx \fln EVENT: settled his pack more comfortably on his tired shoulders \ftx \fln EVENT: and stepped out from the shelter of a large chestnut \ftx tree into the middle of the road. \ftx \fln Descriptive: A violent yellow sunset was pouring through a rift \ftx in the clouds to the westward, \ftx \fln Descriptive: but straight ahead over the hills the sky was the \ftx colour of dark slate. \ftx \fln Descriptive: Every tree an blade of grass was dripping, \ftx \fln Descriptive: and the road shone like a river. \ftx \fln Non-sequential: The Pedestrian wasted no time on the landscape \ftx \fln EVENT: but set out at once with the determined stride of a \ftx good walker \ftx \fln Descriptive: who has lately realized that he will have to walk \ftx farther than he intended. \ftx \fln Descriptive: That indeed was his situation. \ftx \fln Non-sequential: If he had chosen to look back, \ftx \fln Non-sequential: which he did not, \ftx \fln Evaluative: he could have seen the spire of Much Nadderby,... \ftx \txt If one were to ask for a synopsis of 'what happened?' in this excerpt, the simple response would be: 'Some Pedestrian stuffed a map into his pocket, settled his pack on his shoulders, stepped out of the shelter of a chestnut tree and started walking.' All the other information sets the scene, or describes other, sometimes hypothetical related situations. \ftx \fln To continue this discussion, highlight and jump to the topic below: \cf Hdisc_cnt_act_proc \ftx \dt 27/Apr/1998 \key Hdisc_cnt_act_proc \shd Action continuity--Procedural Text \txt In a procedural text, the foregrounded clauses would be those that express the steps in the procedure. Backgrounded clauses would be comments on why one does this, descriptions of the materials, tools, etc. In a hortatory discourse, the foregrounded clauses would be those that express the behaviors the speaker is trying to elicit from the hearer (see Hdisc_genr_hort). \ftx \txt Care should be taken in too quickly identifying 'foreground' with the 'most important' information in a text. It may be more accurate to say that the foregrounded material is the framework on which the important information is hung. For example, Linde and Labov (1975) make the distinction between a 'chronicle' and a narrative. A chronicle is essentially a narrative without background material. It is the kind of text one often hears in response to the parental question 'what did you do today?': \ftx \txt I got up. I got dressed. I ate breakfast. I washed up. I walked to school. I played ... \ftx \txt Without evaluative, supportive material, the text has no 'point', other than to fulfill the parents' apparently meaningless question. The real significance of a narrative, as well as any other kind of text, often is carried in the 'backgrounded' clauses. For example, many fables would hardly be worth quoting without the 'moral of the story'. \ftx \txt References: Hopper and Thompson (1979), Givón (1983b), Hopper (1982), Delancey (1982), Wallace (1982). \ftx \dt 27/Apr/1998 \key Hdisc_cnt_prom \shd Episodic Prominence \txt In addition to continuity and discontinuity, discourse also exhibits various kinds of prominence. Some morphosyntactic devices used to highlight or ascribe prominence to clause elements are described in Hprag_iden on pragmatically marked structures. In this section, the fieldworker may want to describe recognizable, preferably grammaticalized, ways in which speakers of the language draw special attention to certain portions of discourse. \ftx \txt To continue this discussion, highlight and jump to the topic below: \cf Hdisc_cnt_prom_intns \dt 27/Apr/1998 \key Hdisc_cnt_prom_clim \shd Climax/peak \txt CLIMAX in narrative describes the point at which rhetorical tension is released. A climax must be preceded by a build-up of tension, and followed by some sort of resolution. 'Climax' does not necessarily refer to the end of a discourse. Climactic points are often characterized by unusual morphosyntactic structures. For example, in the story of 'Little Red Riding Hood' tension builds as Little Red senses something strange about her 'grandmother', who is really Big Bad Wolf in disguise: \ftx \fln (4) a. LR: My what big eyes you have Grandma! \ftx b. BBW: The better to see you with my dear. \ftx c. LR: My what big ears you have Grandma! \ftx d. BBW: The better to hear you with my dear. \ftx e. LR: My what big teeth you have Grandma! \ftx f. BBW: The better to EAT you with my dear! \ftx \txt Clause f. in this sequence can be characterized as a climax. Tension mounts in clauses a through e as the audience, aware that "Grandma" is in fact the wolf in disguise, anticipates this revelation to Little Red. In this example, the only special morphosyntactic device that signals the climax is special intonation. In many languages, other 'emphatic' or stylistically marked devices may occur. For example, in Yagua, objective inflection of intransitive verbs of motion is used at points of episodic climax (T. Payne 1990). Other languages may use cleft constructions, special constituent orders or particles at such points. \ftx \txt PEAK refers to a point in a narrative discourse where events are presented in rapid succession, with little backgrounded material interspersed. It is in some sense 'the important part' of the story. Climax and peak are certainly related phenomena, but they are logically distinct. The same sort of phenomena associated with climax are also commonly associated with peaks. Additional morphosyntactic correlates of narrative peaks include unusually short sentences, longer paragraphs, neutralization of tense/aspect marking and increased frequency of 'emphatic' particles. \ftx \fln References: Longacre (1986). \ftx \dt 27/Apr/1998 \key Hdisc_cnt_prom_intns \shd Intensification \ftx \shd2 Rhetorical Questions \txt RHETORICAL QUESTIONS are clauses that have the morphosyntactic form of questions, but which do not expect a literal answer. Rather, they function to 'highlight' or intensify an assertion. Rhetorical questions are particularly common in persuasive and expressive discourse, and are typically characterized by unusual intonation patterns. For example, a politician attempting to generate support for a war might culminate his speech by saying something like: \ftx \fln (5) Who will save our great nation from this dire threat? \ftx \txt Of course, the politician is not naïvely questioning his audience concerning some information they have and he does not. Rather, he is trying to elicit the response 'we will!'. If the build-up portion of his discourse has been successfull, this response will be obvious to all concerned. Rhetorical questions that seem to require 'yes' or 'no' responses (see Hprag_ndecl_q_ynq) are sometimes used to assert that certain propositions ought to be obvious. For example: \ftx \fln (6) Shall we continue to be humiliated by this foreign despot? \ftx \txt The obvious answer to this 'question', assuming that the discourse has been effective to this point, is 'No!'. In other situations a rhetorical question might be used to convey the speaker's emotional commitment to a proposition. For example, in an argument if I take offense at something my interlocutor says, I might respond: \ftx \fln (7) How can you say that?!?! \ftx \txt The combination of question marks and exclamation points is one graphic means sometimes used in English writing to represent the intonation of 'incredulity' associated with such rhetorical questions. Of course the function of this clause in the context of an argument is not at all to request information. Rather it is an intensive way of saying 'I disagree with what you just said'. \ftx \txt Occasionally even in narrative discourse, rhetorical questions will be used to convey intensive assertions. For example: \ftx \fln (8) Did he ever yell! \ftx \txt This clause is in the grammatical form of a question, but with the right intonation is really an intensive assertion meaning 'he really yelled.' \ftx \shd2 Negation \ftx \fln Negation is also sometimes used in this way: \ftx \fln (9) a. Did that bother you? \ftx b. Oh no, nothing like that. \ftx \txt When uttered with appropriate 'sarcastic' intonation, 9b can convey an intensive assertion: 'of course that really bothered me.' For some languages this is much more integral a part of normal discourse than it is in English. \ftx \txt Sometimes, as in English, rhetorical questions are posed in the negative: \ftx \fln (10) Didn't I tell you to take out the trash? \ftx \txt Again, with the appropriate intonation and in the right context, this clause would be likely to convey an intensive assertion: 'I told you to take out the trash!', or imperative: 'Take out the trash!'. \ftx \txt To continue this discussion, highlight and jump to the topic below: \cf Hdisc_cnt_prom_clim \dt 27/Apr/1998 \key Hdisc_cnt_ref \shd Topic (referential) continuity \txt The kinds of structures that are likely to function in the domain of topic continuity are: \ftx \ftx 1. Anaphoric zeros \ftx 2. Verb coding (or anaphoric agreement, or grammatical agreement) \ftx 3. Unstressed (clitic) pronouns \ftx 4. Stressed (independent) pronouns \ftx 5. Demonstrative pronouns \ftx 6. Full noun phrases \ftx 7. Specified noun phrases \ftx 8. Modified noun phrases \ftx 9. Special constituent orders, e.g., fronting \ftx 10. 'Voice' alternations, e.g., active, passive antipassive and inverse \ftx 11. 'Switch reference' systems \ftx \txt Of course, these structures also are sensitive to functional influences other than topic continuity and discontinuity. Nevertheless, a full fledged topic continuity study of a language would need to take into account any and all of these options that exist in the language. \ftx \txt To continue this discussion, highlight and jump to the topic below: \cf Hdisc_cnt_ref_refd \ftx \dt 27/Apr/1998 \key Hdisc_cnt_ref_ind \shd Indices of Topic continuity \txt There are three `indices' of topic continuity suggested by Givón: \ftx \txt REFERENTIAL DISTANCE measures the gap between the current mention of a referent and its previous mention in the discourse in terms of number of clauses. \ftx \txt PERSISTENCE is based on the notion that some referents are 'destined', as it were, to figure more prominently in the ensuing discourse than are others (see previous record). Recent studies have measured the persistence of a referent by counting the number of mentions of the same referent in a continuous span of text, usually 10 clauses, following the reference in question. \ftx \txt AMBIGUITY relates to the fact that whenever a referent is mentioned, there may be other referents on stage that compete for interpretation as the referent of the particular coding device employed. \ftx \txt Ref: Clancy (1980), Du Bois (1980), Givón (1983a,b), Hopper and Thompson (1984), T. Payne (1985) (on how to handle quoted material). \ftx \dt 27/Apr/1998 \key Hdisc_cnt_ref_inher \shd Topic continuity--Inherent versus context imparted factors \txt All of the above observations serve to illustrate that there is more to referential coding choices than simply the given vs. new distinction. \ftx \txt Givón (1983a, b, c) proposes a SCALAR notion of topic continuity. That is, any referent ('topic' in Givón's terminology) is more or less likely to be mentioned (more or less continuous) at any given point in the discourse. There are several factors that enter into the degree of continuity evidenced by a referent. These factors can be divided into INHERENT characteristics of the referents themselves and CONTEXT IMPARTED factors. These are briefly outlined below: \ftx \ftx 1. Inherent continuity (topic worthiness). Humans are \ftx inherently more likely to be mentioned in human discourse \ftx than are non-humans; entities that control events than \ftx entities that are passively affected by events, etc. \ftx \ftx 2. Context-imparted continuity. SAPs (Speech Act \ftx Participants) are highly likely to be mentioned in any \ftx discourse. Entities that are visible to speaker and hearer \ftx at the moment of speaking are more likely to be mentioned \ftx than random objects in the world. Referents that have \ftx already been mentioned are more likely to be mentioned again \ftx than are random referents in the world. \ftx \txt The quantitative methodology developed by Givón (1983) and colleagues is a way of determining how continuous any referent is at any given point in a text according to the last mentioned kind of context-imparted continuity. Although the methodology only measures one kind of continuity, it has the advantage of being rigorous and non-circular. Once continuity is measured, referential devices can be ranked in terms of the average continuity values of the referents they code. \ftx \txt To continue this discussion, highlight and jump to the topic below: \cf Hdisc_cnt_ref_ind \ftx \dt 27/Apr/1998 \key Hdisc_cnt_ref_other \shd Topic continuity--Other referential devices \txt In addition to the brute binary distinction between initial mention and subsequent mentions, there are many more functional principles that are known to impinge on referential systems. Prince (1981) provides a good framework within which a field linguist might couch a more detailed description of the topic-continuity-related devices of a language. \ftx \txt One function not developed by Prince is the notion of 'deployability' or 'importance'. This notion reflects the fact that not all initial mentions are created equal. Some referents are 'destined' to figure prominently in the subsequent discourse whereas others are just passing through. Languages typically possess alternative coding devices to reflect this distinction. For an obvious example, Wright and Givón (1987) demonstrate that in spoken North American English a noun phrase preceded by the demonstative this serves to introduce referents that are destined to figure prominently in the subsequent discourse. For example, 1 below sounds odd because the use of 'this guy' sets the reader up to expect the subsequent discourse to involve the indicated referent: \ftx \ftx (1) I was sitting there reading a newspaper when this guy \ftx walks up to me. It was the New York Times, and I was \ftx fascinated by a front-page story about linguistics. After I \ftx finished reading it I went home. \ftx \fln Example 2 is a more natural use of this expression: \ftx \ftx (2) I was sitting there reading a newspaper when this guy \ftx walks up to me and says, "Hey lady, you got a quarter for a \ftx cuppa coffee?" He looked familiar somehow, so I asked him ... \ftx \txt On the other hand, the standard expression using the article a tends to refer to new referents that are not going to be particularly important in the subsequent text (cf. a newspaper, a quarter and a cuppa coffee in 2). \ftx \txt To continue this discussion, highlight and jump to the topic below: \cf Hdisc_cnt_ref_inher \ftx \dt 27/Apr/1998 \key Hdisc_cnt_ref_refd \shd Topic continuity--Referential devices \txt At the grossest level, referential devices have one of two functions in the domain of topic continuity: they either code initial appearances of a referent on the discourse stage or they code appearances of a referent that is already on the discourse stage. In other words, whenever a referent is mentioned in a text it is either already `on stage' or it is being brought `onto stage'. It is safe to say that languages always possess distinct structures that characteristically code these two functions. Any grammar sketch should describe at least this aspect of the referential system. Some of the terminology that has been used for this distinction is: \ftx \ftx Initial appearance Subsequent appearance \ftx -------------------- --------------------- \ftx Coming onto stage Already on stage \ftx New Given (Halliday 1967) \ftx Switch Continuing \ftx Previously inactivated Activated (Chafe 1987) \ftx Discontinuous Continuous (Givón 1983) \ftx \txt This distinction may be relevant within a clause or within a higher level unit. For example, switch-reference systems (see Hcplx_med_sw1) typically indicate whether a referent is the same or different from a referent in a neighboring clause even though all referents are 'on stage' in terms of the discourse as a whole. Very different structures are used to indicate that a referent is being mentioned for the very first time in the discourse ('brand new' in terms of Prince 1981). \ftx \txt It should be noted that the crucial criterion is whether the referent is 'on stage', not whether it has already been mentioned in the text. There are various ways in which a referent can be made `available' on the discourse stage without necessarily being overtly mentioned. Thus the first actual textual mention of a referent need not be an introduction of that referent. Referents are often treated as 'given' when they are first mentioned. Some ways in which referents can be brought `onto stage' without explicit mention are: \ftx \fln Presence in the discourse context: \ftx I hope he's not vicious. (One pedestrian to another as a large dog \ftx approaches) \ftx \fln Perpetually present: \ftx The sun was out. (No need to say 'there is a sun that \ftx sometimes comes out.') \ftx \ftx I love you. (Speech act participants are always 'on \ftx stage') \ftx \fln Part of a discourse 'frame': \ftx We had dinner at Fat (The restaurant frame includes \ftx City last night. The waiters/waitresses) \ftx waitress spilled coffee \ftx on me. \ftx \fln Rhetorical suspense: \ftx The executioner smiled. (First line of a short story) \ftx \txt In spite of the many extra-grammatical ways in which referents get onto the discourse stage, all languages also provide morphosyntactic devices to explicitly accomplish this function. \ftx \txt Also, a referent that has appeared earlier in the text may subsequently have been removed from the discourse stage. In this case it may have to be 're-introduced'. \ftx \txt To continue this discussion, highlight and jump to the topic below: \cf Hdisc_cnt_ref_other \ftx \dt 27/Apr/1998 \key Hdisc_cnt_thm \shd Thematic continuity \ftx \shd2 Hierarchy of speech \txt Speech is necessarily linear, since sounds are uttered one at a time in a continuous stream. Ideas, on the other hand, are hierarchically structured. They are grouped topically, and some thoughts can be broken down into lots of sub-thoughts or concepts. Not all thoughts are related to one-another `head-to-tail', conceptually speaking, the way linguistic units are in the speech stream. One problem that the grammar of any language must deal with is how to represent such hierarchically structured ideas by means of a linear speech stream. Often morphosyntactic devices sprinkled in the speech stream help speaker's express and hearer's recover the hierarchical nature of discourse. For example, indentation is a way of indicating certain 'high-level' boundaries in written text, as are chapter headings, section headings, etc. These are formal devices (as formal as any morphological device in spoken language) present in the morphology of the text that delimit the hierarchical structure of the text. Natural spoken texts also exhibit morphological signals of hierarchical structure usually in the form of 'pesky particles' and special intonational patterns. \ftx \txt To continue this discussion, highlight and jump to the topic below: \cf Hdisc_cnt_thm_infr \ftx \dt 27/Apr/1998 \key Hdisc_cnt_thm_ex \shd Thematic continuity--Examples \txt In the following brief text, there are no overt markers of the semantic relation that holds between the two parts: \ftx \fln (3) I'm hungry. Let's go to the Fuji gardens. \ftx \txt However, most English speakers readily identify the first part as representing a 'problem' for which the second part is a 'solution' (Mann and Thompson 1987). How do we know this? We infer it from our understanding of the propositional content of each part. This inference is similar to the inferential process that takes place in languages (such as Sierra Popoluca) that rely heavily on pragmatics in order to distinguish grammatical relations (see Hgrel_intro). Longer texts can be assigned hierarchical structures based on just this kind of inferences. For example, Mann and Thompson (1987) assign the following structure to a brief newspaper article: \ftx \fln 1. Farmington police had to Volitional \ftx help control traffic recently, result \ftx \fln 2. when hundreds of people \ftx lined up to be among the first Background \ftx applying for jobs at the \ftx yet-to-open Marriot Hotel. \ftx \fln 3. The hotel's help wanted \ftx announcement - for 300 Circumstance \ftx openings - was a rare \ftx oportunity for many unemployed. \ftx \fln 4. The people waiting in line \ftx carried a message, a refutation, \ftx of claims that the jobless could \ftx be employed if only they showed \ftx enough moxie. \ftx \fln 5. Every rule has exceptions, Concession \ftx \fln 6. but the tragic and too-common \ftx tableaux of hundreds or even \ftx thousands of people snake-lining Evidence \ftx up for any task with a paycheck \ftx illustrates a lack of jobs, \ftx \fln 7. not laziness. Antithesis \ftx \txt To continue this discussion, highlight and jump to the topic below: \cf Hdisc_cnt_thm_strc \ftx \dt 27/Apr/1998 \key Hdisc_cnt_thm_infr \shd Thematic continuity--Inference \txt In spite of the fact that every language employs some morphosyntactic devices to convey hierarchical thematic structure, the most common `device' is inference. In a certain sense, all of the information gleaned from a text is inferred. Recall the view of discourse production and comprehension outlined in Qintro_embod_play_A, in which a 'message' is perceived as a mental model constructed cooperatively or competitively by all discourse participants. Participants want their discourse to succeed (i.e. they want communication to take place, which means they want the mental models entertained by the various participants to be substantially similar), and so they use every resource at their disposal to make their version of the model transparent, and to construct a coherent version of what the other participants are trying to say. The morphosyntactic structures of a shared language constitute one set of tools for accomplishing these tasks. However, assumptions regarding context, attitudes, perceptions and knowledge of the interlocutors also help the process along. Inference is the process of reconstructing the mental model intended by another discourse participant using all tools available, including but not limited to linguistic structure. The message is not the words; the words (and other linguistic units) are merely tools that aid in constructing and deconstructing mental models. \ftx \txt When we say that thematic structure is normally inferred, however, we are not using 'infer' in this general sense that all information in discourse is inferred. Rather, we mean that thematic structure is often not coded by overt morphosyntactic cues. Hearers use their knowledge of the propositional content of the text alone to infer the thematic structure. \ftx \fln To continue this discussion, highlight and jump to the topic below: \cf Hdisc_cnt_thm_ex \ftx \dt 27/Apr/1998 \key Hdisc_cnt_thm_strc \shd Thematic continuity--Structural frameworks \txt There are several frameworks within which the thematic structure of the message content of a text can be diagrammed including: RHETORICAL STRUCTURE THEORY (Mann and Thompson 1985, 1987), LOCATIONAL STRUCTURE (T. Payne 1985), STORY GRAMMAR (Rummelhart (1975). Individual languages, genres or particular texts may be more or less amenable to one or another of these frameworks. \ftx \txt Of course, this work would be considered discourse interpretation rather than analysis given the characterization provided in Hdisc_cnt. The interpretive nature of applying hierarchical thematic structures to texts is confirmed by the fact that different observers will apply different hierarchical structures to the same text. However, there is definitely a place for interpretation in discourse analysis. For instance, after having diagrammed a text, the fieldworker might look at the distribution of some pesky particle or other morphosyntactic device whose meaning has not been adequately identified. In many cases, one's understanding of the functions of that device will be enhanced greatly, if not nailed down definitively, by the exercise of interpretation. \ftx \txt In this section of the descriptive grammar, the fieldworker should describe some semantic principles on which texts are hierarchically structured in the language, and any morphosyntactic devices that contribute to the interpretation of that hierarchical structure. Finally, one may want to diagram the thematic structure of a few short texts of various genres, and give clear, non-circular answers to the following question: How is the thematic structure of this text encoded in the morphosyntax, if at all? Be especially careful in this section to support all assertions with concrete examples. \ftx \dt 27/Apr/1998 \key Hdisc_genr \shd Genres \txt The following sections constitute a possible list of genres, with well-known examples from the English tradition, and/or questions the fieldworker might ask to elicit texts of various genres. Beware, however, of texts elicited in this manner. It is always better to record a text in its natural setting, e.g. when a father actually is exhorting his son prior to marriage, rather than in a hypothetical context. However, such opportunities are frustratingly rare. \ftx \ftx This topic contains the following subsections: \cf Hdisc_genr_cnvs Conversation \cf Hdisc_genr_nar Narrative \cf Hdisc_genr_hort Hortatory \cf Hdisc_genr_proc Procedural \cf Hdisc_genr_exp Expository \cf Hdisc_genr_desc Descriptive \cf Hdisc_genr_rit Ritual Speech \ftx \dt 27/Apr/1998 \key Hdisc_genr_cnvs \shd Conversation \txt Conversation is probably the universal default discourse type. However, conversation does not easily qualify as a 'genre' in that there is no consistent and obvious 'organizational parameter' that structures it. Rather, any and all the organizational parameters that define the other genres are used cooperatively or competitively by the various participants in a conversation as each sees fit. In fact, most discourse employs a combination of organizational principles. Conversation, however, takes this truism to an extreme. \ftx \txt The most obvious structural feature of conversation is the turn. A turn is a contiguous portion of a conversation in which one participant speaks. A coherent conversation consists of a series of turns occupied by the various conversational participants. Communities (if not languages) typically employ various morphosyntactic and gestural devices to initiate, nurture, yield and hold a turn. \ftx \txt Turn initiating devices are signals used to indicate that a participant has a contribution to make. Another way of stating this is that the participant 'wants the floor' (Sacks, Schegloff and Jefferson 1974). In English such signals include inhaling audibly, raising the eyebrows, and using interjections such as "but . . .", "well . . ." etc. \ftx \txt Turn nurturing devices are used by conversational participants who do not have the floor to encourage those who do have the floor to continue. In English such devices include the ubiquitous "Uh huh . ." as well as less stylized expressions such as "Really?", "Cool", or just an interested look. In many speech communities nurturing devices appear to be more central to all kinds of discourse than they are in European communities. For example, in various indigenous languages of the Americas, it is common for an interlocutor to repeat the entire expression, or portion of the contribution of another to encourage the other to continue. For example, the following was recorded in Panare. A and B represent the two participants in the exchange: \ftx \fln (11) A: He was picking fruit. \ftx B: Ummmmmm \ftx A: Big fruit. \ftx B: Big fruit. \ftx A: Bigger than around here. \ftx B: Ummmmmm \ftx A: Like this big. \ftx B: Big. \ftx \txt McLendon (19##) has even written an article on 'the role of the whatsayer,' where the 'whatsayer' is a designated member of the audience who has a crucial role in the development of traditional narrative. Panare consultants often had difficulty recounting stories unless there was another Panare speaker who could provide encouragement and ask appropriate questions to keep the turn-taking rythm of the narrative alive. \ftx \txt Turn yielding devices include special intonation patterns and even grammatical particles. These are used to signal that a participant is finished with a particular contribution, and that the floor is open for others. Question intonation is often used to elicit a response from an interlocutor, even if the clause is not an actual question. \ftx \txt Finally, floor holding devices indicate that a speaker is not finished with his or her contribution. Often speakers need a chance to formulate their thoughts, but don't want to `give up the floor' while they are thinking, so they use 'fillers' or 'hesitation particles' to occupy their turn while they are pondering the rest of their contribution. In English such particles include: er, um, well uh, etc. \ftx \txt By far the majority of linguistic research to date on conversation has been based on English or related languages. Therefore, it is not yet clear whether findings can be at all extended to other languages. There may be communities in which conversation does not primarily consist of alternating turns. However, no clear examples have yet been documented. \ftx \dt 27/Apr/1998 \key Hdisc_genr_desc \shd Descriptive \txt People occasionally want to describe the characteristics of something, someone or some abstract concept. This is another uncommon genre, and hence fieldworkers may have difficulty eliciting clear examples. Some examples of descriptive discourse are: a classified advertisement for a house, the scene-setting section of a novel or short story. \ftx \fln Questions to ask: \ftx \ftx a. 'What is the place you grew up in like?' (or 'what is community X like?' \ftx \ftx b. 'What is your house like?' \ftx \ftx c. 'What is your father/brother like? (Be careful here. Male researchers especially should be careful not to appear nosy with respect to females). \ftx \dt 27/Apr/1998 \key Hdisc_genr_exp \shd Expository \txt EXPOSITORY discourse is an attempt at explaining something. This is another uncommon genre. Attempts to elicit expository texts, especially around topics related to cosmology, are likely to result in folklore or mythology. Expository discourse may be organized according to location, if the subject matter is concrete, or logic, if the subject matter is an abstract or technical concept. Examples of expository discourse include: expository sermons, technical articles/textbooks. \ftx \fln Questions to ask: \ftx \ftx a. 'Why do you hunt when the moon is full?' \ftx \ftx b. 'Which animals do you hunt at night? Why?' \ftx \ftx c. 'Where are your gardens located? Why?' etc. \ftx \ftx d. `What is this thing? What is it for' (Demonstrating some complex \ftx object, idea or organism) \ftx \dt 27/Apr/1998 \key Hdisc_genr_hort \shd Hortatory \txt HORTATORY discourses are attempts on the part of the speaker to get the hearer to do something, or to act in a certain way. Languages differ as to how hortatory discourse is handled. Some (especially in West Africa) have specific 'hortatory' constructions. Others use commands. Still others use first person plural forms. For example a North American parent is likely to be heard saying something like the following to a child: 'we don't throw food at mommy.' Some examples of hortatory discourses are: didactic sermons, and scoldings or parental lectures. \ftx \fln Questions to ask: \ftx \ftx a. 'What would you tell your daughter/son just before marriage?' \ftx \ftx b. 'My kid is doing terribly in school. What should I tell him?' \ftx \dt 27/Apr/1998 \key Hdisc_genr_nar \shd Narrative \txt NARRATIVES are stories. That is, they are texts in which the speaker recounts a set of events in the real world or some imagined world. The events of a narrative are usually (but not necessarily) related to one another according to time, i.e., chronologically prior events are described before other events, etc. The following sections describe some common subtypes of narrative. There may be some types that have not been listed. \ftx \txt To continue this discussion, highlight and jump to the topic below: \cf Hdisc_genr_nar_pers \ftx \dt 27/Apr/1998 \key Hdisc_genr_nar_flk \shd Folk stories \txt Technically, folklore consists of stories about real or imagined ancestors. Folk stories may contain supernatural elements, but are not primarily concerned with explaining natural phenomena. They are the stories that define a community. Counterparts in Western culture would be: Daniel Boone, Davy Crocket, Robin Hood, and the book of Exodus. \ftx \fln Questions to ask: \ftx \ftx a. 'Do you know any stories about the ancestors?' \ftx \ftx b. 'Do you know any stories about animals that talk? \ftx \txt To continue this discussion, highlight and jump to the topic below: \cf Hdisc_genr_nar_myth \ftx \dt 27/Apr/1998 \key Hdisc_genr_nar_hist \shd Historical \fln Some examples of HISTORICAL NARRATIVES are: \ftx \txt a. War stories. Most areas of the world have experienced significant political conflicts within the lifetime of living individuals. These are especially rich sources of personal experience and historical narratives. Stories that 'go the rounds' may have become polished and stylized. Such stories may provide valuable insights into the characteristics of planned speech - the precursor to a written tradition. However, these stories are not likely to reflect everyday narrative style. \ftx \txt b. 'What was life like under the colonial government?' This is especially relevant for Africa and insular Asia. However, the results may be politically sensitive, especially if the person says that life was better under the colonial system. \ftx \txt c. 'How was this community founded?' \ftx \txt To continue this discussion, highlight and jump to the topic below: \cf Hdisc_genr_nar_flk \ftx \dt 27/Apr/1998 \key Hdisc_genr_nar_myth \shd Mythology \txt This genre may merge with folklore. In some communities there is a recognizable distinction. In such systems mythology would consist of stories that rely heavily on the supernatural and which typically deal with explanations for the current state of the world. Folklore, on the other hand, would consist of tales which rely less on the supernatural, and don't necessarily purport to explain anything about the world. Some examples of mythological narratives are: the Greek myths, Paul Bunyan, and the first eight chapters of Genesis. \ftx \fln Questions to ask: \ftx \ftx a. 'Was there ever a time when animals could talk?' \ftx \ftx b. 'What is the origin of X? (where X is a culturally significant plant, \ftx animal, body-part, geographic landmark or group of people). \ftx \ftx c. 'How did the world begin?' \ftx \ftx d. 'Was the world ever covered with water?' \ftx \dt 27/Apr/1998 \key Hdisc_genr_nar_pers \shd Personal experience \fln Some examples of PERSONAL EXPERIENCE NARRATIVES are: \ftx \ftx a. How I spent my summer vacation. \ftx \ftx b. What happened on my hunting trip. \ftx \fln Ways to elicit personal experience narratives: \ftx \ftx a. 'Did you ever have an experience where you almost died?' \ftx \ftx b. 'Tell me about your trip to . . .' \ftx \ftx c. Take advantage of significant events in the community, e.g. 'Tell me \txt about the fire at Vicente's house,' or 'where were you when the lights went out/earthquake struck/hurricane hit?', 'what did you do for Carnaval?' etc. Try to be as specific as possible, and to focus on activities that are especially important to the consultant. For example, 'You killed eleven monkeys? That's amazing. How did it happen?' is better than 'Tell me about what you did yesterday.' \ftx \txt To continue this discussion, highlight and jump to the topic below: \cf Hdisc_genr_nar_hist \ftx \dt 27/Apr/1998 \key Hdisc_genr_proc \shd Procedural \txt PROCEDURAL discourses are instructions on how to do something. This is seldom a natural genre. Beware of elicited procedural discourses. Attempts to elicit procedural discourses are likely to result in hortatory speech. Procedural discourse, like narrative, is usually organized according to time. The foregrounded portions of a procedural text are the clauses that refer to the 'steps' in the procedure. \ftx \txt Examples of procedural discourse include: recipes, instructions on how to assemble a swing set. \ftx \fln Questions to ask: \ftx \ftx a. 'How do you make a blowgun?' \ftx \ftx b. 'Great meal! How did you cook it?' \ftx \dt 27/Apr/1998 \key Hdisc_genr_rit \shd Ritual speech \txt RITUAL SPEECH consists of prescribed discourse types used in religious or other ceremonial contexts. This is a very common genre, but can be difficult to elicit. Some examples of ritual speech are: prayers, religious liturgy such as might be heard at weddings, funerals, coming of age celebrations, healing rituals and rituals employed in conflict resolution. In the Philippines and Indonesia, as well as other areas of the world, poetry and song are commonly employed as means of resolving local conflicts. \ftx \dt 27/Apr/1998 \key Hdisc_idm \shd Idiomatic expressions / proverbs \txt IDIOMATIC EXPRESSIONS are turns of phrase which mean more than the actual words contained in the expression would lead one to expect. There may be a fine line between idiomatic expressions and proverbs. Some languages (especially in Africa) place great importance on proverbs. People use them much more readily than we are used to in the English tradition. For this reason they may function almost as idiomatic expressions. Some common English idiomatic expressions are: 'to get dolled up', 'to fathom' ('I can't fathom that'), etc. \ftx \txt PROVERBS: 'look before you leap', 'a stitch in time saves nine', 'the calm before the storm', 'the grass is always greener', 'birds of a feather', 'haste makes waste' etc. \ftx \fln Some examples from Sùpyìré, West Africa: \ftx [incomplete] \ftx \dt 27/Apr/1998 \key Hdisc_intro_intrp \shd Discourse analysis versus interpretation \txt An important distinction to make when considering the scientific study of discourse is one which I will describe as the difference between interpretation and analysis. Much of what passes as 'discourse analysis' should more appropriately be termed 'discourse interpretation.' For example, if I look at a text and divide it up into 'paragraphs' based on my understanding of the propositional content of the text, e.g., when the speaker finishes talking about one thing and begins talking about another, then I am interpreting the text. However, if I look at the same text and divide it up according to the presence of certain particles, referential devices, pauses, and intonational patterns, I am engaged in linguistic analysis of the text. \ftx \txt Interpretation certainly has a role in linguistic analysis, but the two are not the same thing. For example, I may interpret the paragraphing in a text based on the propositional content alone. Then, I may analyze the text according to the morphosyntactic clues sprinkled within it. If I am able to successfully keep my interpretation independent of my analysis, then some meaningful, scientifically valid, generalizations may result, e.g. 'particle X marks paragraph boundaries'. However, if I let interpretation into my analysis, e.g., by defining particle X as a marker of paragraph boundaries, or deciding ahead of time where paragraph boundaries are by looking for particle X, then no meaningful generalization is possible. \ftx \txt The point at which a sociologist who studies discourse becomes a sociolinguist studying language is when he or she tries to make scientifically valid generalizations (predictions, explanations) using empirical data as evidence. This is not to say that there is no place for discourse interpretation. Much wisdom regarding the functions of morphosyntactic devices is based on interpretation, and hypotheses are generated through interpretive judgements. Nevertheless, linguistics as a discipline derives its independence from its distinctively empirical methodologies and perspectives. Sociologists and literary scholars do not need linguists if interpretation is their only concern. \ftx \txt In this sense, the entire manual to this point has been concerned with discourse, as a basic assumption has been that function is distinct from form, but that the functions of morphosyntactic structures affect and to a great extent explain their forms. Analysis of form necessarily involves analysis of function. Put another way, analysis of language necessarily involves analysis of discourse. \ftx \txt In this section, we will first describe some general properties of discourse that tend to be reflected in language. In Hdisc_genr a compendium of discourse genres will be presented. Finally, in Hdisc_idm several topics that a field linguist may wish to include in a concluding section of a grammatical sketch are suggested. \ftx \txt To continue this discussion, highlight and jump to the topics below: \cf Hdisc_cnt Continuity (cohesion) and discontinuity \cf Hdisc_genr Genres \cf Hdisc_idm Idiomatic expressions / proverbs \cf Hdisc_symb [incomplete] \cf Hdisc_typ [incomplete] \cf Hdisc_misc [incomplete] \cf Hdisc_bib Bibliography or references cited \ftx \dt 27/Apr/1998 \key Hdisc_intro_ling \shd Discourse analysis and linguistic analysis \txt Discourse is human interaction. Much of human interaction involves language, therefore the study of discourse often involves the study of language. However, discourse and language are two potentially independent fields of investigation. Because they are independent, each can provide evidence for claims made in the other -- if they were identical, or notational variants of the same phenomenon, then generalizations made in one domain based on evidence from the other would be tautologous. \ftx \txt For example, AGENT is a concept that is useful in human interaction. AGENTs exist quite apart from language (see Hgr_v_cl_sem_com_ag). Subject (as defined in this manual), on the other hand, is a linguistic concept. It does not exist apart from its role as a category in linguistic structures. If AGENT and subject were simply two names for the same concept, generalizations such as `in this sentence the AGENT is the subject', or `AGENT is the primary candidate for subjecthood' would be meaningless. One couldn't meaningfully explain anything about AGENT in terms of subject or vice versa. \ftx \txt To continue this discussion, highlight and jump to the topic below: \cf Hdisc_intro_intrp \ftx \dt 27/Apr/1998 \key Hdisc_misc \shd Miscellaneous and Conclusions \fln Other discourse (disc) sections provide some suggested headings. \ftx \ftx [incomplete] \dt 21/Aug/1997 \key Hdisc_symb \shd Sound symbolism \ftx \ftx [incomplete] \dt 20/Aug/1997 \key Hdisc_typ \shd Typological findings \ftx \ftx [incomplete] \dt 20/Aug/1997 \key Heth \shd Demographic and ethnographic information \ftx \txt This topic contains the following subsections. Highlight and jump to the topic of interest: \shd2 Name of language \cf Heth_nam_ext External \cf Heth_nam_int Internal \cf Heth_nam_orig Origins \shd2 Economic Activity \cf Heth_eth_econ Economic Activity \cf Heth_eth_ecosys Ecosystem \cf Heth_eth_cult Material Culture \cf Heth_eth_cosm Cosmology \shd2 Demography \cf Heth_dem_loc Locale \cf Heth_dem_other Other Groups \cf Heth_dem_intct Interaction \shd2 Genetic Affiliation \cf Heth_affil_fam Family \cf Heth_affil_rel Relatives \cf Heth_affil_src Published Sources \shd2 Sociolinguistic Situation \cf Heth_socl Intro \cf Heth_socl_mult_perc Multilingualism -- Percentages \cf Heth_socl_mult_lang Languages \cf Heth_socl_cntxt Contexts of Use \cf Heth_socl_via_use Viability -- Use by Children \cf Heth_socl_via_pres Pressures \cf Heth_socl_via_semi Semi-speakers \cf Heth_socl_loan Loan Words \shd2 Others \cf Heth_dlect Dialects \cf Heth_prevres Previous Research \dt 12/May/1998 \key Heth_affil_fam \shd Genetic affiliation--Family \txt It is very important to be aware of all work that has been done on a particular language or language family. If at all possible, the fieldworker should get to know personally the prominent scholars in the field. True scholars are always eager to interact with anyone who shows a sincere interest in their work. The fieldworker should become thoroughly familiar with all historical/comparative work done on this language and/or its family. There are few, if any, languages for which no previous work exists. Diachronic and comparative observations will then inform the sketch at every point, and the fieldworker will have a good idea of where his or her own work fits within the general scheme of investigation on this language. However, previous work must be evaluated closely before one assumes that the linguistic work has 'been done'. \ftx \dt 27/Apr/1998 \key Heth_affil_rel \shd Genetic affiliation--Relatives \fln See Help screen under: \cf Heth_affil_fam \ftx \key Heth_affil_src \shd Genetic affiliation--Published sources \txt The following sources will provide a good general introduction to the languages and language families of the world. These should be seen as starting points for detailed and exhaustive research into the specific literature relating to the language being studied. \ftx \fln Languages and language families: \ftx Voegelin and Voegelin 1977. \ftx B. Grimes 1988. \ftx \fln Establishing genetic relationships: \ftx Antilla 1972. \ftx \dt 27/Apr/1998 \key Heth_dem_intct \shd Demography--Interaction \fln See comment under: \cf Heth_dem_loc Demography--Locale \ftx \key Heth_dem_loc \shd Demography--Locale \txt In order to gather this information, it may be necessary to carry out a proper sociolinguistic survey of the area. \ftx \dt 27/Apr/1998 \key Heth_dem_other \shd Demography--Other groups \fln See comment under: \cf Heth_dem_loc Demography--Locale \ftx \key Heth_dlect \shd Dialects \txt Variation is a fact of every language. Variation can be individual, as in variant pronunciations of the word 'economics' or 'either' in English. It can be sociological, e.g. I might pronounce the word 'often' with the t, avoid certain terms or sentence structures (such as stranded prepositions), etc. in certain formal contexts. Occasionally it can be geographical, e.g. in Britain the term 'pants' typically refers to what in America is referred to as 'underpants', whereas the same word 'pants' in America corresponds, in its most common sense, to the British term 'trousers'. \ftx \txt The common socio-linguistic distinction between language and dialect is the following: two speech varieties are said to be dialects of one language if speakers of the two varieties can understand one another immediately, i.e. with no period of familiarization on the part of either interactant. Conversely, two speech varieties are said to be two distinct languages if speakers cannot understand one another (i.e. communication is severely impaired) until they have adjusted their production and comprehension to allow for the variation. As yet there is no standard definition that is more explicit than this (though one could easily imagine a paradigm by which such a definition might be devised -- it simply is not an important enough distinction for linguists to be very concerned about). \ftx \txt There are many reasons other than geographic separation for the rise of speech varieties. However, the term 'dialect of L' usually refers to geographically defined varieties of language L. For the purposes of a grammar sketch, it is helpful to include some brief information on known geographically defined dialects. Nevertheless, the fieldworker should be aware that most variation is not primarily defined geographically. That is, it is usually not the case that a particular variant form is restricted to or characteristic of a particular geographic region, though many are. In support of this claim we need only reflect on the source of variation. People's speech (and behavior in general) tends to become more like that of those people with whom they identify and interact. Even within a geographic region there are typically several sociologically distinct groups (defined by age, occupation, culture, interests, etc.). Members of each group unconsciously (and occasionally consciously) imitate the norm for their particular group and ignore the norms of other groups. This is true of traditional societies, as well as more complex societies. To the extent that such an 'identification group' centers in a geographic area to the exclusion of others, the speech that characterizes a group can be said to be a geographically defined dialect. Geographical variation, therefore, can be considered to be but one kind of sociological variation. \ftx \txt Language variation can occur at any of the traditional levels of linguistic analysis, i.e. phonetics, phonology, morphology, syntax and discourse. If dialects are mentioned in the sketch, it helps to specify at which of these levels the major observed differences lie, and give examples. In any case, the fieldworker should not dwell on this section. Again, there is usually enough complexity in the area of linguistic variation to constitute a monographic study in and of itself. \ftx \fln References: Chambers and Trudgill 1980, Nelson 1983, Simons 1983, \ftx Trudgill 1986. \ftx \dt 27/Apr/1998 \key Heth_eth_cosm \shd Ethnography--Cosmology \fln See comment under: \cf Heth_eth_econ Ethnography--Economic activity \ftx \key Heth_eth_cult \shd Ethnography--Material culture \fln See comment under: \cf Heth_eth_econ Ethnography--Economic activity \ftx \key Heth_eth_econ \shd Ethnography--Economic activity \txt The linguistic fieldworker may be tempted to spend a lot of time describing the material culture and cosmology of the people who speak the language under study. Careful ethnographic notes should be taken throughout a linguist's time on the field, since an essential aspect of getting to know a language is getting to know the people who speak that language. However, the amount of space dedicated to this topic in a grammatical description should be limited. A detailed ethnography is a worthy topic for a future monographic study. \ftx \dt 27/Apr/1998 \key Heth_eth_ecosys \shd Ethnography--Ecosystem \fln See comment under: \cf Heth_eth_econ Ethnography--Economic activity \ftx \key Heth_nam_ext \shd The name of the language--External \txt Give the generally accepted term or terms by which the group is known to the world at large. \ftx \key Heth_nam_int \shd The name of the language--Internal \txt SELF-REFERENT is the anthropological term for the name a group of people uses to refer to themselves. Often this term can only be translated as 'people', or 'human beings'. It may also have hierarchically related meanings. For examples, the word e'ñapa in Panare (a Carib language of Venezuela) means 'person' when used in opposition to the term në'na 'animal'. The same term means 'Indian' when used in opposition to the term tato 'non-Indian'. Finally, the term can also refer strictly to Panare Indians, when used in opposition to terms referring to neighboring Indian groups. Only the context can disambiguate. \ftx \dt 27/Apr/1998 \key Heth_nam_orig \shd The name of the language--Origins \txt The terms by which language groups are known to outsiders are usually drawn from the outsiders' language, and are often derogatory in nature, e.g., in Perú the group now known as the Urarina used to be called the Chimaco, a term from Quechua meaning 'unreliable.' Such terms are often not recognized by the people themselves, and, as in the case with Urarina, the self-referent can sometimes be substituted for the outsiders' term. On the other hand, the term Panare mentioned above is a Tupí word meaning 'friend'. So the outsiders' form of reference isn't always derogatory. If there is a well-established tradition in the literature of using the outsiders' term, the linguistic researcher should not try to change it, unless the people themselves are offended by the general term and clearly would prefer to be known by the self-referent. \ftx \dt 27/Apr/1998 \key Heth_prevres \shd Previous research \txt There may have been anthropological, sociological, political, demographic, health, agricultural, or other studies carried out by other researchers. To find them may take some searching in bibliographies and libraries, but you should make a point of being familiar with these if at all possible. \ftx \dt 27/Apr/1998 \key Heth_socl \shd Sociolinguistic situation-Intro \txt Much of this material may be available in published sources including government reports and censuses. Although this material may be very helpful and save you much work, be careful to evaluate it for accuracy before placing too much reliance on it. \ftx \dt 10/May/1998 \key Heth_socl_cntxt \shd Contexts of use \txt References: Sherzer (1977), Bauman and Sherzer 1974. Besnier 1986, Baugh and Sherzer 1984. \ftx \key Heth_socl_loan \shd Loan words \txt Languages 'borrow' words from other languages under various circumstances. The references provide extensive discussions of the sociological circumstances of linguistic borrowing. Multilingual speakers will, of course, borrow words for the nonce from another language if a native term is not readily available. This practice is termed CODE SWITCHING, and is extremely common in multilingual societies. However, a vocabulary item from another language can be considered a borrowing only if it is recognized by otherwise monolingual speakers in appropriate native contexts. In other words, a word has been borrowed into another language only when it has become part of the lexical system of that language, as recognized by monolingual speakers. For example, the word canoe can be considered an English word, even though it is of Carib origin, because English speakers who have no knowledge of a Caribean language will use the term freely with no sense that it is a "foreign" word. However, if I use a term like samirya, meaning 'OK', in the middle of a discourse that is otherwise in English, this is not a borrowing. It can only be used between people who have some knowledge of Yagua, hence it would more appropriately be classified as code switching. \ftx \fln References: Burling (1970). \ftx \dt 27/Apr/1998 \key Heth_socl_mult_lang \shd Multilingualism--Languages \fln References: Sankoff 1980, Baugh and Sherzer 1984. \ftx \key Heth_socl_mult_perc \shd Multilingualism--Percentage \fln Treat men and women separately. \ftx \key Heth_socl_via_pres \shd Viability--Pressures \txt See Dorian's treatment of East Sutherland Gaelic (1981). There are many other discussions of this topic in the sociolinguistic literature. \ftx \dt 27/Apr/1998 \key Heth_socl_via_semi \shd Viability--Semi-speakers \txt The topic of language death and viability relates to the question of whether someone can be a 'partially competent' native speaker of a language. It is clearly possible to have a native-like knowledge of one part of a language system and be lacking in another part. For example, one can have native-like phonology and syntax, but have holes in vocabulary and imperfect gender and case morphology, or satisfactory phonology and morphology but gaps in the syntax and vocabulary. It also appears that items heard in early childhood can persist in long-term memory and reappear in consciousness only decades later (Wayles Browne, p.c.). Also, there have been cases (none documented that I know of) where individuals appear to lack full flency in any language. For example, among the Yagua people of Northeastern Peru, we found that certain younger women who were partially culturally assimilated did not have full command of Spanish or Yagua. These women stood out in comparison to a) older culturally assimilated women who would speak Yagua fluently and some Spanish, b) non- assimilated women of all ages who spoke only Yagua fluently and c) all men, who would speak Yagua fluently and Spanish to varying degrees of fluency. In all situations that we were able to observe, including conversations with their husbands, these younger, assimilated women would not speak Yagua at all, and would only speak rudimentary Spanish, even though their husbands and others would address them in Yagua. It is hard to imagine that these women were fluent but 'latent' speakers of Yagua. \ftx \dt 27/Apr/1998 \key Heth_socl_via_use \shd Viability--Use by children \txt Here are some impressionistic rules of thumb for evaluating the viability of a language that may be on the verge of extinction. These should not be considered definitive by any means, since a language's viability may be affected by any number of extralinguistic factors. On the negative side these include imposed and chosen cultural assimilation, disease, genocide and insensitive government policies. On the positive side literacy campaigns, nationalistic movements within the group, and progressive government policies may improve a language's viability. With these qualifications in mind, here are some rules of thumb: \ftx \ftx 1) If there are no, or extremely few children under the age of \ftx ten who are learning the language as their only language, \ftx the language will become extinct in their lifetime \ftx (i.e. 60 - 70 years). \ftx \ftx 2) If there is more than a handful of 10 year old children who are \ftx monolingual in the language, and who have regular contact with \ftx each other (i.e. they live in the same community), the language \ftx will be taught to the next generation. This means that in \ftx 60 years the language will still be used as a regular means of \ftx everyday conversation in some communities. The viability of the \ftx language may still improve or deteriorate depending on \ftx sociological and other factors. \ftx \ftx 3) If many children are learning the language monolingually and \ftx essential economic activity (e.g., buying, selling and/or \ftx distributing of essential goods) is conducted in the language, \ftx extinction is not imminent -- the language could persist \ftx indefinitely. \ftx \dt 27/Apr/1998 \key Hgr \shd Grammatical Categories \ftx \txt This topic contains the following subsections. Highlight and jump to the topic of interest: \cf Hgr_cat Grammatical categories \cf Hgr_cat_min **incomplete \shd2 Adverbs \cf Hgr_adv_gen General \cf Hgr_adv_vscpl vs. Complement-taking Verbs \cf Hgr_adv_man Manner \cf Hgr_adv_time Time \cf Hgr_adv_dir/loc Direction/Location \cf Hgr_adv_ev/epis Evidential/Epistemic \shd2 Descriptive Adjectives \cf Hgr_mod_adj_form Form \cf Hgr_mod_adj_agr **Agreement \cf Hgr_mod_adj_sem Semantic Classes \cf Hgr_mod_adj_size **Size \cf Hgr_mod_adj_shp **Shape \cf Hgr_mod_adj_col **Color \cf Hgr_mod_adj_val **Value \cf Hgr_mod_adj_txt **Texture \cf Hgr_mod_adj_func **Functionality \cf Hgr_mod_adj_other **Other \shd2 Numerals \cf Hgr_mod_non-num **Non-numeral Quantifiers \cf Hgr_mod_num_sys System Used -- Numerals \cf Hgr_mod_num_ext Extent \cf Hgr_mod_num_agr **Agreement \shd2 Nouns \cf Hgr_n_gen General discussion \cf Hgr_n_msyn Morphosyntactic Criteria \cf Hgr_n_dist Distributional Characteristics \cf Hgr_n_strc Structural Characteristics \cf Hgr_n_ex Changing Grammatical Category \cf Hgr_n_ex_other Other structural characteristics of the example \cf Hgr_n_ex_sol Solutions to the example \cf Hgr_n_ex_sol_rec Recommendation \cf Hgr_n_ex_sol_1 Discussion of Solution 1 \cf Hgr_n_ex_sol_2 Discussion of Solution 2 \cf Hgr_n_ex_sol_3 Discussion of Solution 3 \cf Hgr_n_typ Types of Nouns \cf Hgr_n_typ_count **Count nouns \cf Hgr_n_typ_mass **Mass nouns \cf Hgr_n_typ_other Other subcategories of noun. \cf Hgr_n_typ_prop Proper names \cf Hgr_n_n/npstr The Structure of the Noun or Noun Phrase \shd2 Pronouns and/or Anaphoric Clitics \cf Hgr_n_prn/clt_intro General Discussion \cf Hgr_n_prn/clt_dist Relevant Distinctions \cf Hgr_n_prn/clt_dist_pers Person \cf Hgr_n_prn/clt_dist_num Number \cf Hgr_n_prn/clt_dist_gen **Gender \cf Hgr_n_prn/clt_dist_grel **Grammmatical Relations \cf Hgr_n_prn/clt_dist_role Semantic Roles \cf Hgr_n_prn/clt_dist_def **Definiteness/Specificity \cf Hgr_n_prn/clt_dist_hon Honorifics \shd2 Verbs \cf Hgr_v_cl General Discussion \cf Hgr_v_cl_sem Semantic Roles \cf Hgr_v_cl_sem_com common Semantic roles \cf Hgr_v_cl_sem_com_ag Agent \cf Hgr_v_cl_sem_com_exp Experiencer \cf Hgr_v_cl_sem_com_force Force \cf Hgr_v_cl_sem_com_gua Agent in Guaymí \cf Hgr_v_cl_sem_com_instr Instrument \cf Hgr_v_cl_sem_com_pat Patient \cf Hgr_v_cl_sem_com_recip Recipient \cf Hgr_v_cl_sem_encod Encoding Semantic Roles \cf Hgr_v_cl_sem_univ Universal semantic roles \cf Hgr_v_cl_sem_vsrel Semantic roles versus grammatical relations \cf Hgr_v_cl_weath Weather Verbs \cf Hgr_v_cl_st States \cf Hgr_v_cl__inv Involuntary Processes \cf Hgr_v_cl_bod Bodily Functions \cf Hgr_v_cl_bod_loc Locomotion \cf Hgr_v_cl_bod_mot Motion \cf Hgr_v_cl_bod_mot_op Motion--verbal operators \cf Hgr_v_cl_pos Position \cf Hgr_v_cl_actn Actions \cf Hgr_v_cl_actproc Action Processes \cf Hgr_v_cl_fact Factives \cf Hgr_v_cl_cogn Cognition \cf Hgr_v_cl_sens Sensation \cf Hgr_v_cl_emot Emotion \cf Hgr_v_cl_utt Utterance \cf Hgr_v_cl_manip Manipulation \cf Hgr_v_cl_ex1 Verb classes-Examples from Guaymí \cf Hgr_v_cl_ex2 Verb classes-Examples from Gujarati \cf Hgr_v_cl_ex3 Verb classes-Examples from Tagalog \dt 12/May/1998 \key Hgrel \shd Grammatical relations \ftx \txt This topic contains the following subsections. Highlight and jump to the topic of interest: \ftx \shd2 Introduction \cf Hgrel_intro Grammatical relations--Introduction \cf Hgrel_intro_prag pragmatic status \cf Hgrel_intro_func functional approach \cf Hgrel_intro_lprot less prototypical cases \cf Hgrel_intro_core core categories \cf Hgrel_sum Summary \shd2 Syntactic ergativity \cf Hgrel_erg Syntactic ergativity \cf Hgrel_erg_cnj conjunction reduction \cf Hgrel_erg_1 the process I \cf Hgrel_erg_2 the process II \cf Hgrel_erg_ex1 Example I \cf Hgrel_erg_ex2 Example II \cf Hgrel_erg_sum Summary \shd2 Split systems \cf Hgrel_splt Split systems \cf Hgrel_splt_intr intransitive \cf Hgrel_splt_intr_ex1 intransitive examples I \cf Hgrel_splt_intr_ex2 intransitive examples II \cf Hgrel_splt_intr_ex3 intransitive examples III \cf Hgrel_splt_intr_fluid intransitive--Fluid-S languages \cf Hgrel_splt_intr_other intransitive--other possibilities \cf Hgrel_splt_tr_agr transitive--agreement and pronouns \cf Hgrel_splt_tr_ag/emp transitive--agentive hierarchy \cf Hgrel_splt_tr_a/p transitive--agent/patient distinction \cf Hgrel_splt_tr_ex1 transitive--examples I \cf Hgrel_splt_tr_ex2 transitive--examples II \cf Hgrel_splt_tr_ex3 transitive--examples IV \cf Hgrel_splt_tr_ex4 transitive--examples VI \cf Hgrel_splt_tr_np noun phrases in transitive clauses \cf Hgrel_splt_tr_sol transitive--solutions \cf Hgrel_splt_tr_sum transitive--summary \cf Hgrel_splt_tr_top transitive--topicality \cf Hgrel_splt_t/a tense/aspect \cf Hgrel_splt_t/a_expl tense/aspect--Explanation \cf Hgrel_splt_t/a_ex1 tense/aspect--Examples I \cf Hgrel_splt_t/a_ex2 tense/aspect--Examples II \cf Hgrel_splt_t/a_ex3 tense/aspect--Examples III \cf Hgrel_splt_sum Summary \shd2 System Type \cf Hgrel_typ_erg ergativity characterized \cf Hgrel_typ_erg_eng ergativity in English \cf Hgrel_typ_erg/abs ergative/absolutive \cf Hgrel_typ_ex1 examples I \cf Hgrel_typ_ex2 examples II \cf Hgrel_typ_ex3 examples III \cf Hgrel_typ_nom/acc nominative/accusative \cf Hgrel_typ_ord constituent order \cf Hgrel_typ_ord_other other possibilities \cf Hgrel_typ_rat1 rationale-I \cf Hgrel_typ_rat2 rationale II \shd2 Other topics \cf Hgrel_case Case marking \cf Hgrel_ag/emp **Agentivity/Empathy Hierarchies \cf Hgrel_ord **Constituent order \cf Hgrel_vcode **Verb coding \cf Hgrel_val Valency \cf Hgrel_val_enc Encoding grammatical relations \dt 14/May/1998 \key Hgrel_ag/emp \shd Agentivity/Empathy Hierarchies \ftx \ftx \key Hgrel_case \shd Case marking \txt Morphological cases are overt noun phrase operators, usually suffixes, that express something about the semantic role, pragmatic status and/or grammatical relation of the noun they are associated with. Typically, morphological cases have been thought of as directly indicating grammatical relations, but this is far from general. Often there are other, more convincing, indications of what a nominal's grammatical relation is. Further, morphological cases are sometimes more sensitive to semantic or pragmatic roles. Finally, many languages simply lack morphological cases. It is certainly not the case that such languages lack grammatical relations. Hn_case discusses the difference between case marking and adpositions. The forms of the cases should have been described in Hn_case. Here the fieldworker should be concerned with case as a system for organizing relations between nouns and verbs in clauses, rather than as simply morphosyntactic categorization of noun phrases. \ftx \txt Often languages will employ case marking for certain kinds of nominal elements but not others. The following is a hierarchy of the kinds of nominal elements likely to formally express case distinctions: \ftx \ftx Personal Demonstrative Full NP Other \ftx Pronouns > Pronouns > Patients > Full NPs \ftx \dt 27/Apr/1998 \key Hgrel_erg \shd Syntactic ergativity \txt A syntactic process is said to operate on an ergative-absolutive basis if in some way it responds to S and P as 'the same' and A differently. For example, as mentioned earlier, noun-verb incorporation in English might be said to operate on an ergative-absolutive basis because S and P can be incorporated but A cannot: \ftx \fln (42) a. bird chirping Incorporation of S \ftx b. fox hunting Incorporation of P \ftx c. *doctor recommending Incorporation of A \ftx \txt This might be considered to be a case of SYNTACTIC ERGATIVITY in English. One could say that noun-verb incorporation in English operates on an ergative-absolutive basis. Of course, noun-verb incorporation is a very marginal syntactic process in English, but there are a few languages in the world for which processes that are more central to the morphosyntax do seem to operate on an ergative-absolutive basis. Not surprisingly, perhaps, these languages also exhibit ergativity in case marking and/or verb coding (see Hgrel_splt_t/a). \ftx \txt To continue this discussion, highlight and jump to the topic below: \cf Hgrel_erg_1 \ftx \dt 27/Apr/1998 \key Hgrel_erg_cnj \shd Syntactic ergativity--conjunction reduction \txt One other syntactic process that has been claimed to be sensitive to the ergative-absolutive system in some languages is 'conjunction reduction'. This is the process whereby one element of a conjoined clause can be omitted when coreferential to an element in the previous clause. For example, in English, an argument of the second of a conjoined pair of clauses can be omitted when it is coreferential with an element of the first clause: \ftx \fln (52) a. John greeted Mary and coughed. \ftx \ftx b. John grabbed Mary and slapped him. \ftx \txt In 52a, we understand that John is the person who coughed. If we want to specify that Mary coughed, we must mention her explicitly: \ftx \fln (53) John greeted Mary and she coughed. \ftx \txt With no further specification, 52a must be interpreted such that 'John', and not 'Mary', coughed. In 52b we again understand that 'John' is the omitted element of the second of the conjoined pair of clauses and 'him' refers to an unnamed third participant. Even though it would be pragmatically quite natural for someone named Mary to slap someone who might grab her, it is extremely difficult to interpret 'Mary' as the subject and 'John' as the object of the second conjoined clause. Since the omitted element, 'John', is the S (52a) or the A (52b) argument, and cannot be the P argument, we can say that conjunction reduction in English is sensitive to the nominative-accusative distinction. \ftx \txt For more examples, jump to: \cf Hgrel_erg_ex2 \ftx \dt 27/Apr/1998 \key Hgrel_erg_ex1 \shd Syntactic ergativity--Example I \txt Tagalog (Philippines) is a language in which the syntactic processes of relativization and floating quantifiers have been claimed to be sensitive to the ergative-absolutive distinction. The following is a brief overview of this argument. \ftx \txt Examples 47 and 48 illustrate simple intransitive and transitive clauses. The prenominal case marker 'ang' marks the absolutive case (S in example 47 and P in example 48), while 'ng' (pronounced nan) marks the ergative case (48): \ftx \fln (47) Lumalapit ang babae. (ABS=S) \ftx INTRNS:came ABS woman \ftx 'The woman came.' \ftx \fln (48) Binasa ng lalaki ang diyaryo. (ABS=P) \ftx TRANS:read ERG man ABS newspaper \ftx 'The man read the newspaper.' \ftx \txt In Tagalog, only absolutive arguments can undergo relativization (see Hcplx_relcl). That is, only an absolutive argument can be the relativized NP within a relative clause. This is illustrated for an intransitive relative clause in 49 and a transitive relative clause in 50. In 49 the head of the RC is 'woman' (as indicated by the cliticized relativizer '-ng'), and this NP is omitted ('gapped') from the relative clause itself (indicated by 0): \ftx \fln (49) Matalino ang babae-ng lumalapit 0. \ftx inteligent ABS woman-REL INTRNS:came \ftx 'The woman who came is intelligent.' \ftx \txt In 50 the head of the RC is 'diyaryo'-- 'newspaper'. This too is then omitted from the relative clause: \ftx \fln (50) Interesante ang diyaryo-ng binasa ng lalaki 0. \ftx interesting ABS newspaper-REL ABS=P:read ERG man \ftx 'The newspaper that a/the man is reading is interesting.' \ftx \txt In example 51 the head of the relative clause is 'lalaki'-- 'man' (as indicated by the cliticized relativizer '-ng'). 'Lalaki' is also missing from the relative clause itself. However, this example is unacceptable because 'lalaki' is the ergative nominal in the relative clause, i.e. he is the one that does the reading: \ftx \fln (51) *Matalino ang lalaki-ng binasa ang diyaryo. \ftx intelligent ABS man-REL ABS=P:read ABS newspaper \ftx \txt This restriction that relativization apply only to absolutive nominals is fairly strict in Tagalog and several other Philippine languages. \ftx \txt It has been claimed that Tagalog illustrates syntactic ergativity in quantifier float, as well as relativization (T. Payne 1982). However, it is not clear that quantifier float is a truly productive syntactic process in Tagalog. It is only allowed by some speakers in a colloquial context. Therefore I now consider it to be a marginal process, like noun-verb incorporation in English. However, to the extent that it exists in Tagalog, quantifier float also is sensitive to an ergative/absolutive distinction. \ftx \txt In addition to these syntactic processes that exhibit an ergative-absolutive distinction, Tagalog possesses at least two major syntactic processes that are sensitive to the nominative-accusative distinction. These are complement argument omission and reflexivization (see chart #7.7 above). As mentioned above, these are processes that appear to be universally sensitive to the nominative-accusative distinction. \ftx \dt 27/Apr/1998 \key Hgrel_erg_ex2 \shd Syntactic ergativity--Example II \txt Even though Tagalog exhibits morphological ergativity, and syntactic ergativity in at least one major domain (relativization), conjunction reduction does not operate on an ergative-absolutive basis in Tagalog. However, in Yup'ik Eskimo, another morphologically ergative language, conjunction reduction, to the extent that it is a syntactic process at all, most certainly does operate on an ergative-absolutive basis. \ftx \fln (54) Tom-am Doris-aq cinga-llru-a tua-llu quyi-llru-u-q. \ftx Tom-ERG Doris-ABS greet-PAST-3SG>3SG then-and cough-PAST-INTR-3SG \ftx 'Tom greeted Doris and (she) coughed.' \ftx \txt In this example the second conjunct is only specified for a third person singular subject. Nevertheless, the only possible understanding is that Doris is the one who coughed. This must be because 'Doris' in the first conjunct is in the absolutive case. Thus we can say that conjunction reduction in Yup'ik is sensitive to the category absolutive. \ftx \txt Thus, a language may exhibit an ergative system for organizing grammatical relations in its morphology (case marking and verb coding), in constituent order, or in more complex syntactic processes. To date no languages have been documented which exhibit an ergative system in all aspects of its grammar. In fact, syntactic processes, even in languages which are highly ergative in their morphology, tend to be sensitive to the nominative-accusative distinction. The few counterexamples, e.g. relativization in Tagalog, conjunction reduction in Yup'ik, illustrate that ergativity need not necessarily be merely a morphological phenomenon. \ftx \dt 27/Apr/1998 \key Hgrel_erg_sum \shd Syntactic ergativity--Summary \txt There are certain syntactic processes (such as complement argument omission) that appear to be universally sensitive to the nominative-accusative distinction. Other syntactic processes are sensitive to the ergative-absolutive distinction in some languages, and the nominative-accusative distinction in others. Still other processes that affect subject and object arguments are not sensitive to either of the major distinctions. There are apparently no processes that are universally sensitive to the ergative-absolutive distinction. Chart #7.7 illustrates this pattern: \ftx \txt Chart #7.7: Syntactic processes sensitive to the ergative/absolutive or the nominative/accusative system \ftx \ftx NOM/ACC ERG/ABS None \ftx \txt Complement-argument X Omission \ftx \fln Reflexivization X \ftx \fln Relativization X X X \ftx \fln Floating quantifiers X X X? \ftx \fln Conjunction formation X X X \ftx \dt 27/Apr/1998 \key Hgrel_erg_1 \shd Syntactic ergativity--the process I \txt In order to understand what it means for a syntactic process to respond to S and P as 'the same', it may be helpful to illustrate some processes in some familiar languages that respond to S and A as the same. 'Complement subject omission' is a process that is sensitive to S and A as a category in all languages investigated to date. For example, in English one may say: \ftx \fln (43) a. John wants to leave. INTRANSITIVE COMPLEMENT \ftx \ftx b. John wants to kiss Mary. TRANSITIVE COMPLEMENT \ftx \fln or \ftx \fln (44) a. John wants Mary to leave. INTRANSITIVE COMPLEMENT \ftx \ftx b. John wants Mary to kiss him. TRANSITIVE COMPLEMENT \ftx \txt In 43a the S argument of the intransitive complement clause (see Hcplx_cplcl) 'X leave' is understood to be the same entity as the subject of 'want'. In 43b the A of the transitive complement clause 'X kiss Mary' is also understood to be the same entity as the subject of 'want'. Under these conditions, the S or the A of the complement clause may be omitted in English. In 44a and b, however, the subject of each complement clause is different from the subject of 'want', therefore the subject of the complement clause cannot be omitted. The important fact is that the object of 44b cannot be omitted either: \ftx \fln (45) *Bob wants Mary to kiss. 'Bobi wants Mary to kiss Bobi.' \ftx \txt By allowing omission of S and A arguments of the complement clause when coreferential with an argument (for the verb 'want' it's the subject) of the matrix clause, but not allowing omission of P arguments, this syntactic process treats S and A alike and P differently, and therefore it manifests a nominative-accusative system. \ftx \txt To continue this discussion, highlight and jump to the topic below: \cf Hgrel_erg_2 \ftx \dt 27/Apr/1998 \key Hgrel_erg_2 \shd Syntactic ergativity--the process II \txt Is it possible for a language to have a similar process that is sensitive to the ergative-absolutive distinction? Certainly it is logically possible, but to date no clear examples have been documented. In order to illustrate such a system, we will have to call the process 'complement argument omission' so as not to prejudice the issue towards the familiar subject-based systems. If complement argument omission in English were sensitive to the ergative-absolutive distinction, 46a and b would be grammatical, but 46c would be ungrammatical: \ftx \fln (46) a. Bob wants to leave. 'Bobi wants Bobi to leave.' \ftx \ftx b. Bob wants Mary to kiss. 'Bobi wants Mary to kiss Bobi.' \ftx \ftx c.*Bob wants to kiss Mary. 'Bobi wants Bobi to kiss Mary.' \ftx \txt To see more examples, highlight and jump to: \cf Hgrel_erg_ex1 \ftx \dt 27/Apr/1998 \key Hgrel_intro \shd Grammatical relations--Introduction \txt Grammatical relations (GRs) are often thought of as the relations between arguments and predicates in logical form, i.e. subject, direct object, indirect object and oblique. Grammatical relations are the most difficult relations between nouns and verbs to intuitively justify, yet in some ways they are the most useful. For example, it is intuitively obvious why a language should pay attention to the difference between AGENT and PATIENT -- it is an important part of many communication situations to distinguish entities that act from those that are acted upon. So if a language were to express a sentence with the verb meaning 'kill', and two associated noun phrases, say 'John' and 'lion' it would be a strange situation indeed if there were no automated, easy way of communicating which of the two noun phrases refers to the AGENT of killing and which refers to the PATIENT. Similarly, though possibly less obviously, it is important for languages to be able to communicate pragmatic statuses, such as topicality or identifiability (see Hprag_stat). We structure our messages so as to identify important items or activities that we sense the hearer cannot accurately identify on his or her own. However, in the interests of efficiency, we refrain from over specifying elements of our message. In many aspects of life we see this balance between accuracy and efficiency at work (see Zipf 1949, Grice 1975 and other references on pragmatics). Hence it makes sense that languages should have automated, well-oiled systems of expressing the pragmatic statuses of nominals in discourse. \ftx \txt However, no such intuitive or pre-theoretical justification for grammatical relations, as they are commonly thought of among linguists, is forthcoming. Why should a language pay particular attention to relations between arguments and predicates in the logical form of sentences? What use is logical form per se to speaker and hearer in a communication situation? If grammatical relations are seen as means of expressing semantic roles and/or pragmatic statuses (and conceivably other functional roles as yet undiscovered), then their existence is validated in terms of the function of language as a device for human communication. But as simply labels on nodes in the logical form of sentences, they have no obvious communicational value. \ftx \txt To continue this discussion, highlight and jump to the topic below: \cf Hgrel_intro_prag \ftx \dt 27/Apr/1998 \key Hgrel_intro_core \shd Grammatical relations--core categories \txt It is an empirical observation that languages tend to have about three distinct core grammatical relation categories (subject, object, and indirect object). [NOTE: Oblique is also a grammatical relation, but it is not considered to be a core grammatical relation. Oblique arguments can be thought of as those that are objects of prepositions in English. They are optional sentence adjuncts, akin to adverbial elements (see Hgr_adv_gen and Hcplx_advcl).] This probably reflects human cognitive limitations on keeping track of participant roles in a given situation and/or the number of participant roles necessary to express the kinds of messages human beings normally care to express. In other words there are two, and possibly three role categories necessary to keep participant roles distinct in normal human interaction without overburdening the mind. The actual communicational statuses (i.e. pragmatic and semantic roles) of members of any one of these categories may vary widely. \ftx \txt So the formal category of subject in any given language is like the use of my car's turn signal. It is a formal category. A noun phrase either is a subject or it is not (in languages that have a category of subject). Similarly, I can either use my turn signal or not use it in a given traffic situation. The conjunction of the functional notions AGENT and topic is to the formal category of subject what the conjunction of 'warning cars to the rear' and 'turning' is to the use of my turn signals. Since I only have two choices, operate turn signals or don't operate them, my decisions as to when to operate them are based on my (largely unconscious) assessment of how close a given situation approximates the prototypical situation for the use of the turn signals. Does changing lanes constitute turning? Does a car 300 meters behind need to be warned? Once signalling becomes automated behavior, I might do it even when there are no cars for miles around, or when I just think about turning. On the other hand I may forget or be too busy to use them in some particular prototypical turn signalling situation. The point is that even though there is no direct 100% relationship between turn signalling (form) and any particular traffic situation (function), we cannot deny that turn signals have a function, or pretend we can't figure out what that function is. This is essentially what autonomous syntactic theories do with the relationship between semantic roles and grammatical relations. \ftx \fln References: Comrie 1978, Dixon 1979, Planck 1982. \ftx \dt 27/Apr/1998 \key Hgrel_intro_func \shd Grammatical relations--functional approach \txt From a functional point of view, the obvious though inexact relationship between pragmatic/semantic roles and grammatical relations is motivated in terms of the notion of PROTOTYPE. Noun phrases with certain roles in clause structure are quite likely to be encoded in particular grammatical relations, though there is no necessary requirement. For example a noun phrase that is both a very good semantic AGENT and a very good pragmatic topic is likely to be encoded as a grammatical subject. A functionalist would say that such a noun phrase is a PROTOTYPICAL subject. It is the best example of the kind of noun phrase, in terms of pragmatic/semantic role, likely to be encoded as a subject. Such core elements give rise to the formal category of subject in the first place. This constitutes a very useful and common relational category of noun phrase, and therefore it makes sense that languages should have a highly automated (grammaticalized) way of identifying them. \ftx \txt NOTE: Just as McDonald's has a highly automated way of making cheesburgers with everything. This is a useful category because so many North Americans want cheeseburgers with everything. However, once this is made into an automated category, it drives even more people to choose cheeseburgers with everything because to ask for something slightly different takes longer. So the category, once grammaticalized, becomes a kind of self-fulfilling prophecy where people choose it just because it is a category, and not necessarily because it is the best solution to the gastronomic problem they are facing. \ftx \txt To continue this discussion, highlight and jump to the topic below: \cf Hgrel_intro_lprot \ftx \dt 27/Apr/1998 \key Hgrel_intro_lprot \shd Grammatical relations--less prototypical cases \txt Now, what happens when a noun phrase is a slightly less prototypical AGENT, or a less prototypical topic? As mentioned above, languages tend to have only about three grammatical relations. This indicates that role deviation can be quite significant before a nominal must be excluded from a particular grammatical relation. It would seem unreasonable and inefficient to have a distinct grammatical category for every conceivable nuance in semantic/pragmatic role that might obtain between a noun phrase and a verb. Therefore 'clustering' of pragmatic/semantic roles occurs. Particular nominals that are 'close enough' to the prototype are placed in the same formally instantiated category as more prototypical nominals. Since this notion of 'close enough' is a judgement call on the part of language users, there is variability from language to language (and even, in some languages, from situation to situation) as to how the roles cluster. For example, in the English sentence 'John likes beans' the person who experiences the liking is treated the same grammatically as the AGENT of an agentive verb like 'kill' or 'eat'. In other languages, notably Spanish, the person who likes something is treated as an indirect object: \ftx \fln (2) me gusta la yuca 'I like manioc.' \ftx 1SG:DAT like:3SG ART manioc \ftx \txt In this sentence the subject is 'yuca', as evidenced by 3rd person verb agreement. The 'liker' occurs as a dative adjunct. Hence it appears that in Spanish the experiencer of the sensation of liking is treated as a member of the same grammatical category as RECIPIENTs or PATIENTs of more agentive verbs. Whereas in English the 'liker' clusters grammatically with AGENTs. In summary, grammatical relations are automated (overlearned) formal categories that allow languages to deal with an infinite range of variability in the realm of semantic roles and pragmatic statuses. Note that this is not to say that languages have no way of communicating many degrees of agentivity or topicality. It's just that they do not have automated, well-oiled grammatical means of doing so. \ftx \txt To continue this discussion, highlight and jump to the topic below: \cf Hgrel_intro_core \ftx \dt 27/Apr/1998 \key Hgrel_intro_prag \shd Grammatical relations--pragmatic status \txt There are problems with asserting too quickly that grammatical relations are mere representations of semantic roles and/or pragmatic statuses, and therefore have no independent theoretical validity. For example, as mentioned in Hgr_v_cl_sem_com, there are many (potentially an unlimited number of) semantic roles and pragmatic statuses, whereas there are only a few widely recognized grammatical relations (around three). How can three formal categories represent an infinity of categories in the message world? This question is answered in the common observation that a given grammatical relation in any language typically may encode a variety of semantic roles. For many linguists (e.g. Rosen 1983), this fact is prima-facie evidence that grammatical relations cannot derive from semantic roles. Therefore grammatical relations must have independent status (be 'autonomous') in universal grammar. \ftx \txt Attempts to derive grammatical relations from pragmatic statuses are similarly unproductive. For example, some linguists have proposed that the 'subject' category in language is the linguistic manifestation of a pragmatic status such as 'topic'. Now topic is a term that is even more difficult to independently define than is AGENT (see Hprag_intro_top). However, most traditional definitions assume some form of the statement 'the topic is what the sentence is about.' If we try to apply this definition to real data, we run into similar problems as above. For example in the following sentences it is hard to imagine a way of identifying 'I' as 'the noun phrase the sentence is about' independently of its status as grammatical subject: \ftx \fln (1) a. I just married the most beautiful girl in the world. \ftx b. Now BEANS I like. \ftx c. As for democracy, you'll agree, I think, it's not all it's cracked \ftx up to be. \ftx \txt These sorts of examples make it clear that there is no direct 'mapping' or 'derivational' relationship from the intuitively important notions of semantic role or pragmatic status to grammatical relations. Nevertheless, languages do seem to have grammatical relations. These notions have proven useful to linguists for centuries, even though there has been much debate and little agreement as to why they should exist. They seem natural because languages do have them, but their functional status in language as a system of communication has been difficult to explain satisfactorily. \ftx \txt NOTE: that this problem has been the impetus behind all autonomous theories of syntax. Such theories simply beg the question of what the purpose of grammatical relations is by positing a level of syntactic structure that is not subject to the communicational purposes of language. In this structure, grammatical relations simply exist. They are either primitives of the theory, as in Relational Grammar or Arc-pair Grammar, or notions derived from other purely structural primitives such as linear order and dominance within a phrase-structure tree. This solution to the problem of why languages should have grammatical relations makes syntax much less messy, but is of little value to the linguist who is interested in language as a tool for communication. Autonomous syntactic approaches to grammatical relations are analogous to an 'explanation' of automotive turn signals independent of their function as warning devices, simply because there is no direct mapping between form and function. That is, a) sometimes turn signals are used when there is no car to be warned, and b) sometimes they are not used even when there IS a car to be warned. \ftx \txt To continue this discussion, highlight and jump to the topic below: \cf Hgrel_intro_func \ftx \dt 27/Apr/1998 \key Hgrel_ord \shd Constituent order \ftx \ftx \key Hgrel_splt \shd Split systems \txt Languages may manifest both a nominative-accusative and an ergative-absolutive system within a single area of the morphosyntax, or in different areas. Such systems are called SPLIT-ERGATIVE systems (Silverstein 1976, DeLancey 1982, Merlan 1985 and the references cited above). In most split-ergative systems the distinction between the nominative-accusative subsystem and the ergative-absolutive subsystem is related either to intransitive verbal semantics (Hgrel_splt_intr), relative semantic-pragmatic characteristics of the noun phrases in the transitive clause (Hgrel_splt_tr_a/p), or to the tense and/or aspect of the clause (Hgrel_splt_t/a). \ftx \dt 27/Apr/1998 \key Hgrel_splt_intr \shd Split systems--intransitive \txt It has often been observed that there are two kinds of intransitive clauses; those in which the subject acts with volition and control and those in which the subject exhibits no volition or control: \ftx \fln (12) a. Agentive, active, volitional 'John left.' \ftx b. Non-agentive, stative, non-volitional 'John died.' \ftx \txt Some languages treat these two types of intransitive clause differently, either in case marking or verb agreement. Subjects of Agentive verbs take A-like morphology, and subjects of non-agentive verbs take P-like morphology. Sometimes such languages are termed stative-active, active, or split-S languages. \ftx \txt To illustrate this phenomenon, we will imagine that English exhibited a Split-S system. Subjects of active, volitional intransitive verbs would take the pronouns common to subjects of transitive verbs (13a), whereas subjects of stative, non-volitional intransitive verbs would take objective pronouns (13b): \ftx \fln (13) a. He went. \ftx \ftx b. Him died. \ftx \txt To continue this discussion, highlight and jump to the topic below: \cf Hgrel_splt_intr_ex1 \ftx \dt 27/Apr/1998 \key Hgrel_splt_intr_ex1 \shd Split systems--intransitive examples I \txt The following illustrates very similar examples from Lakhota (a Siouan language of the upper Midwestern United States): \ftx \fln (14) a-ma-ya-phe 'you hit me.' \ftx DIR-1SG-2SG-hit \ftx \txt Example 14 illustrates that the prefix 'ma-' refers to the first person singular P argument of a transitive clause. Intransitive verbs that do not involve volition on the part of the subject, such as 'to fall', 'to die' and 'to shiver' also take 'ma-' to refer to first person subjects: \ftx \fln (15) a. ma-hîxpaye 'I fell' \ftx 1SG-fall \ftx \ftx b. ma-t'e' 'I died' \ftx 1SG-die \ftx \ftx c. ma-c'âca 'I shivered' \ftx 1SG-shiver \ftx \txt Intransitive verbs that are normally carried out volitionally, e.g. 'to play', 'to swim' and 'to sing', take a different prefix, 'wa-', for first person subjects: \ftx \fln (16) a. wa-skate 'I played' \ftx 1SG-play \ftx \ftx b. wa-nûwe 'I swam' \ftx 1SG-swim \ftx \ftx c. wa-lowâ 'I sang' \ftx 1SG-sing \ftx \fln Thus Lakhota exhibits a split-S system of verb agreement. \ftx \txt To continue this discussion, highlight and jump to the topic below: \cf Hgrel_splt_intr_ex2 \ftx \dt 27/Apr/1998 \key Hgrel_splt_intr_ex2 \shd Split systems--intransitive examples II \txt Guaymí, a language of Costa Rica and Panama, illustrates a complex split-S system of case marking of free noun phrases. Example 17 illustrates a transitive clause in which the A is marked with the suffix '-gwe' and the P is left unmarked: \ftx \fln (17) Toma-gwe Dori dêma-ini. 'Tom greeted Doris.' \ftx Tom-ERG Doris greet-PAST:1 \ftx \txt Subjects of intransitive verbs that involve volition can also take this 'gwe-' suffix: \ftx \fln (18) Dori-gwe blit-ani. 'Doris spoke.' \ftx Doris-ERG speak-PAST:1 \ftx \txt Subjects of verbs that do not involve volition may not take the 'gwe-' suffix: \ftx \fln (19) Nu nat-ani. 'The dog died.' \ftx dog:ABS die-PAST:1 \ftx \ftx (*Nu-gwe nat-ani) \ftx \txt We will see in Hgrel_splt_t/a (tense/aspect) that the split-S system in Guaymí only operates in the perfective aspect. Nevertheless, these examples do show that a split-S system for organizing grammatical relations can be manifested in nominal case marking as well as in verb agreement. \ftx \txt To continue this discussion, highlight and jump to the topic below: \cf Hgrel_splt_intr_other \ftx \dt 27/Apr/1998 \key Hgrel_splt_intr_ex3 \shd Split systems--intransitive examples III \txt In Chickasaw, A Muskogean language of the Southeastern United States, the subject of some verbs can be expressed in any of three ways, depending on the semantics. For example, the root 'chokma'-- 'good', can be inflected for A when the subject acts volitionally (21a), for P when goodness is a property exhibited by the subject (21b) and for a Dative participant when goodness is a feeling experienced by the subject (21c): \ftx \fln (21) a. Chokma-li. 'I act good.' (volitional) \ftx good -1SGA \ftx S \ftx \ftx b. Sa-chokma. 'I am good.' (non-volitional) \ftx 1SGP-good \ftx S \ftx \ftx c. An-chokma. 'I feel good.' (experiential) \ftx 1SGD-good \ftx S \ftx \dt 27/Apr/1998 \key Hgrel_splt_intr_fluid \shd Split systems--intransitive--Fluid-S languages \txt Fluid S languages exhibit a variation of split ergativity in that intransitive verbs treated as agentive plus transitive verbs constitute a nominative/accusative system, while intransitive verbs treated as non-agentive plus transitive verbs constitute an ergative/absolutive system: \ftx \ftx +---------+ \ftx | Sa | \ftx | +-----+ V[intransitive] \ftx | |Sp | \ftx | | | \ftx | A | P | V[transitive] \ftx +---+-----+ \ftx \txt As with Split-S systems, there are no Fluid-S systems that treat agentive subjects of intransitive verbs like P arguments, while non-agentive intransitive subjects are treated like A arguments. That is, there are no systems that exhibit the following sort of pattern: \ftx \fln (22) a. He hit him \ftx b. Him fell. (on purpose) \ftx c. He fell. (accidentally) \ftx \txt As might be expected, Split-S and Fluid-S languages do not constitute two mutually exclusive language types. Typically, a given language will have some intransitive verbs that require agentive subjects, others that require non-agentive subjects and still others that allow either agentive or non- agentive subjects. Merlan (1985) calls this general phenomenon split intransitivity, and proposes several universal statistical tendencies regarding such systems. \ftx \dt 27/Apr/1998 \key Hgrel_splt_intr_other \shd Split systems--intransitive--other possibilities \txt Split-S marking of intransitive verbs results in one kind of split-ergative GR marking system. That is, transitive verbs plus agentive intransitive verbs constitute a nominative/accusative system, whereas transitive verbs plus non-agentive intransitive verbs together constitute an ergative absolutive system. \ftx \ftx +--------+ +--------+ \ftx | Sa | V[agentive] | Sp | V[non-agentive] \ftx | +---| |----+ | \ftx | A | P | V[transitive] | A | P | V[transitive] \ftx +--------+ +--------+ \ftx \txt There are no languages for which subjects of non-agentive intransitive verbs are treated like A arguments while subjects of agentive intransitive verbs are treated like P arguments. This universal makes sense in terms of the identifying function of GR marking systems (see above). Agentive intransitive subjects are functionally more like prototypical A arguments, in that both act with volition and control. Similarly, non-agentive intransitive subjects are functionally more like Prototypical P arguments in that both receive or undergo the action expressed by the verb. It would be dysfunctional for agentive intransitive subjects to share morphology with transitive patients, while non-agentive intransitive subjects shared morphology with transitive agents. \ftx \txt Some languages can treat the intransitive subject of certain verbs as either A or P depending on the semantics desired. For example, the concept expressed in English as 'to fall' can either be conceived as something the subject does or as something that the subject undergoes. In some languages this distinction is grammaticalized in the case marking of the subject nominal. In Bats (a Caucasian language, Comrie 1981:53) the subject of the verb 'qitra'-- 'fall' appears in the absolutive case (the case of transitive objects) when the action is non-volitional (ex. 20a). It appears in the ergative case (the case of transitive subjects) to express that the action is volitional: \ftx \fln (20) a. TXo naizdraX qitra. \ftx we:abs to:ground fell \ftx 'We (unintentionally) fell to the ground.' \ftx \ftx b. AtXo naizdraX qitra. \ftx we:erg to:gound fell \ftx 'We (intentionally) fell to the ground.' \ftx \txt To continue this discussion, highlight and jump to the topic below: \cf Hgrel_splt_intr_ex3 \ftx \dt 27/Apr/1998 \key Hgrel_splt_sum \shd Split systems--Summary \txt There are three respects in which languages can exhibit split systems for organizing their grammatical relations. Individual languages may be sensitive to one or a combination of these types of split systems. These systems have described as 'split-intransitivity', 'split-transitivity' and splits based on tense-aspect. \ftx \txt A split intransitive system is one in which the only argument of intransitive clauses can be morphosyntactially coded either as an A argument of a transitive clause or as a P argument of a transitive clause. It is always the case, in such systems, that agentive, active, volitional intransitive subjects take A-like morphology while non-agentive, stative, non-volitional intransitive subjects take P-like morphology. \ftx \txt A split transitive system is one which is based on a hierarchy of animacy, empathy, potential of agency etc. of nominal participants in transitive clauses. In such systems, arguments that are higher on this hierarchy will condition the nominative-accusative system and arguments that are lower on the hierarchy will condition the ergative-absolutive system. \ftx \txt Languages may also have split ergativity based on tense/aspect. If so, then a nominative-accusative system will be used in non-past tense and/or non-perfective aspects, while an ergative-absolutive system will be used in past tense and/or perfective aspects. \ftx \dt 27/Apr/1998 \key Hgrel_splt_tr_agr \shd Split systems--transitive--agreement and pronouns \txt Further up the hierarchy the relation between the points on the hierarchy and potential for agentivity is indirect. For example, how are things that happen to be referred to with verb agreement markers more likely to be agentive than things referred to with pronouns? Pronouns than proper names? etc. The answer to these questions resides in the notions of topicality and empathy. The notion of topicality will be elaborated and clarified at several points throughout this manual. For now, let us examine the notion of empathy (Kuno 1976). There is an empathy principle of human communication that can be expressed as follows: \ftx \txt Human beings tend to select as topics entities with whom they empathize, first of all themselves, then the person they are speaking to, then other human beings, then non-human animate beings, and finally the inanimate world. Therefore, morphosyntactic expressions whose function is to refer to topical entities indirectly tend to refer to entities that speakers empathize with. \ftx \dt 27/Apr/1998 \key Hgrel_splt_tr_ag/emp \shd Split systems--transitive--agentive hierarchy \txt The previous examples illustrate that some verbal arguments are inherently more likely to be agentive than others. The best way to describe this fact is in terms of an AGENTIVITY HIERARCHY. The following hierarchy is one that derives from empirical study of a number of languages. The entire hierarchy is not necessarily represented in any given language, but neither is the hierarchy a theoretical model based on a pre-empirical notion of 'agentivity'. Rather, it derives from a survey of languages that rely, at least partially, on pragmatics to distinguish A from P in some subset of their transitive clauses. In all such languages, the argument further to the left on the hierarchy will automatically be interpreted as AGENT, unless some specific marker signifies otherwise. What is crucial about this hierarchy is that the arrows always (in languages studied to date) go from left to right. That is, there is no language for which an argument lower on the hierarchy will automatically be interpreted as acting upon an argument higher on the hierarchy. Morphological or grammatical devices must always be employed to express such a situation: \ftx \ftx 1st/2nd \ftx Agr > pronouns > 3rd PROs > proper-names > human > animate > inanimate \ftx (Automatically definite) > (definite or indefinite) \ftx \ftx Typical As are: Typical Ps are: \ftx Animate Inanimate \ftx Definite Indefinite \ftx Thematic (TOPIC) Rhematic \ftx \txt Sometimes, this hierarchy (and other similar ones) has been called an ANIMACY HIERARCHY. This is not really a very accurate term, as many of its elements have nothing to do with animacy in the usual sense. For example, verb agreement, pronouns and proper names can refer to biologically animate or inanimate entities. Even 'agentivity hierarchy' is not really accurate in so far as items coded by agreement are not necessarily more agentive than items coded by pronouns, etc. A better term would be something like 'propensity to encode an AGENT hierarchy'. However, since this term is so long, and has not been used in the previous literature, we will continue to use the less precise term 'agentivity hierarchy'. The functional reality that this hierarchy reflects is the fact that certain kinds of linguistic expressions are more likely to refer to AGENTs than are others. For example, animate beings (people and animals) are more likely to act volitionally than are inanimate things. \ftx \dt 27/Apr/1998 \key Hgrel_splt_tr_a/p \shd Split systems--transitive--agent/patient distinction \txt Some languages rely strictly on pragmatic inference to distinguish A from P in most circumstances. In order to explain this possibility, I will ask the reader to imagine a language for which neither case marking, verb-coding nor constituent order distinguishes A from P in transitive clauses. In such a language, how could speakers express and hearers understand the crucial information concerning who acts upon whom? The answer to this question stems from the fact that in the vast majority of transitive situations, one of the participants is pragmatically more likely to be the A than the other one. The pragmatically more probable A argument is automatically understood to be the A, unless specific cues to the contrary are provided. \ftx \txt For example, in most situations involving humans and non-humans, the humans are more likely to be the controlling actors than are the non-humans. (In the following examples, we will imagine a verb-initial language that happens to have the same vocabulary as English): \ftx \fln (23) a. Throw John rock \ftx A P \ftx \ftx b. Throw rock John \ftx P A \fln (24) a. Ate John food \ftx A P \ftx \ftx b. Ate food John \ftx P A \ftx \txt If the situation involves 'eating' and one argument is a person and the other argument is food, the chances of food being the AGENT and the person being the PATIENT are nil -- people eat food but food does not eat people. Our pragmatic knowledge of the world is sufficient to allow us to infer which of the arguments is the A and which is the P. Therefore no special marking, either verb-agreement, nominal case marking or constituent order is necessary to express this fact, though many languages mark the distinction anyway. \ftx \dt 27/Apr/1998 \key Hgrel_splt_tr_ex1 \shd Split systems--transitive--examples I \txt Let us make these notions relating to agentivity concrete with some more hypothetical examples, and then some real ones. What about a situation where both arguments of a transitive clause are equally likely to be agentive? Then true ambiguity may result: \ftx \fln (25) a. Killed John Lion \ftx ? ? \ftx \ftx b. Killed Lion John \ftx ? ? \ftx \txt Even these types of clauses are normally ambiguous only when divorced from their contexts. For example, imagine the sentences in 25 occurring in two stories involving John and a Lion: \ftx \fln (26) Context 1: John lay quietly in the grass as the lion slept. Slowly \ftx he raised his spear. Suddenly, he threw it. It found its mark directly \ftx in the lion's heart. The lion let out a mighty roar, and died. From \ftx then on it was said that: \ftx \ftx Killed lion John / Killed John lion. \ftx \fln (27) Context 2: The lion lay quietly in the grass as John slept. Slowly the \ftx lion raised his body and began to approach. When he was within 20 feet, \ftx suddenly he pounced. He grabbed John by the throat with his teeth. John \ftx never knew what hit him. And that's how: \ftx \ftx Killed lion John / Killed John lion. \ftx \txt In both of these situations the context makes it clear which of the two arguments is the A and which is the P. This is what might be called context imparted agentivity. There are really only relatively few potential situations in human communication where determination of A and P cannot be made either by inherent or by context imparted agentivity. Some languages rely on context-imparted agentivity more than others. \ftx \dt 27/Apr/1998 \key Hgrel_splt_tr_ex2 \shd Split systems--transitive--examples II \txt Managalasi, a highland language of Papua New Guinea, employs an ergative/absolutive case-marking system for pronouns and a nominative-accusative system for verb-coding: \ftx \fln (31) a. a va¿-ena 'You will go.' \ftx 2SG go-FUT:2SG \ftx \ftx b. na va¿-ejo 'I will go.' \ftx 1SG go-FUT:1SG \ftx \ftx c. na-ra a an-a¿-ejo 'I will hit you.' \ftx 1SG-ERG 2SG hit-2SG-FUT:1SG \ftx \ftx d. a-ra na an-i¿-ena 'You will hit me.' \ftx 2SG-ERG 1SG hit-1SG-FUT:2SG \ftx \txt Note that '-ra' marks the pronominal A of the transitive clauses in 31c and d while '0' marks the S of the intransitive clauses in 31a and b and the P of the transitive clauses. This grouping of S with P as opposed to A illustrates a classic ergative-absolutive system of pronominal case marking. Verb agreement, however, operates on a nominative-accusative system. '-ena' agrees with second person singular S and A arguments (31a and d), and '-ejo' agrees with first person singular S and A arguments (31b and c). The agreement markers for P arguments are 'a¿' (2SG, ex. 31c) and 'i¿' (1SG, ex. 31d). Thus in verb agreement S and A are treated alike and P is treated differently. \ftx \txt In Dyirbal, an aboriginal language of Australia, first and second person pronouns operate on a nominative/accusative basis (32): \ftx \fln (32) a. ngana-0 banaganyu 'We returned.' \ftx 1pl-NOM returned \ftx S \ftx \ftx b. nyura-0 banaganyu 'You returned.' \ftx 2pl-NOM returned \ftx S \ftx \ftx c. nyura-0 ngana-na buran 'You saw us.' \ftx 2pl-NOM 1pl-ACC saw \ftx A P \ftx \ftx d. ngana-0 nyura-na buran 'We saw you.' \ftx 1pl-NOM 2pl-ACC saw \ftx A P \ftx \txt Third person pronouns and all other noun phrases in Dyirbal, however, operate on an ergative/absolutive basis. The following chart illustrates the relevant markers: \ftx \ftx Chart #7.2: Dyirbal case marking \ftx \ftx | S | A | P | \ftx ----------+-------+---------+------| \ftx 1,2 pro | -0 | -0 | -na | Nominative-accusative \ftx ----------+-------+---------+------+----------------------- \ftx 3 pro | -0 | -ngu | -0 | Ergative-absolutive \ftx ----------+-------+---------+------| \ftx Names | -0 | -ngu | -0 | \ftx ----------+-------+---------+------| \ftx Comon Np | -0 | -ngu | -0 | \ftx \txt To continue this discussion, highlight and jump to the topic below: \cf Hgrel_splt_tr_ex3 \ftx \dt 27/Apr/1998 \key Hgrel_splt_tr_ex3 \shd Split systems--transitive--examples IV \txt Cashinawa, a Panoan language of Peru, exhibits a nominative/accusative system for case marking of first and second person pronouns, a tripartite system (system V of Chart #7.1 above) for third person pronouns and an ergative absolutive system for all other noun phrases. These data are summarized below: \ftx \ftx Chart #7.3: Cashinawa \ftx | S | A | P | \ftx ---------+------+----------+------+----------------------- \ftx 1,2 pro | -0 | -0 | -a | Nominative-accusative \ftx ---------+------+----------+------+----------------------- \ftx 3 pro | habu | habû | haa | Tripartite \ftx ---------+------+----------+------+----------------------- \ftx full NP | -0 | -~ | -0 | Ergative-absolutive \ftx \txt The fact that the system for marking third person pronouns is tripartite can be seen as the intersection of the nominative/accusative and the ergative/absolutive systems. That is, a tripartite system marks both the A and the P in order to maximally distinguish them (see example 29 earlier JOHNp LIONa EXAMPLE). In particular, Cashinawa employs nasalization from the ergative/absolutive system to mark the A and a vowel '-a' from the nominative/accusative system to mark the P in the redundantly marked tripartite system for third person pronouns. \ftx \txt Even English is consistent with the universal, even though there is no ergativity in the basic system. The only case marking of main clause constituents in English is evident in the pronominal system -- 'subjective' (A and S) pronouns take one form and 'objective' (P) pronouns take a distinct form. This represents a clear nominative accusative system. Full noun phrases, on the other hand, take no case marking whatsoever. Therefore it is a 'neutral' case marking system. If the neutral system is understood as being in some sense a step closer to an ergative/absolutive system than the standard pattern, then English is consistent with the universal that in a split GR marking system, elements higher on the agentivity hierarchy will represent a nominative/accusative system while elements lower on the hierarchy are more likely to express an ergative/absolutive system. Chart #7.4 illustrates the English system. Note that if the situation were the opposite, that is, if English pronouns were neutral and case marking of full noun phrases operated on a nominative/accusative basis, then English would be in violation of the universal expressed in 30. \ftx Chart #7.4: English \ftx \ftx | S | A | P | \ftx |-------+-------+-------| \ftx | I | I | me | \ftx | we | we | us | \ftx Pronouns: | he | he | him | Nominative-accusative \ftx | she | she | her | \ftx | they | they | them | \ftx -------------+-------+-------+-------+----------------------- \ftx Other nouns: | -0 | -0 | -0 | Neutral \ftx \txt To continue this discussion, highlight and jump to the topic below: \cf Hgrel_splt_tr_ex4 \ftx \dt 27/Apr/1998 \key Hgrel_splt_tr_ex4 \shd Split systems--transitive--examples VI \txt Spanish also supports 30 in an interesting and subtle way. In Spanish most clausal arguments that are referred to with full noun phrases are not case marked. However, direct objects that refer to specific, human participants take an accusative case marker: \ftx \fln (33) a. Estoy buscando una empleada. \ftx be:1SG looking:for indef housekeeper \ftx 'I'm looking for a housekeeper (don't have one in mind).' \ftx \ftx b. Estoy buscando a una empleada. \ftx be:1SG looking:for CM def housekeeper \ftx 'I'm looking for a housekeeper (have a particular one in mind).' \ftx \txt In other words, the case marker 'a' is used only when a specific and human direct object is referenced. All other nominal direct objects are not case marked. Therefore case marking for specific human direct objects manifests a nominative/accusative GR marking system, but case marking for all other classes of full noun phrases is neutral. It so happens that verb agreement and pronominal case marking also operate on a nominative accusative basis. Therefore, Spanish is nominative/accusative in the highest ranges of the agentivity hierarchy right up to the point of non-specific human direct objects. Then it becomes neutral. These data are illustrated in chart #7.5: \ftx \ftx Chart #7.5: Spanish \ftx \ftx | S | A | P | \ftx ---------+-------+---------+--------| \ftx 1,2,3 verb|-a, -o,|-a, -o, | lo/le, | \ftx coding | etc. | etc. |te, me, etc. \ftx ---------+-------+---------+--------| Nominative/accusative \ftx pronouns | 0 | 0 | a | \ftx ---------+-------+---------+--------| \ftx definite | | | | \ftx human NPs | 0 | 0 | a | \ftx ---------+-------+---------+--------+---------------------- \ftx def non- | | | | \ftx human NPs | 0 | 0 | 0 | Neutral \ftx indef NPs | 0 | 0 | 0 | \ftx \txt Again, note that if the a case marker happened to occur on non-human or non-specific direct objects, the universal expressed in 30 would be violated. \ftx \txt Many languages case mark direct objects only when they are 'definite' or identifiable. Turkish, Hebrew and Farsi are well know examples of such languages. Following are some examples from Farsi (the national language of Iran): \ftx \fln (34) a. Man domboli kitob. 'I'm looking for a book.' \ftx I look:for book \ftx \ftx b. Man domboli kitob-ro . 'I'm looking for the book.' \ftx I look:for book-CM \ftx \txt In these examples the suffix '-ro' appears only on direct objects, and only on direct objects that are identifiable. Again, therefore, this fact is consistent with the universal stated in 30. \ftx \txt To continue this discussion, highlight and jump to the topic below: \cf Hgrel_splt_tr_sum \ftx \dt 27/Apr/1998 \key Hgrel_splt_tr_np \shd Split systems--noun phrases in transitive clauses \txt A third type of split ergativity is based on the potential for agentivity of the noun phrases in a transitive clause. In any given transitive clause, one argument will probably be inherently more likely to be an AGENT than the other. The 'neutral' or 'natural' state of affairs is for the argument that is the most likely AGENT to in fact be the AGENT. As discussed above, some languages employ no overt coding of A or P status when this normal flow of action from high potential agentivity to low potential agentivity is adhered to. However, some of those languages do use a special case marker on either the A or the P of a transitive clause only in those unusual circumstances when the P is high in potential agentivity, or when the A is low in potential agentivity. \ftx \txt For example, suppose our imaginary language employed a case marking system to render sentences such as 25a (THE JOHN AND LION EXAMPLES DISCUSSED EARLIER) less ambiguous. There would be two possible case marking solutions to the ambiguity problem. One would be to mark the A with a special case marker (28a) and the other would be to mark the P (28b): \ftx \fln (28) a. Kill John lion-a 'The lion killed John.' \ftx -A \ftx P A \ftx \ftx b. Kill John-p lion 'The lion killed John'. \ftx -P \ftx P A \ftx \txt Assuming that intransitive subjects remain unmarked, the solution represented in 28a results in an ergative/absolutive case marking system with '-a' functioning as the ergative case marker. The solution represented in 28b results in a nominative/accusative case marking system, with '-p' functioning as the accusative case marker. Each of these solutions is equally as efficient in accomplishing the task of making the identity of the A and P arguments explicit. Theoretically (and occasionally in reality) both solutions may be used. This results in a tripartite system of grammatical relations, and is thus overly explicit: \ftx \fln (29) Kill John-p lion-a 'The lion killed John.' \ftx -P -A \ftx \txt To continue this discussion, highlight and jump to the topic below: \cf Hgrel_splt_tr_sol \ftx \dt 27/Apr/1998 \key Hgrel_splt_tr_sol \shd Split systems--transitive--solutions \txt Functionally, the choice of which solution to employ in marking agentivity is influenced by a principle of iconicity (see Qintro_what_symb_G). When the flow of action is in the normal, expected direction from high to low on the agentivity hierarchy, no special marking is needed to make A and P explicit. However, when P is unexpectedly high on the hierarchy, it may need to be marked. Similarly, when A is unexpectedly low on the hierarchy, it may need to be marked. Placing a special marker on P results in a nominative/accusative system. Therefore, the nominative/accusative solution (ex. 28b) is likely to be employed when direct objects are high on the agentivity hierarchy. The ergative/absolutive system is likely to be used for A arguments that are low on the agentivity hierarchy. In fact, this is a universal of split-ergativity based on semantic/pragmatic content of the noun phrases in transitive clauses. This universal can be expressed as follows: \ftx \fln (30) Universal: If a language exhibits split ergativity based on the \ftx semantic/pragmatic content of the noun phrases in the transitive \ftx clause, it is always the case that the nominative/accusative system \ftx will be manifested for nominal arguments that are high in potential \ftx agentivity and the ergative/absolutive system will be manifested for \ftx nominal arguments that are low in potential agentivity. \ftx \txt In other words, when split ergativity is based on the semantic and/or pragmatic characteristics of the noun phrases, it will always be the case that noun phrases higher (to the left) on the agentivity hierarchy introduced in Hgrel_splt_t/a_expl will trigger a nominative-accusative subsystem, while noun phrases lower (to the right) on the hierarchy will trigger an ergative-absolutive subsystem. This hierarchy will be repeated here for convenience: \ftx \ftx 1st/2nd \ftx Agr > pronouns > 3rd PROs > proper-names > human > animate > inanimate \ftx \txt The actual location of the split on the hierarchy varies from language to language, but the general arrangement of high = nominative-accusative and low = ergative-absolutive is invariant. \ftx \txt To continue this discussion, highlight and jump to the topic below: \cf Hgrel_splt_tr_ex2 \ftx \dt 27/Apr/1998 \key Hgrel_splt_tr_sum \shd Split systems--transitive--summary \txt Chart #7.6 summarizes and compares the data presented so far. Note that in every case, the nominative/accusative system extends from the left of the agentivity hierarchy and the ergative/absolutive or neutral system extends from the right of the hierarchy. No language has been documented to date that violates this pattern. Cashinawa is particularly interesting in that it illustrates a system when the ergative/absolutive and nominative/accusative systems actually overlap. \ftx \ftx Chart #7.6: Summary of Split ergativity based on semantic/pragmatic \ftx characteristics of NPs \ftx \ftx Agr | 1/2 pro | 3 pro | Def-Human | Def-NPs | other NPs \ftx Managalasi --->|<--------+-------+-----------+---------+--------- \ftx | | | | | \ftx Dyirbal: ....|-------->|<------+-----------+---------+--------- \ftx | | | | | \ftx Cashinawa: ....|---------+------>| | | \ftx | |<------+-----------+---------+--------- \ftx | | | | | \ftx English: ----+---------+------>| | | \ftx | | | | | \ftx Spanish: ----+---------+-------+---------->| | \ftx | | | | | \ftx Farsi: ....|---------+-------+-----------+-------->| \ftx nomin-accus. erg-absol \ftx (*erg-abs *nom-acc.) \ftx \dt 27/Apr/1998 \key Hgrel_splt_tr_top \shd Split systems--transitive--topicality \txt Agreement markers and pronouns are morphosyntactic devices whose function is to refer to relatively highly topical entities (see Hdisc_cnt_top). In languages for which agreement and pronouns function along the continuum of topicality, agreement is used for more topical arguments than are free pronouns (Givón 1984). The empathy principle in effect states that agreement markers and pronouns code topical arguments, and topical arguments are arguments that people empathize with, therefore agreement markers and pronouns code arguments that people empathize with. \ftx \txt A less technical way of expressing this fact is to say that people identify with and like to talk about things that act, move, control events and have power. Therefore sentences in communication tend to make AGENTs highly topical, and action tends to flow from the highly topical, agentive, entities to less topical entities. This is not to say that action always flows from highly topical and agentive arguments to less topical and patientive arguments. In fact, the ways in which languages deal with deviations from the natural flow of action are some of the more fascinating aspects of linguistic structure. It's just that the normal, or most common, situation is for people to choose powerful, dynamic and controlling entities as topics for their sentences. \ftx \txt To continue this discussion, highlight and jump to the topic below: \cf Hgrel_splt_tr_ex1 \ftx \dt 27/Apr/1998 \key Hgrel_splt_t/a \shd Split systems--tense/aspect \txt Some languages manifest a nominative-accusative system in one tense-aspect category and an ergative-absolutive system in another. In all such languages, the ergative-absolutive system occurs in the past tense or perfective aspect, while the nominative-accusative system occurs in the non-past tense(s) or imperfective aspect (Delancey 1982). To date no clear exceptions to this universal have been attested. \ftx \txt To continue this discussion, highlight and jump to the topic below: \cf Hgrel_splt_t/a_ex1 \ftx \dt 27/Apr/1998 \key Hgrel_splt_t/a_expl \shd Split systems--tense/aspect--Explanation \txt DeLancey (1982, 1990) proposes functional explanations for this universal phenomenon based on the notion of starting point vs. endpoint perspective. We have already noted that an ergative system for organizing grammatical relations can be thought of as being 'patient oriented'. Similarly, a nominative-accusative system can be thought of as Agent-oriented. Now, past tense and perfective aspect provide ways of expressing situations as completed events. The result of a completed transitive event is likely to be recorded in the patient. In fact, for DeLancey, the definition of PATIENT as a semantic role has more to do with its status as the endpoint of an action, rather than as the participant most likely to undergo a change in physical state. Hence, the past tense and perfective aspect are more patient-oriented than are non-past tense and/or imperfective aspect. \ftx \txt Perhaps the best way of illustrating this principle is through example. Example 36 is a past tense, perfective aspect clause of English. \ftx \fln (36) John hit (has hit) Bill. \ftx \txt If true, this statement describes the event from the point of view of a completed act, perhaps with Bill lying on the floor with a bloody nose. Bill is definitely and wholly involved in the action from this point of view. Example 37, on the other hand, is in the future tense: \ftx \fln (37) John will hit Bill. \ftx \txt This statement can be construed as saying more about John than about Bill. That is, from this point of view, Bill is not yet involved in the event, though John is very much involved, i.e. John is probably angry, and may be storming down the hall with boxing gloves on, while Bill is going happily about his everyday business. The clause is presented as viewing the event as from the perspective of its (potential) inception, and the event begins with John. Hence, we can say that the future is agent oriented. Similarly, though less obviously, the imperfective aspect presents the situation as ongoing, and therefore the result is unknown. \ftx \dt 27/Apr/1998 \key Hgrel_splt_t/a_ex1 \shd Split systems--tense/aspect--Examples I \txt The following example is from Georgian, the language of the Republic of Georgia: Georgian (from Comrie): \ftx \fln (35) a. Student-i midis. 'The student goes.' \ftx -NOM goes \ftx \ftx b. Student-i ceril-s cers. 'The student writes the letter.' \ftx -NOM letter-ACC writes \ftx \ftx c. Student-i mivida. 'The student went.' \ftx -ABS went \ftx \ftx d. Student-ma ceril-i dacera. 'The student wrote the letter.' \ftx -ERG letter-ABS wrote \ftx \txt In these examples, the case marker '-i' marks S and A nominals in the 'present' tense (examples 35a and b). Therefore, it is appropriate to refer to this case marker as marking 'nominative' case. The same operator, however, marks S and P nominals in the 'past' tense (examples 35c and d). In these clauses, then, it is appropriate to describe '-i' as an absolutive case marker. \ftx \txt To continue this discussion, highlight and jump to the topic below: \cf Hgrel_splt_t/a_expl \ftx \dt 27/Apr/1998 \key Hgrel_splt_t/a_ex2 \shd Split systems--tense/aspect--Examples II \txt Newari (Nepal and Tibet) exhibits a very interesting combination split case marking system. In Newari the ergative case marker '-nâ' is required in the perfective aspect for all transitive verbs (38a, 39a and 40a). For the verb 'ha'-- 'sing' the ergative case marker is not used in the imperfective aspect (38b) and 'optional' in the future tense (38c). \ftx \ftx Newari (Givón 1984:155): \fln (38) a. wo mânu-nâ me ha-lâ 'The man sang (a song)' \ftx the man-ERG song sing-PERF \ftx \ftx b. wo mânu me ha-yi cô-gu de \ftx the man:ABS song sing-INF be-NOM be \ftx 'The man is singing (a song)' \ftx \ftx c. wo mânu-(nâ) me ha-yi 'The man will sing (a song).' \ftx the man-ERG/ABS song sing-FUT \ftx \txt For the verb 'tó' 'drink', the ergative case marker is 'optional' in both imperfective (39b) and future (39c) clauses: \ftx \fln (39) a. wo mânu-nâ wo lâ tó-nâ 'The man drank the water.' \ftx the man-ERG the water drink-PERF \ftx \ftx b. wo mânu-(nâ) wo lâ tó-ni cô-gu du \ftx the man-ERG/ABS the water drink-INF be-NOM be \ftx 'The man is drinking the water.' \ftx \ftx c. wo mânu-(nâ) wo lâ tó-ni \ftx the man-ERG/ABS the water drink-PERF \ftx 'The man will drink the water.' \ftx \txt Finally, with the verb 'täjya'-- 'break' the ergative case marker is required in all three tense/aspects: \ftx \fln (40) a. wo mânu-nâ wo jhya täjya-tâ \ftx the man-ERG the window break-PERF \ftx 'The man broke the window.' \ftx \ftx b. wo mânu-nâ wo jhya täjya-yi cô-gu du \ftx the man-ERG the window break-INF be-NOM be \ftx 'The man is breaking the window.' \ftx \ftx c. wo mânu-nâ wo jhya täjya-yi \ftx the man-ERG the window break-INF \ftx 'The man will break the window.' \ftx \txt This pattern makes sense in terms of the semantic differences that obtain among these verbs. 'To sing (a song)' does not result in a major change in the state of the world, hence the ergative case marking system is relatively under-represented for this verb, even being disallowed in the imperfective aspect. 'To drink (water)', on the other hand, does have some concrete, physical effect on the state of the world (some water is consumed), though not much. From the point of view of semantic transitivity (Hopper and Thompson 1980) 'drink' would be more transitive than 'sing'. Correspondingly the ergative case marker has a wider distribution for this verb. Finally, 'to break (a window)' clearly has a concrete, physical effect on the state of the world. This is the most semantically transitive of the three verbs illustrated here. Consequently, the ergative case marker is required in all the tense/aspect categories for this verb. \ftx \txt So again we see that there is a correspondence between change of state/patient/endpoint/completed action orientation and ergative case marking. \ftx \dt 27/Apr/1998 \key Hgrel_splt_t/a_ex3 \shd Split systems--tense/aspect--Examples III \txt Before leaving the topic of split-ergativity, we will present one more example of a combination split-ergative case marking system. This is from the Guaymí language of Panama and Costa Rica. As illustrated earlier, Guaymí has a straightforward Split-S case marking system in both of the 2 past tenses. This is illustrated in 41a, b and e. The ergative case marker '-gwe' can only occur on the A arguments of transitive verbs or agentive intransitive verbs in one of the past tenses: \ftx \ftx Guaymí: \fln (41) a. Dori-gwe blit-ani. 'Doris spoke.' \ftx Doris-ERG speak-PAST:1 \ftx \ftx b. Toma-gwe Dori dêma-ini . 'Tom greeted Doris.' \ftx Tom-ERG Doris greet-PAST:1 \ftx \ftx c. Dori blit-e. 'Doris is speaking.' \ftx Doris speak-PRES \ftx \ftx d. Toma Dori dêma-e. 'Tom is greeting Doris.' \ftx Tom Doris greet-PRES \ftx \ftx e. Nu nat-ani. 'The dog died.' \ftx dog die-PAST:1 \ftx \ftx f. Nu nat-e. 'The dog is dying.' \ftx dog die-PRES \ftx \txt In the non-past tenses, however, the ergative marker never occurs, regardless of the semantics or transitivity of the verb (41c, d and f). One could say that Guaymí has two quite distinct case marking systems, one for the past tenses and one for the present. Consistent with the universal expressed at the beginning of this section, the only system that contains even a trace of ergativity is that which is used in the past tenses: \ftx \ftx Past Present \ftx \ftx +---------+ +---------+ \ftx | Sa | So | | S | \ftx | | | | | \ftx | A | O | | A O | \ftx +---------+ +---------+ \ftx \dt 27/Apr/1998 \key Hgrel_sum \shd Summary \txt The following chart summarizes the definitions of the terms S, A, P, nominative, accusative, ergative and absolutive as we have presented them in this chapter: \ftx \fln Chart #7.8: Core semantico-syntactic roles (Comrie 1978) \ftx +---+ +---+ \ftx Nominative-| A | | A |-Ergative A = Most AGENT-like argument \ftx (or subject)| | +---+ of a transitive clause \ftx | | +---+ \ftx | S | | S | S = Only argument of an \ftx +---+ | | intransitive clause \ftx +---+ | | \ftx Accusative-| P | | P |-Absolutive P = Non-AGENT-like argument \ftx (or object) +---+ +---+ of a transitive clause \ftx \txt Languages can manifest a nominative-accusative or an ergative-absolutive system in any one or more of the following areas of the morphosyntax: \ftx \ftx 1. Case marking of full noun phrases. \ftx 2. Pronominals \ftx 3. Verb coding (agreement or anaphoric clitics) \ftx 4. Constituent order \ftx 5. Syntactic processes (complement subject omission, \ftx reflexivization, conjunction reduction, relativizability etc.) \ftx \txt If a language consistently manifests an ergative-absolutive system in case marking of full noun phrases, it is likely to be called an 'ergative language' (a non-technical term). This is the case even though such languages always (as attested to date) manifest a nominative-accusative system in one or more of the other areas mentioned above. Mayan languages do not have case marking of full noun phrases, yet they are also widely considered to be 'ergative' because their verb coding systems are solidly organized on an ergative-absolutive basis. \ftx \dt 27/Apr/1998 \key Hgrel_typ_erg \shd System type--ergativity characterized \txt At this point we are prepared to provide a broader characterization of ergativity than simply morphological identity or non-identity. The broadest possible definition of ergativity is the following: \ftx \txt An ergative-absolutive system is any morphosyntactic system which unites S and P as opposed to A. \ftx \txt Note that this definition refers to systems (i.e. case marking, verb agreement, etc.) not languages. The term 'ergative language' is simply an informal term that refers to languages that have an ergative case marking system on full noun phrases in basic clauses, or exhibits an ergative system of verb agreement and no case marking on noun phrases. Ergativity itself is merely a convenient means of conceptualizing some aspects of the clause structure of some languages. It is not a wholistic typology that makes a wide range of predictions concerning other aspects of a language's grammatical structure. \ftx \txt To continue this discussion, highlight and jump to the topic below: \cf Hgrel_typ_erg_eng \ftx \dt 27/Apr/1998 \key Hgrel_typ_erg_eng \shd System type--ergativity in English \ftx \txt In view of this broad definition, we may ask the question whether English has any ergativity. The following examples illustrate some marginal evidence of ergativity in English: \ftx \fln (11) a. Escap-ee \ftx S \ftx \ftx b. Employ-er c. Employ-ee \ftx A P \ftx \ftx d. bird-chirping \ftx S \ftx \ftx e. fox-hunting f. *doctor-recommending \ftx P A \ftx \txt The suffix '-ee' (a reflex of the French past participle) forms a nominalization that refers to the S of an intransitive verb (ex. 11a) and the P of a transitive verb (11c). A morphologically distinct suffix, '-er', must be employed to form a nominalization that refers to the A of a transitive verb (11b). In this respect, then, S and P are treated alike, while A is treated differently. Similarly, when an argument of a verb is incorporated into the present participle of the verb in English, it is either the S of an intransitive verb (11d) or the P of a transitive verb (11e) that is incorporated. Although English does allow marginal incorporation of A arguments (this medicine is doctor-recommended), this type of incorporation is not available for present participles (*I was doctor-recommending aspirin all evening'). This is further illustrated by the fact that fox-hunting cannot mean 'hunting that foxes do'. Rather, it must mean 'hunting that is done to foxes'. \ftx \txt These facts illustrate some processes in English that might be said to operate on an ergative/absolutive basis. In many other languages ergativity spreads much further through the grammatical system. Very few, if any, languages can be said to be 'purely' ergative. \ftx \dt 27/Apr/1998 \key Hgrel_typ_erg/abs \shd System type--ergative/absolutive \txt Some languages organize their grammatical relations on a different basis. Instead of grouping S and A as a category, they group S and P, and treat A distinctly: \ftx \ftx +---------+ \ftx | S | \ftx |----+ | \ftx | A | P | \ftx +---------+ \ftx ergative | absolutive \ftx \txt This system is referred to as an ERGATIVE/ABSOLUTIVE system. The case that encodes only A arguments is referred to as the ERGATIVE case, while the case that encodes S and P arguments is referred to as the ABSOLUTIVE case. This system often seems unnatural to speakers of Indo-European languages since it is very rare in these languages. However, it is extremely common in other areas of the world, therefore all linguistic fieldworkers need to be aware of it. Ergativity occurs as a basic system for marking verbal arguments throughout the Austronesian family (Philippines, Indonesia, Malaysia, Madagascar, the coasts of Papua New Guinea, and the Pacific Islands), Australia, Central Asia, Basque and in many languages of the Americas. It occurs as a partial case marking system in South Asia (Nepal, Tibet, India, Pakistan, Bangladesh, Bhutan), the non-Sino-Tibetan languages of Southeast Asia and in most other languages of the Americas. In short, virtually the only areas of the world in which ergativity is not known to occur to any appreciable extent are Europe and Africa. \ftx \txt To continue this discussion, highlight and jump to the topic below: \cf Hgrel_typ_ex1 \ftx \dt 27/Apr/1998 \key Hgrel_typ_ex1 \shd System type--examples I \txt The following examples from Yup'ik Eskimo (Alaska) illustrate an ergative/absolutive case marking system on free noun phrases: \ftx \fln (6) a. Doris-aq ayallruuq. 'Doris travelled.' \ftx -ABS travelled \ftx S \ftx \ftx b. Tom-am Doris-aq cingallrua. 'Tom greeted Doris.' \ftx -ERG -ABS greeted \ftx A P \ftx \txt In these examples the case marker '-aq' occurs on the S argument of an intransitive clause (6a) and the P argument of a transitive clause (6b). The case marker '-am' marks only the A of a transitive clause. For this reason we can term '-aq' the absolutive case marker and '-am' the ergative case marker. \ftx \txt Other possibilities In addition to morphological case marking on pronouns or free noun phrases, languages manifest ergative/absolutive or nominative/accusative systems in other areas of the morphosyntax. \ftx \txt To continue this discussion, highlight and jump to the topic below: \cf Hgrel_typ_ex2 \ftx \dt 27/Apr/1998 \key Hgrel_typ_ex2 \shd System type--examples II \txt We have seen that Quechua manifests a nominative/accusative system in case marking of free noun phrases. Verb-coding in Quechua also manifests a nominative/accusative system for organizing grammatical relations: \ftx \fln (7) a. Aywa-n. 'He goes.' \ftx go-3SG \ftx S \ftx b. Aywa-a. 'I go.' +---------+ \ftx go-1SG | S | \ftx S | +----| \ftx | A | P | \ftx c. Maqa-ma-n. 'He hit me.' +---------+ \ftx hit-1SG-3SG \ftx P A \ftx \txt In example 7a the third person singular S of an intransitive verb is referred to by the suffix '-n'. In 7b the first person S argument is coded by the suffix '-a' (actually length on the final vowel of the root). Example 7c shows that the suffix '-n' is also used for third person A arguments of transitive verbs. Hence A and S are treated morphologically 'the same' by the verb coding system of Quechua. The fact that, in 7c, the first person suffix for P arguments is '-ma' rather than '-a' illustrates that P and S are treated as 'different' by the verb coding system of Quechua. This system of marking S and A alike and P differently has been defined as a nominative/accusative system. \ftx \txt To continue this discussion, highlight and jump to the topic below: \cf Hgrel_typ_ex3 \ftx \dt 27/Apr/1998 \key Hgrel_typ_ex3 \shd System type--examples III \txt As might be expected, languages can also manifest an ergative/absolutive GR system in verb-coding. Yup'ik will again serve as our example of such a system: \ftx \fln (8) a. Ayallruu-nga. 'I travelled.' \ftx travelled-1SG \ftx S \ftx b. Ayallruu-q. 'He travelled.' +---------+ \ftx travelled-3SG | S | \ftx S |----+ | \ftx | A | P | \ftx c. Cingallru-a-nga. 'He greeted me.' +---------+ \ftx greeted-3SG-1SG \ftx A P \ftx \txt In example 8a the suffix '-nga' indicates a first person singular S argument of an intransitive verb. In 8b the suffix '-q' marks the third person S. In 8c the suffix '-nga' marks the first person P argument of a transitive clause. Since this is the same marker that is used for first person S arguments, this suffix groups S and P together morphologically into an absolutive category. The third person singular A argument of a transitive clause is coded by a suffix '-a'. Since this suffix is different from the third person S suffix it can be said to identify ergative arguments. This treatment of S together with P as distinct from A is defined as an ergative/absolutive system. \ftx \dt 27/Apr/1998 \key Hgrel_typ_nom/acc \shd System type--nominative/accusative \txt Many languages treat S and A the same, and P differently. The following English examples illustrate this fact with pronominal case forms -- one form, 'he', is used for 3rd person singular masculine pronouns in both the S and the A roles. A different form, 'him', is used for 3rd person masculine singular pronouns in the P role: \ftx \ftx +---------+ \fln (4) a. He left. | S | \ftx | +----| \ftx b. He hit him. | A | P | \ftx +---------+ \ftx nominative | accusative \ftx \txt This system is often referred to as a NOMINATIVE/ACCUSATIVE system. The case that marks both S and A roles is the NOMINATIVE case, while the case that marks the P role is the ACCUSATIVE case. The Quechuan languages (a group of languages spoken throughout the Andes mountains in South America) manifest this system in morphological case marking on free noun phrases. In the following examples the same case marker, '0' (zero), occurs on noun phrases in both the S (ex 5a) and A (5b) roles. Another case marker, '-ta', occurs on noun phrases in the P role: \ftx \fln (5) a. Juan-0 aywan. 'Juan goes.' \ftx -NOM goes \ftx S \ftx \ftx b. Juan-0 Pedro-ta maqan. 'Juan hits Pedro.' \ftx -NOM -ACC hits \ftx A P \ftx \txt This system seems very reasonable to speakers of Indo-European languages since these languages, to the extent that they have case marking at all, do it this way. \ftx \txt NOTE: The terms nominative and accusative are from the grammars of classical languages. To a large extent their use in those grammars corresponds to the definitions given here. However, the terms in the classical languages refer strictly to morphological cases. The markers that signal those cases are often used in many other ways in addition to marking A, S and P arguments. For example the 'accusative' case in Latin occurs on objects of certain prepositions. Here we are using the terms nominative and accusative to describe groupings of semantico-syntactic roles, no matter how those roles are instantiated in the morphosyntax. So we may, for example, refer to a particular noun phrase as a 'nominative' noun phrase if it is an S or an A argument, whether or not it is marked by a distinct nominative case marker. More commonly, certain syntactic rules may refer to the nominative category, whether or not the case marking system of the language is organized on a nominative-accusative basis (Anderson 1976). This same sort of practice was observed in the section on noun phrases where it was noted that possessors are often referred to as 'genitives' even in languages that have no morphologically distinct genitive case. \ftx \txt Related to this issue, case markers (or any morphosyntactic units for that matter) that are glossed with terms from familiar languages should not necessarily be expected to always correspond exactly to the operators that bear those glosses in the familiar languages. For example, if a language has a marker that consistently marks nominative arguments (according to the definition given here) such that the fieldworker feels justified in glossing that marker as nominative case, it is not to be expected that that case marker will have the same range of functions that the nominative case in Greek or Latin has. We have given a strictly linguistic definition of nominative and accusative case: S together with A = nominative case. P alone = accusative case. \ftx \dt 27/Apr/1998 \key Hgrel_typ_ord \shd System type--constituent order \txt Since constituent order is univerally one major means of expressing grammatical relations, one might ask whether ergative-absolutive and/or nominative-accusative systems can be manifested in constituent order. Of course the answer is 'yes'. English, consistent with its strong nominative-accusative orientation, treats S and A alike in that both subjects of intransitive verbs and subjects of transitive verbs most neutrally occur in preverbal position. Objects of transitive verbs, on the other hand, are treated differently in that they occur in post-verbal position. \ftx \txt However, manifestation of constituent order nominativity or ergativity is only possible for languages in which the verb occurs in between the two core arguments of a transitive verb (i.e. AVP or PVA languages). In all other logically possible constituent order types both P and A occur most neutrally on the same side of the verb. Therefore either both are in the same position or both are in different positions from the intransitive S argument. Hence there can be no 'grouping' of S with A against P or S with P against A. \ftx \txt Since most languages with ergatively organized GR systems are of the APV or VAP typologies, there are few languages which could potentially manifest an ergative system in constituent order. Besides, there is no a priori reason to expect that the system of organizing GR roles in one area would necessarily correspond to the system in another area (see Hgrel_splt on 'split ergativity'). Nevertheless, there is at least one language, Kuikúro (Cariban, Brazil), that exhibits an ergative-absolutive system in constituent ordering. Perhaps not surprisingly, this language also has an ergative-absolutive case marking system (examples from Franchetto 1990): \ftx \fln (9) a. karaihá kacun-tárâ 'The non-indian is working.' \ftx non-indian work-CONT \ftx S \ftx \ftx b. kuk-aki-sâ ta-lâígo léha karaihá-héke \ftx 1INC-word-POSS hear-FUT ASP non:indian-ERG \ftx P A \ftx 'The non-indian will hear our words.' \ftx \txt In 9a the S argument of an intransitive verb occurs in preverbal position. In 9b the P argument of a transitive verb occurs in preverbal position, and the A argument occurs in post-verbal position. Since both S and P occur in the same position, we can say that this language manifests an ergative-absolutive system in constituent order. According to the analysis presented by Francetto, these are the most neutral constituent orders of the language. \ftx \txt To continue this discussion, highlight and jump to the topic below: \cf Hgrel_typ_ord_other \ftx \dt 27/Apr/1998 \key Hgrel_typ_ord_other \shd System type--other possibilities \txt We have seen that languages can organize their systems of encoding grammatical relations in at least two distinct ways. In this section we will see that in fact there are three other logically possible ways, but that all three of these are far less common than the ergative-absolutive or nominative-accusative systems. This observation leads us to investigate possible functional motivations for the various patterns. \ftx \txt Chart #7.1 illustrates the five logically possible ways in which languages could conceivably group S, A, and P. Under each type an informal estimate of the relative number of languages that instantiate each type is given (reference?). \ftx \ftx Chart #7.1: Logically possible GR marking systems for S, A and P \ftx \ftx I II III IV V \ftx nominative/ ergative/ tripartite \ftx accusative absolutive \ftx +---------+ +---------+ +---------+ +---------+ +---------+ \ftx | S | | S | | S | | S | | S | \ftx | +----| |----+ | | | |---------| |---------| \ftx | A | P | | A | P | | A P | | A P | | A | P | \ftx +---------+ +---------+ +---------+ +---------+ +---------+ \ftx Lots Lots Rare Non-existant Rare \ftx \txt To continue this discussion, highlight and jump to the topic below: \cf Hgrel_typ_rat1 \ftx \dt 27/Apr/1998 \key Hgrel_typ_rat1 \shd System type--rationale-I \ftx \txt Why should so many languages employ systems I and II as their basic system for organizing grammatical relations, while so few languages employ systems III, IV and V? Could it not have been simply random which system a given language would 'choose'? The answer to this question stems, at least partially, from the DISCRIMINATORY FUNCTION of GR marking systems. The most important distinction to make among A, S and P is between A and P. This is because A and P are the only arguments (among these three) that are instantiated in the same clause, and it is very important from the point of view of communication to identify which argument is acting upon which other argument. On the other hand, the two other distinctions, S vs. A, and S vs. P, are communicationally irrelevant. This is because by definition these combinations of roles never occur in the same clause, therefore there is no possibility that they would ever be confused. Systems I and II are equally efficient from this point of view. They both make the important distinction between A and P, and don't unnecessarily distinguish S. \ftx \txt All of the other systems, however, are dysfunctional in one way or another. In particular, system IV ignores the important distinction between A and P and makes a useless one between S on the one hand and A and P on the other. System III ignores all distinctions, while V overdistinguishes them. From the point of view of the discriminatory function of GR marking, then, it is understandable why systems I and II should be about equally as common, and systems III, IV and V should be far less common. \ftx \txt To continue this discussion, highlight and jump to the topic below: \cf Hgrel_typ_rat2 \ftx \dt 27/Apr/1998 \key Hgrel_typ_rat2 \shd System type--rationale II \txt In addition to the discriminatory function, however, GR marking also has an IDENTIFYING FUNCTION. That is, in addition to keeping nominal elements that have different roles distinct, GR marking also serves to unite nominal elements whose roles are similar. This is another manifestation of the general principle that formal similarity (or ISOMORPHISM) results from functional similarity (Haiman 1980). From this point of view the treatment of S and A as 'the same' in system I leads us to expect some functional similarity between these roles. Similarly, the treatment of S and P as 'the same' in system II would indicate some functional commonality to these two roles. \ftx \txt The functional similarity between S and A is agentivity: if a clause has an AGENT, it will be the S or the A argument, depending on whether the clause is transitive or intransitive. The similarity between S and P is change of state. If any argument in a clause changes state, it will be either the S or the P: \ftx \fln (10) a. The bomb exploded. S changes state. \ftx b. Lucretia broke the vase. P changes state. \ftx \txt It is very difficult to conceive of a transitive clause in which the A changes state independently of the P. Possible examples are quite marginal, e.g. 'John underwent surgery', 'Paul received a blow to the head', etc. \ftx \txt The GR marking system of any given language will apparently center on one or the other of these principles. GR marking in ergative languages can be thought of as paying more attention to change of state, whereas GR marking in nominative-accusative languages pays more attention to agentivity. The less common systems (systems III and V) may quite possibly represent GR systems in transition. \ftx \txt To continue this discussion, highlight and jump to the topic below: \cf Hgrel_typ_erg \ftx \dt 27/Apr/1998 \key Hgrel_val \shd Valency \txt The terms S, A and P describe three basic semantico-syntactic roles. This terminology is from Comrie (1978). Similar terms are used by Silverstein (1976) and Dixon (1979). These terms presuppose two prototypical clause types: \ftx \fln (3) a. Single argument, 'John left.' \ftx S V \ftx \ftx b. Multi-argument, 'John kissed Mary.' \ftx A V P \ftx \txt The S is defined as the only nominal argument of a single argument clause. Sometimes this type of clause is referred to as an INTRANSITIVE clause (see Hval_p-calc_tr for a discussion of different kinds of transitivity). The A is defined as the most AGENT-like core argument of a multi-argument clause. [NOTE: By AGENT-like here, we mean the argument which most closely approximates the ideal of the initiating, controlling actor in the particular scene described by the proposition. We do not refer to inherent properties of the verbal arguments, as described in the 'agentivity hierarchy' in Hgrel_splt_tr_ag. That is, in a sentence like: \ftx \fln (55) The change in schedule made me late \ftx \txt The phrase 'the change in schedule' is the A because it is presented as being the entity that controls the specific event of being late, even though a first person is much higher on the hierarchy of inherent agentivity. The notion of inherent agentivity refers to the fact that certain kinds of entities are inherently quite likely to be coded as AGENTS, but does not by any means preclude the fact that such entities are often found in other roles as well.] If there is no argument that is a very good AGENT, the A is the argument that is treated morphosyntactically in the same manner as prototypical AGENTs are treated. Usually there will be one nominal in every verbal clause that exhibits this property, though there may not be. More complex systems are described below. P is the 'most PATIENT-like' core argument of a multi-argument clause. Again, if none of the core arguments is very much like a PATIENT, then the argument that is treated like a prototypical PATIENT is considered to be the P. \ftx \txt For help on encoding grammatical relations, jump to the topic below: \cf Hgrel_val_enc \ftx \dt 27/Apr/1998 \key Hgrel_val_enc \shd Encoding grammatical relations \txt While S, A and P may be universally identifiable clause elements, individual languages will typically treat two or more of them as 'the same' morphosyntactically. This is similar to the tendency for languages to treat various universally identifiable semantic roles, say AGENT and EXPERIENCER as 'the same' for purposes of grammatical encoding. In the following extended discussion, we will discuss the various systems for grouping S, A and P and the morphosyntactic means languages employ to express them. \ftx \txt There are basically four possible answers to this question, plus combinations thereof: \ftx \ftx a) morphological case on NPs (Hgrel_case) \ftx b) verb coding (Hgrel_vcode) \ftx c) constituent order (Hgrel_ord) \ftx d) some pragmatic hierarchy (Hgrel_ag/emp) \ftx \dt 27/Apr/1998 \key Hgrel_vcode \shd Verb coding \ftx \ftx \key Hgr_adv_dir/loc \shd Direction/location \txt up/downriver, up/downhill, up/down(ward), north(ward), south(ward), east(ward), west(ward), left(ward), right(ward), hither, thither, etc. \ftx \txt To continue this discussion, highlight and jump to the topic below: \cf Hgr_adv_ev/epis \dt 29/Jun/1998 \key Hgr_adv_ev/epis \shd Evidential/epistemic \ftx \txt What evidential/epistemic adverbs are there? Evidential adverbs indicate the source of the information contained in the sentence (e.g. hearsay, first hand observation, second hand observation, pure conjecture). Epistemic adverbs indicate the degree to which the speaker is committed to the truth of the sentence. English does not have a clear class of evidential adverbs. Instead it uses what appear to be matrix verbs of saying or perception to accomplish this function, e.g. 'I understand', 'they say', 'I hear' etc. That these are not really prototypical matrix verbs is evidenced by the fact that these locutions distribute like adverbs in a sentence, e.g., 'Democracy is, I understand, the best form of government', or 'They are going to dedicate, I hear, a new linguistics building'. In many languages that have lexicalized evidential adverbs, these can be traced etymologically to matrix verbs of utterance or perception. English does have epistemic adverbs, e.g.: possibly, definitely, clearly etc. English also uses erstwhile matrix verbs of cognition for this purpose, e.g. I think, I know, etc. \ftx \dt 27/Apr/1998 \key Hgr_adv_gen \shd Adverbs--General \txt If you posit a distinct class of adverbs, argue for why these forms should not be treated as nouns, verbs or adjectives. For each kind of adverb listed in this section, list a few members of the type, and specify whether there are any restrictions relative to that type, e.g. where they can come in a sentence, any morphemes common to the type, etc. For each class, examples are given from English, though the fieldworker should be aware that the language may not exhibit all of these classes, or may have classes not represented here. Notice that '-ly' signals epistemic and manner adverbs and '-ward' signals directional adverbs in English (though this is not an absolute rule). \ftx \txt To continue this discussion, highlight and jump to the topic below: \cf Hgr_adv_vscpl \dt 29/Jun/1998 \key Hgr_adv_man \shd Manner \txt This is the largest subcategory of adverbs in every language: quickly, slowly, patiently, etc. \ftx \txt To continue this discussion, highlight and jump to the topic below: \cf Hgr_adv_time \dt 29/Jun/1998 \key Hgr_adv_time \shd Time \txt Yesterday, today, tomorrow, next/last year/week/ month, early, late etc. Yagua has an adverb 'tââriy' which means something like 'long time'. With a clause in the past tense it means 'a long time ago', in other clauses it means 'for a long time' or 'late in the afternoon' \ftx \txt To continue this discussion, highlight and jump to the topic below: \cf Hgr_adv_dir/loc \dt 29/Jun/1998 \key Hgr_adv_vscpl \shd Adverbs versus complement-taking verbs \txt Adverb is a 'catch all' category. Any word with lexical content (i.e., other than grammatical particles) that isn't clearly a noun, a verb or an adjective is usually put into the class of adverb. Semantically, forms that have been called adverbs cover an extremely wide range of concepts, hence they cannot be identified in terms of time-stability or any other well-defined semantic parameter. Also, they typically function on the clause or sentence level, i.e., their semantic impact (scope) is relevant to entire clauses or sentences rather than simply to phrases. As with adjectives, there are no prototypical adverbs. Formally, adverbs can be characterized primarily in terms of their distribution. They are typically the most unrestricted grammatical category in terms of their position in sentences. Some adverbs of English are signalled by the suffix '-ly', e.g. quickly, slowly, finally, adverbially etc. \ftx \fln Reference: Givón (1984:77). \ftx \txt To continue this discussion, highlight and jump to the topic below: \cf Hgr_adv_man \dt 29/Jun/1998 \key Hgr_cat \shd Grammatical categories \txt In traditional grammar, GRAMMATICAL CATEGORIES; are called 'parts of speech'. Every language has at least two major grammatical categories -- noun and verb. Two other major categories, adjective and adverb, may or may not be instantiated in any given language, though they usually are to some extent. Most languages also have minor grammatical categories such as conjunctions, particles and prepositions. In this chapter of the grammar, the field worker should describe the grammatical categories that exist in the language, and give evidence for treating them as distinct. Depending on the scope of the grammar, the fieldworker may wish to describe subclasses of the major grammatical categories as well. \ftx \txt Grammatical categories are distinct from formal relational categories such as subject, object and predicate, or functional categories such as agent, topic, or definite. They are the building blocks of linguistic structure. They are sometimes called 'lexical categories' since most forms in the lexicon of any language must be specified for their grammatical category. \ftx \txt It is important to present empirical evidence for each grammatical category posited. In the sections on nouns and verbs the questions lead the reader to discover the formal properties that distinguish these major categories. For the other categories, however, there are too many language-specific properties to offer a compendium of all possibilities. \ftx \dt 27/Apr/1998 \key Hgr_cat_min \shd Minor Grammatical categories \ftx \fln (incomplete) \dt 20/Aug/1997 \key Hgr_mod \shd Modifiers \txt For discussion of the various kinds of modifiers and their characteristics, see the following sections. \ftx \key Hgr_mod_adj_agr \shd Descriptive adjectives--agreement \ftx \ftx \key Hgr_mod_adj_col \shd Color \ftx \ftx \key Hgr_mod_adj_form \shd Descriptive adjectives--form \txt If you posit a distinct category of adjectives, give evidence for not lumping these forms with the verbs or nouns. \ftx \key Hgr_mod_adj_func \shd Functionality \ftx \ftx \key Hgr_mod_adj_other \shd Other \ftx \ftx \key Hgr_mod_adj_sem \shd Descriptive adjectives--semantic classes \txt Adjectives are problematic in almost every language. Unlike nouns and verbs, adjectives cannot be defined in terms of a prototype. This is because there is no semantically definable class of concepts that universally falls into a category that we would want to call adjectives. Rather adjectives stand 'between' nouns and verbs, lexicalizing concepts that are indeterminate in terms of time-stability. Some languages have no formally distinct category of adjectives. In such languages concepts that are expressed as adjectives in other languages are expressed as either nouns or verbs. The following diagram illustrates the continuum of time stability: \ftx \fln Time stability: \ftx \fln Unstable -------------------------------------------- Stable \ftx active > (stative > (adjectives) > nouns \ftx verbs verbs) \ftx \txt Any given language will divide up this functional continuum differently in terms of its formally distinct grammatical categories. Some languages have neither a formally distinct class of stative verbs nor a class of adjectives. All languages, however, formally distinguish at least nouns from verbs. A prototypical adjective can be used in a noun phrase to specify some quality of the head noun of the phrase. GET EXAMPLES FROM SCHACHTER. \ftx \dt 27/Apr/1998 \key Hgr_mod_adj_shp \shd Shape \ftx \ftx \key Hgr_mod_adj_size \shd Size \ftx \ftx \key Hgr_mod_adj_txt \shd Texture \ftx \ftx \key Hgr_mod_adj_val \shd Value (good/bad) \ftx \ftx \key Hgr_mod_non-num \shd Non-numeral quantifiers \ftx \ftx \key Hgr_mod_num_agr \shd Numerals--agreement \ftx \ftx \key Hgr_mod_num_ext \shd Numerals--extent \txt Some languages are reported to have number systems consisting of as few as four terms: one, two, three and many. Other languages have native terms for the first few numbers (e.g. 1, 2, and 3) but then resort to terms borrowed from a trade language for the higher numbers. Still others have native terms that can be used to count almost infinitely. English uses native terms from one to 999,999. Million is a borrowing from French meaning 'big thousand'. In many parts of the world, different vocabulary is used to refer to numerical concepts depending on the context. For example, most Philippine minority languages have native terms for numbers into the thousands. These terms are used in most everyday situations. However, in the domains of money and time, Spanish terms are used. This is in spite of the fact that Spanish is no longer a trade language in the Philippines. \ftx \txt To continue this discussion, highlight and jump to the topic below: \cf Hgr_mod_num_agr \ftx \dt 27/Apr/1998 \key Hgr_mod_num_sys \shd Numerals--system used \txt Different languages employ different number systems. Almost all natural number systems are either base five (quintenary) or base ten (decimal). This is probably because people universally keep track of quantities of items using their fingers and toes. In many languages, the word for 'five' is the same as or etymologically related to the word for 'hand'. 'Ten' may be a compound related to 'two hands'. In languages of Papua New Guinea, it is common for the word for 'twenty' to be a compound meaning 'one man', i.e. the number of fingers and toes of a man. A base-five (quintenary) number system is one where the words for 6, 7, 8, and 9 consist of the word for 'five' plus the word for 1, 2, 3, and 4 respectively. There may be phonological reduction and/or some extra morphological trapping. For example, 6 may be expressed as 'one on top of five', 7 as 'two on top of five' etc. 11, 12, etc. may then be expressed as 'one on top of two fives', 'two on top of two fives' etc. Some languages employ more than one system. In Papua New Guinea some languages use an alternative system of counting which enumerates body parts. These usually start with the fingers of one hand, move up the arm, across the face and down the other arm to the second hand. These systems may have a total of forty to fifty terms. \ftx \fln DESCRIBE THE PANARE QUINTENARY SYSTEM HERE? \ftx \txt To continue this discussion, highlight and jump to the topic below: \cf Hgr_mod_num_ext \ftx \dt 27/Apr/1998 \key Hgr_n_dist \shd Nouns--Distributional characteristics \txt The HEAD of a noun phrase is the one word within the phrase that refers to the same entity as the whole phrase refers to. For example, an English phrase like 'garbage man' refers to a time stable concept, hence we suspect it is a noun phrase. However, it also contains two words, each of which also refers to a time stable concept. So the question arises as to which of the two nouns is the head of the noun phrase. The answer in this case is easy: the whole phrase refers to a man, not to garbage, hence the noun 'man' is the head of the noun phrase. \ftx \txt For languages that either have no grammatical category of adjective, or for which adjectives and nouns are very similar formally, the identification of the head of a noun phrase can be more complex. In such a language, in the phrase 'the red hen' the words for 'red' and 'hen' could equally refer to the entity that the whole phrase refers to. That is, the color terms and other descriptive terms can in many languages function just like nouns. \ftx \txt For these languages, noun phrases such as 'the red hen' are often considered to be examples of apposition, i.e., 'the red one, the hen'. D. Payne (1990) has proposed a method for determining headship even in this type of language. In her system, the head of a noun phrase is the element that persists; in the continuing discourse. So when a participant in a story is mentioned in a phrase like 'the red hen', if it is subsequently referred to as 'the red one', then 'red' is the head & vice versa. \ftx \dt 27/Apr/1998 \key Hgr_n_ex_chgcat \shd Nouns--Changing Grammatical Category, an Example \txt The English suffix '-ing' will serve as an illustrative example of how to determine the categorical status of a derived form. It is very difficult to think of a term such as 'walking' as referring to a time stable concept -- 'walking' inherently involves motion and change. However, let us look at the distributional and structural characteristics of this form. First, can it distribute like a noun? i.e. can it be the subject or object of a sentence? The following examples show that 'walking' passes the distributional test for nounhood: (1) SUBJECT a. 'Walking is good for you.' OBJECT b. 'I like walking.' \ftx \txt To continue this discussion, highlight and jump to the topic below: \cf Hgr_n_ex_other \ftx \dt 27/Apr/1998 \key Hgr_n_ex_other \shd Nouns--Other structural characteristics of the example \txt Structural characteristics of nouns include: 1) the possibility of taking adjectival modifiers, and 2) the use of genitive case pronouns for central arguments. Again, the form 'walking' passes both of these tests: \ftx \fln (2) a. DESCRIPTIVE MODIFIERS 'slow walking' \ftx b. GENITIVE CASE PRONOUNS 'his walking' (*he walking) \ftx \txt However, verbs with the '-ing' suffix are not the best examples of nouns. This is illustrated by the fact that certain other properties of prototypical nouns do not hold for verbs with '-ing'. For example, ability to pluralize, take determiners and to take a wide range of descriptive modifiers are all characteristics of prototypical nouns. It is rare or odd-sounding for verbs with the '-ing' suffix to display these properties: \ftx \fln (3) a. PLURALIZATION: '?many walkings' \ftx b. DETERMINERS: '?the walking to school' \ftx c. DESCRIPTIVE MODIFIERS: '?red/little/pretty walking' \ftx \fln Also, '-ing' verbs sometimes can take adverbial modifiers that nouns cannot: \ftx \fln (4) a. ?I like slowly walking. c. *I like slowly cars. \ftx b. I like walking slowly. d. *I like cars slowly. \ftx \txt To continue this discussion, highlight and jump to the topic below: \cf Hgr_n_ex_sol \ftx \dt 27/Apr/1998 \key Hgr_n_ex_sol \shd Nouns--Solutions to the example \txt Solutions to this sort of 'problem' (though it should not be problematic, since it is a characteristic of every language) vary from analysis to analysis. Such solutions include: \ftx \ftx 1. Taking some nominal properties as criterial. For instance, we \ftx could simply define noun for English as a form that can refer \ftx to its only argument with a genitive pronoun. In this case \ftx 'walking' is a noun. However, if we decided that ability to \ftx pluralize or take a wide range of descriptive modifiers were \ftx the criterion for nounhood, then walking would not be a noun. \ftx \ftx Discussion of this solution: Hgr_n_ex_sol_1 \ftx \ftx 2. Making up a different syntactic category for each complex of \ftx nominal features instantiated by some form or forms in the \ftx language. In this case only those lexical items that have all \ftx nominal properties would be considered nouns. Forms such as \ftx 'walking' would be considered something else, such as a present \ftx participle. \ftx \ftx Discussion of this solution: Hgr_n_ex_sol_2 \ftx \ftx 3. Acknowledging that the difference between nouns and verbs is a \ftx continuum, and that verbs with the '-ing' suffix fall somewhere \ftx in between the two extremes. \ftx \ftx Discussion of this solution: Hgr_n_ex_sol_3 \ftx \ftx Recommendation: Hgr_n_ex_sol_rec \ftx \dt 27/Apr/1998 \key Hgr_n_ex_sol_rec \shd Nouns--Recommendation \txt The recommended approach is to combine solution #2 and solution #3 in something like the following manner: forms that are indeterminate as to their grammatical category membership (such as verbs with '-ing' suffixes in English) can be given strictly formal labels (e.g. '-ing participles'), with an explanation given of their characteristic functions and key morphosyntactic properties. In most cases it is just not worth the effort to be more explicit than this. The payoff in terms of clarity of description is too miniscule. Participle is a relatively widely understood term for verb forms that have reduced verbal properties, but which are not full nominalizations. Sentences formed with participial verbs are often referred to as PARTICIPIAL PHRASES. However, since languages normally have more than one such form and, as mentioned above, the term participle has no more specific universal definition, it is important to clarify that the label is simply a shorthand way of referring to the formal class as a whole, and that it should not imply that the form is directly comparable to forms that have been called participles in other languages. In other words, to call a verb form a participle is only to give it a, hopefully insightful, label. It is not a theoretical claim (see Qintro_dling). \ftx \fln References: Givón (1984, ch. 3), \ftx Hopper and Thompson (1984), \ftx Schachter (1985). \ftx \dt 27/Apr/1998 \key Hgr_n_ex_sol_1 \shd Nouns--Discussion of Solution 1 \txt Solution #1 above is problematic because a) it cannot be applied universally and b) it ignores the obvious but inexact functional basis for the grammatical category of nouns. Criterial definitions are inherently questionable for supposedly universal categories. This is because there is always the possibility that some language may not display a particular criterion. We do not want to say that such languages lack universal categories (if so the category is not universal, so our universal definition goes out the window anyway). It is invariably the case that criterial definitions (even those found lurking in the pages of this manual) are biased towards the well-known languages of the world. Some of the most interesting discussions in linguistics have been those involving languages in which important grammatical categories display different morphosyntactic properties than those of the well-studied languages. \ftx \txt Secondly, picking an arbitrary morphosyntactic feature as criterial, however closely that feature seems to correlate with pre-theoretical notions of what the category should include, ignores the question of why the category exists in the first place. For example, to define nouns as all forms that refer to their arguments with genitive pronouns is like defining the class of human beings as all featherless bipeds. Though this 'definition' may succeed in distinguishing to a large extent the category of human beings from all other animals, it focuses on incidental rather than definitional (or necessary) properties. In other words, it begs the question of why one would even consider featherless bipeds as a class apart from all other potentially arbitrary classes of items in the natural world, say red socks, or broken sticks. Certainly if we were to take a feathered biped and remove all of its feathers, it would not become a human being. It would not even come one whit closer to being a human being than before its unfortunate defeathering. On the other hand, if we dye a white sock red it does enter the category of red socks. \ftx \txt Furthermore, we can conceive of a world in which human beings were not featherless bipeds, whereas it is correspondingly difficult to conceive of a world in which red socks were really green. This shows that the category 'human' consists of more than merely the conjunction of the features 'featherless' and 'biped'. Another way of saying this is that 'featherless bipeds' is a sufficient but not a necessary definition of the category 'human being'. \ftx \txt The best criterial definitions are those which include both necessary and sufficient criteria. Unfortunately such definitions are extremely uncommon in linguistics. If a criterial definition such as ability to take genitive case pronouns is suggested for the category of nouns, one should also ask: What is it about items that take genitive case pronouns that makes them adhere as a category? Why should they have that property, and not some other (say high tone on the first syllable)? \ftx \dt 27/Apr/1998 \key Hgr_n_ex_sol_2 \shd Nouns--Discussion of Solution 2 \txt Solution #2 is the approach traditionally taken in descriptive linguistics. It has the advantage of giving the analyst pigeon holes within which to place the various word types in the language. The basic problem with this approach is that, like solution #1, it is not universally applicable. The categories derived from various clusters of morphosyntactic properties are a) not necessarily related to one another in any systematic way, and b) not comparable from one language to the next. This situation makes for a grammatical description that is less readable to someone with no previous experience with the language. For example, the term PARTICIPLE is found in many grammar descriptions. Nevertheless, what constitutes a participle in language A may or may not have any commonality with what is called a participle in language B. Therefore someone who knows language B may be misled when reading the description of language A. \ftx \dt 27/Apr/1998 \key Hgr_n_ex_sol_3 \shd Nouns--Discussion of Solution 3 \txt Solution #3 reflects most accurately the true nature of linguistic categorization. This in itself is a point in favor of this approach. However, it also has its disadvantages to the field linguist attempting to present information about a language clearly and precisely in a grammar sketch. These disadvantages include: \ftx \txt 1. There is no explicit way of determining exactly where on the continuum between noun and verb a particular category falls. One could conceivably count nominal properties and verbal properties, and assign items with more verbal properties to a position closer to the verb extreme and vice versa for nouns. However, this approach assumes that all properties are weighted equally in terms of their effect on the category membership of the form. There is no a priori reason to accept this assumption. In fact, solution #1 above is based on the opposite assumption, namely that there exists one and only one property that is important enough to distinguish the categories, all the other properties being incidental. In any case, it is futile to try to sort out a relative ranking of importance of morphosyntactic properties in determining class membership. \ftx \txt 2. It is often the case that a fieldworker just doesn't know what all of the relevant properties are for a given form. For example, some verbs with '-ing' might take plurals more easily than others: 'his many failings' vs. '?his many eatings'. This fact puts 'failing' closer to the noun end of the continuum than 'eating'. These subtle differences between the behavior of various forms are probably not available to the fieldworker faced with a list of 16,000 forms based on 1,000 verb roots and 12 properties that distinguish nouns from verbs. \ftx \txt 3. The point of a grammatical sketch is to help the reader understand how particular constructions function within the grammatical system of the language. It is clear that a detailed ranking of structures according to their relative nounhood would be of limited use in accomplishing this task. Given the observation above that such a task would also potentially be of unlimited complexity, it is not likely that many field linguists would attempt such a ranking. \ftx \dt 27/Apr/1998 \key Hgr_n_gen \shd Nouns--General \txt The class of NOUNS in any language is that grammatical category that includes forms which embody the most TIME-STABLE concepts, e.g. 'rock', 'tree', 'mountain', 'house', etc. (Givón 1984:51). These are concepts that characteristically do not vary appreciably over time. The time-stability criterion only defines prototypical nouns. The prototype of any category is the member or set of members of a category that best characterize the category as a whole. Prototypical nouns, then, would be those nouns that are very and obviously time-stable. \ftx \txt For example, the concept 'fist' is less than prototypical in that it does not characteristically persist over a long period of time, hence some languages do not express this concept with a simple noun. In determining whether any given form is a noun or not, one must first determine the morphosyntactic characteristics of prototypical nouns. Then the grammatical category of a questionable form can usually be determined according to how closely the form follows the morphosyntactic pattern of prototypical nouns. There will usually be some truly ambiguous examples as well as forms that function sometimes as a verb and sometimes as a noun. A discussion of some ways of dealing with ambiguous cases will be provided in subsequent screens. \ftx \dt 27/Apr/1998 \key Hgr_n_msyn \shd Nouns--Morphosyntactic criteria \txt Morphosyntactic criteria for nounhood fall into two categories: distributional (or configurational) and structural. Distributional characteristics of nouns have to do with how nouns are distributed in noun phrases, sentences and texts. For example, nouns can serve as heads of noun phrases (see below), subjects and objects of clauses (see Hgrel_val), and topics of texts (see Hprag_top). Structural characteristics have to do with the internal structure of the noun itself. For example, in some languages nouns exhibit case marking, number marking, gender marking etc. whereas other grammatical categories do not exhibit these characteristics. \ftx \dt 27/Apr/1998 \key Hgr_n_n/npstr \shd The structure of the noun or noun-phrase \txt This question may be adequately answered in chapter 4. If so, make reference to that section here. However, if the language is at all polysynthetic, it would be good to provide an overview of the structure of the noun word here. For English this may be something like the following: \ftx \ftx (8) HEAD-PL \ftx \txt The notion of noun phrase would also be relevant for most languages. A description of the noun phrase for English might be: \ftx \ftx (9) DET QUANT ADJ* HEAD-PL RELCL \ftx \txt (DET includes possessors. Everything is 'optional' except HEAD. * means this item is repeatable.) \ftx \txt This diagram leaves out a lot of detail, but it does give the reader of the sketch a general idea of how words are put together in phrases. Be careful to not be overly concerned with the formalism, as this is a trap out of which there is no escape. Again, if this is adequately covered in the section on constituent order typology, cross-reference that section here. \ftx \dt 27/Apr/1998 \key Hgr_n_prn/clt_def_clt \shd Definitions--Anaphoric clitics \txt ANAPHORIC CLITICS are not free morphologically -- they must attach (cliticize) to another word. However, like pronouns they are in complementary distribution with full noun phrases. That is, typically either a noun phrase or a clitic, but not both, can refer to an entity in a given clause. For example, in Yagua, a reference to a subject can be either a full noun phrase (ex 10) or a proclitic (ex 11), but not both (ex 12): \ftx \fln (10) Manungo murróóy 'Manungo sings.' \ftx M. sing \fln (11) sa-murróóy 'He sings.' \ftx 3SG-sing \fln (12) *Manungo sa-murróóy \ftx \txt That 'sa-' is not a pronoun is evidenced by the fact that it cannot stand alone. For example, you cannot answer a question like 'Who's singing?' simply with the word 'sa'; there is a distinct 3rd person singular pronoun, 'níí', that is used in such contexts. Also, it can only appear immediately before the verb stem, whereas pronouns (such as 'níí') have the same distributional privileges as full noun phrases (i.e. they can occur pretty much anywhere in a sentence). Furthermore, there is good morphological evidence that 'sa-' is a clitic. With certain verb classes it enters into morphophonemic rules that do not cross word boundaries. For example, when the verb begins with the syllable 'ha', the 'h' is lost, and a long nasalized vowel occurs: \ftx \fln (13) sa '3SG' + hatu 'drink' = sªªtu 'he drinks' \ftx \txt Example 14 shows that this process does not cross word boundaries. Note that the final 'a' of the first word and the initial 'ha' of the second do not coalesce to 'ªª': \ftx \fln (14) vuryªªnuma hachipííyªª 'Now we must think.' \ftx vurya-ha-numa hachiy-píí-hªª \ftx 1INCL-AUX-now heart-VRBLZR-DIST \ftx \txt In summary, 'sa-' in Yagua is an anaphoric clitic because it is morphologically bound, yet is in complementary distribution with noun phrases. \ftx \txt For a discussion of anaphoric clitics in Romance languages, jump to: \cf Hgr_n_prn/clt_def_clt_ex \ftx \dt 27/Apr/1998 \key Hgr_n_prn/clt_def_clt_ex \shd Anaphoric clitics in Romance languages \txt Accusative and dative 'pronouns' in Romance languages fulfill our definition of anaphoric clitics because a) they must be phonologically bound, and b) they distribute differently than full noun phrases -- the clitics occur before inflected verbs and after non-inflected verbs in most cases, whereas there are no such distributional restrictions on full noun phrases. Note that you cannot answer a question such as 'whom did you see?' in Spanish simply with the form 'la'-- 'her'. Rather a full pronoun is required. \ftx \txt Cross-linguistically anaphoric clitics often attach to either the first or the last element of a clause, whether that element is a noun, a verb, an adverb, an auxiliary or whatever. If the clitic occurs on the first element, it can attach either to the front or the end of that element. If it attaches to the end of that element it can be termed a 'second position' clitic. Clitics occurring on the last element of a clause will always cliticize to the end (Klavens 1985). \ftx \dt 27/Apr/1998 \key Hgr_n_prn/clt_def_pron \shd Definitions--Pronouns \txt PRONOUNS are free forms (as opposed to affixes) that function alone to fill the position of a noun phrase in a sentence. They normally have all the distributional characteristics of noun phrases. \ftx \dt 27/Apr/1998 \key Hgr_n_prn/clt_dist \shd Relevant Distinctions \txt There are a number of different distinctions which may be relevant for the pronoun/anaphoric clitic paradigm. Not all of these will be applicable to all languages. A few possibilities are considered in the following segments, but there may be additional examples which will need to be analysed and described. \ftx \dt 27/Apr/1998 \key Hgr_n_prn/clt_dist_def \shd Pronoun/Anaphoric clitic distinctions--Definiteness/specificity \fln Definiteness/specificity \ftx \key Hgr_n_prn/clt_dist_gen \shd Pronoun/Anaphoric clitic distinctions--Gender \fln Gender. \ftx \key Hgr_n_prn/clt_dist_grel \shd Pronoun/Anaphoric clitic distinctions--Grammatical relations \fln Grammatical relations. \ftx \key Hgr_n_prn/clt_dist_hon \shd Pronoun/Anaphoric clitic distinctions--Honorifics \txt Very often different pronouns are used depending on the relative social statuses of the speech act participants. In English, there are some unusual situations where special forms are used in place of the standard second person pronouns. For example, when addressing a judge in a courtroom situation it is still customary to use the term 'your honor' in place of 'you'. Many other languages use honorifics on an everyday basis. For example, standard Spanish uses 'tu' and 'te' for the second person subject and non-subject pronouns when speaking in a familiar manner. In a more formal situation 'Usted' and 'le' are more appropriate. \ftx \fln References: Givón (1983a, 1984), Weismann (1986). \ftx \dt 27/Apr/1998 \key Hgr_n_prn/clt_dist_num \shd Pronoun/Anaphoric clitic distinctions--Number \ftx Number. \ftx \key Hgr_n_prn/clt_dist_pers \shd Pronoun/Anaphoric clitic distinctions--Person \txt Person;. 'First person' refers to the speaker. 'Second person' refers to the hearer. First and second persons are sometimes collectively referred to as SPEECH ACT PARTICIPANTS. 'Third person' usually refers to any non- speech act participant. The term 'fourth person' has been used in a variety of ways in the literature. None of the previous uses of this term describe a function that is not covered by some other term employed in this manual. Therefore we will not attempt to survey the various uses of this term. \ftx \txt Many languages have an INCLUSIVE/EXCLUSIVE distinction within the category of first person. First person inclusive includes speaker and hearer and may or may not include a non-speech act participant. Some languages have an 'inclusive dual' form, even though dual may not be specified in any other part of the grammar. This form refers to only speaker and hearer and excludes a non-speech act participant. First person exclusive includes the speaker and a non-speech act participant, but excludes the hearer. \ftx \fln Diagram of First Person Non-singular Categories: \ftx \ftx +-----------------+ \ftx ¦ ¦ \ftx ¦ +-------+ ¦ \ftx ¦ ¦ ¦ ¦ \ftx Inclusive --¦ +---+---+ ¦ ¦ \ftx Plural ¦ ¦ ¦ 1 ¦ ¦ ¦ \ftx ¦ ¦ ¦ ¦ +--+-- Exclusive \ftx ¦ ¦ 2 ¦ ¦ 3 ¦ ¦ (Dual and plural, depending \ftx Inclusive --+--¦ +---+---+ ¦ on whether 3 is singular \ftx Dual ¦ ¦ ¦ ¦ or non singular) \ftx ¦ +-------+ ¦ \ftx ¦ ¦ \ftx +-----------------+ \ftx \dt 27/Apr/1998 \key Hgr_n_prn/clt_dist_role \shd Pronoun/Anaphoric clitic distinctions--Semantic roles \fln Semantic roles. Agent, Patient, etc. See: \cf Hgr_v_cl_sem \ftx \key Hgr_n_prn/clt_intro \shd Pronouns and/or anaphoric clitics \txt These are distinct from grammatical agreement. Agreement will be discussed later. Also, the functions of pronouns and clitics will be discussed later. \ftx \txt For many languages it is difficult to distinguish PRONOUNS from AGREEMENT (or concord) affixes. Here we will give strictly formal definitions, though it must be kept in mind that there is no direct correlation between the function of a particular device in one language and formally similar devices in other languages (though there are generalities that can be made -- see Hdisc_cnt_ref -- Topic Continuity). So, for example, free pronouns in language X might function like verb agreement does in language Y, etc. \ftx \txt For an example, jump to: \cf Hgr_n_prn/clt_vsagr \ftx \txt For a definition of pronouns, jump to: \cf Hgr_n_prn/clt_def_pron \ftx \txt For a definition of anaphoric clitics, jump to: \cf Hgr_n_prn/clt_def_clt \ftx \dt 27/Apr/1998 \key Hgr_n_prn/clt_vsagr \shd Example of anaphora versus agreement \txt As an illustration of the fact that the correlation of pronouns and verb agreement do not necessarily match across languages, let us compare English and Spanish. Spanish has two ways of referring to subjects--pronouns and verb agreement. English also has pronouns and (very rudimentary) verb agreement. However, verb agreement in English functions differently from verb agreement in Spanish. In English, verb agreement cannot constitute the only reference to a participant, e.g. 'walks' is not a well-formed sentence of English, even though the '-s' suffix in some sense 'refers to' a third person singular subject. In Spanish, on the other hand, the third person singular form of the verb is sufficient to stand as a complete sentence, e.g. 'anda'-- 'he/she walks'. Hence we want to say that in Spanish, verb agreement is an ANAPHORIC DEVICE, whereas in English it is not. \ftx \txt Now, let us compare the pronouns. Spanish free pronouns are used very rarely in discourse, and are usually described as 'emphatic' or 'contrastive', whereas English pronouns are much more common. When we look at English pronouns more closely, however, we find that there are really two types -- stressed and unstressed. Most pronouns in English discourse are unstressed. If they are stressed, they function very similarly to the Spanish pronouns, i.e. to signal contrastiveness of some sort (see Hpnom_p-adj_ncontr). So a Spanish sentence with a pronoun, e.g. 'ellos vinieron', roughly corresponds in function to an English sentence with a stressed pronoun, 'THEY came' (as opposed to someone else). The Spanish sentence without a pronoun, 'vinieron', corresponds more or less to the English sentence with an unstressed pronoun, 'they came'. Hence it appears that English and Spanish each have two anaphoric devices functioning within the domain of participant reference. Spanish verb agreement corresponds functionally to English unstressed pronouns while Spanish pronouns correspond to English stressed pronouns (roughly speaking). This illustrates the fact that devices that seem similar formally (e.g. pronouns in English and Spanish), can function very differently in discourse. For this reason, we will define our terms purely formally. Once the anaphoric devices are identified formally, the investigator must strive to understand how the various devices function within the system of participant reference. \ftx \dt 27/Apr/1998 \key Hgr_n_strc \shd Nouns--Structural characteristics \txt Every language has ways of changing the grammatical category or subcategory of a linguistic form, either to make it more or less noun-like. Such devices typically are sensitive to the demands of discourse, i.e. a device which renders a form more like a noun is used when the concept referred to has independent DEPLOYABILITY in the discourse (see Hopper and Thompson 1984). Hence there can be nouns that are based on verb roots. For such nouns the time stability criterion is difficult, if not impossible to apply. Therefore the only criteria available are the distributional and structural characteristics of the form. \ftx \dt 27/Apr/1998 \key Hgr_n_typ \shd Types of nouns \txt Typically languages have 'subclasses' of nouns and verbs. For example, English makes a distinction between proper names and common nouns. Proper names are distinguished in that they do not (easily) take articles, quantifiers or other modifiers: \ftx \fln (5) Proper names Common nouns \ftx \ftx Mt. Rushmore car \ftx ?the Mt. Rushmore the car \ftx ?several Mt. Rushmores several cars \ftx ?an outlandish Mt. Rushmore an outlandish car \ftx ?Mt. Rushmore that has four the car that has four \ftx Presidents' faces carved in it. Presidents' faces carved in it. \ftx \txt Within the category of common nouns, there is the distinction between COUNT nouns and MASS nouns. Mass nouns do not pluralize (unless used in a special, count, sense). Furthermore, mass and count nouns take a distinct, but partially overlapping, class of articles and quantifiers: \ftx \fln (6) Mass nouns Count nouns \ftx \ftx sand car \ftx ?many sands many cars \ftx much sand ?much car \ftx some sand ?some car \ftx ?a sand a car \ftx ?some sands some cars \ftx \txt Note that these distinctions are functionally (semantically) based, but evidence for their existence is formal properties. There is potentially an infinite number of "noun subclasses" based on semantic properties, but these subclasses are only significant for the grammar if they have some overt consequences. It is interesting to note how the grammaticalized formal properties can be used to produce special effects. For example, some of the expressions marked with ? above may be used to accomplish specialized communicative tasks: \ftx \fln (7) We'll have three waters please. (Mass noun being used as count noun to \ftx refer to a quantity of the mass concept). \ftx \ftx That's a lot of car you've got. (Count noun being used as mass noun to \ftx refer to a quality of the count concept.) \ftx \dt 27/Apr/1998 \key Hgr_n_typ_count \shd Count nouns \ftx \ftx \key Hgr_n_typ_mass \shd Mass nouns \ftx \ftx \key Hgr_n_typ_other \shd Other subcategories of noun. \txt The actual list of subcategories of nouns may be shorter or longer depending on the language and the degree of completeness of the grammatical description. \ftx \dt 27/Apr/1998 \key Hgr_n_typ_prop \shd Proper names \ftx \ftx \dt 01/May/1998 \key Hgr_v \shd Verbs \txt The class of VERBS in any language is that grammatical category that includes lexemes which refer to the least time-stable concepts, e.g. events such as die, run, break etc. (Givón 1984:51,55). This time-stability criterion only defines prototypical verbs. In determining whether any given form is a verb or not, one must look at the prototypical verbs and determine what their morphosyntactic properties are. Then the grammatical category of a questionable root can usually be determined according to how closely the root follows the morphosyntactic pattern of prototypical verbs. \ftx \fln Morphosyntactic criteria for verbhood fall into two categories: \ftx \ftx distributional (or configurational) \ftx structural. \ftx \txt Distributional characteristics of verbs have to do with how verbs function in verb phrases, sentences and texts. For example, verbs can serve as heads of verb phrases, predicates of clauses, and they code events in a text. Structural characteristics have to do with the internal structure of the verb itself. For example, in some languages verbs exhibit subject agreement, tense/aspect/mode marking etc., whereas other grammatical categories do not. \ftx \txt The functions of some major verb or verb phrase operations (e.g. tense/aspect/mode) will be discussed in depth in later sections. Here, the basic structure of the verbal word or verb phrase should be described. \ftx \dt 27/Apr/1998 \key Hgr_v_cl \shd Verb classes-general \txt The following is a list of some verb classes that might evoke distinct morphosyntactic treatment. Not every grammar will require a section dealing with each type. The point is to describe any distinctive morphosyntactic characteristics of any of these verbs, e.g., unexpected case marking patterns, restrictions on TAM marking, etc. If all (or the overwhelming majority) of a particular verb type follow the paradigm verb, then that class should receive no special section. Possible paradigm verbs: \ftx \ftx Univalent, agentive: dance \ftx Univalent, patientive: die \ftx Divalent: kill (if this is a causative based on die, \ftx try eat) \ftx Trivalent: give, put \ftx \txt For examples, jump to: \cf Hgr_v_cl_ex1 \ftx \txt Other useful and more elaborate systems for semantic classification of verbs can be found in Chafe (1970), Dowty (1987) and Foley and Van Valin (1984). \ftx \dt 27/Apr/1998 \key Hgr_v_cl_actn \shd Actions \txt Verbs which describe voluntary acts, but which do not involve an affected patient participant, e.g. dance, sing, speak, sleep/rest, look (at), read, deceive, care for, carry(?). Note that action verbs can be either dynamic (dance, sing, speak), non-dynamic (rest, look at) or somewhere in between. These answer the question 'What did X do?', but less easily 'What happened to X?', unless a slightly ironic, sarcastic or extended meaning is desired: \ftx \txt What did Lucretia do? She danced the tango. What happened to Lucretia? ??She danced the tango. What happened to the tango? ??Lucretia danced it. \ftx \txt What did Wimple do? She read War and Peace. What happened to Wimple? ??She read War and Peace. What happened to the book? ??Wimple read it. \ftx \txt For help on action processes (involving both a voluntary actor and a distinct affected PATIENT) highlight and jump to the topic below: \cf Hgr_v_actproc \ftx \dt 27/Apr/1998 \key Hgr_v_cl_actproc \shd Action processes \txt These are verbs that involve both a voluntary actor and a distinct affected PATIENT, such as kill, hit, stab, shoot, spear (and other violent verbs), plus the divalent (transitive) versions of break, melt, smash, change, etc. These verbs easily answer both of the questions 'What did X do?' and 'What happened to X?': \ftx \ftx What did Zelmo do? He melted the ice. \ftx What happened to the ice? Zelmo melted it. \ftx \ftx What did Chryseum do? She broke Arno's nose. \ftx What happened to Arno's nose? Chryseum broke it. \ftx \dt 27/Apr/1998 \key Hgr_v_cl_bod \shd Bodily functions \txt These are like _invuntary processes except they do not involve a change of state. Some languages may treat these in special ways morphosyntactically, often as onomatopoeic expressions. Cough, sneeze, hiccup, burp, bleed, sweat, vomit, expectorate, urinate, defecate, sleep/awaken(?), \ftx \dt 27/Apr/1998 \key Hgr_v_cl_bod_loc \shd Locomotion \txt Many languages make a distinction between LOCOMOTION, i.e., change of place, and simple motion. For these languages, oddly enough, the locomotion verbs tend to be more stative than verbs that simply describe motion. For example, in Wappo (Northern California) the locomotion verbs fall into the stative class. Examples 33a and b illustrate simple stative verbs with their characteristic suffix '-khi¿': \ftx \fln (33) a. mey-i söy'iya-khi¿ 'The water is hot.' \ftx water-NOM hot-STAT \ftx \ftx b. lel-i ceta wil-khi¿ 'The rock is over there.' \ftx rock-NOM there sit-STAT \ftx \fln For processes, se¿ contrasts with -khi¿: \ftx \fln (34) a. cephi hincatk:-se¿ 'She's waking up.'(process) \ftx 3SG:NOM wake:up-DUR \ftx \ftx b. cephi hincatel-khi¿ 'She's awake.' (state) \ftx 3SG:NOM wake:up-STAT \ftx \txt But with verbs of locomotion, '-khi¿' also occurs, even though these embody past punctual actions: \ftx \fln (35) a. ah pawata¿ te-hew'i-khi¿ \ftx 1SG:NOM once DIR-jump-STAT \ftx 'I jumped down once.' \ftx \ftx b. met'e-t-i me¿a i-thu nat'o¿ah-khi¿ \ftx woman-PL-NOM many 1SG-DAT come-STAT \ftx 'Many women came to me (i.e. to my house).' \ftx \ftx c. cephi te-piyola-khi¿ 'She sneaked in.' \ftx 3SG:NOM DIR-sneak-STAT \ftx \txt It is significant that all of these motion verbs that take the stative suffix describe locomotion, not simple motion. That is, each involves movement from out of one scene and into another. A sentence like 'he ran into the kitchen' would be locomotive, whereas 'he ran by' would not. It is very common for languages to treat predicates of locomotion as statives, even though other predicates of motion are active. Other languages actually treat predicates of locomotion as non-verbal predicates (see Hpnom_p-loc). \ftx \dt 27/Apr/1998 \key Hgr_v_cl_bod_mot \shd Motion \txt The basic, unmarked motion verb(s), come/go, may have different morphosyntactic properties from more explicit motion verbs like swim, run, walk, crawl, fly, jump, etc. Furthermore, while all motion verbs have an affected participant, some attribute more control to that 'affectee' than others do. The language may make a morphosyntactic distinction here. All those listed so far are voluntary. The following are likely to be _invuntary: fall, drop, flow, spew/squirt, etc. \ftx \txt Some languages employ verbal operators that change a verb from one class to another. For help on verbal operators, highlight and jump to the topic below: \cf Hgr_v_bod_mot_op \ftx \txt Some languages distinguish motion and locomotion. For help on locomotion, highlight and jump to the topic below: \cf Hgr_v_bod_loc \ftx \dt 27/Apr/1998 \key Hgr_v_cl_bod_mot_op \shd Motion--verbal operators \txt Some languages employ verbal operators that change a verb from one class to another. For example, many languages of the Americas employ verbal markers that transform a non- motion verb into a motion verb. In Yagua, the suffixes '-nuvîî', '-nuvaa' and a few others designate that the action expressed by the verb they attach to occurs relative to a particular locational scene and trajectory of movement (T. Payne 1990): \ftx \fln (36) a. Naani-ipeni-yââ 'They dance all over the place.' \ftx 3DL-dance-DIST \ftx \ftx b. Naani-ipeni-yââ-nuvîî 'They dance all over on arrival \ftx 3DL-dance-DIST-on:arrival1 (current scene).' \ftx \ftx c. Naani-n··-ñuvee 'They look on arrival (new scene).' \ftx 3DL-look-on:arrival2 \ftx \txt Both '-nuvîî' and '-nuvaa' indicate that the action expressed by the verb occurs upon arrival on some scene. The opposition between the two is determined by whether that scene is the currently activated one or if it implies the activation of a new scene. The complex verb stem consisting of a verb root plus locomotion suffix then patterns just like a verb of locomotion. For example, it has inflectional possibilities characteristic only of other verbs of locomotion. Other similar locomotion suffixes in Yagua include: '-rîî'-- 'passing by', '-ja'-- 'moving horizontally, across water or land', '-jasúmiy'-- 'moving upwards', '-siy'-- 'departing'. \ftx \txt Some verbs of motion specify a portion of a trajectory of movement, rather than the whole trajectory. Such verbs include 'depart', 'arrive' 'to crash', etc. These may or may not be treated like other verbs of motion. \ftx \dt 27/Apr/1998 \key Hgr_v_cl_cogn \shd Cognition \txt Know, think, understand, learn, remember, forget(?). In many languages all or many of these are based on the same root, often the name of an internal body part, e.g. heart, liver, stomach (see Hgr_v_cl_emot -- Emotion verbs). \ftx \dt 27/Apr/1998 \key Hgr_v_cl_emot \shd Emotion \txt Fear, like/love, be angry/sad/mournful, be happy/joyful/pleased, grieve/mourn. This is another verb class that is often based on a nominal root that refers to an internal body part such as `heart'. For example, in many Papuan languages, the center of thinking and feeling is the liver. Hence expressions of emotion and cognition are compounds based on the root for `liver'. This example is from the Orya language of Irian Jaya, courtesy of Phil Fields: \ftx \fln (37) Ano en-lala-na beya-na. \ftx 1SG:GEN liver-liquid-TOP much-BE \ftx `I am worried.' (Lit: `My bile is much,' or `I have much bile.') \ftx \txt In Yagua there are two verbs that express cognitive processes. One, 'dáátya', refers to mental processes, like knowing or understanding. The other, 'jachipiy', refers to more 'emotional' or reflective processes, such as 'ponder' or `meditate'. The second is transparently related to the noun root 'jachiy' meaning 'heart'. The following is a single example that illustrates both the nominal and derived verbal uses of this root: \ftx \fln (38) Naanajachipíyâânúúyanú jíjeechitya. \ftx naana-jaachiy-píy-yââ-núúy-janu jíy-jaachiy-tà \ftx 3DL-heart-VBLZR-DIST-IMPERF-PAST3 COR-heart-INSTR \ftx 'She pondered in her heart.' (Lit: `They two went all over the place \ftx thinking in their heart.' Women who have had children are always \ftx referred to as dual in Yagua. T. Payne,to appear). \ftx \dt 27/Apr/1998 \key Hgr_v_cl_ex1 \shd Verb classes-Examples from Guaymí \txt Some languages allow various case marking patterns depending roughly on the semantic roles of the core nominals. So, for example, as briefly illustrated earlier, in Guaymí (a Chibchan language of Costa Rica and Panama), EXPERIENCERs appear in the dative case: \ftx \fln (24) Davi-e Dori gare (25) Toma-e Dori tîrî. \ftx David-DAT Doris know:pres Tom-DAT Doris remember:pres \ftx 'David knows Doris.' 'Tom remembers Doris.' \ftx \fln (26) Ti-e ru hatu-aba. (27) Ti-e tîmëna nib-i. \ftx I-DAT airplane see-PAST I-DAT thirst feel-PRES \ftx 'I saw the airplane.' 'I feel thirst.' ('I'm thirsty.') \ftx \txt Certain other verbs that describe _invuntary actions place one central participant in a LOCATIVE case: \ftx \fln (28) a. Jose-biti Maria köinigwi-ani-nggö. \ftx José-LOC Maria forget-PAST1-ASP \ftx 'José forgot Maria.' \ftx \ftx b. Köinigwit-ani-nggö i-biti. \ftx forget-PAST1-ASP I-LOC \ftx 'I forgot it.' (or 'It was forgotten upon me.') \ftx \fln (29) Davi-bötö Dori hurö rîb-aba. \ftx David-LOC Doris fear feel-PAST2 \ftx 'David was afraid of Doris.' \ftx \fln (30) Ti-bötö kö nibi tibo. \ftx I-LOC place feel:PRES cold \ftx 'I'm cold.' \ftx \txt To see another example, highlight and jump to: \cf Hgr_v_cl_ex2 \ftx \dt 27/Apr/1998 \key Hgr_v_cl_ex2 \shd Verb classes-Examples from Gujarati \txt In Gujarati (and many other Indo-Aryan languagaes), AGENTIVE participants are treated differently in the `past tense' (ex. 31a) than in the present tense (31b) (examples from Adenwala 1979): \ftx \fln (31) a. raju-e kam kar-y-uN `Raju did work.' \ftx Raju-AGT work do-PAST-SG \ftx \ftx b. raju kam kar-e ch-e `Raju does/is doing work.' \ftx Raju work do-SG AUX-SG \ftx \txt Note that the the AGENT, Raju, takes the case marker '-e' in 31a but no case marker in 31b. In 31a 'Raju' is more AGENTIVE than in 31b in that a completed act describes an accomplished change of state in the world, whereas the result of the ongoing action represented in 31b is still not known. Since a prototypical AGENT engenders some concrete, visible change in the world, the AGENT of 31a is a 'better' (or more prototypical) AGENT than the one in 31b. \ftx \txt To continue with a third example, highlight and jump to the topic below: \cf Hgr_v_cl_ex3 \ftx \dt 27/Apr/1998 \key Hgr_v_cl_ex3 \shd Verb classes-Examples from Tagalog \txt Other languages express information concerning the semantic roles of the arguments of the clause via markers on the verb. Philippine languages are famous for this. For example, in Tagalog, there are several verb forms depending on the semantic role of the absolutive nominal: \ftx \fln (32) a. Humiram ang babae ng pera sa bangko. \ftx A:borrow ABS woman OBL money OBL bank \ftx 'The woman borrowed money from a bank.' \ftx \ftx b. Hiniram ng babae ang pera sa bangko. \ftx P:borrow ERG woman ABS money OBL bank \ftx 'The woman borrowed the money from a bank.' \ftx \ftx c. Hiniraman ng babae ng pera ang bangko. \ftx L:borrow ERG woman OBL money ABS bank \ftx 'The woman borrowed money from the bank.' \ftx \txt Note the different verb forms and prepositional case markers on the noun phrases in each of these sentences. The prepositional case marker 'ang' occurs before absolutive nominals, and the verb inflects for the semantic role of the absolutive. In 32a the verb form 'humiram' says in effect 'the absolutive nominal is the AGENT'. The form 'hiniram' says 'the absolutive nominal is the PATIENT', while the form 'hiniraman' says 'the absolutive nominal is the LOCATION.' Some verbs in Tagalog are said to have up to seven different forms that indicate seven different semantic roles for the absolutive nominal. Many of these constructions may be insightfully analyzed as applicatives; (Hval_incr_appl). \ftx \fln References: Fillmore 1968, 1977, 1977, J. Anderson 1977. Dowty 1987, \ftx Vendler 1967, Schachter 1977. Foley and VanValin 1984. \ftx \dt 01/May/1998 \key Hgr_v_cl_fact \shd Factives \txt Factive verbs are those that describe the coming into existence of some entity, e.g. build, ignite, form, create, make, gather as in 'a crowd gathered', etc. \ftx \dt 27/Apr/1998 \key Hgr_v_cl_manip \shd Manipulation \txt Force, oblige, compel, urge, make, cause, let, allow, permit, forbid, (see Hval_incr_caus -- Causatives). \ftx \key Hgr_v_cl_pos \shd Position \fln Stand, sit, crouch, kneel, lie down, hang, float, etc. \ftx \key Hgr_v_cl_sem \shd Verb classes-Semantic roles \txt SEMANTIC ROLES are conceptual relationships in the 'message world' (see Qintro_embod_play_A). Though they influence the morphosyntax profoundly, they are not primarily morphosyntactic categories. They are part of the 'content' of linguistic messages rather than categories of linguistic form. Semantic roles are the roles that participants play in message world situations, quite apart from the linguistic encoding of those situations. So, for example, if in some real or imagined situation, someone named John purposely hits someone named Bill, then John is the AGENT and Bill is the PATIENT of the hitting event, regardless of whether any observer ever utters a sentence like 'John hit Bill' to describe that event. \ftx \txt Recent works that have influenced linguists' thought on semantic roles most profoundly have been those within the framework of CASE GRAMMAR (Fillmore 1968, Anderson 1977, inter alia). A great many works build upon the insights in these core articles. In this section I will present the concepts of 'classical' case grammar, embedded within a general functionalist perspective. Field workers with interest in pursuing specific proposals for formulating the relationship between semantic roles and grammatical relations are heartily encouraged to consult the references provided. \ftx \txt In the tradition of case grammar, semantic roles are referred to as 'cases' or DEEP CASES. We will avoid this terminology as it conflicts with our notion of cases as being morphosyntactic rather than semantic categories of nominals (Hgrel_val_en). In the tradition of generative grammar they have come to be called THEMATIC ROLES or simply THETA ROLES. Again, this terminology conflicts with the term 'thematic structure' that we will use in a very different sense in discussing the structure of discourse (Hdisc_cnt_thm). The term 'semantic role' is the most unambiguous and widely understood terminology available. Nevertheless, the fieldworker must be aware of the alternate terminologies. \ftx \dt 27/Apr/1998 \key Hgr_v_cl_sem_com \shd Verb classes-common Semantic roles \txt Here we will describe some semantic roles languages are usually sensitive to. These are also described by Comrie 1981:52-53. The semantic roles to be discussed are AGENT, FORCE, INSTRUMENT, EXPERIENCER, RECIPIENT and PATIENT. These are the semantic roles most often embodied by the grammatical relations of subject, object and indirect object in natural languages. Other semantic roles, e.g. LOCATION, DIRECTION, SETTING, PURPOSE, TIME, MANNER etc. are more likely to be embodied in oblique (adpositional) phrases or adverbials, though even these can at times be embodied by subjects or objects, e.g. 'He swam the channel' (channel = LOCATION of swimming), 'We did Norway last summer' (Norway = SETTING). \ftx \txt To continue this discussion, highlight and jump to the topic below: \cf Hgr_v_cl_sem_ag \ftx \dt 27/Apr/1998 \key Hgr_v_cl_sem_com_ag \shd Verb classes-common Semantic roles-Agent \ftx \shd2 Agent in English \ftx \txt An AGENT is 'the typically animate perceived instigator of the action' (Fillmore 1968). \ftx \txt NOTE: See Delancey (1990) for an alternative definition of AGENT. I believe Delancey's definition of AGENT as 'the first CAUSE in the clause' is esse ntially compatible with Fillmore's definition plus the notion of 'message world'. That is, the clause is the linguistic unit within which message wor ld scenes are perspectivized. In so far as the 'instigator of the action' is equivalent to the 'first CAUSE', and the message world 'scene' is equiva lent to the 'clause', the two definitions become near restatements of one another. Foley and Van Valin (1984) describe a functional continuum between two 'macroroles', ACTOR and UNDERGOER. The prototypical ACTOR is an AGENT and the prototypical UNDERGOER a PATIENT in the classic case grammar sense . This is their method of preserving an objectivist definition of AGENT and PATIENT while still accounting for the variability in grammatical coding of these roles. \ftx \txt In the scenes embodied by the following sentences, Percival is the AGENT: \ftx \fln (15) a. Percival ate beans. \ftx b. Percival ran around the block. \ftx c. That vase was broken by Percival. \ftx d. Percival saw the accident. \ftx e. Who did Percival kiss? \ftx f. It was Percival who deceived the President. \ftx \txt A prototypical AGENT is conscious, acts with volition; (i.e. on purpose), and performs an action that has a physical, visible effect. It is a powerful controller of an event. According to this characterization, Percival in 15a and c is a near prototypical AGENT. In b, although Percival is conscious and presumably acts with volition, there is no visible change in the world that results from Percival's act. The same sort of observation can be made for 15d, e and f. Hence, Percival is a less than prototypical AGENT in 15b, d, e and f. \ftx \shd2 Agent in Guaymí \ftx \txt Some languages may actually grammaticalize actions where there is no visible effect on the world differently from the way it treats those such as the AGENT in 15a above: \ftx \fln (16) Guaymí: a. Dorí-gwe Toma dëma-ini \ftx Doris-AGT Tom greet-PAST \ftx `Doris greeted Tom.' \ftx \ftx b. Toma-e Dori hatu-aba \ftx Tom-DAT Doris see-PAST \ftx `Tom saw Doris.' \ftx \txt Prototypical AGENTS in Guaymí are often coded with the morphological case signalled by '-gwe'. The most agent-like argument of the verb 'hatu'-- 'see', however, takes the dative case. This is because the entity that receives the visual impression in a scene described by the verb 'hatu' is not a prototypical AGENT. Rather it is an EXPERIENCER (see below). Other verbs whose most agent-like arguments appear in non-AGENTIVE morphological cases in Guaymí include `forget', `remember', `go' and `come'. \ftx \dt 27/Apr/1998 \key Hgr_v_cl_sem_com_exp \shd Verb classes-common Semantic roles-Experiencer \txt An EXPERIENCER neither controls nor is visibly affected by an action. Normally an EXPERIENCER is an entity that receives a sensory impression, or in some other way is the locus of some event or activity that involves neither volition nor a change of state. For example, in the following English sentences, Lucretia is an EXPERIENCER: \ftx \fln (19) a. Lucretia saw the bicycle. \ftx b. Lucretia broke out in a cold sweat. \ftx c. The explosion was heard by Lucretia. \ftx d. What did Lucretia feel? \ftx e. It was Lucretia who smelled the bacon first. \ftx \dt 27/Apr/1998 \key Hgr_v_cl_sem_com_force \shd Verb classes-common Semantic roles-Force \txt A FORCE is an entity that instigates an action, but not consciously or voluntarily. For example, the wind is a FORCE in the following sentences: \ftx \fln (17) a. The wind opened the door. \ftx b. The wind blew in through the open window. \ftx c. That vase was broken by the wind. \ftx d. What did the wind knock over? \ftx e. It was the wind that formed those rocks. \ftx \dt 27/Apr/1998 \key Hgr_v_cl_sem_com_gua \shd Verb classes-common Semantic roles-Agent in Guaymí \txt An AGENT is 'the typically animate perceived instigator of the action' (Fillmore 1968). \ftx \txt NOTE: See Delancey (1990) for an alternative definition of AGENT. I believe Delancey's definition of AGENT as 'the first CAUSE in the clause' is esse ntially compatible with Fillmore's definition plus the notion of 'message world'. That is, the clause is the linguistic unit within which message wor ld scenes are perspectivized. In so far as the 'instigator of the action' is equivalent to the 'first CAUSE', and the message world 'scene' is equiva lent to the 'clause', the two definitions become near restatements of one another. Foley and Van Valin (1984) describe a functional continuum between two 'macroroles', ACTOR and UNDERGOER. The prototypical ACTOR is an AGENT and the prototypical UNDERGOER a PATIENT in the classic case grammar sense . This is their method of preserving an objectivist definition of AGENT and PATIENT while still accounting for the variability in grammatical coding of these roles. \ftx \txt In the scenes embodied by the following sentences, Percival is the AGENT: \ftx \fln (15) a. Percival ate beans. \ftx b. Percival ran around the block. \ftx c. That vase was broken by Percival. \ftx d. Percival saw the accident. \ftx e. Who did Percival kiss? \ftx f. It was Percival who deceived the President. \ftx \txt A prototypical AGENT is conscious, acts with volition; (i.e. on purpose), and performs an action that has a physical, visible effect. It is a powerful controller of an event. According to this characterization, Percival in 15a and c is a near prototypical AGENT. In b, although Percival is conscious and presumably acts with volition, there is no visible change in the world that results from Percival's act. The same sort of observation can be made for 15d, e and f. Hence, Percival is a less than prototypical AGENT in 15b, d, e and f. \ftx \ftx Verb classes-common Semantic roles-Agent in Guaymí \txt Some languages may actually grammaticalize actions where there is no visible effect on the world differently from the way it treats those such as the AGENT in 15a above: \ftx \fln (16) Guaymí: a. Dorí-gwe Toma dëma-ini \ftx Doris-AGT Tom greet-PAST \ftx `Doris greeted Tom.' \ftx \ftx b. Toma-e Dori hatu-aba \ftx Tom-DAT Doris see-PAST \ftx `Tom saw Doris.' \ftx \txt Prototypical AGENTS in Guaymí are often coded with the morphological case signalled by '-gwe'. The most agent-like argument of the verb 'hatu'-- 'see', however, takes the dative case. This is because the entity that receives the visual impression in a scene described by the verb 'hatu' is not a prototypical AGENT. Rather it is an EXPERIENCER (see below). Other verbs whose most agent-like arguments appear in non-AGENTIVE morphological cases in Guaymí include `forget', `remember', `go' and `come'. \ftx \dt 27/Apr/1998 \key Hgr_v_cl_sem_com_instr \shd Verb classes-common Semantic roles-Instrument \txt An INSTRUMENT is an entity that instigates an action indirectly. Normally an AGENT acts upon an INSTRUMENT and the INSTRUMENT affects the action. For example, in the following sentences, a hammer is an INSTRUMENT: \ftx \fln (18) a. Prescott broke the window with a hammer. \ftx b. A hammer broke the window. \ftx c. That window was broken by a hammer. \ftx d. What did Prescott break with a hammer? \ftx e. It was a hammer that Prescott broke the window with. \ftx \dt 27/Apr/1998 \key Hgr_v_cl_sem_com_pat \shd Verb classes-common Semantic roles-Patient \txt PATIENT is the unmarked semantic role. If an entity does not act with volition, instigate an event, receive something, or experience a sensory impression, it is probably a PATIENT. In the following sentences, Joaquin is the PATIENT: \ftx \fln (22) a. Montezuma stabbed Joaquin. \ftx b. Joaquin fell from the third floor. \ftx c. Joaquin was amazed by the mosquito. \ftx d. Who wanted Joaquin? \ftx e. It was Joaquin that the republicans believed. \ftx \txt A prototypical PATIENT undergoes a physical, visible change in state. In 22a and b Joaquin is a fairly prototypical PATIENT. In 22c-e Joaquin does not undergo a change in physical state. English, however, treats these as 'the same' as more prototypical PATIENTs. Some languages (e.g. Guaymí and Gujarati, see below) treat non-affected PATIENTs differently than affected PATIENTs. \ftx \dt 27/Apr/1998 \key Hgr_v_cl_sem_com_recip \shd Verb classes-common Semantic roles-Recipient \txt A RECIPIENT is the typically animate destination of some moving object. The difference between RECIPIENT and DESTINATION is similar to, but more subtle than, the difference between AGENT and FORCE. Consequently in many languages, the forms used for DESTINATIONS are similar to those used for RECIPIENTS. For example, English uses the preposition 'to' to mark both roles: \ftx \fln (20) a. I sent the book to Mary. (Mary = RECIPIENT) \ftx b. I sent the book to France. (France= DESTINATION) \ftx \txt Evidence that the two roles have been grammaticalized separately (made into separate structural categories) in English is that 21a can undergo a 'dative shift' operation, whereas 21b cannot: \ftx \fln (21) a. I sent Mary the book. \ftx b. *I sent France the book. \ftx \dt 27/Apr/1998 \key Hgr_v_cl_sem_encod \shd Verbs--Encoding Semantic Roles \txt The linguistic encoding of semantic roles The central contribution of case grammar was to observe that semantic roles do not correspond directly to grammatical relations. For example, in the following sentences the formal category of subject (as manifested by preverbal position, pronominal form and potentially verb agreement in English) realizes three distinct roles: \ftx \fln (23) a. I opened the door with the key. SUBJECT = AGENT \ftx b. The key opened the door. SUBJECT = INSTRUMENT \ftx c. The door opened. SUBJECT = PATIENT \ftx \txt Furthermore, the argument represented by the noun phrase 'the key' is an oblique phrase in sentence a and a subject in sentence b, even though it fills the same semantic role in both sentences. Similarly, 'the door' is the direct object in a and b, but subject in c, even though it is the semantic PATIENT in all three sentences. The determination of which participant becomes subject, then, is a matter of perspectivization; (Fillmore 1976). For example, sentences 22a, b, and c, could all be descriptions of the same message world situation, but from different perspectives. \ftx \dt 27/Apr/1998 \key Hgr_v_cl_sem_univ \shd Verb classes-Universal semantic roles \txt In the broad tradition of case grammar, there have been several attempts to formulate a list of universal semantic roles that languages pay attention to morphosyntactically, e.g., the references cited above and Chafe 1970, among others. Such lists have consistently proven inadequate -- as soon as a list is published, some language is demonstrated to grammaticalize a new semantic role. This problem stems largely from confusion between conceptual and morphosyntactic categories. Semantic roles are conceptual, hence infinitely variable. Languages respond to this continuum morphosyntactically in many, if not an unlimited number of, different ways. The question of determining a list of universal semantic roles, then, becomes one of how fine a level of analysis is appropriate. Ultimately, every semantic role played by every participant in every message world situation is subtly different from every other one. However, an infinitely long list of semantic roles is as valueless as no list at all. The important question for descriptive linguists is how is the morphosyntax of the language being described sensitive to semantic roles? \ftx \dt 27/Apr/1998 \key Hgr_v_cl_sem_vsrel \shd Verb classes-Semantic roles versus grammatical relations \txt As field linguists we should not be surprised that semantic roles do not correspond directly to grammatical relations. This is because semantic roles are conceptual notions whereas grammatical relations are morphosyntactic. A theoretical principle often reiterated in this manual is that morphosyntax 'discretizes' (imposes discrete categories upon) conceptual space. Whereas conceptual (sometimes 'functional') domains are continuous and not necessarily inherently categorized, language of necessity exhibits discrete categorization. This is because the human mind cannot function adequately with infinite variability (see the references cited for further discussion). From this point of view, a semantic role such as AGENT is not a discrete category. Rather it defines one extreme of a continuum. Any given participant in any given situation may be more or less AGENT-like. The following diagram illustrates this continuum: \ftx \fln Conceptual domain: AGENT PATIENT \ftx |-------------------------------------| \ftx \fln Morphosyntactic |--¦--------¦----------------¦--------| \ftx domain: ( A )( B )( C ) \ftx \txt In this diagram, '¦' indicates the focal coding points of three hypothetical morphosyntactic categories that function along the AGENT-PATIENT continuum. Parentheses indicate inexact boundaries. That is, formal structures are not direct 'mappings' from conceptual categories. Rather, morphosyntactic devices are associated with certain conceptual focal points. Concepts that occur right at these points are very 'easy' to code. As concepts diverge from the focal point (or 'prototype') the choice of which coding device to use becomes less clear. \ftx \dt 27/Apr/1998 \key Hgr_v_cl_sens \shd Sensation \fln see, hear, feel, taste, sense, observe, smell, perceive.... \ftx \key Hgr_v_cl_st \shd States \txt These may very well have already been covered under predicate adjectives or predicate nominals. This section is only for languages that have a class of STATIVE VERBS, e.g. (be) hot/cold (not weather), broken, rotten, melted, skinned, dead, alive, born, unborn. \ftx \dt 27/Apr/1998 \key Hgr_v_cl_utt \shd Utterance \txt Speak, talk, say, tell, ask, answer, shout, yell, whisper, call, assert, imply, state, affirm, declare, murmur, babble, converse, chat, discuss, sing(?). These verbs (and others) are often onomatopoeic expressions (see Hdisc_symb). As such they may exhibit irregular phonological and/or morphological characteristics. \ftx \dt 27/Apr/1998 \key Hgr_v_cl_weath \shd Weather verbs \txt To rain, (be) wind(y), (be) cold, (be) hot, to thunder, to flash (as lightning), (get) cloudy, (be) dark, (be) light, to dawn. \ftx \key Hgr_v_cl__inv \shd Involuntary processes \fln These are uni-valent verbs whose subjects \ftx \ftx a) undergo a change in state \ftx b) do not act with volition, \ftx c) do not necessarily move through space, and \ftx d) are not the source of some moving object, \ftx \txt For example, the intransitive versions of grow, die, melt, wilt, dry up, explode, rot, tighten and break would belong to this class. These verbs answer the question, 'What happened to X?', but less easily 'What did X do?': \ftx \ftx What happened to Sylvan? He died. \ftx What did Sylvan do? ??He died. \ftx What happened to the mustard? It dried up. \ftx What did the mustard do? ??It dried up. \ftx \dt 27/Apr/1998 \key Hgr_v_strc_A \shd Verb or verb-phrase structure \fln Give charts of the various paradigms, e.g. person marking, \ftx tense/aspect/mode (TAM) etc. Indicate major allomorphic variants. \ftx \key Hgr_v_strc_B \shd Verb or verb-phrase structure \ftx \ftx \key Hgr_v_strc_C \shd Verb or verb-phrase structure \ftx \ftx \key Hgr_v_strc_D \shd Verb or verb-phrase structure \ftx \ftx \key Hgr_v_strc_E \shd Questions to answer for all verbal operations \txt The following is a list of operations that are typically associated with verbs. References are made to the sections of the manual in which these operations are discussed in more depth. \ftx \ftx Verb agreement/concord (Hgrel_vcode) \ftx Semantic role markers (applicatives) (Hval_incr_appl) \ftx Valence Increasing Devices (Hval_incr) \ftx Valence Decreasing Devices (Hval_decr) \ftx Tense/aspect/mode (Hv_tam) \ftx Evidentials (Hv_evid_intro) \ftx Location and direction (Hv_loc) \ftx Speech-act markers (Hprag_ndecl) \ftx Verb(-phrase) Negation (Hprag_neg_intro) \ftx Subordination/nominalization (Hcplx_intro, Hv_nz) \ftx Switch-reference (Hcplx_med_sw1) \ftx \dt 27/Apr/1998 \key Hintro_cncsl \shd Conclusion \txt In summary, language is both an adaptive functional organism and a formal system of representation. Any approach to linguistic description must be cognizant of both of these properties. The process of coming to understand the grammatical system of a language involves several interchanges, all of which can be characterized as a manifestation of the adaptive functional and formal systematic properties of language. The particular interchanges that have been mentioned in this chapter include 1) function and form 2) assigning language-specific linguistic phenomena to universal categories and elaborating the set of possible universal categories, and 3) text and elicited data. As a linguistic fieldworker my understanding of the formal systematic properties of language must be informed by an understanding of the purposes to which language is applied and the human environment in which it exists. Similarly, my understanding of the functions of particular morphosyntactic forms must be informed by an understanding of the ways in which those forms relate to one another in the formal system of the language. My understanding on either front is enriched as I concentrate on understanding the other. \ftx \txt This concludes this discussion, close each jump window until the main calling window is active, then jump to the next topic. \dt 27/Apr/1998 \key Hintro_data \shd On Elicited and Text Data \txt Both text and elicited data are essential to good descriptive linguistics. They each have advantages and disadvantages. The field linguist needs to be aware of these in order to make the best use of all the data available. Even as a fork is no good for eating soup, and a spoon is awkward for eating steak, so elicited and text data each have their own areas of usefulness. The field linguist will be handicapped in conceptualizing a linguistic system if he or she attempts to use one type of data to accomplish a task best performed by the other type. \ftx \txt In the following sections, I will first define and present some characteristics of text and elicited data. Then I will list the areas of linguistic analysis that each type of data is best suited to. Finally, I will suggest some ways in which text and elicited data might be managed in the course of a linguistic field program. \txt To continue this discussion, highlight and jump to the topic below: \cf First Set: Hintro_data_def \cf Second Set: Hintro_data_dataman \txt To conclude this discussion, highlight and jump to: \cf Hintro_cncsl \dt 27/Apr/1998 \key Hintro_data_dataman \shd Suggestions for managing text and elicited data \txt In all of these areas there should properly be an 'interchange' between elicitation and text. One excellent method of learning a language is to start with a well-transcribed text (sometimes this is not obtainable until the phonological system has been assimilated, i.e. several months into the field program). The linguist and the consultant then go over the text sentence by sentence, with the consultant commenting on the meanings of each sentence (this scenario assumes a bilingual but not necessarily literate consultant). The linguist takes notes on these comments in the margins of the printed text and elicits utterances around the sentences that appear in the text. For example, if the meaning of a particular morpheme is not clear, the linguist may ask if the sentence is possible without that morpheme. What, according to the consultant's interpretation, semantic nuances change when the morpheme is removed? Can different word orders be employed? What would the speaker have meant if he/she had said ACB instead of ABC? \ftx \txt All utterances elicited in this way should be clearly marked as elicited in whatever filing system is employed. Proposed semantic or pragmatic nuances should also be checked carefully with other consultants. The first inclination for many consultants regarding grammatically acceptable variants of a sentence is to say 'they mean the same thing'. Needless to say, the linguist should not take a consultant's first attempt at contrastive semantic shadings as definitive. Some consultants are better than others at introspecting about their language and operating in hypothetical communicative situations. Also, some linguistic alternations have no consistent semantic effects. They either really do 'mean the same thing' or their semantic differences vary from context to context, speaker to speaker, or even day to day for the same speaker. \ftx \txt I would suggest beginning fieldwork in a language with a heavy emphasis on elicitation, moving towards a greater reliance on text material as the fieldworker begins to internalize the systematic properties of the language. Perhaps a rule of thumb would be to begin with 90% elicited data, and 10% text data, then move gradually to 90% text data and 10% elicited data sometime in the second year. Consistent with this progression, the fieldworker should begin by studying the systematic aspects of language and gradually move toward the less systematic, more idiosyncratic aspects (see above). \ftx \txt Text data should be distinguished from elicited data in whatever cataloguing system is employed. The functions of these two types of data are so different that they should be kept formally distinct as much as possible. In an automated filing system, one can either mark each record as elicited or text, or one can keep elicited data in a completely different database from text data. I have done it both ways. In my text database I have 'comment' records interspersed with the records that constitute the body of the text. Each comment has the same record number as the record it is a comment about, with the addition of the characters 'cm N' where N is a number. The characters 'cm' simply identify the record as an elicited sentence - not part of the text - while the number allows multiple comments on any given text record. For example, the reference field containing 'FAO16.1 cm 1' indicates that this record is the first comment record attached to the record FAO16.1. If I want to just look at or print the text, I can filter out all records that contain 'cm' in the reference field. I also have another entire database set up for elicited data. These files are distinguished by their filenames from the files containing text data. \ftx \txt Text and elicited data are both essential to a well-rounded field program. Each is useful for particular purposes. This functional difference makes a formal distinction between the two types of data essential. \txt This concludes this discussion, close each jump window until the main calling window is active, then jump to the next topic. \dt 27/Apr/1998 \key Hintro_data_def \shd Definitions \txt I will use the word 'text' to mean any sample of language that accomplishes a non-hypothetical communicative task. By contrast, 'elicitation' (or 'elicited data') refers to samples of language that accomplish hypothetical communicative tasks. \ftx \txt The task of elicited language samples is to fulfill a meta-linguistic request on the part of the linguist, e.g. 'How do you say "dog"?' The response would not actually refer to any concept, either referential or non-referential. No particular dog or characteristic of dogs in general would be communicated. The task of the response would be to accommodate the inquirer by providing a reasonable analog to some hypothetical utterance in another language. So elicited utterances, like all intentional human behavior, do fulfill tasks. It's just that the communicative tasks they fulfill are 'hypothetical', in the sense just described. 'Text' would include, then, some single-sentence utterances, for example greetings. Similarly, 'elicitation' could include multi-sentence language samples. Length of utterance is simply not a defining characteristic of either elicitation or text. Longer utterances are more likely to qualify as text, but there is no necessary connection. My experience is that longer utterances, even when in response to meta-linguistic queries, tend to evolve into real text, as it is difficult for most speakers to maintain a hypothetical perspective on their speech for an extended period of time. Most people need to be taught if they are to speak in terms of hypothetical knowledge. Meta-linguistic queries tend to be interpreted as non-hypothetical, especially when a language consultant is new on the job. For example, I once asked a consultant 'How do you say "Yero kissed Dena"?'. She responded with 'He would never do that!'. Scribner (1979) is a fascinating empirical study of the relation between speech based on general knowledge and speech based on hypothetical knowledge. \ftx \txt To continue this discussion, highlight and jump to the topic below: \cf Hintro_data_prop \dt 27/Apr/1998 \key Hintro_data_prop \shd Properties of text and elicited data \ftx \txt Good text data is uncontrolled, open-ended and dynamic. A text will contain forms that never appear in elicitation. It will also contain forms that appear in elicitation, but in sometimes obviously and sometimes subtly different usages. There is much idiosyncrasy in text. That is, forms are used in novel ways in order to accomplish very specific communicative tasks. Sometimes this is referred to as 'nonce' usages. For example, a sentence like 'he psycho-babbled away our two hour appointment' might arise in a particular communication situation, even though the verb 'to psycho-babble' is probably not a part of the lexicalized vocabulary of most English speakers. One wonders how such a sentence could possibly be elicited! Such idiosyncrasy in text is more common than one might expect and often provides great insights into speakers' ways of thinking and conceptualizing their experience. \ftx \txt In addition to learning the uncontrolled, flexible, idiosyncratic aspects of a language, the fieldworker also needs to be aware of its regular, systematic and predictable aspects. Elicited data is controlled, limited and static. Phonology is probably the most rule-governed and systematic area of language, though even in phonology there is communicationally based and idiosyncratic variation. \ftx \shd2 The value of elicited material \ftx \txt The controlled, systematic and rule-dominated parts of language are best approached with an emphasis on elicited data. These would include: \ftx \ftx 1. Phonology (excluding intonation). \ftx \ftx 2. Morphophonemics \ftx \ftx 3. Inventory of derivational morphology (which derivational \ftx operations apply to which roots, etc.) \ftx \ftx 4. Inflectional inventory (determining the range of \ftx inflectional possibilities for person and number 'agreement' \ftx and case marking) \ftx \ftx 5. Pronoun inventory (isolating the entire set of free pronouns) \ftx \ftx 6. Lexical inventory (acquiring the words for a large number \ftx of culturally significant things and activities). \ftx \shd2 Shortcomings of text \ftx \txt Note that in elicitation there is an emphasis on obtaining 'inventories' of various coding possibilities. This is because languages typically employ a small number of forms in text, though many more forms are possible. For example, a declarative sentence with a second person subject is very rare in texts. This is because people don't often inform other people concerning activities of the person spoken to, e.g., 'You are making bread'. Questions are much more natural in such a context. Nevertheless, a description of the language would be incomplete if the second person declarative forms were missing. \ftx \txt Elicitation is essential to the completion of paradigm charts. Often the meaning of a particular operator is not clear until the entire set of operators that it is in a paradigmatic relationship with is identified. Entire paradigms are rarely obtained by inspection of texts. The same observation can be applied to syntactic constructions. For example, whether a particular transitive construction is a passive or an ergative depends at least partially on whether there exists a corresponding 'active' construction. Similarly, the precise function of SVO word order may not be apparent until minimal pairs with VSO order are obtained. Text data may exhibit SVO and VSO orders, but in text examples there are usually enough other formal differences that the precise contribution of word order to the observed semantic differences is obscured. True minimal pairs are usually obtainable only through elicitation. \ftx \shd2 The value of text \ftx \txt The more pragmatic, semantic and subtle parts of language are best analyzed via a large body of text data, supplemented by elicitation where necessary. This would include: \ftx \ftx 1. Intonation. \ftx \ftx 2. Constituent order. \ftx \ftx 3. Inflectional morphology (determining the precise functions, including \ftx tense/aspect/mode). \ftx \ftx 4. Voice (alignment of grammatical relations and semantic roles of verbal \ftx arguments) \ftx \ftx 5. Sentence level particles (evidentials, validationals and pragmatic \ftx highlighting particles). \ftx \ftx 6. Clause combining (including relativization, complementation, adverbial \ftx clauses and clause chaining) \ftx \ftx 7. Lexical semantics (determining the nuances associated with various \ftx lexical choices, including derivational morphology and pronouns). \ftx \txt To continue this discussion, close each jump window until the main calling window is active, then jump to the topic code indicated beside the "Second Set:" label. \dt 27/Apr/1998 \key Hintro_dling \shd Descriptive linguistics \txt Descriptive linguistics is a technical and a theoretical task. It is technical in so far as it involves interpreting linguistic phenomena and assigning them to appropriately defined categories, e.g., 'first person singular pronoun', 'passive', 'perfective aspect' etc. It is theoretical insofar as it involves the elaboration and definition of the categories themselves. \ftx \shd2 The technical task \txt The technical task of assigning structures to particular theoretical categories thus goes beyond simply labeling units with insightful terminology. Rather, it is a matter of relating a particular structure to a lot of other structures in other languages that fall into the same theoretical category. In some frameworks, such assignment of language particular linguistic units to universal theoretical categories is given the label 'explanation'. But regardless of one's theoretical orientation, any attempt at linguistic description involves this interchange between data and theory. The data control the development of the theory and the theory provides a grid through which data are interpreted and represented. To this extent, then, descriptive linguistics is a technical and a theoretical task. \shd2 Linguistic theory \txt Any linguistic theory aims to provide a minimal set of categories and concepts that characterizes the entire range of structures possible within the phenomenon known as human language. [NOTE: Such categories and concepts can be primitive units such as noun and verb, configurational units such as noun phrase and sentence, relational units such as subject and object, processes such as left-dislocation or principles such as iconicity or the projection principle (Chomsky 1982).] \ftx \txt For example, Relational Grammar (RG) provides a list of well-formed 'relational networks'. The structure of any clause in any language should be represented by one element of this set of networks (Perlmutter 1980). The task of the linguistic technician working in an RG framework, then, is to determine which relational network (best) represents the structure of each clause type in the language being described. The technician becomes a theoretician exactly when he or she realizes that the set of theoretical categories (in this case relational networks) currently available is not adequate. That is, when linguistic data require the sanction of a relational network previously considered not well-formed, or when a new technical entity is needed to express some morphosyntactic fact, then the theory must be revised. Once the theory (i.e. the set of categories and concepts available to the linguistic technician) has been revised, it is taken back to the data for verification and further elaboration. A theory is validated to the extent that it provides the categories, and the mechanisms for developing and evaluating new categories, that consistently represent the structures of human language, without requiring extraneous mechanisms. \txt To proceed to the next topic, use Database Search (ALT-D,S) to go to: \cf Hintro_man_nature \dt 27/Apr/1998 \key Hintro_embod \shd Embodiment \txt In this manual we will often refer to the distinction between CONCEPTS and the FORMS that EMBODY them. \ftx \txt NOTE: The notions presented here are based largely on Lakoff 1988. At times such terms as 'code', 'realize' and even 'represent' may be used more or less synonymously with 'embody'. We trust that this practice will enrich the clarity and readability of the presentation rather than detract from it. \ftx \txt This terminology stems from our assumption that language is a symbolic system. We assume that language consists of elements of form, e.g. sounds, morphemes, words, and sentences, inextricably bound to concepts, i.e. 'meanings'. The concepts that language represents consist of any and every thing human beings care to discuss -- they may be drawn from 'real world' categories, to the extent that such categories are useful in communication. However, they are also drawn from analogical, metaphorical and iconic extensions of basic physical-world categories. This is because human beings universally find a need to reason in terms of categories that aren't necessarily self-evident in the natural world, e.g. 'sincerity', 'government', 'love' etc. We will follow Lakoff (1987) in using the verb 'embody' to refer to the relationship between a form and the concept or concepts it stands for. This terminology acknowledges that form is only loosely associated with meaning. That is, the form-meaning bond is fixed enough to allow communication, but loose enough to provide for creativity and extension of meaning to new areas of conceptual space. \ftx \txt To continue this discussion, highlight and jump to the topic below: \cf First Set: Hintro_embod_msyn \cf Second Set: Hintro_embod_play_A \cf Third Set: Hintro_embod_bldg_A \txt To proceed to the next topic, use Database Search (ALT-D,S) to search for \cf Hintro_data \dt 27/Apr/1998 \key Hintro_embod_bldg_A \shd Building a message is like building a building--the concept \txt At several points in this manual we will describe three methods by which languages accomplish certain tasks. These methods are LEXICAL, MORPHOLOGICAL and ANALYTIC (or PERIPHRASTIC). In describing how these methods are used, it is sometimes helpful to think of the process of building a mental model as similar to the process of building a building. In the rest of this section we will briefly explore this metaphor. \ftx \txt Every building has a unique function. The proposed function of a particular building affects its form from the very earliest design stages and throughout the construction process. If I am building a building, I have at my disposal certain raw materials, and an idea (perhaps a blueprint) of what the building is supposed to become. For our purposes, let us suppose the raw materials are irregularly shaped stones. The construction process consists of taking stones from a resource pile and placing them judiciously into locations in the emerging wall of the building, being conscious at all times of the blueprint (i.e., I don't randomly pile stones together). The locations have certain shapes because of the stones I have already placed there, and the stones in my resource pile also have various shapes. My task is to match resource stones to locations in such a way that the result is a building that serves the intended purpose. \ftx \txt To continue this discussion, highlight and jump to the topic below: \cf Hintro_embod_bldg_B \ftx \dt 27/Apr/1998 \key Hintro_embod_bldg_B \shd Building a message is like building a building--the processes \txt So, what do I do in the many situations where a particular stone does not quite fit the current location? The three main processes, it seems, would be: \ftx \ftx 1. Look for another stone. \ftx 2. Reshape the current stone or the location. \ftx 3. Combine stones. \ftx \txt Other possible processes that have occurred to members of various classes I have taught are: \ftx \ftx 4. Get a stone from a different pile. \ftx 5. Use a lot of mortar. \ftx 6. Change the blueprint. \ftx 7. Give up. \ftx \ftx Perhaps there are others. If this is a good metaphor for the process of \txt building a message, then all (or at least as many as possible) of these procedures should have analogs in message building. \ftx \txt To continue this discussion, highlight and jump to the topic below: \cf Hintro_embod_bldg_C \dt 27/Apr/1998 \key Hintro_embod_bldg_C \shd Building a message is like building a building--the function \txt All messages have a function. This follows from the assumption that language behavior is intentional. People in general do not use language randomly (though I think I know some who do). The form a message takes is affected by the function of the message at every stage. The resources available to the message builder are an idea of what is to be communicated and a store of conventionalized structures, for example vocabulary items. The task of message building involves judiciously fitting together existing structures in a unique way to create the particular message that is needed. At any given point in the process there is a partially completed message and a range of possible structures available to build the message in the intended direction. What processes are available to the builder if a particular unit, say a lexical item, does not fit the message context?: \ftx \ftx 1. Get another lexical item (LEXICAL) \ftx 2. Modify the first one, or the context (MORPHOLOGICAL) \ftx 3. Combine lexical items (PERIPHRASTIC) \ftx \txt To continue this discussion, highlight and jump to the topic below: \cf Hintro_embod_bldg_D \ftx \dt 27/Apr/1998 \key Hintro_embod_bldg_D \shd Building a message is like building a building--example \txt If my message requires the semantic notion of CAUSE(x, DIE(y)), i.e. something caused something else to die (see Hval_p-calc for an explanation of the predicate calculus notation), English offers me two strategies for accomplishing that function. I can use the verb kill or I can use the expression 'cause to die.' Kill is a lexical solution to the problem because it is a lexical item that embodies all of the needed information within its conventionalized semantic makeup. 'Cause to die', on the other hand, is a periphrastic strategy for accomplishing essentially the same task (though see Hval_incr_caus for a discussion of the functional differences between periphrastic and lexical causatives). Some languages, e.g. Turkish, use a special morpheme to indicate causation. This is added to the root meaning 'die' to form a new root meaning 'kill'. This is a morphological causative. This triad of lexical, morphological and periphrastic strategies is relevant to many different functional tasks in language. Some tasks that are typically accomplished by one strategy in one language, may be accomplished by one of the other strategies in the next language. For example, past tense is encoded morphologically in English via a verb suffix. In other languages the time of a situation is expressed periphrastically via temporal adverbial phrases such as 'two days ago', etc. Furthermore, languages invariably allow certain functions to be accomplished by more than one coding strategy. This is the case with kill and cause to die in English. Usually, however, when such a choice exists there is some slight difference in function between the various coding possibilities. \ftx \txt To continue this discussion, highlight and jump to the topic below: \cf Hintro_embod_bldg_E \ftx \dt 27/Apr/1998 \key Hintro_embod_bldg_E \shd Building a message is like building a building--other strategies \txt What about the other possible strategies for building a building? Are there relevant analogies in message building to the strategies suggested by students for building a stone building? How good is this metaphor anyway? \ftx \txt 1. Go to a different pile. How about going to the lexicon of some other language? Sometimes the right English word just doesn't come to mind at the right time, but there is a perfect Spanish word just waiting to be used. If I judge my interlocutor will understand, I may just use that Spanish word. This is referred to as CODE SWITCHING and is probably more common than we might suspect around the world, as most societies are multi-lingual. \ftx \txt 2. Use a lot of mortar. Well, er, um, I dunno, maybe. . . I mean it's like you just kinda slop your message together a little, ya know? \ftx \txt 3. Change the blueprint. Conceivably I may have so much trouble expressing a particular message, that I may just decide to say something a little different instead. \ftx \txt 4. Give up. This strategy has a direct analog in message building. \ftx \dt 27/Apr/1998 \key Hintro_embod_grnd \shd Grounding \txt In describing the semantics of morphosyntactic operations, we refer to some operations as GROUNDING the situation described by the clause. This term refers to operations that establish the time, location or actuality of a situation according to some reference point. For temporal operations the reference point is normally 'now', i.e. the time of utterance. For spatial operations the reference point is normally 'here', i.e. the place of utterance. Some modal and evidential operations ground a situation with respect to 'actuality'. That is, these operations convey, perhaps among other things, the speaker's perception of how likely it is that the situation described by the verb did or will actually take place (Chafe and Nichols 1985). \ftx \txt To continue this discussion, close all jumps until the main edit window is active, then highlight the topic code indicated beside the "Second Set:" label. \ftx \dt 27/Apr/1998 \key Hintro_embod_msyn \shd Morphosyntactic operations and operators \txt A morphosyntactic OPERATION describes a dynamic relation between one linguistic form and another. \ftx \txt NOTE: This concept and terminology is reminiscent of Hockett's (1954) 'item and process' model of grammatical description. Along with most current approaches to grammatical description, we consciously adopt the item and process model, and accept the implications of teleology that go along with it. I have tried to be consistent with terminology used by most linguists in adopting the term 'operation' for particular morphosyntactic devices with conceptual content in specific languages, and reserving the term 'process' for broad strategies for encoding those operations, e.g. plural formation in English is a morphological operation, whereas suffixation is a kind of morphological process. Though terminology varies widely in this area, I believe this distinction, though perhaps confusing to some, will be generally understandable to linguists of various theoretical persuasions. The distinction between an operation and a process becomes more difficult to maintain as one transcends morphology and begins to examine analytic (syntactic) structures. This is due to the fact, mentioned earlier, that syntax is more determined by universal functional principles (such as iconicity) than is morphology. For example, left-dislocation is an appropriate syntactic analogy to prefixation in morphology. It is defined entirely in terms of the structural change involved, and thus is not necessarily tied to any conceptual content in any given language. Nevertheless, there is a distinct universal tendency for left-dislocation to code a particular function, namely contrastive focus (see Hprag_foc). This tendency is so strong that linguists may actually think of left-dislocation as synonymous with the functional term 'contrastive focus'. Many formal devices, such as 'passive', have even been given functional sounding labels precisely because they have such strong associations with particular functions. For these reasons, we will in practice dispense with the operation/process distinction in syntax. \ftx \txt The relation is ordered from morphosyntactically (and hence conceptually) simpler to more complex forms. The simplest forms can be termed ROOTS. In most cases the presence of a morphosyntactic operation is embodied by a formal OPERATOR, e.g. a prefix, a suffix, a stress shift or a combination of two or more of these (see chapter 2, section #2.3 for a detailed exposition of the morphological means that languages employ to code operations). However, some operations exist independently of any overt coding. One method of noting the existence of a morphosyntactic operation that has no overt realization is to posit a zero morpheme. \ftx \txt To continue this discussion, highlight and jump to the topic below: \cf Hintro_embod_grnd \ftx \dt 27/Apr/1998 \key Hintro_embod_play_A \shd Discourse is like a play: scenes, scripts and mental models--frames \txt Throughout this manual we will be using such terms as MESSAGE WORLD and DISCOURSE STAGE in describing the content of linguistic expressions. In other words, we can understand these terms as metaphors for the domain symbolized by language. In this view, discourse depicts a world that can metaphorically be described as a stage on which a play is being acted out. Much research on discourse comprehension and production has used some form of this metaphor to formulate substantive hypotheses and claims. For example, Minsky (1975) used the term FRAMES to refer to stereotyped situations in which knowledge is categorized and stored in memory, e.g., the 'restaurant' frame consists of tables chairs, waiters/waitresses, food, a check, etc. Schank and Abelson (1977), building on Schank (1972), introduced the notion of SCRIPTS. Whereas a frame is a static set of entities in a particular arrangement, e.g., a restaurant, a script is a potentially dynamic set of events and situations, e.g, the process of sitting down, ordering and dining at a restaurant. Fillmore (1976, 1977) suggests that verbs with their unique case frames activate SCENES in the minds of language users. Lakoff's (1987) notion of COGNITIVE MODEL is an extension and elaboration of the notion of scene. What frames, scripts, scenes and cognitive models have in common is that all are idealized mental structures, 'pictures' if you will, that the human mind uses to categorize and store experience and knowledge. These approaches capture the fact that all knowledge is acquired and stored relative to a context. One way of thinking about that context is in terms of a dramaturgical metaphor. \ftx \txt To continue this discussion, highlight and jump to the topic below: \cf Hintro_embod_play_B \ftx \dt 27/Apr/1998 \key Hintro_embod_play_B \shd Discourse is like a play: scenes, scripts and mental models--mental models \txt Johnson-Laird (1981) uses the term 'mental model' to refer to the specific 'picture' that is developed in the process of discourse. Mental models are different from frames, scenes, scripts and cognitive models in that mental models are specific and detailed. The technical term for this kind of model is 'enriched'. For example, a speaker may wish to describe a recent personal experience. Since the experience occurred in a restaurant, the speaker can use an idealized restaurant scene to ground his discourse, since he assumes the hearer also has mental access to a roughly comparable scene. So this shared mental background serves as the 'stage' on which the speaker produces a drama. Similarly, if the experience involved some standard events, the speaker can use an idealized restaurant 'script' as the backbone for his specific message. However, the message would hardly be worth communicating if it consisted of nothing more than idealized event sequences in a stereotypical scene. The specific details of the experience must be superimposed on the idealized, shared, cognitive images. It is those details and variations that make each particular message communicationally valid. A mental model, in the view of Johnson-Laird, is the individualized model of the message being constructed in a specific communicational act. \ftx \txt To continue this discussion, highlight and jump to the topic below: \cf Hintro_embod_play_C \ftx \dt 27/Apr/1998 \key Hintro_embod_play_C \shd Discourse is like a play: scenes, scripts and mental models--summary \txt The following chart summarizes some of the major contributions to this area of research: \ftx \ftx Idealized Enriched \ftx \ftx Frames (Minsky 1975) Mental Models (Johnson-Laird 1981) \ftx Scenes (Fillmore (1976) \ftx Scripts (Schank and Abelson 1977) \ftx Cognitive Models (Lakoff 1987) \ftx \txt To continue this discussion, highlight and jump to the topic below: \cf Hintro_embod_play_D \ftx \dt 27/Apr/1998 \key Hintro_embod_play_D \shd Discourse is like a play: scenes, scripts and mental models--comprehending \txt The processes of producing and comprehending discourse can be understood as building and inferring mental models. Simply put, the speaker has a picture (an image) in mind that he or she intends to establish in the hearer's mind. Similarly, a cooperative hearer intends to recreate in his own mind the mental picture being constructed by the speaker. Communication is successful when the hearer's mental model is roughly equivalent to the speaker's. Both sides of the process use stereotypical mental structures to assist in their respective tasks, and both tasks involve active participation. In fact, it is arguably the case that mental models evolve in the discourse process. That is, there is no predetermined mental model that 'the speaker' tries to construct for 'the hearer'. Rather all discourse participants cooperate (or compete) to create a unique mental model. This view is particularly appealing in the analysis of conversation. \ftx \txt To continue this discussion, highlight and jump to the topic below: \cf Hintro_embod_play_E \ftx \dt 27/Apr/1998 \key Hintro_embod_play_E \shd Discourse is like a play: scenes, scripts and mental models--cross-cultural implications \txt If comprehension indeed depends on stereotypical mental images shared within a community, it should be obvious that study of scenes, scripts and cognitive models would be invaluable in cross-cultural studies and translation. Cross-cultural misunderstanding often occurs when communicators implicitly (and unconsciously) assume that their interlocutors share their own basic conceptualizations. For example, when dealing with Yagua medical problems, I assumed that my conceptualization that medicine ingested through the mouth would enter the bloodstream and circulate throughout the body would be shared by the Yaguas. It did not occur to me that anyone would have any other concept. However, I discovered that more than once oral medicines I had prescribed were being crushed and applied topically to the area of infection or pain. Even though I would try to explicitly instruct patients to take the medicine orally, their cognitive model of sickness and healing would not admit the possibility that this mode of administration would be effective. \ftx \txt To continue this discussion, close all jumps until the main edit window is active, then highlight the topic code indicated beside the "Third Set:" label. \ftx \dt 27/Apr/1998 \key Hintro_man_approach \shd The Linguistic Field Manual--which approach should we use? \txt A partial solution to the dilemma inherent in a choice between form-driven and function-driven modes of linguistic description is available. That is, begin with a form-driven approach in the areas where languages most obviously vary, namely the traditional "lower levels" of structural analysis. Then adopt a function-driven approach for those particular areas for which structural parameters have been well-established. These typically occur at the "higher" levels of analysis, the clause, sentence and discourse. \ftx \txt To continue this discussion, highlight and jump to the topic below: \cf Hintro_man_over \ftx \dt 09/May/1998 \key Hintro_man_early \shd The Linguistic Field Manual--Earlier descriptive grammars \txt Up until the advent of typological-functional linguistics, a descriptive grammar would typically consist of chapters representing various structurally defined categories, e.g., word, phrase, clause, sentence, or alternatively, 'phonology', 'morphology', and 'syntax'. The earlier grammars concentrated on the 'lower level' phenomena (phonology and morphology) while later grammars increasingly paid attention to syntax and even discourse structure. Each chapter would describe processes that occurred at these various levels of structure. For example, all morphological operations would be treated in the 'word' or 'morphology' chapter. All analytic operations would be treated in the 'phrase', 'clause' or 'syntax' chapters. \ftx \txt This organizational principle was based on the idea that 'comparison and contrast' were the only reliable means of determining what categories were relevant to a given language. These notions were of central importance to classical structuralism throughout the social sciences. In order for units to be compared and contrasted meaningfully, they must be as nearly identical as possible. The idea is that within any system of units in relationship with one another, a unit can only be considered 'real' if it can be demonstrated to contrast with all other units, other factors being equal. In linguistic description, one had to isolate just those differences among units that were important to the underlying grammatical system. \ftx \txt To continue this discussion, highlight and jump to the topics below: \cf Hintro_man_early_ex-I \cf Hintro_man_early_ex-II \txt To Return to the main topic for this, highlight and jump to: \cf Hintro_man_nature \dt 09/May/1998 \key Hintro_man_early_ex-I \shd The Linguistic Field Manual--Example of earlier techniques-I \txt The idea of contrast proved very useful in phonological analysis. For example, one would not compare the sound [p] with the word 'antidisestablishmentarianism', as these are units of different 'levels' of structure, and hence not directly comparable. [p] could only be compared with other segmental sounds, e.g., [b]. If [p] and [b] contrast, this would indicate that they belong to different units in the phonological system and that 'voicing' is an important distinction within that system. The only sure way to prove that sounds are distinct is to show two different words that contrast only with respect to the two sounds in question. Such is the case with [p] and [b] in English: \ftx \fln (1) [pæt] 'pat' \ftx [bæt] 'bat' \ftx \txt If there are no such pairs, the two different sounds may simply be two random or conditioned variants of the same underlying unit. Such is the case of [p] and aspirated [ph] in English: \ftx \fln (2) [tæp] 'tap' \ftx [tæph] 'tap' \ftx \txt To continue this discussion, highlight and jump to the topic below: \cf Hintro_man_early_ex-II \ftx \txt To Return to the main topic for this, highlight and jump to: \cf Hintro_man_early \dt 09/May/1998 \key Hintro_man_early_ex-II \shd The Linguistic Field Manual--Example of earlier techniques-II \txt The insight which comparison and contrast provided proved so useful in phonology that it was extended to the analysis of other areas of linguistic description. For example, in morphology, minimal word forms were compared and contrasted to see if there were any internal or external differences: \ftx \fln (3) [dog] 'dog' \ftx [dogz] 'dogs' \ftx \ftx [bæg] 'bag' \ftx [bægz] 'bags' \ftx \txt These examples would be considered evidence for the following kind of generalization: \ftx \fln (4) NOUN+0 = singular NOUN \ftx NOUN+z = plural NOUN \ftx \txt In other words, -0 and -z are contrasting 'morphemes' because 0 consistently goes with singular and -z consistently goes with plural, all environmental factors being equal. \ftx \txt In syntax, the following two structures would both be analyzable as 'possible clause structures' of English: \ftx \fln (5) NOUN VERB \ftx NOUN 'be' NOUN \ftx \txt Because they contrast structurally and do not freely vary, these represent distinct 'clause types' of English. \ftx \txt Comparison and contrast continue to be important analytic tools for the field linguist. For example, the way we distinguish nouns from verbs in chapter #3 below is by comparing the structural characteristics of a questionable form to those known to hold for forms that are not questionable. \ftx \txt To continue this discussion, close all jumps until the main edit window is active, then highlight the topic code indicated beside the "Second Set:" label. \ftx \txt To Return to the main topic for this, highlight and jump to: \cf Hintro_man_early \dt 09/May/1998 \key Hintro_man_form \shd The Linguistic Field Manual--Form driven grammars \txt A grammatical description that approaches language from the point of view of the different contrasting units at various levels of structure can be termed a form-driven description (Lehmann 19## uses the term #################). \ftx \txt Form-driven grammatical description was consistent with the view prevalent in linguistics and anthropology in the early twentieth century that each language should be analyzed on its own terms. We should have no expectations about what we 'ought' to find upon investigating a newly discovered language. As outsiders, we are prone to impose categories from our own internalized linguistic system and to be blind to categories and distinctions that our own system lacks. This danger certainly was realized in a very obvious way in linguistic descriptions that appeared before the advent of structural linguistics. For example, many of the grammars of Amerindian languages done by missionaries between 1600 and 1900 described the realizations of the seven 'cases' of Latin, even though in fact the languages lacked morphological case marking altogether. \ftx \txt To continue this discussion, highlight and jump to the topic below: \cf Hintro_man_form_prob \txt To Return to the main topic for this, highlight and jump to: \cf Hintro_man_nature \dt 29/Jun/1998 \key Hintro_man_form_prob \shd The Linguistic Field Manual--Problems with form driven grammars \txt There were problems with purely form-driven grammatical descriptions. For example, structuralism in its heyday led to grammatical descriptions that were quite unintelligible to even a highly educated person who was not familiar with the language being described. In striving to keep one's analysis as independent of such 'subjective' variables as meaning, some descriptive linguists assigned numbers rather than meaningful glosses to the morphemes in a language. \ftx \txt Second, similar kinds of meanings are often expressed at different levels of structure. For example, tense in English is accomplished partially morphologically and partially analytically. Because of the requirement that levels of analysis be kept distinct in a structuralist paradigm, present and past tense would have to be treated in the word-level chapter while future tense would be treated in the phrase level or syntax chapter. Thus the notion of a unified 'tense system' would be lost. Similarly, the commonality between anaphoric affixes and anaphoric free pronouns from one language to another, and even within the same language is lost in a purely structuralist treatment. \ftx \txt Third, some important areas of linguistic analysis have ramifications for both morphology and syntax. For example, the notion of 'case' cannot adequately be described on either the word level or the clause level alone. Rather, reference to both must be made in order to capture the generalization, or even write a grammatical rule, that accurately describes the case system of a language. If case is treated as a noun-phrase opposition, the result is a list of the morphological case forms (the "declensions") for the various types of nouns. If case is treated as a syntactic phenomenon alone, then important details about morphological case forms and variations in the general system are lost. The fact is that case is part and parcel of the complex interaction among semantics, pragmatics, syntax and morphology. To try to describe it from any one of these points of view to the exclusion of the others would be to ignore its global significance. \ftx \txt Finally, recent research has shown that the boundary between morphology and syntax is in fact quite arbitrary, and need not necessarily have any implications for linguistic description. \ftx \txt To continue this discussion, highlight and jump to the topic below: \cf Hintro_man_fuzzy \ftx \dt 09/May/1998 \key Hintro_man_func \shd The Linguistic Field Manual--Function driven grammar \txt Within this sub-discipline, a new mode of grammatical description increasingly began to be adopted. This new mode I will term FUNCTION-DRIVEN (corresponding to Lehmann's, 19##, ############). This approach proposes that there are certain 'jobs' that linguistic structures perform (see Qintro_what for further explanation). The particular structural (or astructural) means that languages use to perform those jobs may vary from language to language. Therefore the logical method of approaching linguistic description is to start with a list of jobs (say 'temporal grounding') and describe how each of those jobs is accomplished in the language being described. The jobs are primary. In this kind of grammatical description, past, present and future tense in English would all be treated in the same section of the grammar because all accomplish the task of 'temporal grounding.' Other languages, e.g., Indonesian, accomplish the same task, but in structurally incomparable ways. \ftx \txt To continue this discussion, highlight and jump to the topic below: \cf Hintro_man_func_dangr \ftx \dt 09/May/1998 \key Hintro_man_func_dangr \shd The Linguistic Field Manual--Dangers of the function driven approach \txt There are dangers in a 'pure' function-driven approach. In particular, if we grab a function "out of the air", we are likely to fall into the trap of imposing artificial analyses. For example, if we take the notion of 'causation' as being something that all languages must express in some way, and then look for how that notion is expressed in English, we may include both of the following structures: \ftx \fln (8) a. He made him eat turnips. \ftx b. He ate turnips because the bus was late. \ftx \txt These expressions both contain the notion of 'cause', but something tells us that they are quite different. And in fact, in most (all?) languages there are two different structures that instantiate ideas that approximate these English expressions. The problem is defining what the list of jobs is. It is important to restrict oneself to functions that are well-defined and documented in a variety of languages insofar as possible. `Causation' in this general a sense does not seem to be a very useful concept to descriptive linguists because the structures used to perform this job (e.g. 8a and b) are so varied that there are few, if any, generalizations that hold for all of its instantiations. Hence we have an inherent dilemma: we don't want to be so tied to structure that we fail to recognize significant generalizations that transcend structural boundaries, yet we don't want to be so tied to functions that we artificially suppose commonality even when there is no empirically based reason for doing so (like the early field linguists who found the seven cases of Latin in every language they investigated). \ftx \txt To continue this discussion, close all jumps until the main edit window is active, then highlight the topic code indicated beside the "Fourth Set:" label. \ftx \dt 09/May/1998 \key Hintro_man_fuzzy \shd The Linguistic Field Manual--Example of fuzzy morphology/syntax boundary \txt In Yagua (Peba-Yaguan, Peru) there is a set of forms that function syntactically with whatever follows but which attach morphologically to the item to the left: \ftx \fln (6) Sa-suutá-ra mijay 'He/she washed the mosquito net.' \ftx 3SG-wash-INAN mosquito.net \ftx \txt That the '-ra' morpheme really goes syntactically with the object 'mijay' and is not an 'agreement' marker on the verb is illustrated by the facts that 1) nothing can intervene between '-ra' and 'mijay' and 2) '-ra' attaches to anything that comes after the verb: \ftx \fln (7) Sa-suuta waturá-ra mijay 'The unmarried woman washed the \ftx 3SG-wash unm.woman-INAN m.net mosquito net.' \ftx \txt That the '-ra' morpheme really does attach to the previous word, and is not a free particle, is evidenced by the tone pattern. Note that in 6 the presence of '-ra' triggers penultimate high tone on 'suuta'. In 7 'suuta' takes the normal neutral tone pattern and 'waturára' takes penultimate high tone. These, and many other similar examples from around the world, plus the commonly observed fact that word boundaries in many languages are quite difficult to determine, suggest that morphology and syntax are not universally distinct structural domains in the grammars of languages. Yet due to the 'all-other-things-being-equal' requirement, a strict comparison and contrast mode of linguistic description requires that the two levels be treated separately. \ftx \txt To continue this discussion, close all jumps until the main edit window is active, then highlight the topic code indicated beside the "Third Set:" label. \ftx \dt 09/May/1998 \key Hintro_man_lingstruc \shd The Linguistic Field Manual--possible linguistic structures \txt As linguists increasingly adopted a universalist perspective, a picture began to emerge of the kinds of structures likely to be found in languages. With advances in transportation and communication technology, the structures of many more languages have become known to linguists, to the point where we now have a fairly accurate picture of the kinds of linguistic structures found in every part of the world. This is not to say that there are no surprises lurking out there in the many languages that have yet to be described. In fact a major purpose of this manual is to help field linguists uncover those residual surprises. However, it is certainly the case that the overall picture of what Language is like is a great deal clearer now than it was even twenty years ago. No professional linguist, I believe, would say anymore that languages vary randomly. We know from experience that languages vary along well-defined parameters, and that there are limits to that variation. \ftx \txt To continue this discussion, highlight and jump to the topic below: \cf Hintro_man_approach \txt To Return to the main topic for this, highlight and jump to: \cf Hintro_man_nature \dt 29/Jun/1998 \key Hintro_man_nature \shd The Linguistic Field Manual--It's Nature \txt The outline accompanying this system, beginning with "ethno", represents one possible system of categorizing and describing linguistic structures ("ethno" represents important ethnolinguistic and other background information). The system of organization suggested here may be used to structure the grammatical description of any language, in so far as descriptive linguists find it helpful. However, this system of organization is quite different from that of a traditional descriptive grammar. It differs primarily in that, at least beginning with "nomop" -- Noun and Noun Phrase Operations, it is FUNCTION-DRIVEN rather than FORM-DRIVEN. In the next sections, I will describe this distinction, and the importance it holds within the present state of the art of descriptive linguistics. \txt To continue this discussion, highlight and jump to the topic below: \cf First Set: Hintro_man_early \cf Second Set: Hintro_man_form \cf Third Set: Hintro_man_univ \cf Fourth Set: Hintro_man_lingstruc \cf Fifth Set: Hintro_man_rel \txt To proceed to the next topic, use Database Search (ALT-D,S) to go to: \cf Hintro_what_assump \dt 27/Apr/1998 \key Hintro_man_other \shd The Linguistic Field Manual-- relation to other models \txt The particular organization suggested is one which is consistent with general principles of late 20th century linguistic science. That is, the terms and concepts should be understandable to linguists from a variety of theoretical orientations. As the field linguist works through the grammar of a language using the outline of this manual as a guide, questions will undoubtedly arise as to the appropriateness of particular definitions and interpretations to the language being described. This is good. It is only through honest interaction with data that we learn where our conceptions concerning universal characterizations of linguistic structure need to be revised. In fact, it might be said that one purpose of this manual is to encourage field linguists to find holes in current theoretical understandings of linguistic structures. To the extent that it makes such understandings accessible to the linguistic technician, then it has accomplished its task. \ftx \dt 09/May/1998 \key Hintro_man_over \shd The Linguistic Field Manual-- overview \txt The present manual attempts to provide exactly this sort of combination of form-driven and function-driven approaches to linguistic description. Chapters 2, 3 and 4 are largely form-driven: chapter 2 asks for a description of the overall morphological characteristics of the language, without a detailed exposition of the functions of any particular morphological structures. Chapter 3 suggests a framework for recognizing the structurally defined grammatical categories of the language, again without asking for a detailed exposition of the functions of the morphosyntactic operations used to differentiate grammatical categories. In chapter 4 a framework is provided for determining the "basic constituent order" of a language, according to the classic studies (e.g., Greenberg 1966), but without asking for a detailed description of the function of constituent order variation. Chapter 4 truly represents the interface between the form- and function-driven sections of the grammar. Traditional constituent order typology is quite structural, but the striking generalizations that this line of research uncovered led naturally to a quest for functional explanations for constituent order patterns. \ftx \txt Beginning with chapter 5, the emphasis shifts to a more function-driven perspective. Chapter and section headings reflect common "jobs" that languages are known to perform, and questions lead the user of the manual to discover how the language being described accomplishes those jobs. The help sections contain brief descriptions of how other languages around the world are known to accomplish the function in question. Finally, references are provided to more comprehensive treatments of specific topics, should the user of the manual have interest in developing a particular idea more fully. \ftx \txt To continue this discussion, close all jumps until the main edit window is active, then highlight the topic code indicated beside the "Fifth Set:" label. \ftx \dt 09/May/1998 \key Hintro_man_rel \shd The Linguistic Field Manual--relevancy \txt Every language has its particularly well-elaborated areas of interest and complexity. Some sections of this manual may not be relevant, or may be only marginally relevant to certain languages. Similarly, there may be aspects of a particular language that are not mentioned at all in the manual. Hence it is inevitable that certain sections of a grammar sketch based on this manual be more fully articulated than others. The particular areas of elaboration will vary from sketch to sketch depending on the characteristics of the language and the interests of the fieldworker. This is desirable in that it is what distinguishes each grammar sketch from all the rest. In most cases the areas of a language that are likely to be of especial interest to the linguistic community at large are exactly those areas that the fieldworker finds personally interesting. Hence, the fieldworker is encouraged to use this manual as a general guide as far as there is a felt need for guidance. However, the outline of the manual should not be taken as a prescriptive norm that all grammar sketches should slavishly adhere to. \ftx \txt To continue this discussion, highlight and jump to the topic below: \cf Hintro_man_techn \txt To Return to the main topic for this, highlight and jump to: \cf Hintro_man_nature \dt 29/Jun/1998 \key Hintro_man_sem \shd The Linguistic Field Manual--The importance of semantics \txt Since about 1970, with great involvement of field linguists, semantics and other previously "taboo" subjects increased in importance to linguists of the universalist persuasion. Explanations for structural universals in terms of a purely structural notion of Universal Grammar proved unsatisfactory to those who were engaged in learning and describing "exotic" languages. These field linguists realized that there were commonalities from one language to the next, but rather than explain those commonalities in terms of purely structural principles (e.g., the "Specified Subject Condition", Chomsky 1977), they found explanation in terms of the purposes, or functions, that language serves to be more satisfactory. This movement towards a more functional view of language gave birth to what is known as TYPOLOGICAL-FUNCTIONAL LINGUISTICS. \ftx \txt To continue this discussion, highlight and jump to the topic below: \cf Hintro_man_func \ftx \dt 09/May/1998 \key Hintro_man_strc \shd The Linguistic Field Manual--structure and use \txt Under many of the headings and subheadings there appear questions such as 'how are relative clauses formed?' Answers to these questions could constitute the substantive portions of a grammar sketch or full reference grammar. If the fieldworker understands a question and can relate it immediately to some specific data in the language being described, he or she can simply answer the question and provide examples. In many cases, however, the fieldworker will not necessarily be able to answer the questions completely without consulting some additional reference material. Paragraphs labeled 'help' are designed to provide a more detailed description of the particular linguistic system treated in each subsection of the manual. These help paragraphs normally provide illustrations of various ways in which languages are known to accomplish the particular function in question. For example, there are three broad ways in which languages are known to form predicate nominal clauses. Under 'help' in the section on predicate nominals, each of these three strategies and their subtypes are briefly explained and exemplified. Then the reader is referred to relevant resources for additional information. \ftx \dt 09/May/1998 \key Hintro_man_techn \shd The Linguistic Field Manual--working techniques \txt Some fieldworkers may be daunted by the size and apparent complexity or technicality of this manual -- I've heard it said that a big book is a big nuisance. The same might be said of this on-line presentation. Wherever possible the helps have been broken down into sections one or two screens long. Not all of the entries in the so-called question database are questions. Frequently they are tags to point to the discussion of a particular theoretical point, or to examples of a particular phenomena. In cases like that, the user can skip on to the next entry without reference to the corresponding help screen if he wishes. That way he will only have to look at the material which is relevant to his particular needs. \ftx \txt A rule of thumb might be that if, after having worked intensely with a field language for a year or more, some section of the manual still appears incomprehensible, chances are that section can be omitted or given brief treatment in the fieldworker's own grammar sketch. It is anticipated that as readers discover that previously opaque sections of the manual bear on central questions in the language he or she studies, those sections will suddenly make sense and come alive with interest. \ftx \txt As should be clear from the above discussion, a basic assumption of the manual is that the best way to understand Language, as well as any particular language, is intense interaction with data. Hence, extensive examples are provided from various languages for illustration and the reader is encouraged to compare the illustrations (and the principles they are supposed to illustrate) with linguistic data from a language he or she is attempting to describe. It is through such comparison and interaction that an understanding of the language develops, and almost as a by product, theoretical principles of linguistic structure and categorization emerge. \ftx \txt To Return to the main topic for this, highlight and jump to: \cf Hintro_man_nature \dt 09/May/1998 \key Hintro_man_univ \shd The Linguistic Field Manual--The search for a Universal Grammar \txt By about the middle of the 20th century linguists became interested in the principles that underlie the organization of all languages, rather than simply the possibly idiosyncratic grammars of individual languages. Thus the quest for a UNIVERSAL GRAMMAR became a major thrust of research in linguistics. Since structuralism was so tied to comparison, contrast and levels of structure, it lost much of its usefulness during this period. For example, as shown above, a purely form-driven description would miss the generalization of a 'tense-system' for English. This is because 'tense' in English is expressed in more than one way, at more than one level of structure. This is a problem for a form-driven approach to grammatical description just within the description of a single language. The deeper issue of comparability of structures from one language to the next is even more problematic - when we look at the way in which tense is expressed in many unrelated languages we of course find even more variety than we do within the grammar of English. Hence, many more generalizations would be lost if we attempted to devise a form-driven grammar that was applicable to many, if not all, languages. \ftx \txt The first attempts to describe characteristics of Universal Grammar were still quite structural, or at least they tried to be. It was still felt that, in order to make scientifically valid generalizations, one had to refrain from making subjective judgements. Since semantics always involves an element of subjectivity (i.e. I can never "get inside" a speaker's head to determine and measure the meanings attached to linguistic forms - I have to rely on interpretations, which are really just subjective judgements), evidence for claims concerning Universal Grammar had ideally to be based on more concrete, structural, evidence. \ftx \txt To continue this discussion, highlight and jump to the topic below: \cf Hintro_man_sem \ftx \dt 29/Jun/1998 \key Hintro_what_assump \shd What is Language?--Introduction to Assumptions \txt The following are descriptions of some assumptions that I believe modern descriptive linguists commonly make concerning the nature of language, whether they consider themselves theoreticians or not. These assumptions do not necessarily arise by logical deduction from other, more primitive principles, though they may. In any case, most descriptive linguists do not consider it their task to be concerned about the logical grounds for these assumptions -- they are normally considered to be self-evident. We will use these assumptions to build a 'theoretical framework', i.e. a set of metaphors, that will be useful in conceptualizing the incredibly complex phenomenon known as human language. \ftx \txt Once one has interacted with linguistic data from a variety of languages, one realizes that languages are not arbitrarily difficult to understand, neither are they arbitrarily complex. Rather, they exhibit a substantial degree of commonality. An underlying assumption of much descriptive linguistics is that such commonality is due to common FUNCTIONAL pressures that impinge upon language. Such functional pressures stem from general characteristics of the human mind (e.g. limitations on memory and attention), characteristics of social interaction (e.g.dominance, deference and respect), general characteristics of functional systems (e.g. the principle of least effort), and specific characteristics of symbolic systems (e.g.iconicity, metaphor and analogy). This is the ECOLOGY OF LANGUAGE, i.e. the interdependent environment within which language exists and functions. \ftx \txt There are many similarities apparent among languages and as we look at these similarities, we find that they also make sense. For example, we are not surprised to find that people living in the Amazonian rain forest as well as those living in Central Africa use series of vocal articulations (consonants and vowels) as the primary modulators for communicating with each other. This is because the human vocal apparatus is especially suited for producing signals with enough complexity and precision to symbolize the sorts of concepts that human beings normally have a need to express. For a more meaningful example, we find that left-dislocated noun phrases in English and Yagua both refer to items that are contrastive in the discourse. This fact makes sense because the human mind assigns special prominence to the first stimulus (primacy effects) in a series of stimuli. Because Yagua and English speakers have the same cognitive make-up, both tend to (unconsciously) exploit primacy to accomplish similar communicative tasks. We use the term 'explanation' in this manual as the establishment of a (potentially teleological) relationship between a particular linguistic structure and what it does within the symbolic system of which it is a part. In other words, to explain a structure is to articulate why it has the form it does in terms of the purpose it serves. This theoretical orientation will be elaborated and illustrated in section #0.2.1 below. \ftx \txt Along with their commonalities, languages also exhibit vast differences. Thus the field linguist must constantly strike a balance between assuming that functional similarity leads to structural similarity, and being open to the parameters along which languages vary. Part of the job of the theoretical linguist is to determine the range and limits of linguistic variation, as well as the extent of linguistic universality. \txt To continue this discussion, highlight and jump to the topic below: \cf First Set: Hintro_what_hbehav \cf Second Set: Hintro_what_hbehav_rout \cf Third Set: Hintro_what_symb_A \cf Fourth Set: Hintro_what_symb_E \cf Fifth Set: Hintro_what_symb_I \txt To proceed to the next topic, use Database Search (ALT-D,S) to go to: \cf Hintro_embod \dt 27/Apr/1998 \key Hintro_what_hbehav \shd Language as human behavior. \txt The first implicit assumption of descriptive linguistics is that linguistic behavior is an integral part of human behavior in general. To the 'naïve' field linguist there is no reason to assume anything but the obvious: language is people behaving. From this it follows that linguistic behavior is sensitive to principles common to human behavior in general, and has many of the general properties of other kinds of human behavior. Some of these properties will be described and illustrated in the following subsections. \ftx \txt To continue this discussion, highlight and jump to the topic below: \cf Hintro_what_hbehav_goal \ftx \dt 29/Jun/1998 \key Hintro_what_hbehav_func \shd Any given language is functional \txt Another assumption of descriptive linguistics is that any form of speech used by a community of speakers for everyday communication is FUNCTIONAL. That is, it adequately meets the needs (serves the purposes) to which those speakers apply it. It may be the case that some forms of speech in common use are not functional, but descriptive linguists assume they are functional until proven otherwise. This assumption allows the fieldworker to presume that the observed structures in fact accomplish communicative tasks. If it were a matter of empirical proof rather than a priori assumption that a given language 'does the job', then tests would have to be devised to determine the relative functionality of communicative systems. For example, one might devise a list of tasks that all languages should be able to perform, using some universal criterion. This would be similar to the lists of supposedly universal vocabulary items that are often used in linguistic surveys and comparative work. Then, some method would be employed to ascertain whether those tasks were fulfillable (or actually fulfilled in common use) in the particular communicative system under investigation. As of yet no such tests have been a felt need of descriptive linguists. At least since Boas (1911) descriptive linguists have assumed that if a communicative system is in use for everyday communication by a community, then that system must be capable of expressing every concept felt to be necessary to express within that community. \ftx \txt To continue this discussion, close all jumps until the main edit window is active, then highlight the topic code indicated beside the "Second Set:" label. \ftx \dt 09/May/1998 \key Hintro_what_hbehav_goal \shd Language is goal-directed \txt Descriptive linguists also assume that language is INTENTIONAL behavior. This assumption is evident in such common locutions as 'X means Y' or 'form X refers to Y' etc. Linguistic structures do not simply exist and have random effects on the world for good or for ill in terms of the survival and evolution of our species. Rather they exist and are used to accomplish specific human-engineered effects on the world. People use language in order to accomplish purposes and goals, just as they use their legs, skills, memory or clothes to accomplish purposes and goals. Language is one important tool that people use to communicate with one another. As such, its form will tend to adapt to the functions it performs, within the bounds of human cognitive/perceptual capacity, the principle of least effort etc. (See Zipf 1949 on the principle of least effort and the tools and jobs metaphor). \ftx \txt To continue this discussion, highlight and jump to the topic below: \cf Hintro_what_hbehav_func \ftx \dt 09/May/1998 \key Hintro_what_hbehav_rout \shd Language is subject to routinization \txt Because language is a subset of human behavior, it is subject to habitualization. This is the process whereby patterns of behavior become standardized and routinized though frequency of use. This process is perhaps best illustrated via an example from another symbolic system most readers of this manual will be familiar with -- the turn signals of an automobile. Driving a car is goal-directed behavior. In order to accomplish the goal of getting from place A to place B (and perhaps to accomplish such subsidiary goals as being polite) it is advantageous to a driver to avoid being hit by cars from the rear. One way to mitigate this possibility is to alert drivers to the rear whenever one is about to make a turn: \ftx \ftx Function = alert other drivers of an intended turn \ftx \ftx Form = flashing red light on rear of car, on the side that \ftx corresponds to the direction of the intended turn. \ftx \txt NOTE: The turn signal example was developed at a place and time when turn signals really were all red. In recent years, American laws and automobile manufacturers have adopted the eminently reasonable international practice of allowing amber as well as red turn signals. Since the addition of amber turn signals complicates the metaphor, we shall engage in a time-honored practice among academic linguists -- we will ignore the complication (but see note Qintro_what_symb_H). Note, however, that it is not a complication of the data that is being ignored (though that also might be a reasonable interim practice at times). Rather, it is a complication in the system employed metaphorically to elucidate the data. \ftx \fln The process \ftx \txt When I first learn to drive, I am very conscious of alerting drivers behind me when I plan to turn. So I may consciously think, "Oh, I am about to turn. I don't want the car behind to run into my car (or I don't want the driver to get mad at me), so I will activate the turn signal." This is what psychologists would call an ATTENDED process -- I must allocate attention (ATTEND) to the required function and the form that is available to serve that function. After a while (i.e. after the form has been used to serve, or 'code', the function a number of times in actual instances of use), I may get into the habit of using the turn signal whenever I am about to make a turn; I do it automatically, without thinking. This is what psychologists would call a HABITUALIZED (also AUTOMATED or ROUTINIZED) process. \ftx \txt To continue this discussion, highlight and jump to the topic below: \cf Hintro_what_hbehav_rout2 \ftx \dt 29/Jun/1998 \key Hintro_what_hbehav_rout2 \shd Language is subject to routinization--grammaticalization \ftx \txt The more the various patterns of behavior that constitute the process of driving become habitualized, the more I can concentrate on other matters, such as the best route to get from A to B, what I will fix for dinner, or a conversation with my passengers. Finally, I begin to use the turn signal every time I am about to make a turn, even though there may be no cars for miles around. Technically, when there are no other drivers to be alerted of the turn, using the turn signal has no function. It does not 'code' the turning of the corner, even though it correlates with that activity at nearly the 100% level. That is, it isn't essential to accomplishing the act of turning the way manipulating the steering wheel is. It is a reflex of another function that operating the turn signal does normally code. The use of a particular form in situations where it is no longer directly functional is a sure sign of habitualization, or, in more linguistic terms, of GRAMMATICALIZATION. \ftx \shd2 Overlearning \ftx \txt Though descriptive linguists assume that language is inherently functional, there are many instances in which language does not directly encode a function. That is, there are a great many situations, like the use of the turn signal when there are no other drivers to be warned, in which the function of a specific linguistic form is not transparent, and/or where one form seems to function in several different ways. This common observation follows from the assumption that linguistic structures become associated with functions via a process of overlearning (Langacker 1987). Overlearned behaviors persist beyond the point where they are directly functional -- they take on a life of their own, and are no longer directly subject to changes in functional pressure. Thus coding relationships ('mappings') between functions and formal units are seldom exact. From this point of view we can understand linguistic patterns to be habitualized behaviors intended to accomplish communicative tasks or functions. \ftx \shd2 Fluency \ftx \txt The tendency for behavioral patterns to progress from attended to automated processes through frequency of use is evident in many areas of life. For example, becoming 'fluent' in a particular job, say assembling sandwiches in a fast food restaurant, has much in common with becoming fluent in a language. There is a learning period during which the various motor and mental activities must be carried out carefully and thoughtfully. During this period pace is slow, errors are common and activities are restricted to a well-defined set. After habitualization sets in, the pace of execution is faster, errors are less common and the worker is able to effectively 'bend' the normal sequences of activities to accommodate new situations. One might say that at this point the process has become so 'ingrained' that the worker or language learner can use the norms in creative ways to more efficiently accomplish a wider range of intended functions. \ftx \txt Habitualization of repeated behavioral patterns is advantageous in that it allows one to master common tasks so that attention can be allocated to new and possibly more complex tasks. It also allows one to accomplish more than one task at a time; it is possible to engage in several automated activities at once, but one can only attend to one activity at a time. Finally, several habitualized patterns can combine to make up one complex attended process, which may in its turn become habitualized and enter into an even higher level attended process, and so on. This 'layering' of habitualized processes may be the source of what linguists have observed as HIERARCHICAL STRUCTURE in language. \ftx \txt To continue this discussion, close all jumps until the main edit window is active, then highlight the topic code indicated beside the "Third Set:" label. \ftx \dt 09/May/1998 \key Hintro_what_symb_A \shd Language is a symbolic system--use of forms \txt Field linguists commonly assume that language consists of elements of form that people employ to `express', 'code', 'represent' or 'refer to' other things. \ftx \txt NOTE: Here I wish to interpret the term 'other things' in the broadest sense possible, i.e. I make no claims as to whether the signifié (the item signalled, or coded by language form) is a real world item, a 'message world' item, a 'mental concept', a connotation, a denotation, etc. The characteristic of language that is of interest to this discussion is that it is representational, i.e. it represents something else (even if the something else happens to be language itself). The precise nature of what it is that language represents is not at issue here. My personal view is that the notion of cognitive model proposed by Lakoff 1987 will prove to be an extremely fruitful source of insight into the nature and structure of the content of linguistic expressions, and the ways in which form is shaped by that content. However, it is not necessary to share this view in order to agree that linguistic units represent something. \ftx \txt Although linguists (even good descriptive linguists) often imply that the linguistic forms themselves express concepts, this must be taken as a shorthand way of saying that speakers use linguistic forms (among other things) to accomplish acts of expressing, referring, representing, etc. (Brown and Yule 1983:27ff). \ftx \txt Because linguistic behavior is goal directed, the bond between symbol and signified item is INTENTIONAL. That is, the language user intends to establish a representational link between form and function. From this it follows that the forms used to represent (or 'code') concepts will be structured so as to optimize the transparency of the link, within limits of cognition, memory etc. This is not to deny the possibility that certain aspects of language may actually serve to conceal concepts. However, we make it a working assumption that in general language users expect and want linguistic forms to represent ideas to be communicated. \ftx \txt To continue this discussion, highlight and jump to the topic below: \cf Hintro_what_symb_B \ftx \dt 27/Apr/1998 \key Hintro_what_symb_B \shd Language is a symbolic system--form versus function \txt In any symbolic system, form and function cannot be randomly related to one another. In other words a system is not a symbolic system at all if there is no consistency in the relationship between the symbols and categories or dimensions in the symbolized realm. Ideal symbolic systems (e.g. computer 'languages') maximize this principle by establishing a direct, invariant coding relationship between every form and its function or functions. However, real language is not an ideal symbolic system in this sense. It exists in an environment where variation and change are the rule rather than the exception. New functions for language appear every day in the form of new situations and concepts that speakers wish to discuss. Vocal and auditory limitations cause inexact articulations and incomplete perceptions of messages. These and many other factors lead to variation in the form of language, even in the speech of a single speaker. The bond between form and function in real language, then, is neither rigid nor random; it is direct enough to allow communication, but flexible enough to allow for creativity, variation and change. \ftx \txt To continue this discussion, highlight and jump to the topic below: \cf Hintro_what_symb_C \ftx \dt 27/Apr/1998 \key Hintro_what_symb_C \shd Language is a symbolic system--transparency \txt Here we will explore the turn signal example further to illustrate some properties of intentional symbolic systems. The turn signals of a vehicle consist of forms that are intentionally related to functions. As such, the forms have evolved under the requirement that they result in a sufficiently transparent relationship to their functions. \ftx \txt NOTE: The term 'sufficiently' here means 'close enough for the human sensory and cognitive apparati to discern the connection'. The similarity may be physical, but it also may be analogical, metaphoric or iconic. This is one respect in which human cognition departs from computer 'cognition' -- computers can recognize two entities as 'the same' only if they share certain computer-observable physical attributes, e.g. patterns of electronic impulses. In contrast, the human mind can recognize, for example, a phone call, a 'hit' in baseball and a piece of rope as all being 'long' because of metaphorical connections among them (Lakoff 1987). \ftx \txt The constraints that operate on the form of symbolic systems fall into three categories: convention, perception and cognitive principles. \ftx \txt To continue this discussion, highlight and jump to the topic below: \cf Hintro_what_symb_D \ftx \dt 27/Apr/1998 \key Hintro_what_symb_D \shd Language is a symbolic system--conventionality \txt The form of the turn signal is CONVENTIONAL, i.e. established by convention. One could think of any number of other equally efficient ways to accomplish the function, but the one that has been 'agreed upon' is flashing lights on the side of the car closest to the intended direction of the turn. This system is in some sense arbitrary -- an artifact of the history of the development of automotive turn signal devices. With another possible history, the modern day devices could have evolved in very different ways, ranging from using green lights instead of red to sending up puffs of smoke or using microwave emissions. Of course, there are variations within the general unity of form, such as the varied shapes the lights take from one vehicle to the next. But the core assumption, established by social convention, is that when I operate a device that causes a red light to flash on the left rear of my car, any observer will be alerted that I intend to turn left. I could conceivably be mistaken in this assumption, e.g. in a case where the driver of the car behind me is from a country that lacks this particular social convention. \ftx \txt NOTE: Or where automotive turn signal devices have undergone a different history. To the extent that my device and the device employed by the hypothetical other driver share portions of their histories (including the social/conventional portions), we expect intelligibility between my system and his. \ftx \txt Nevertheless, if we view the relationship between form and function as entirely conventional, there is no explanation other than historical accident for the rise of one form as opposed to another. On the other hand, if we view the relationship between form and function as not totally conventional but at least partially motivated by speakers intentions, we are allowed to ask the question 'Why this convention and not some other logically possible one?' \ftx \txt To continue this discussion, close all jumps until the main edit window is active, then highlight the topic code indicated beside the "Fourth Set:" label. \ftx \dt 27/Apr/1998 \key Hintro_what_symb_E \shd Language is a symbolic system--perceptibility \txt The first answer to the question about how conventions arise comes from perceptual principles -- the signal must be PERCEPTIBLE to the intended audience in order for the function to be accomplished. It is no accident, for example, that the signalling devices are on the rear of the car, rather than exclusively on the front. It would be dysfunctional for a system designed to alert drivers to the rear for the signals to be imperceptible from the rear. Similarly, the choice of lights, rather than flags or smoke signals, can be seen as a consequence of perceptual considerations -- the signals, in order to accomplish their function, must be discernable at night, on windy days, etc. On the other hand, the system need only be perceptible to a certain audience -- namely persons who happen to be within 100 meters or so of my car. Hence it would be dysfunctional for my turn signal device to alert the entire city of my intention, e.g. by means of a fireworks display, a siren or radio blitz. \ftx \txt To continue this discussion, highlight and jump to the topic below: \cf Hintro_what_symb_F \ftx \dt 27/Apr/1998 \key Hintro_what_symb_F \shd Language is a symbolic system--cognitive principles \txt The form of devices used to accomplish intended functions is constrained by GENERAL COGNITIVE PRINCIPLES. For example, it is no accident that the turn signal convention insists that a flashing light signals a turn in the same direction as the side of the car on which the light appears. We could imagine a conventionalized system in which a flashing light on the right indicated a turn to the left and vice versa. Sometimes I think certain people actually operate under this alternative social convention! From a strictly conventional point of view this is a perfectly efficient system -- a one-to-one correspondence between form and function, -- but it is not intuitive. \ftx \txt To continue this discussion, highlight and jump to the topic below: \cf Hintro_what_symb_G \ftx \dt 27/Apr/1998 \key Hintro_what_symb_G \shd Language is a symbolic system--iconicity \txt In the case of the 'light-in-direction-of-turn' convention we see a principle of ICONICITY. This principle, in essence, states that variation in code replicates variation in the coded domain. In this case the code variable under discussion is the distinction between lights on the left and right of a car. The coded domain is direction of intended turn -- also divisible into left and right. The relevant iconicity principle might be stated along the lines of the following: \ftx \txt Given a coded domain delimited by two extremes, A and B, and a range of coding devices whose variation between device Da and device Db replicates the variation between A and B respectively, then Da will code A and Db will code B. \ftx \txt To continue this discussion, highlight and jump to the topic below: \cf Hintro_what_symb_H \ftx \dt 27/Apr/1998 \key Hintro_what_symb_H \shd Language is a symbolic system--analogy \txt It also might be argued that the choice of red as the color for turn signal lights is also not entirely conventional. In other domains within the same social system red is conventionalized as the color for warning. To the extent that a turn signal is perceived as an instance of warning it is 'natural', within this social system, that red should be chosen as the color of the turn signal lights. This is an argument of ANALOGY, i.e. to the extent that function A is perceived to be an instance of function B, then the form of devices used to code B may be superimposed on devices used to code A. \ftx \txt NOTE:I am told that the use of amber rather than red for turn signals on newer automobiles is motivated by the fact that red turn signals might be confused with brake lights. Both brake lights and turn signals are warnings of different types. However, in this case the need to keep the two devices formally distinct (to avoid confusing them) overrides the pressure from analogy to keep them the same. Here we have a good example of 'competing motivations' (Du Bois 1985). \ftx \txt Pushing this explanation one step further, one might ask 'Why is red the color of warning?' To this there are a couple of speculative answers that might actually be testable. First, red may be the color that is the most perceptually salient against the kinds of backgrounds in which humans normally find themselves. This hypothesis could be tested by an experiment that measures reactions to signals of varying wavelengths of light. Second, one might appeal to a natural and perhaps archetypical warning sign -- the sight of arterial blood. The fact that this sight has been associated with pain, danger and death throughout human history perhaps has resulted in an unconscious tendency to use red to signal alarm. This is a purely speculative explanation, but one that may be deserving of empirical investigation. \ftx \txt To continue this discussion, close all jumps until the main edit window is active, then highlight the topic code indicated beside the "Fifth Set:" label. \ftx \dt 27/Apr/1998 \key Hintro_what_symb_I \shd Language is a symbolic system--summary \fln In summary, in order for a form to 'code' a function it must be: \ftx \fln 1. Conventional (i.e. agreed upon by both signer and audience) \ftx \fln 2. Perceptible under potentially relevant conditions. \ftx \fln 3. Consistent with general cognitive principles such as iconicity and analogy. \ftx \txt It seems that, of these three characteristics, the only necessary one is perceptibility. The other two complement one another -- a given form/function link will typically make more use of one or the other. There is no doubt that there are form-function relationships in symbolic systems that are purely conventional, i.e. that could be accomplished by any number of arbitrary alternative means. Lexical semantics is perhaps the domain where conventionality reigns supreme -- the forms of most words in any language bear no direct diagrammatic or iconic relation to their meanings. \ftx \txt NOTE: Though even here there are significant elements of iconicity. See Zipf (1949) on the diagrammatic relationship between length of words and complexity of the semantic concepts coded. \ftx \txt For example, any arbitrary string of phonemes could refer to the conceptual domain signalled by the English word 'fall'. As English speakers we have simply 'agreed upon' this particular string. Any number of other possible strings could equally well have served the same function. \ftx \txt NOTE: Of course this 'agreement' has been arrived at historically, through the etymology of the word and general developments in the history of English. In this sense the form is not arbitrary. However, in the semiotic sense, i.e. as a sign used to refer to a concept, it is arbitrary. \ftx \txt To continue this discussion, highlight and jump to the topic below: \cf Hintro_what_symb_J \ftx \dt 27/Apr/1998 \key Hintro_what_symb_J \shd Language is a symbolic system--unconventional systems \txt There are also form-function relationships that are not conventional. Coinages are perhaps the most obvious examples of these. When someone 'makes up' a new word, there is no possibility that the word communicates strictly because of conventionality. For example, I recently heard someone use the verb 'to plastic' to mean 'to cover with a sheet of plastic'. I immediately understood what was meant because I was familiar with the context the speaker was referring to and I drew an analogy based on other contexts in which the noun 'plastic' is normally used and on other English noun-verb derivations, e.g. 'to panel', 'to oil', 'to wallpaper'. However, there is no point in arguing that 'plastic' had a conventionalized meaning as a verb. It is possible that this was the first time that particular expression had been used in the history of the English language. \ftx \txt To continue this discussion, highlight and jump to the topic below: \cf Hintro_what_symb_K \ftx \dt 27/Apr/1998 \key Hintro_what_symb_K \shd Language is a symbolic system--final word \txt In general, what varies from one language to another are the conventional relations. What is universal are the iconic relations. As the level of analysis moves from lexical semantics, through morphology, syntax and discourse structure, the nature of the relationships between form and function tends to become more iconic and less conventionalized. Hence, the area where languages most obviously and widely vary is vocabulary. This is why vocabulary comparison has been the most reliable method of determining genetic relationships between languages -- the possibility that two languages might have similar vocabularies by pure chance is very remote. Syntax, on the other hand, is much more iconic. Therefore to note that two languages have similar syntaxes says little about whether or not they are genetically related -- there are too many universal cognitive principles that may have given rise to the similarity. Even in vocabulary studies, however, one must be careful of possible universal cognitive principles that may cause unrelated items to appear cognate. For example, Swadesh (1971:194) cautions against using words such as 'blow', 'suck' and 'mother' in vocabulary comparisons because such words universally tend to contain bilabial consonants. \ftx \dt 27/Apr/1998 \key Hmor \shd Morphological Typology \ftx \txt This topic contains the following subsections. Highlight and jump to the topic of interest: \shd2 Introduction \cf Hmor_intro General \cf Hmor_intro_cncpt Conceptualization \cf Hmor_intro_def Definitions \cf Hmor_intro_where Where it fits \shd2 Typology \cf Hmor_typ Morphological Typology \cf Hmor_typ_synth Synthesis \cf Hmor_typ_fus Fusion \shd2 Inflection and Derivation \cf Hmor_infl_gen General \cf Hmor_infl_drv Derivational \cf Hmor_infl_infl Inflectional \shd2 Others \cf Hmor_rt Roots and Stems \cf Hmor_mproc Morphological Processes \cf Hmor_hd/dep Head/Dependent Marking \dt 12/May/1998 \key Hmor_hd/dep \shd Head / dependent marking \txt The head of a phrase is (under most definitions) the element that determines the syntactic function of the whole phrase. So, in a noun phrase the head is the noun that refers to the same entity that the whole phrase refers to, e.g. 'crown' in 'the Queen of England's crown'. An adposition is the head of an adpositional phrase because the presence of the adposition is what gives adpositional phrases their particular syntactic properties -- without the adposition the phrase would be a simple noun phrase. Other elements in a phrase are sometimes referred to as dependents. \ftx \txt Some languages tend to mark the relationship between a head and a dependent on the head, while others tend to mark the relationship on the dependent. For example, English is predominantly a dependent marking language. This is illustrated by the fact that in possessive noun phrases, the head noun is not marked to indicate that it is possessed. Rather, the possessor is marked: 'John's book.' Other languages, e.g. Farsi, typically mark the head, e.g. \ftx \ftx (5) Zhon kitab-é 'John's book.' \ftx John book-POSS \ftx \txt Languages will typically follow one pattern or the other across various phrase types, i.e. NP, VP, PP etc. Head marking languages are common throughout the Americas, Africa, Asia, Siberia and in the mideast. Dependent marking languages are less common, and are found primarily in Europe, though some Amerindian languages are dependent marking. Most languages are of a mixed type. Once you know the marking type of a language, several other typological characteristics are predicted. \ftx \fln References: Nichols (1986). \ftx \dt 27/Apr/1998 \key Hmor_infl_drv \shd Inflection and derivation--Derivational operations \txt Derivational operations consist primarily of the following: \ftx 1. Operations that change the grammatical category of a root, e.g. \ftx denominalization and nominalization (see Hn_dnom and Hv_nz). \ftx 2. Operations that change the valence (transitivity) of a root, \ftx e.g. detransitivization, causativization and desiderative \ftx (see subjects under Hval). \ftx 3. Operations which in other ways significantly change the basic \ftx concept expressed by the root, e.g. distributive, diminutive \ftx (see Hv_misc_dist). \ftx \txt Non-definitional characteristics of derivational operations include: \ftx 1. They are 'optional' insofar as they are employed in order to \ftx shape the basic semantic content of roots and are not themselves \ftx governed by some other operation or element in the syntactic \ftx configuration. \ftx 2. They tend to be idiosyncratic and non-productive. \ftx 3. They do not occur in well-defined 'paradigms'. \ftx \txt To continue this discussion, highlight and jump to the topic below: \cf Hmor_infl_infl \ftx \key Hmor_infl_gen \shd Inflection and derivation--General \txt A convenient and fairly unequivocal distinction can be made between two types of morphological operations based on formal behavior and straightforward semantic criteria. This distinction is commonly referred to as INFLECTION vs. DERIVATION. In this manual DERIVATIONAL operations are defined as operations which derive an inflectable stem from a root or an intermediate stem. However, a root plus derivational operations alone may or may not be a fully-formed word (see Anderson 1985 for a definition of the term 'word'). \ftx \txt As with all functional distinctions, inflection vs. derivation is less a distinction than a continuum. Some operations fall in between the prototypical extremes, and operations tend to migrate diachronically from inflection to derivation (very rarely do they migrate in the opposite direction). Nevertheless, this distinction does seem to be instantiated in many languages, and is worth elaborating briefly in a grammar sketch or reference grammar. \ftx \fln References: Anderson 1982. D. Payne 1985b. \ftx \txt To continue this discussion, highlight and jump to the topic below: \cf Hmor_infl_drv \ftx \dt 27/Apr/1998 \key Hmor_infl_infl \shd Inflection and derivation--Inflectional operations \txt INFLECTIONAL operations are those which are required by the syntactic environment in which a root appears. Inflectional operations do not substantially affect the basic semantic content of a root. Rather they ground the semantic content of a root according to place, time and participant reference. That is, they specify when the event or situation took place, who or what were the participants, and sometimes where, how or whether an event or situation really took place. They don't normally shape the character of the event or situation itself. Typical inflectional operations include: \ftx 1. Person, number, gender, etc. 'agreement' (see Hv_agr). \ftx 2. Tense, Aspect, Mode (see Hv_tam). \ftx \txt Non-definitional characteristics of inflectional operations include: \ftx 1. They are grammatically required in certain syntactic \ftx environments, e.g. the main verb of an English sentence must \ftx be inflected for subject and tense. \ftx 2. They tend to be regular and productive (at least in comparison \ftx to derivational operations). \ftx 3. They tend to occur in 'paradigms', i.e. sets of forms of which \ftx one form must be selected in certain environments. For example, \ftx there are two 'tenses' in English, one of which must be \ftx specified for all independent verbs. \ftx \txt Inflectional operations create forms that are fully grounded (see the introduction, section Qintro_embod_grnd) and able to be integrated into discourse, whereas derivational operations create stems that are not necessarily fully grounded, and which may still require inflectional operations before they can be integrated into discourse. \ftx \dt 27/Apr/1998 \key Hmor_intro \shd Synthesis and Fusion--General \txt MORPHOLOGY is the study of shapes. One can talk about the morphology of camels -- different species of camels have different morphologies, i.e. they have different body shapes. Morphology in linguistics has to do with the shapes of words. How are words shaped in such and such a language? What systematic rules determine when and how they may adjust their shapes? \ftx \fln The Morpheme \ftx \txt A MORPHEME is a minimal shape. The classical definition of a morpheme is the minimal formal unit that embodies meaning. For example the English word 'dogs' contains two morphemes, 'dog' which embodies the main semantic content of the expression, and '-s' which embodies the meaning of plurality. The form 'dog' itself is not further divisible into meaningful component pieces, therefore it is a morpheme -- a minimal shape. In most situations this definition works fine. However, more current approaches acknowledge the fact that particular meanings are not necessarily directly linked to particular chunks of form. The meaning contributed by a morpheme may vary depending on other morphemes in the immediate environment. The whole message may be more, less than or simply different from the sum of the 'meanings' of all the morphemes in the message. \ftx \fln To continue this discussion, highlight and jump to the topic below: \cf Hmor_intro_cncpt \dt 27/Apr/1998 \key Hmor_intro_cncpt \shd Morphological Typology--Conceptualization \txt In this manual we will conceptualize morphology as a system of adjustments in the shapes of words that contribute to adjustments in meaning. This is the source of the operator/operation terminology introduced in the introduction, Qintro_embod_msyn. The forms of words in messages are shaped in a variety of ways, including but not limited to the addition of morphemes, in order to convey intended meanings. The precise meaning conveyed is developed via interaction among the linguistic context, the extralinguistic context and the morphological operator or operators employed. It does not reside completely in the morphological operators themselves. This approach is consistent with the 'Word and Paradigm' approach to morphology proposed by Anderson (1982). Anderson, and others working in the Word and Paradigm framework, conceives of morphemes as rules involving the linguistic context rather than as isolated 'chunks' of linguistic matter. Ideally, the descriptive linguist should be able to go beyond the linguistic context to specify how a morpheme interacts with the non-linguistic context as well. Unfortunately, such a worthy undertaking is beyond the scope of most reference grammars. In a reference grammar one should correctly and clearly specify the conventionalized effects each morpheme has in its most common linguistic contexts. \ftx \txt To continue this discussion, highlight and jump to the topic below: \cf Hmor_intro_def \ftx \dt 27/Apr/1998 \key Hmor_intro_def \shd Morphological Typology--Definitions \txt A BOUND MORPHEME (also referred to as an AFFIX) is a morpheme that must attach to some other morpheme. The morpheme '-s' in 'dogs' is an example of a bound morpheme, since it could never be uttered by itself. 'Dog', on the other hand, is a FREE MORPHEME since it does not have to attach to some other form. A clitic is a bound morpheme that can attach to other morphemes of a variety of grammatical categories. Clitics and affixes are discussed in more detail in Hmor_rt. \ftx \fln Allomorphs \ftx \txt An ALLOMORPH is a variant pronunciation of a morpheme. For example the plural morpheme usually written as '-s' in English has at least three allomorphs: [-s] as in [hæts] 'hats', [-z] as in [dögz] 'dogs', and [-ïz] as in [báksïz] 'boxes'. Sometimes morphemes are conceived of as sets of allomorphs. Determining the 'underlying' form of a morpheme is important for orthography development and for satisfactory and consistent glossing of texts, but it is not a major theoretical issue that should occupy a great deal of space in a grammar sketch or reference grammar. The references cited below contain many suggestions on determining underlying forms and rules for deriving surface forms. Sometimes in the literature the terms 'morph' or 'formative' are used to refer to particular morphological shapes if it is unknown or unimportant whether they constitute morphemes or not. \ftx \fln Morphophonemic rules \ftx \txt MORPHOPHONEMIC RULES have the form of phonological rules, but are restricted to particular morphological contexts. The allomorphs of a particular morpheme are derived from phonological rules and any morphophonological rules that may apply to that morpheme. For example, the allomorphs of the plural morpheme in English are determined by a phonological rule that can be represented in generative notation as follows: \ftx \ftx /s/ --> [ïz] / [+sibilant] ___ \ftx \ftx [z] / [+voice] ___ \ftx \ftx [s] / ___ \ftx \txt This is a phonological rule because it applies to the phoneme /s/ whether it happens to also be the plural marker or not. The allomorphs [il] and [ir] of the prefix /in-/, on the other hand, are determined by a morphophonemic rule. This is illustrated by the fact that not all examples of /n/ become [l] or [r] when occurring before /l/ or /r/. It is only the /n/ of the prefix /in-/: \ftx \ftx (1) irregardless /in+rigardles/ \ftx irrespective /in+rispektiv/ \ftx irresponsible /in+rispansibl/ \ftx \ftx illegal /in+ligl/ \ftx illogical /in+lajikl/ \ftx illiterate /in+litërët/ \ftx \ftx etc. \ftx \txt The prefixes /un-/ and /non-/, for example, do not exhibit this regular pattern: \ftx \ftx (2) unresponsive (*urresponsive) \ftx unreliable (*urreliable) \ftx unreached (*urreached) \ftx \ftx unlimited (*ullimited) \ftx unleash (*ulleash) \ftx \ftx non-lethal (*nol-lethal) \ftx \ftx etc. \ftx \txt Therefore the rules n --> l / __ l and n --> r / __ r are not phonological rules. They must be specified as occurring only with the prefix /in-/. Whenever such morphological information is required to specify the environment for an allophonic rule, the rule is morphophonemic. \ftx \txt Morphophonemic rules may be presented in the sections dealing with particular morphological operations. The notation used to represent the rules is normally compatible with the notation used to express phonological rules. The Word and Paradigm framework provides a notational system that is uniquely suited to languages with a great deal of morphophonemic variation. \ftx \fln References: Pike (1947), Elson and Picket (1988), Anderson (1985, 1985b) \ftx \txt To continue this discussion, highlight and jump to the topic below: \cf Hmor_intro_where \ftx \dt 27/Apr/1998 \key Hmor_intro_where \shd Morphology -- Where it fits \txt The section on morphology in a modern grammar sketch should provide an overview of the morphological typology and characteristic morphological processes evident in the language. In the era of structuralism in descriptive linguistics, the chapter on 'morphology' consisted of lists of all the morphemes, their allomorphs and the morphophonemic rules that derive the allomorphs from hypothesized underlying forms. In a functionally oriented sketch, this information will be found in the chapters on the various morphosyntactic functions. For example, the various forms and idiosyncracies of the causative morpheme (or morphemes) will be found in the section on causativization, the forms of the tense/aspect operators will be found in the section on tense/aspect, etc. \ftx \dt 27/Apr/1998 \key Hmor_mproc \shd Morphological processes \txt There are six basic morphological processes by which stems can be formally altered to adjust their meanings to fit their syntactic and communicational context. These six processes are 1) PREFIXATION, 2) SUFFIXATION, 3) INFIXATION, 4) STEM MODIFICATION, 5) REDUPLICATION and 6) SUPRAFIXATION (i.e. suprasegmental modification). A seventh process, SUPPLETION, may not appropriately be called morphological since it involves the replacement of one stem with another. Nevertheless any of the operations that are typically coded by the six basic processes can also be coded by suppletion, hence it deserves at least passing mention in this section. \ftx \txt Prefixes are bound morphemes that attach to the front of stems, e.g. 'un-' in 'unselfish'. Suffixes are bound morphemes that attach to the ends of stems, e.g. the '-ed' past tense marker of English. Infixes are bound morphemes that occur within stems. There are none of these in English, but they are fairly common in Austronesian languages, e.g. Bontok 'um' in 'fumikas'-- 'become strong' (c.f 'fikas'-- 'be strong'). Stem modification is what happens in the 'sing', 'sang' 'sung' paradigm of English. Reduplication is where a piece of a root (possibly the whole root) is repeated, e.g. Ilokano 'pingan'-- 'dish', 'pingpingan'-- 'dishes'; 'talon'-- 'field', 'taltalon'-- 'fields'. Suprasegmental modification is where the tone or stress pattern of a word signals a particular morphological operation. English makes some use of stress patterns to signal the difference between certain nouns and related verbs, e.g. 'pérmit' (noun), vs. 'permít' (verb), 'cónvert' (noun), vs. 'convért' (verb), etc. Some languages, especially in Africa, use tone modification to signal very common morphological operations like tense and aspect. As mentioned above, suppletion is the replacement of one stem for another. In English the verb 'be' is notoriously suppletive. 'is', 'am', 'are', 'were', 'was', and 'be' are all distinct stems used to embody various combinations of morphological operations applied to this verb. \ftx \txt Most languages that are at all agglutinative employ suffixes. Some of these also employ prefixes. Some of these also employ infixes. Very few languages employ only prefixation, and none employ only infixation or any of the other types of morphological processes mentioned above. \ftx \fln References: Greenberg (1978, vol. 3), Anderson (1985). \ftx \dt 27/Apr/1998 \key Hmor_rt \shd Roots and stems \txt Closely related to the distinction between derivational and inflectional morphology is the distinction between root and stem. A root is an unanalyzable form that embodies the basic lexical content of the word. \ftx \txt A STEM consists minimally of a root, but may be analyzable into a root plus derivational morphemes. Yet a stem does not necessarily constitute a fully-grounded understandable word in and of itself. An inflectional operation, usually involving a prefix or a suffix, is required. \ftx \txt In the following example from Panare (Carib, Venezuela), the verb root is aamë 'raise'. In this context, this root has two derivational morphemes, s- 'DETRANS' and o- 'INTRans'. These are clearly derivational because they are not required (yaamëñe alone means 'he raises it'), and when they do occur, they are not sufficient to allow the verb form to be integrated into discourse (*saamë, *osaamë): \ftx \ftx (4) Tée y-o-s-aamë-n e'ñapa tyityasa' \ftx s.w.appear 3-INTR-DETRANS-raise-NONSPEC:I person one \ftx `There someone stood up.' (i.e. raised himself) (RMW.01.2) \ftx \txt In effect, then, these derivational prefixes take the root meaning 'raise' and convert it into another verb meaning 'to stand up'. This is a very common sort of function for derivational morphology. \ftx \txt The prefix-suffix combination y- ... -n consists of inflectional operators. First, they (or other members of the same 'paradigm') are required in order for the verb to be properly integrated into discourse. Second, they are sufficient to allow the verb to be integrated into discourse. No other affixation is necessary. \ftx \dt 27/Apr/1998 \key Hmor_typ \shd Morphological Typology \txt One of the first items of business in analyzing the grammar of a language is determining what sort of language it is in terms of its morphology. This section will provide a framework and suggestions for understanding the morphological typology of a language. \ftx \txt A TYPOLOGY is simply a division of a range of phenomena into types. To 'typologize' a phenomenon is to categorize it into types. For example, a typology of motorized vehicles might consist of a list containing trucks, automobiles, buses, etc. The value of a typology to those who study natural phenomena, such as language, is dependent on the extent to which the proposed typology makes predictions concerning a range of important characteristics of the individual types proposed. For example, a cluster of important formal and functional properties distinguishes buses from automobiles. \ftx \txt On the other hand it would make little sense to typologize motor vehicles according to color. There are no important properties that correlate with (are predicted by) the color of a motorized vehicle. That is to say, there is far more coherence and substance to the concept of 'bus' than there is to the concept of, for example, 'maroon colored motor vehicle'. \ftx \txt Several typologies of language have been proposed in the history of linguistic science. The first typology that has maintained lasting interest is morphological typology. This typology refers primarily to the extent to which words in the language are divisible into clearly individuated morphemes. The first proposals recognized three morphological language types 1) isolating, 2) agglutinative and 3) inflectional or fusional. In this section we will suggest an approach to morphological typology presented by Comrie (1981). \ftx \txt There are some significant linguistic facts that correlate with morphological typology. For example, agglutinating languages tend to place the verb first or last in main, declarative clauses. In general, however, knowing the morphological typology of a language does not allow one to make many predictions concerning other possible characteristics that the language is likely to have. \ftx \txt To continue this discussion, highlight and jump to the topic below: \cf Hmor_typ_synth \ftx \dt 27/Apr/1998 \key Hmor_typ_fus \shd Fusion \txt The index of FUSION (Comrie 1981) has to do with the degree to which units of meaning are 'fused' into single morphological shapes. A highly FUSIONAL language (sometimes called 'inflectional', but since this has other connotations, we will stick with fusional) is one in which one form can have several meanings, e.g. Spanish '-ó' in 'habló' simultaneously codes indicative mode, third person, singular, past tense, and perfective aspect. If any one of these meaning components changes, the form of the verbal suffix must change. Turkish is a language for which each component of meaning is represented by its own morpheme. Hence Turkish is a highly AGGLUTINATIVE language. Again, there is no quantitative method for precisely establishing the index of fusion for a given language. For highly isolating languages, the index of fusion just doesn't apply. If anything, English is agglutinative rather than fusional, e.g., in 'anti-dis-establish-ment-ari- an-ism' each morpheme has a specific and fairly straightforward meaning. But then, such words are all of Latin origin. One hint of fusion in English is certain 'strong' verb forms, e.g. sang, thought, brought, etc., in which a past tense morpheme cannot be strictly separated out from the root. \ftx \fln References: Comrie (1981:ch. 2). \ftx \dt 27/Apr/1998 \key Hmor_typ_synth \shd Synthesis \txt The index of SYNTHESIS (Comrie 1981) has to do with how many morphemes tend to occur per word. The index defines a continuum from isolating languages at one extreme to highly polysynthetic languages at the other. A strictly isolating language is one in which every word consists of only one morpheme. Chinese comes close to this extreme. A highly polysynthetic language is one in which words tend to consist of several morphemes. Quechua and Eskimo are good examples of highly polysynthetic languages. The following is an example of a polysynthetic structure in Yup'ik Eskimo (thanks to Eliza Orr): \ftx \ftx (3) tuntussuqatarniksaitengqiggtuq \ftx tuntu -ssur-qatar-ni -ksaite-ngqiggte-uq \ftx reindeer-hunt-FUT -say-NEG -again -3SG:IND \ftx 'He had not yet said again that he was going to hunt reindeer.' \ftx \txt There is as yet no quantitative method for determining exactly how 'synthetic' a language is (i.e. to what degree it 'puts together' morphemes into single words). This is probably because a precise specification of the morphological type of a language is not important enough for anyone to worry too much about. A general rule however, is that if the language can express a whole sentence with just a verb, it is polysynthetic. If it can't, then it is isolating. Adjectives such as 'somewhat' or 'highly' can then be added depending on the investigator's intuitions, e.g. English is 'somewhat isolating', Chinese is 'highly isolating'. Korean is 'somewhat polysynthetic' while Eskimo is 'highly polysynthetic'. The payoff in terms of predicting other structural aspects of the language is simply not great enough to spend much time being more precise than this. \ftx \fln References: Comrie (1981:ch. 2). \ftx \txt To continue this discussion, highlight and jump to the topic below: \cf Hmor_typ_fus \dt 27/Apr/1998 \key Hn \shd Noun and noun-phrase operations \ftx \txt This topic contains the following subsections. Highlight and jump to the topic of interest: \ftx \cf Hn_intro Nominal operations \shd2 Articles \cf Hn_art Articles \cf Hn_art_like like particles \shd2 Demonstratives \cf Hn_art_dem the class \cf Hn_art_dem_deg degrees of distance \cf Hn_art_dem_other other parameters \shd2 Case \cf Hn_case Contrasting with adpositions \cf Hn_case_ex Examples \cf Hn_case_role Semantic roles \shd2 Class \cf Hn_cls Class (including gender) \cf Hn_cls_other other dimensions \cf Hn_cls_rel relation to other constituents \shd2 Compounding \cf Hn_cmp Compounding \cf Hn_cmp_freq Frequency of occurrence \cf Hn_cmp_iden Identifying \cf Hn_cmp_n-v Noun-verb compounds \cf Hn_cmp_prod Productivity \shd2 Denominalization \cf Hn_dnom Denominalization \cf Hn_dnom_adj adjectives \cf Hn_dnom_adv **adverbs \shd2 Number \cf Hn_num_opt Optional marking \cf Hn_num_other Other distinctions \cf Hn_num_s/pl Singular/Plural distinctions \shd2 Possessors \cf Hn_poss Possessors \cf Hn_poss_al/inal alienable vs inalienable \cf Hn_poss_other other kinds \cf Hn_poss_ord **constituent order \shd Other topics \cf Hn_dim/aug Diminution/augmentation \cf Hn_other Other operations encoded in the noun phrase \dt 29/May/1998 \key Hn_art \shd Articles \txt Operators, whether bound or free, which directly indicate something about the definiteness and/or referentiality of a noun phrase are often called articles. Articles such as the English 'the', and 'a(n)', are relatively rare in the world's languages. More common are demonstratives, such as 'this', 'that', 'these', and 'those'. Some linguists use the term determiner to refer to formatives like 'the' and 'a(n)'. This term usually also includes quantifiers (some, many, a few, each, every, the numerals etc.), possessors, as well as demonstratives. This broad category does not very often exhibit unitary syntactic behavior, e.g. there are few languages that consistently place them all in the same position in the noun phrase. Therefore, this category is not very viable as a universal natural class. \ftx \dt 27/Apr/1998 \key Hn_art_dem \shd Demonstratives--the class \txt Most, if not all, languages have a clear class of demonstratives. These are normally free forms,and usually precede the noun they function with (if such a noun occurs at all). A fairly consistent universal is that if any noun phrase element precedes the head noun at all, it will be the demonstrative. Demonstratives also typically can function anaphorically, i.e. alone as the head of an NP ('demonstrative pronouns'). Demonstratives imply 'pointing', i.e. 'demonstrating', the object they refer to, e.g. 'THAT house' (said while pointing to a house), or 'I'll take three of THOSE' (said while pointing at some group of objects). \ftx \txt To continue this discussion, highlight and jump to the topic below: \cf Hn_art_dem_deg \ftx \dt 27/Apr/1998 \key Hn_art_dem_deg \shd Demonstratives--degrees of distance \txt In addition to coding the features common to the pronoun system of the language (number, gender, etc.) demonstratives often code distance, or orientation with respect to the speaker/hearer. For example, the English system has two degrees of distance, represented by the forms this and that (these and those in the plural). Other languages may have three degrees of distance. \ftx \txt To continue this discussion, highlight and jump to the topic below: \cf Hn_art_dem_other \ftx \dt 27/Apr/1998 \key Hn_art_dem_other \shd Demonstratives--other parameters \txt If there appear to be more than three degrees of distance, chances are there is some other parameter that the system is sensitive to in addition to the distance parameter. Some languages make a distinction between items close to the hearer, vs. items close to the speaker, vs. items distant from both. Others code the difference between visible items and non-visible items. When two or more of these parameters interact within a single system, the results can appear very complex. \ftx \dt 27/Apr/1998 \key Hn_art_like \shd Article-like particles \txt Most operators that embody pragmatic or semantic information about a noun will tend to occur more often with nouns of particular syntactic roles (e.g. definitizers in English are more frequent in discourse on nouns that occur in the subject role). In many cases these tendencies have become grammaticalized. One common phenomenon, though somewhat 'exotic' from the point of view of European languages, is articles that mark something like definiteness only for direct objects. Hebrew and Farsi are unrelated languages that exhibit this property: \ftx \fln HEBREW: \ftx (3) a. et ha séfer \ftx \ftx b. \ftx \fln FARSI: \ftx (4) a. man domboli kitob 'I'm looking for a book.' \ftx I look:for book \ftx \ftx b. man domboli kitob-ro 'I'm looking for the book.' \ftx I look:for book-DEF \ftx \txt This definiteness distinction is not morphologically manifested for noun phrases in any other syntactic roles. In a few languages this grammaticalization has gone so far as to render it difficult to determine whether a given particle or affix is a noun phrase or verb phrase operator. For example, Panare (Carib, Venezuela) has a set of pre-nominal particles that function very much like articles in that they encode information about the definiteness/specificity as well as animacy and location of the noun that follows. However, they only occur before nouns that function as subject of the sentence, and then only subjects that come immediately after the verb, and then only in non-past tenses! Because these Panare particles have so many characteristics of verb phrase operators (i.e. consistent position directly after the verb, occurring only in certain tenses) it is difficult to determine whether they should be regarded as functioning more closely with the verb to their left or with the subject nominal to their right. \ftx \fln References: Givón (1984, ch. 11). \ftx \dt 27/Apr/1998 \key Hn_case \shd Case--Contrasting with adpositions \txt It is sometimes difficult to distinguish case marking from adpositions (the latter consist of prepositions and postpositions). This is undoubtedly because there is no necessary universal distinction between the two; like most structural distinctions, the two categories describe extremes of a scale that continually varies over a range of actual phenomena. The following is a rule that probably works 90% of the time. However, the distinction described by this rule is rather subtle, and hence may not be obvious to the fieldworker in early stages of language analysis. Furthermore, in any language there may be exceptions. \ftx \ftx Rule: Case marking is the morphosyntactic categorization of noun-phrases \ftx that is imposed by elements of the structure within which the noun \ftx phrase occurs. Adpositions are free of such configurational \ftx constraints. \ftx \txt So, for example, whether a noun phrase occurs in the dative or accusative case in some languages is determined by the grammatical requirements of the verb (or other case-governing element) with which that noun phrase is in some grammatical relationship. Whether a noun phrase occurs with a locative or benefactive adposition, however, probably depends purely on the communicative intent of the speaker -- it is not imposed by some other grammatical element in the configuration. \ftx \txt For some examples, jump to: \cf Hn_case_ex \ftx \dt 27/Apr/1998 \key Hn_case_ex \shd Case--Examples \txt In Latin, for example, verbs require that their objects occur in one of a few morphological cases. If the object occurs in some other case, either ungrammaticality or a different sense of the verb results. Hence in Latin, verbs are said to govern the case of their objects (e.g. the verb 'servire' governs the dative case, etc.). Prepositions in Latin also govern the case of their objects. So, e.g., 'cum' governs the ablative case, while 'contra' governs the accusative case. Prepositions in Latin are, however, not themselves governed by some other element in the configuration. So no verb requires that its object occur in a 'cum' phrase, for example. adpositional phrases are usually (though not always) 'optional' sentence constituents. \ftx \txt Note that the parameter of morphological binding does not enter into this definition. Under this definition case markers can be free or bound, prepositional or post-positional. The same is true for adpositions. It so happens that in the classical languages the case markers are phonologically as well as functionally distinct from the adpositions. In Quechua (and many other languages), this is not the case. In Quechua the marker '-ta' can be considered to be a case marker, because it must occur on direct objects of transitive verbs: \ftx \fln (1) a. Noqa-ta rika-ma-n `He sees me.' \ftx 1SG-ACC see->1-3 \ftx \ftx b. *Noqa rikaman. \ftx \txt There are, however, many other postpositional operators that are not (normally) governed by a verb: \ftx \fln (2) -chaw Locative \ftx -pita Ablative \ftx -man Goal \ftx -yaq Limit \ftx -paq Purpose \ftx -pa Through, by way of \ftx -wan Commitative, Instrumental \ftx -naw Similarity \ftx etc. \ftx \txt It is true that case markers (as defined above) tend to be more closely tied phonologically to their hosts than are adpositions, but this is not a defining property of case markers. Here we have given a definition based strictly on syntactic function. \ftx \dt 27/Apr/1998 \key Hn_case_role \shd Case--Semantic roles \txt The following is a short list of semantic roles that typically get instantiated as morphological cases. Keep in mind, however, that there is never a direct, one-to-one mapping between semantics and morphosyntax. This list is simply meant as a rule-of-thumb that characterizes general tendencies as represented in the previous literature: \ftx \ftx Semantic Morphological \ftx role case \ftx \ftx AGENT Nominative, ergative (Hgrel_case) \ftx PATIENT Accusative, absolutive \ftx RECIPIENT Dative \ftx POSSESSOR Genitive \ftx \txt Morphosyntactic markers that instantiate other semantic roles can usually be given the same label as the semantic role, e.g. locative, benefactive, instrumental. In so far as possible, cases should be labelled according to their prototypical, or basic function. \ftx \dt 27/Apr/1998 \key Hn_cls \shd Class (including gender) \txt Some languages group nouns into various classes by means of grammatical devices within noun phrases. If the language has a noun class system, it will almost certainly be well installed in the number system. If nothing else in the language agrees with nouns in terms of class, numerals will. Sometimes noun classes correspond (with varying degrees of directness) to semantic classes such as biological gender, physical shape, function etc. The most familiar kinds of noun class systems are the gender systems of Indo-European. In many IE languages nouns are either 'masculine', 'feminine' and sometimes 'neuter'. For example, Spanish codes the difference between masculine and feminine by a suffix '-o/-a': 'niño'-- 'boy' 'abogado'-- 'male lawyer' 'maestro'-- 'male teacher' 'niña'-- 'girl' 'abogada'-- 'female lawyer' 'maestra'-- 'female teacher' etc. Most adjectives must reflect the class of their head nouns, e.g. 'abogado bueno'-- 'good male lawyer', vs. 'abogada buena'-- 'good female lawyer'. The class that a particular noun falls into is fairly clear cut for items that have a biological gender, namely animate entities. However, all nouns in the language are subject to the class system, and most are classified apparently arbitrarily as masculine or feminine (rather than neuter). Romance languages even differ from one another as to the class that particular lexical items fall into, e.g. Italian 'il tavolo' (m) French 'la table' (f)--'the table/board.', Italian 'il mare' (m), French 'la mer' (f)--'the sea.' There is even at least one word in Italian which is masculine in the singular and feminine in the plural 'il uovo'--'the egg', and 'le uova'--'the eggs'. This is simply to show that noun class systems, even those that seem to have a firm natural basis, often exhibit a certain degree of irregularity. \ftx \txt To continue this discussion, highlight and jump to the topic below: \cf Hn_cls_other \ftx \dt 27/Apr/1998 \key Hn_cls_other \shd Class--other dimensions \txt Other noun class systems are based on other dimensions of reality than biological gender, e.g. shape (roundish objects, longish objects, stubby objects, etc.), or function (adornments, items associated with procuring food, items associated with fighting, foods, people, etc.). In every case, however, there are items that seem as though they should belong in one class, but for some apparently idiosyncratic reason, are placed in another class, e.g. in Yagua rocks and pineapples are classed as animates. The most famous noun class systems are those found in the Niger-Kordofanian languages of Africa. They also occur in Australia, Asia and in the Americas. \ftx \txt To continue this discussion, highlight and jump to the topic below: \cf Hn_cls_rel \ftx \dt 27/Apr/1998 \key Hn_cls_rel \shd Class--relation to other constituents \txt References: Dixon (1970), Allen (1977), Craig (1986), Adams and Conklin (1973). Carlson and Payne (1989). \ftx \key Hn_cmp \shd Compounding \txt A compound is a word that is formed from two or more different words. For example the word 'windshield' is composed of the words 'wind' and 'shield'. Needless to say, not every sequence of words is a compound. Hence there must be an explicit way of distinguishing compounds from other simple sequences of words. \ftx \txt To continue this discussion, highlight and jump to the topic below: \cf Hn_cmp_iden \ftx \dt 27/Apr/1998 \key Hn_cmp_freq \shd Compounding--Frequency of occurrence \ftx \ftx \key Hn_cmp_iden \shd Compounding--Identifying \txt The criteria for calling something a compound fall into two groups 1) formal criteria, and 2) semantic criteria. Compounds may exhibit the following formal properties: 1) a stress pattern characteristic of a single word, as opposed to the pattern for two words, e.g. 'bláckbird' (the species), has a different stress pattern than 'black bird' (any bird that happens to be black), cf. also 'líghthouse keeper' versus 'light hóusekeeper'. 2) unusual word order, e.g. 'housekeeper' consists of a noun plus a verb where the noun represents the object rather than the subject of the verb. Normally objects come after the verb in English. 3) Morphophonemic processes characteristic of single words, e.g. the word 'roommate' can be pronounced with a single m, whereas normally if two m's come together accidentally in a sentence both are pronounced, e.g. some mice will be understood as some ice if both m's are not pronounced. 4) Some compounds may have a specific marker, e.g. the '-er' of 'can opener'. To can open is not a verb, *I can opened all evening, but with the instrumental '-er' suffix this stem is treated exactly as though it were a verb (i.e. according to the pattern of slicer, grinder etc.). \ftx \txt The dominant semantic characteristic of compounds is that the meaning of a compound is usually more specific than the meanings of the words which make up the compound would lead one to suspect. For example, the term 'windshield' can't be used for any shield against wind, but only for those specific items made of transparent material used in vehicles of various sorts. So a line of trees along a farmer's field could for the nonce be called a wind shield (though the technical term is shelter belt), it cannot be called a windshield. Similarly, 'blackbird' (the compound) is only appropriately used to refer to particular species of birds, though members of other species, such as crows, vultures, etc. can legitimately be called black birds. Some compounds contain one part which is not a real word, e.g. huckleberry, cranberry, etc. In some languages such compounds may be very difficult to distinguish from roots with an affix. \ftx \fln References: Greenberg (1978, vol. 3). Shopen (1985 vol. III). \ftx \txt To continue this discussion, highlight and jump to the topic below: \cf Hn_cmp_n-v \ftx \dt 27/Apr/1998 \key Hn_cmp_n-v \shd Compounding--Noun-verb compounds \fln e.g. pickpocket, scarecrow. \ftx \txt To continue this discussion, highlight and jump to the topic below: \cf Hn_cmp_prod \ftx \key Hn_cmp_prod \shd Compounding--Productivity \txt Like NOUN-VERB-er in English can-opener. That is, do new compounds of this type continue to appear and to be commonly accepted? \ftx \txt To continue this discussion, highlight and jump to the topic below: \cf Hn_cmp_freq \ftx \key Hn_dim/aug \shd Diminution/augmentation \txt Most languages employ special devices in the noun or noun phrase that indicate unusual sizes. The term for operations that express unusual smallness is diminutive while operations that express unusual largeness are augmentatives. \ftx \txt Typically diminutives also carry an endearing sense, e.g.: \ftx \ftx (6) English: sweety, lambkins \ftx \txt Similarly, augmentatives often carry negative or undesirable connotations: \ftx \ftx Spanish: \ftx (7) durmi-lon 'Sleepyhead/lazybones' \ftx sleep-AUG \ftx \txt There is an apparently universal iconic tendency in diminutives and augmentatives. Diminutives tend to contain high, front vowels whereas augmentatives tend to contain high, back vowels. \ftx \dt 27/Apr/1998 \key Hn_dnom \shd Denominalization \txt The term nominal can be translated 'noun-like'. So to denominalize something is to make it less noun-like. Many languages have morphological processes that derive verbs, adjectives or other grammatical categories from nouns. Such processes are called denominalization. Perhaps the most common type of denominalization makes a possessive verb out of a noun. For example, the Yup'ik Eskimo noun suffix '-ngqerr' means 'to have N' where N is the noun to which the suffix is attached. The following examples (from Reed, et al., 1977) illustrate some common nouns with their denominalized counterparts: \ftx \ftx patu 'lid' patungqerr 'to have a lid' \ftx qayar 'kayak' qayangqerr 'to have a kayak' \ftx irniar 'child' irniangqerr 'to have children' \ftx enr 'bone' enengqerr 'to have a bone' \ftx \ftx Another common denominalization process takes a noun, N, and forms a verb \txt that means 'become N'. These processes are called inchoative (we will distinguish inchoative as a nominal operation from inceptive as a verbal operation, though in the literature the term inchoative sometimes refers to a verbal aspect). For example in Panare the suffix '-ta' when applied to a noun usually means 'to become N': i'yan 'healer', i'yatan 'to become a healer.' \ftx \txt Eskimo is particularly rich in denominalization processes. The meanings of verbs formed by these suffixes include such concepts as the following (in these examples N refers to the noun to which the suffix attaches): \ftx \ftx To go towards N \ftx To be N \ftx To be at N \ftx There is N/there lacks N \ftx To have plenty of N \ftx To be afflicted in one's N \ftx To have cold Ns \ftx To play with N \ftx To hunt for N \ftx To capture N \ftx To eat N \ftx \ftx Sometimes it is difficult to distinguish denominalization from noun-verb \txt compounding (object incorporation, see Hval_decr_ncorp_B). The criterion is that if the denominalizer is independently used as a verb in other contexts with substantially the same meaning, then it is incorporation. If the denominalizer is not attested as a verb (though it probably will be related to a verb), then it is 'true' denominalization. Some of the Eskimo suffixes referred to above are clearly related to verbs, but they are distinct enough in form and meaning from their corresponding verbs to cause us to call them denominalizing suffixes. \ftx \txt Denominalization processes tend to express culturally 'institutionalized' activities (Mithun, 1984). This is illustrated in the Eskimo examples, e.g. hunting, capturing, eating, playing, being cold and being afflicted are undoubtedly concepts that are very common in the Eskimo context. \ftx \fln References: Clark and Clark (1979). \ftx \dt 27/Apr/1998 \key Hn_dnom_adj \shd Denominalization--adjectives \fln incomplete \ftx \key Hn_dnom_adv \shd Denominalization--adverbs \fln incomplete \ftx \key Hn_intro \shd Nominal operations \txt The following questions are pertinent for all nominal operations in the language. They are not repeated each time in the sections which follow, but you will need to take them into account. \ftx \txt a) Is this operation obligatory, i.e. does one member of the paradigm have to occur in every full noun phrase? \ftx \txt b) Is it productive, i.e. can the operation be specified for all full noun phrases and does it have the same meaning with each one? (nothing is fully productive, but some operations are more so than others). \ftx \txt c) Is this operation primarily coded lexically, morphologically or analytically? Are there exceptions? \ftx \txt d) Where in the noun phrase is this operation likely to be located? Can it occur in more than one place? \ftx \dt 27/Apr/1998 \key Hn_num_opt \shd Number--Optional marking \txt Many languages only mark number occasionally. An interesting question for such languages (probably the majority of the world's languages) is when to mark plurality and when to not mark it. \ftx \dt 27/Apr/1998 \key Hn_num_other \shd Number--Other distinctions \txt Other number distinctions are 1) singular vs. dual vs. plural, and 2) singular vs. dual vs. trial vs. plural. The last type is very rare, and the singular, dual, plural type is fairly rare, at least in systems of noun phrase marking; it is more common in verb-coding systems (see Hgrel_vcode). \ftx \dt 27/Apr/1998 \key Hn_num_s/pl \shd Number--Singular/Plural distinctions \txt Nouns and noun phrases are often marked for number. The most common number distinction is between singular and plural. The singular/plural distinction is obligatorily marked for all English nouns that refer to concepts that can be counted (i.e. that consist of individually salient units), e.g. 'dog'-- 'singular' and 'dogs'-- 'plural'. In most number marking systems the singular is unmarked while the plural is marked in some way. Some languages (e.g., all Bantu languages) mark both singular and plural, e.g. Swahili 'umu-ana'-- 'child' vs. 'aba-ana'-- 'children'. Another logical possibility is for the plural to be unmarked while the singular receives a mark of some kind, e.g. 'srik'-- 'dogs', vs. 'sriki-- 'dog' (made up data). This possibility is non-attested as a primary number marking strategy, though it has been rumored to exist as a secondary strategy in some languages. \ftx \dt 27/Apr/1998 \key Hn_other \shd Other operations encoded in the noun phrase \txt Some languages express apparently idiosyncratic functions in the noun phrase. The following is a list of some functions known to occur in noun phrases, but not mentioned in previous sections of this chapter. Do not feel constrained by this list. It is very possible that the language you are describing makes a noun-phrase distinction that has not previously been documented. Include a subection on each important distinction, especially if it is one that is highly unusual or not previously attested: \ftx \ftx Tense/aspect \ftx \ftx Evidentiality \ftx \dt 27/Apr/1998 \key Hn_poss \shd Possessors \txt Languages typically encode many semantic relationships with the same formal construction used to code ownership. We will call such formal constructions possessive constructions, even though the semantic relationship is not always one of possession, e.g. the phrase 'my professor' does not refer to a professor that I 'possess' in the same way that 'my car' refers to a car that I possess. Note that when the relationship is one of true ownership, it is transitive; e.g., 'my car's tires' are also 'my tires'. When the relationship is not ownership, on the other hand, it is not transitive; 'my professor's shoes' are not 'my shoes'. \ftx \txt It is important to distinguish possessive constructions from possessive clauses, discussed in Hpnom_intro_ex. A possessive construction is a noun phrase with two elements -- a possessor and a possessed item. Sometimes the possessor is referred to as the genitive (regardless of whether or not the language has a morphological genitive case). The possessed item is referred to as the possessum or the possessee. \ftx \txt To continue this discussion, highlight and jump to the topic below: \cf Hn_poss_al/inal \ftx \dt 27/Apr/1998 \key Hn_poss_al/inal \shd Possessors--alienable vs inalienable \txt Some languages make a formal distinction based on the semantic difference between alienable and inalienable possession. Alienable possession is the kind of possession which can be terminated, e.g. I can transfer possession of my worldly goods to someone else, hence my relationship to my worldly goods is one of alienable possession. Inalienable possession is the kind of possession that is inherent. Languages which distinguish inalienable possession always include kinship terms and body parts within the class of inalienably possessed items. My head will always be my head, and by brothers and sisters will always be my brothers and sisters. Over and beyond body parts and kinship terms some languages include certain culturally important items within the class of inalienable possessions, e.g. canoes, machetes, etc. Finally, there are usually a few items which semantically seem to go with one class or another, but which are grouped with the other class for no apparent reason, e.g. a language may treat rocks as inalienable and brothers as alienable. (It is always interesting to see how such a language treats husbands and wives). \ftx \txt Ndjuká (Suriname Creole) has what appears to be an inalienable vs. alienable distinction. The following sentence illustrates both types (George Huttar, p.c.): \ftx \fln (5) [ A wagi fu mi ] de gi mi baala. \ftx the vehicle for 1SG COP give 1SG brother \ftx 'My car is for my brother.' \ftx \ftx The bracketed portion of this example illustrates the standard strategy for \txt expressing alienable possession. Note that the possessive pronoun is post-posed and a preposition intervenes between the possessed item and the pronoun. The last NP, 'mi baala', illustrates the standard strategy for expressing inalienable possession. Note that the closer conceptual link between possessor and possessed item in inalienable possession is iconically represented by a closer formal link. \ftx \txt To continue this discussion, highlight and jump to the topic below: \cf Hn_poss_other \ftx \dt 27/Apr/1998 \key Hn_poss_ord \shd Possessors--constituent order \fln References: Ultan (1978). \ftx \key Hn_poss_other \shd Possessors--other kinds \txt Similar but not identical to inalienable possession is inherent possession. Certain items are inherently possessed, e.g. body parts, kinship terms, and items of personal adornment. Other items are not normally possessed, such as trees, the sky etc. Some languages require that inherently possessed items be coded with a possessor. So in such languages you cannot say simply brother or hand. You have to say whose brother or whose hand. Note that there may be inherent possession in a language without an alienable/inalienable distinction. A language with inherent possession may have just one kind of possessor coding, but simply require that some items be possessed, while imposing no such requirement on other items. With alienable/inalienable possession there are actually two (or perhaps more) kinds of possessor coding. \ftx \txt To continue this discussion, highlight and jump to the topic below: \cf Hn_poss_ord \ftx \dt 27/Apr/1998 \key Hord \shd Constituent Order Typology--Contents \ftx \txt This topic contains the following subsections. Highlight and jump to the topic of interest: \shd2 Historical and theoretical background \cf Hord_hist_cl Clauses \cf Hord_hist_ord Constituent order \cf Hord_hist_syn Syntactic correlates \cf Hord_hist_prob Problems with \cf Hord_hist_cncsl Concluding remarks \cf Hord_hist_v-init Distribution of constituent order types around the world--verb initial clauses \shd2 Adpositional phrases \cf Hord_adpp prepositions and postpositions \cf Hord_adpp_vsdepv Adpositional phrases versus dependent verbs \cf Hord_adpp_vsn Adpositional phrases versus nouns \shd2 Constituent order in main clauses \cf Hord_cl_case general case \cf Hord_cl_caut Cautionary note \cf Hord_cl_flex Flexible' constituent order \cf Hord_cl_prag1 Pragmatic factors I \cf Hord_cl_prag2 Pragmatic factors II \cf Hord_cl_prag3 Pragmatic factors III \cf Hord_cl_rigd 'Rigid' constituent order \shd2 Verb phrase \cf Hord_vp_v-aux Verb phrase (Verb, Aux) \cf Hord_vp_v-aux-adv **Verb phrase (Verb, Aux, adverbs) \shd2 Other topics \cf Hord_cmpar Comparatives \cf Hord_dist_com Distribution of constituent order types around the world--common orders \cf Hord_np Noun phrase (Det, Num, Gen, Mod, RC, N) \cf Hord_q_part Question particles \cf Hord_q_wd Question words \shd2 Summary \cf Hord_sum Summary \cf Hord_sum_chart Greenberg Universals Chart \dt 12/May/1998 \key Hord_adpp \shd Adpositional phrases (prepositions and postpositions) \txt See Hord_cl_case for the distinction between adpositions and case markers. Whether an adposition is a preposition or a postposition should be determined on the basis of full noun phrases, not pronouns, anaphoric clitics or agreement markers. For example: \ftx \ftx PREPOSITIONS: \fln (11) Spanish: a. en la mesa b. dentro de la casa \ftx 'on the table' 'inside of the house.' \ftx \ftx c. sobre mi carro d. hasta la tarde \ftx 'over my car' 'until the afternoon.' \ftx \ftx POSTPOSITIONS: \fln (12) Japanese: a. biku no \ftx 'of/inside/near the fishbasket' \ftx \ftx b. kookyu ue \ftx 'Above the palace.' \ftx \txt To continue this discussion, highlight and jump to the topic below: \cf Hord_adpp_vsdepv \ftx \dt 27/Apr/1998 \key Hord_adpp_vsdepv \shd Adpositional phrases versus dependent verbs \txt Adpositions derive historically from nouns or verbs. For some languages, namely serial verb languages (Hcplx_sv), it may be difficult to decide whether a given form is an adposition or a dependent verb. In the following example from Akan (Emmanuel Osam, p.c.), the form 'wô' is a perfectly good verb meaning 'to be at' (13a). It also functions as a locative preposition (13b): \ftx \fln (13) a. ô-wô Eugene 'He is in Eugene.' \ftx 3SG-be:at \ftx \ftx b. o hun no wô Eugene 'We saw him in Eugene.' \ftx 1PL see 3PL in \ftx \txt One indication that 'wô' is a verb in 13a but not in 13b is that in predicate locative clauses like 13a 'wô' can take verb agreement, whereas in 13b, verb agreement is not possible. Such 'tests' show that 'wô' really does belong to different syntactic categories in these two examples. \ftx \txt In Ndjuká, an English based creole language of Suriname, the verb 'gi'-- 'give' also marks a locative, dative or benefactive case (George Huttar, p.c.): \ftx \fln (14) Boo gi den gi mi. `Honk at them for me.' \ftx blow give 3PL give 1SG \ftx \txt For many African and Sino-Tibetan languages there are no formal properties to distinguish serial verbs from prepositions in many cases. One simply has to go by the semantics. For example, in Akan, although there is a clear serial verb pattern that the structure in 15 fits into, there is another pattern that would be more appropriate if the verb 'de' really did refer to an independent act of 'taking': \ftx \fln (15) o de eñkrante; o tya duabasa 'He took a sword and then cut \ftx he take sword he cut branch off a branch.' \ftx \txt Similarly, in Ndjuká there is a distinct way of saying 'honk and then give to me . . .' \ftx \txt To continue this discussion, highlight and jump to the topic below: \cf Hord_adpp_vsn \ftx \dt 27/Apr/1998 \key Hord_adpp_vsn \shd Adpositional phrases versus nouns \txt In some languages the adpositions come from nouns. For example, in English we say 'on top of'. This is a complex preposition consisting partially of the noun 'top'. For many languages the adpositions are body part nouns, e.g. 'back' for 'behind', 'face' for 'in front', 'head' for 'up' and 'foot' for 'down' (Casad 1982). In most languages the set of basic adpositions is rather small, consisting of perhaps five or six forms. Other, more complex relational notions are embodied by complex adpositions built up out of combinations of adpositions and nouns. English is unusually rich in basic prepositions. These are: \ftx \fln (16) at, to, from, in, out, on, over, under, around, though, for, by, with, \ftx along, etc. \ftx \fln Other very common prepositions in English are compounds: \ftx \fln (17) into, upon, toward (to + the suffix ward), on top of, underneath, \ftx behind, below, beneath, out of, next to, etc. \ftx \fln References: Matisoff (1973), DeLancey (1991), Welmers (1973), Casad (1982) \ftx \dt 27/Apr/1998 \key Hord_cl_case \shd Constituent order in main clauses--general case \txt The isolation of one clause type as 'pragmatically neutral' is likely to be problematic for any language. It is especially difficult to find pragmatically neutral clauses that contain one or more full noun phrases. This is because many (if not most) languages avoid the use of full noun phrases in discourse. For these languages the most frequent, pragmatically unmarked clause type has no nouns whatsoever. A general way to distinguish pragmatically neutral clauses is to avoid the following: \ftx \ftx 1. dependent clauses \ftx 2. paragraph-initial clauses \ftx 3. paragraph-final clauses \ftx 4. clauses that introduce participants \ftx 5. clearly contrastive clauses (e.g. clefts, answers to questions, etc.) \ftx 6. Clauses with pronominal arguments. \ftx \txt You can be fairly certain that the clauses that remain are largely pragmatically neutral. If in these remaining clauses there are examples of transitive verbs with two full NPs and if there is consistency of order of those NPs with respect to the verb, then you can probably make a claim of basic constituent order. Unfortunately, this is a rare situation. \ftx \txt If pragmatically neutral clauses tend to have a verb and one or fewer noun phrases, say so. But try to make some statement of neutral constituent order in transitive clauses (i.e. by including some categories mentioned above, though you'll have to hedge by saying 'this isn't really very "neutral"'). Most languages can fairly easily be classified as 'verb-initial', 'verb-medial' or 'verb-final' even if the relative orders of A and P are indeterminate. \ftx \txt Note that the orders of elements in other construction types is not evidence for a particular order in main clauses. Languages are too often inconsistent for us to take non-main-clause orders as heuristics for main clause order. \ftx \txt To continue this discussion, highlight and jump to the topic below: \cf Hord_cl_rigd \ftx \dt 27/Apr/1998 \key Hord_cl_caut \shd Constituent order in main clauses--Cautionary note \txt One final caution. Almost any language can be claimed to allow all possible orderings of A, P and V if constructions from a wide enough range of syntactic statuses are admitted. The following are some examples from English: \ftx \fln (10) a. Fred skins mules. AVP \ftx \ftx b. It's mules that Fred skins. PAV \ftx \ftx c. There he sits, skinning mules, that Fred. VPA \ftx \ftx d. Skins 'em, Fred does to them mules. VAP \ftx \ftx e. Fred's a mule skinner. APV \ftx \ftx f. That mule skinner's Fred. PVA \ftx \txt To determine 'pragmatic ordering principles' that would account for this variation would amount to determining the discourse functions of all of the various clause types represented (clefts, participial clauses, nominalizations, etc.). Such an enterprise is not necessarily unreasonable, as long as one understands that it is not merely a question of determining the functions of various constituent orders. \ftx \txt However, care should be taken in conducting and evaluating studies of constituent order variation in any language. For one thing, linguists who conduct constituent order studies of languages for which they do not understand the basic clause types are particularly prone to bizarre analyses. Second, grammatically marginal clauses such as 10d may be accepted by consultants in preliterate or newly literate societies more readily than they would be by speakers who have more static perceptions of their language. Or the marginality of such clauses may be difficult for pre- literate consultants to express. Finally, and most importantly, the syntactic status of many clause types is often in a state of flux. For example, nominalizations such as 10e and f often function very much like verbal clauses with the nominalizer (-er in English) functioning as a tense, aspect or mode marker (see e.g. T. Payne 1990). In fact, noun and verb morphology often overlap to such an extent that it is difficult to determine whether one is dealing with a verbal clause or a predicate nominal based on a nominalized verb. Whereas it is quite clear (to fully competent native speakers trained in linguistics) that 10a and 10e represent very different construction types, and that order is certainly not the only respect in which they differ, out of context these clauses seem to convey very similar concepts. This sort of functional similarity leads in many languages to a reanalysis of the predicate nominal construction (as in 10e) as a kind of verbal predication. If the language under study is in the process of such a change (and even if it is not), the probability of mistakenly including predicate nominals in a study of constituent order in verbal clauses is quite high (especially if the language uses a zero copular element in predicate nominals, see Hpnom_pnom_incl/eq). On the other hand, if the language has already accomplished the reanalysis from predicate nominal to verbal clause (a process that is one of the major sources of ergative constructions), it would be fully appropriate to conduct a study to determine the discourse pragmatic functions of the two construction types. However, it would still be questionable to consider such a study to be simply one of determining the functions of variant constituent orders. \ftx \fln References: Hawkins 1983; D. Payne 1986, 1987; Mithun 1987 \ftx \dt 27/Apr/1998 \key Hord_cl_flex \shd Constituent order in main clauses--`Flexible' constituent order \txt languages So-called 'flexible' constituent order languages are those in which some principle other than grammatical relations governs the order of nominals in a clause. In Biblical Hebrew the order of noun phrases with respect to the verb is determined largely by pragmatic factors. In general, new, indefinite information occurs preverbally whereas given, definite information occurs post-verbally (Givón 1984:208ff). The following passages from Biblical Hebrew illustrate this variation for both S, P and Oblique objects: \ftx \ftx Biblical Hebrew, VS vs. SV order: \fln (5) a. VS, Already identified(definite)/identifiable subject \ftx \ftx va-yavo'u shney ha-mal'axim Sdom-a b-a-'erev \ftx and-came two DEF-angels Sodom-LOC in-the-evening \ftx 'So the two angels came to Sodom in the evening, \ftx \ftx b. SV, Previously unidentified subject \ftx \ftx vî-Lot yoshev bî-sha ar Sdom; \ftx and-Lot sitting at-gate of Sodom \ftx and Lot was sitting at the city's gate; \ftx \ftx c. VS, Already identified subject \ftx \ftx va-yar' Lot \ftx and-saw Lot \ftx and Lot saw (them) \ftx \ftx d. V-Only, Highly topical subject. \ftx \ftx va-yaqom \ftx and-rose \ftx and rose . . .' (Genesis 19:1) \ftx \ftx Biblical Hebrew, VP vs. PV order: \fln (6) a. VP, continuing, identifiable object: \ftx \ftx . . . va-yiqah 'alohim 'et-ha-'adam \ftx and-took God ACC-DEF-man \ftx '. . . and God took the man \ftx \ftx b. VP(PRO), continuing identifiable object: \ftx \ftx va-yanihe hu bî-gan 'eden \ftx and-put him in-garden:of Eden \ftx and put him in the garden of Eden \ftx \ftx c. Anaphoric agreement (dependent clause) \ftx \ftx lî-'ovd-o u-lî-shomr-o; \ftx to-work-it and-to-guard-it \ftx to work and guard it; \ftx \ftx d. VP, continuing, identifiable object. \ftx \ftx va-yîsav YHWH 'elohim 'al ha-'adam le-'mor: \ftx and-ordered YHWH God unto the-man to-say \ftx and God ordered Adam: \ftx \ftx e. PV, new object in contrast \ftx \ftx "mi-kol 'es ha-gan 'axol to'xel, \ftx from-all tree:of the-garden eating you:eat \ftx "You may eat from all the trees in the garden, \ftx \ftx f. PV, new object in contrast \ftx \ftx "u-me-'es ha-da'at tov ve-ra' lo' to'xel . . . \ftx and-from-tree:of the-knowledge:of good and-evil NEG you:eat \ftx but from the tree of knowledge of good and evil you may not eat..." \ftx (Genesis 2:15-17) \ftx \txt To continue this discussion, highlight and jump to the topic below: \cf Hord_cl_prag1 \ftx \dt 27/Apr/1998 \key Hord_cl_prag1 \shd Constituent order in main clauses--Pragmatic factors I \txt Mithun (1987) questions the notion that every language should be describable in terms of a basic order of constituents determined by grammatical relations. She argues that in at least three languages, Cayuga (Iroquoian of Ontario), Ngandi (Australian of East Arnhem Land) and Coos (of Oregon) grammatical relations have no direct effect on constituent order. Instead, pragmatic status of the nominal constituents is the best determiner of the order of those constituents with respect to the verb. Here we will briefly summarize Mithun's data. \ftx \txt In all three languages some form of the following generalization summarizes constituent ordering: \ftx \ftx New, indefinite or otherwise 'newsworthy' information \ftx is placed early in the clause. \ftx \txt In the following examples from Cayuga, we see PV order when the P refers to a non-specific, newly mentioned entity (7a), and VP order when the P refers to a specific identified item: \ftx \fln (7) a. P V \ftx katsihwá' kihsa:s (lowered o under i of 2nd word) \ftx hammer I-seek \ftx 'I am looking for a hammer.' (said in a hardware store, with \ftx no particular hammer in mind) \ftx \ftx V P \ftx b. to: ti' nika:nô:' nê:kyê katsíhwa'? \ftx how then so-it-costs this hammer \ftx 'How much does this hammer cost?' (holding a specific hammer) \ftx \ftx Ngandi (from Heath 1978:206) follows a similar principle: \fln (8) a. S V \ftx Na¢uweleñ-un gu-jark-yun gu-ja-walk, . . . \ftx then-ABS GU-water-ABS GU-now-go:through \ftx 'Then water passes through,' (First mention of water) \ftx \ftx b. V S \ftx Na¢uweleñ-un gu-ja-geyk-da-ni gu-jark-yun \ftx then-ABS GU-now-throw-AUG-PR GU-water-ABS \ftx 'Then the water rushes through,' (Subsequent mention of water) \ftx \ftx Coos (Frachtenberg 1913:7): \fln (9) a. P A V \ftx TE tc!i'cil yüL is yö'qat . . . \ftx that matting we two split:it \ftx 'Let's split this mat.' (First mention of mat) \ftx \ftx (they did so, and went down to examine the earth. The earth \ftx was still not solid, even . . .) \ftx \ftx b. V P \ftx i lau tci uxhi'touts hE tc!icil. \ftx when that there they:two:put:it:down the matting \ftx 'after they had put down the mat,' (subsequent mention of mat). \ftx \txt To continue this discussion, highlight and jump to the topic below: \cf Hord_cl_prag2 \ftx \dt 27/Apr/1998 \key Hord_cl_prag2 \shd Constituent order in main clauses--Pragmatic factors II \txt In the three languages discussed by Mithun, the positions of all nominal clause constituents (i.e. A, P and Oblique elements) is apparently determined to a large extent by pragmatic factors (though the data are not always clear). For some languages, one nominal element exhibits a fairly fixed position (variable only under extreme pragmatic pressures), while another is more variable. Some languages that we are aware of that operate in this way are: \ftx \ftx Spanish (fixed VP, flexible A) \ftx \ftx Guaymí (fixed PV, flexible A) \ftx \ftx Panare, Nadëb (fixed VA, flexible P). \ftx \ftx Apurinå (fixed AV, flexible P) \ftx \txt The areas of the world in which languages seem particularly sensitive to pragmatic ordering principles are the Americas, Australia and to a lesser extent Austronesian and South Asian languages. The Slavic languages are apparently the most pragmatically sensitive in the Indo-European family, though they do not approach the degree of pragmatic variability demonstrated by such languages as Cayuga (Mithun 1987:309-310). It should be emphasized, however, that pragmatic factors influence constituent order in all languages to one degree or another. For some languages pragmatic factors are so dominant that it is difficult or impossible to describe the 'basic' constituent order in terms of grammatical relations. On the other hand, even languages in which pragmatics clearly dominates constituent order, grammatical relations may still have some correlation with particular clause positions. \ftx \txt To conclude this discussion, highlight and jump to the topic below: \cf Hord_cl_prag3 \ftx \dt 27/Apr/1998 \key Hord_cl_prag3 \shd Constituent order in main clauses--Pragmatic factors III \txt It is interesting to note that all three of the languages discussed by Mithun share several morphosyntactic properties, in addition to pragmatically determined constituent order. Many of these properties also hold for other languages that have been claimed to exhibit pragmatically controlled constituent order (e.g. Panare, Papago, Ute). It remains to be seen whether these properties characterize a substantive linguistic type: \ftx \ftx 1. Polysynthetic morphological typology. \ftx \ftx 2. Agent and/or patient marking on the verb. \ftx \ftx 3. A tendency towards ergative case marking on NPs. \ftx \ftx 4. "Loose" syntactic structure, i.e. nominal elements may \ftx occur under a different intonation contour than the verb, \ftx and adverbial, or other clause-level elements, may freely \ftx intervene between the verb and the nominal elements. \ftx \ftx 5. A tendency to avoid the use of full noun phrases in \ftx discourse. The occurrence of clauses with two or more noun \ftx phrases is extremely rare. \ftx \dt 27/Apr/1998 \key Hord_cl_rigd \shd Constituent order in main clauses--'Rigid' constituent order \txt languages In the following examples, the P argument is emphasized. There are many constituent types that fulfill the P role in terms of order. Some of these constituent types are illustrated for each of the constituent orders exemplified below. In general, AVP and APV languages tend to be more rigid than verb-initial languages. Languages which appear to be verb-initial are commonly sensitive to pragmatic roles rather than grammatical relations. \ftx \ftx English, AVP (SVO): \fln (1) a. The man hit the ball. Direct object \ftx b. The man went to the store. Oblique object \ftx c. The woman knew that her house was on fire. Object Complement \ftx d. The woman wanted to leave. Object Complement \ftx e. She told the man to leave. Object Complement \ftx f. The man was tall. Predicate adjective \ftx g. The man was a teacher. Predicate nominal \ftx \ftx Sherpa, APV (SOV) (Examples from Givón 1984:191): \fln (2) a. ti 'ang-di tu'pe no Adjectival Predicate \ftx the baby-DEF small be \ftx 'The baby is small.' \ftx \ftx b. ti mi-ti dyeken yin Nominal predicate \ftx the man-DEF teacher be \ftx 'The man is a teacher.' \ftx \ftx c. ti mi-ti-gi laga kyaa-sung Direct object \ftx the man-DEF-ERG work do-PERF \ftx 'The man did work.' \ftx \ftx d. ti mi-ti-gi cenyi caaq-sung Direct object \ftx the man-DEF-ERG cup break-PERF \ftx 'The man broke the cup.' \ftx \ftx e. ngaa yambur-la gal-yin Oblique object \ftx I:ABS Katmandu-LOC go-PERF \ftx 'I went to Katmandu.' \ftx \ftx f. ti mi-ti laga ki-tup no-(u)p kyaa-sung Object Complement \ftx the man-DEF work do-INF want-INF do-PERF \ftx 'The man wanted to work.' \ftx \ftx Jacaltec (Mayan), VAP (VSO) (Examples from Craig 1977): \fln (3) a. xa' ix te' hum wet an Direct and \ftx gave CL:she CL:the book to:me PL indirect objects \ftx 'She gave the book to me.' \ftx \ftx b. xahtoj naj yiban no' cheh Oblique object \ftx go:up CL:he on CL:the horse \ftx 'He climbed on the horse.' \ftx \ftx c. sonlom naj Predicate nominal \ftx marimba:player CL:he \ftx 'He is a marimba player.' \ftx \ftx d. c'ul ye ix Predicate adjective \ftx good be CL:she \ftx 'She is fine.' \ftx \ftx Malagasy, VPA (VOS): \fln (4) a. manasa lamba Rasoa Direct object \ftx wash clothes Rasoa \ftx 'Rasoa is washing clothes.' \ftx \ftx b. nanome vola an-Rabe aho Direct and \ftx gave money to-Rabe I indirect objects \ftx 'I gave money to Rabe.' \ftx \ftx c. manaiky manasa ny zaza Rasoa Object Complement \ftx agree wash the baby Rasoa \ftx 'Rasoa agreed to wash the baby.' \ftx \txt To continue this discussion, highlight and jump to the topic below: \cf Hord_cl_flex \ftx \dt 27/Apr/1998 \key Hord_cmpar \shd Comparatives \txt A COMPARATIVE is a construction where two items are compared according to some quality, e.g. 'my father is bigger than your father'. Many languages don't have a syntactically distinct comparative construction. These languages express comparison by simply juxtaposing two (or more) clauses expressing the degree to which the compared entities exhibit the quality in question, e.g. to say 'my father is bigger than your father', one would say something like 'your father is big. My father is very big.' The crucial elements of a grammaticalized comparative construction are: 1) the known STANDARD against which the subject of the clause is compared, 2) the MARKER that signals that the clause is a comparative construction, and 3) the QUALITY by which the subject is compared against the standard. The standard is a noun phrase, the marker can be a special particle, an adposition or an affix and the quality is normally expressed via an adjective. For example: \ftx \ftx JAPANESE, STANDARD-MARKER-QUALITY \fln (18) Inu ga meko yori ookii 'The dog is bigger than the cat.' \ftx dog NOM cat than big \ftx STD MKR QUAL \ftx \txt The 'dog' in this clause is the subject of the comparison. That is, it is the item whose size is being compared to that of a known standard. The position of the subject of a comparative construction is not as typologically significant as the relative positions of the standard, marker and quality. STD-MKR-QUAL order is common in PV languages. \ftx \ftx IRISH, QUALITY-MARKER-STANDARD \fln (19) tá an madadh nios-mó ná an cat 'The dog is bigger than the cat.' \ftx is the dog big -er than the cat \ftx QUAL --MKR-- --STD-- \ftx \txt The order QUAL-MKR-STD is common in VP languages. Comparative constructions often inconsistent with the general constituent order pattern of the language. \ftx \dt 27/Apr/1998 \key Hord_dist_com \shd Distribution of constituent order types around the world--common orders \txt APV/SV is the most common constituent order type; it occurs in virtually every area of the world. APV/SV languages tend to be rather inflexible in allowing alternative orders. Japanese is a nearly prototypical APV/SV language. AVP/SV is the second most common constituent order type. It also occurs in virtually every area of the world. AVP languages tend to allow some flexibility of order, e.g. English, a fairly consistent AVP/SV language, allows alternatives such as PAV ('Beans I like') and VS ('Here comes my bus'). However, these alternatives are clearly pragmatically marked, occurring rarely and only in very well-defined discourse environments. Together APV/SV and AVP/SV languages constitute approximately 80% of the world's languages (Tomlin 19##). \ftx \txt The third most common constituent order type is VAP/VS. This type is well represented in Austronesian languages (Philippines, Pacific Islands, Madagascar and the interior of Indonesia and Malaysia). It is also quite common in the Americas. Verb-initial languages in general tend to allow quite a bit of flexibility of constituent orders. For this reason, if discourse in a given language contains many verb-initial clauses, it may be difficult to determine what the basic constituent order is. \ftx \txt These three common types, APV/SV, AVP/SV and VAP/VS, account for over 90% of the world's languages. For the other 10%, determination of basic constituent order in terms of grammatical relations is likely to be difficult to impossible. What the three major constituent order types have in common is that the A precedes the P in transitive clauses. In the three other logically possible types the P precedes the A. The tendency for A to precede P in basic, pragmatically neutral clauses is so overwhelming that it is extremely unlikely that it could have arisen by chance. This fact has led many researchers to reflect on possible cognitive motivations for the categories of A and P. That is, many have asked the question 'what is it about nominals categorized as A and P that causes languages to practically always order A before P in basic clauses?' Some of this research is discussed in Chapter #8. \ftx \dt 27/Apr/1998 \key Hord_hist_cl \shd Historical and theoretical background-- Clauses \txt For many years linguists have noticed that discourse tends to be expressed in CLAUSES. The notion of a clause seems so intuitive, so central to our conception of language that it is almost incomprehensible to imagine a theory of language that did not include it. There is good reason for this intuition; a significant portion of cognition and reasoning in mature human beings is PROPOSITIONAL. That is, humans mentally combine and manipulate concepts in chunks involving one or two conceptual entities and a relation, activity or property concerning them. Messages are multi-propositional; they consist of groups of such conceptual 'chunks' each contributing some bit of information to the message to be communicated. The clause (or sometimes 'sentence') is the linguistic encoding of a proposition; a proposition is a conceptual notion, whereas a clause is its formal morphosyntactic instantiation. \ftx \txt NOTE: Of course the grammatical clause consisting of a 'subject' and 'predicate' is not the only possible linguistic instantiation of a proposition. In conversation, in particular, propositions are often embodied in hypo-sentential linguistic forms, such as phrases, interjections, incomplete structures, etc. Furthermore, propositions can be instantiated in non-linguistic media as well. For the purposes of a reference grammar, however, the field linguist will want to limit the domain of discussion to conventionalized linguistic instantiations of propositions. \ftx \txt Even as propositions consist of entities and a property, activity or relation, so clauses tend to consist of nouns and a predicating element, either a nominal/adjectival/stative element, or a verb. Given this characterization of propositions, there is no immediately obvious 'natural' order in which the component parts of a proposition should be expressed. And in fact we find that the order in which the predicating element (hereafter 'verb' or simply 'V' for short) and related nouns occur in clauses varies considerably from language to language, and even within the same language. \ftx \txt To continue this discussion, highlight and jump to the topic below: \cf Hord_hist_ord \ftx \dt 27/Apr/1998 \key Hord_hist_cncsl \shd Historical and theoretical background-- Concluding remarks \txt In spite of these rather glaring problems with the Greenberg typology, it is still helpful to a reader of a grammar sketch to have some sense of the basic constituent order type the language represents. Determining the basic constituent order is a good way to begin to present this basic information. However, in this manual we will replace the traditional two-way distinction between subject and object with a three way distinction between subject of transitive verb, subject of intransitive verb and object of transitive verb. These terms are displayed in table 5.1. \ftx \fln Table 4.1: Semantico-syntactic roles (Comrie 1978) \ftx +---+ \ftx | A| Subject of transitive (AGENT-like argument) \ftx Subject -| | \ftx | +-+-+ \ftx | |S| | Subject of intransitive (only argument) \ftx +-+-+ | \ftx Absolutive --| | \ftx |P | Object of transitive (PATIENT-like argument) \ftx +---+ \ftx \txt Within this framework, the SUBJECT category consists of the set of A together with S, while the ABSOLUTIVE category consists of the set of S together with P. These terms are elaborated and given more substance in Hgrel_typ_erg/abs. For purposes of constituent order typology, then, languages can be characterized in terms of A, S, P and V rather than simply S, O and V. This new terminology both provides for languages that treat the intransitive subject like a transitive object in terms of constituent ordering (e.g., Kuikúro, see Hgrel_typ_ord) and provides a bridge into the more detailed treatment of grammatical relations in the Hgrel section. \ftx \dt 27/Apr/1998 \key Hord_hist_ord \shd Historical and theoretical background-- Constituent order \txt Descriptive linguists have long observed that individual languages tend to structure their clauses in characteristic ways; some languages have a propensity to place the verb at the end of a clause, others at the beginning, and still others place it somewhere in the middle. Among the nominal (noun-like) elements, the important categories have traditionally been assumed to be Subject and Object (S and O). From this point of view there are six logically possible constituent orders: SOV, SVO, VSO, VOS, OSV and OVS. Languages can often be categorized according to which of these orders is typical, or 'basic'. Though the assumption that Subject and Object are indeed the universal categories relevant to the ordering of nominal elements in a clause has been seriously questioned (D. Payne, 1986, 1987, Mithun 1987), this typology is often a useful starting point for conceptualizing the syntactic structure and investigating the functions of constituent order variation in a field language. \ftx \txt To continue this discussion, highlight and jump to the topic below: \cf Hord_hist_syn \ftx \dt 27/Apr/1998 \key Hord_hist_prob \shd Historical and theoretical background-- Problems with \txt Greenberg's work Since 1966, much research has revealed problems with Greenberg's original six-way typology. Significant revisions, criticisms and extensions of Greenberg's work are found in D. Payne (1985) Mithun (1987) and Dryer (1988). The two most fundamental problems with the original typology are: 1) the difficulty in identifying the basic constituent order for many languages of the world, and 2) the fact that Greenberg's typology simply assumed that languages order their meaningful elements according to the grammatical relations of Subject and Verb. \ftx \txt These problems are both probably due to a Euro-centric bias among linguists. That is, since the languages that most linguists speak order their main clause elements according to status as subject and object, and since these categories are readily identifiable in most Indo-European languages, it was assumed that all languages must operate in the same way. However, subsequent research has revealed that in a great many languages of the world grammatical relations just are not as clearly identifiable as they are in Indo-European. Furthermore, even when they are identifiable, it is often doubtful whether any significant correlations can be drawn between constituent orders and grammatical relations. While nouns themselves are relatively easy to identify universally, there are many different properties (sometimes 'statuses' or 'roles') that noun phrases exhibit when they enter into syntactic constructions. The roles of subject and object are central in Indo-European languages, but there is no a priori reason to expect that other characteristics, such as agent/patient, definite/indefinite, given/new, animate/inanimate, big/small or abstract/concrete would not also affect the positions of nouns in clauses. Hence, it should be a matter of empirical observation, not a priori assumption, whether or not constituent orders in a language are sensitive to grammatical relations such as subject and object. \ftx \txt Languages which organize their constituent orders according to some principle other than grammatical relations are often called 'free' or 'flexible' constituent order languages. Sometimes they are called NON-CONFIGURATIONAL languages (Hale 1983). More recently this typology has been refined, and the term PRONOMINAL ARGUMENT languages has been used (Jelinek 1984, 1988). Note that from the point of view of language as a symbolic system, it would be highly unlikely that such an obvious and easily manipulable structural variable as the order of words in a clause would not be exploited to code some important functional distinction. Indeed, discourse-based studies of 'free' constituent order languages show that constituent order in such languages is far from random (see Hord_cl_flex below and the references cited therein). \ftx \txt To continue this discussion, highlight and jump to the topic below: \cf Hord_hist_cncsl \ftx \dt 27/Apr/1998 \key Hord_hist_syn \shd Historical and theoretical background-- Syntactic correlates \txt of constituent order Greenberg (1963) observed that several syntactic characteristics tend to correlate with certain of the six basic constituent orders mentioned above. For example, languages whose basic constituent order is SOV almost invariably mark oblique noun phrases with post-positions rather than prepositions. Such observations have come to be known as 'Greenberg's universals', even though very few of them actually describe universal characteristics of language. It is important to note that Greenberg simply observed certain statistically significant correlations. He did not attempt to provide a reason for (i.e. to 'motivate') those correlations. In this sense, Greenberg did not attempt to predict constituent orders in as yet unstudied languages. Nevertheless, Greenberg's work stimulated the sub-field known as 'typological linguistics' and has continued to be very influential. \ftx \txt Much subsequent work on constituent order typology has focussed on motivations for the correlations observed by Greenberg, e.g. Lehmann (1973) Vennemann (1974), and Hawkins (1983). In order to 'motivate' a correlation between two syntactic characteristics (e.g. SOV order and postpositions), the research paradigm has been to first show that the correlation between the characteristics is not random, and then to show that the correlation could not be otherwise, e.g. logically exclude the possibility that the correlation could have been other than the observed facts. In this sense, these subsequent studies attempted to make predictions of constituent orders. That is, they hoped that given certain key constituent orders (say main declarative clause order and order of genitive and head in the noun phrase) for any language they would be able to accurately guess what the other constituent orders would be. \ftx \txt To continue this discussion, highlight and jump to the topic below: \cf Hord_hist_prob \ftx \dt 27/Apr/1998 \key Hord_hist_v-init \shd Distribution of constituent order types around the world--verb initial clauses \txt If a language employs verb-initial clauses quite frequently (approximately 25% or more) in discourse, it will probably be quite difficult to determine a 'basic' order. This will be because of either or both of the following tendencies: verb-initial languages often avoid the use of full noun phrases as much as possible, preferring to rely on pronouns and/or agreement markers, and 2) verb-initial languages often are less sensitive to grammatical relations than are other languages. That is, their basic clause structure can be insightfully described as a verb followed by a string of noun phrases. The order of the noun phrases following the verb tends to be determined by pragmatic or semantic factors that are only indirectly characterizable in terms of grammatical relations. This is true for Philippine languages and many languages of Indonesia and Malaysia. It is also true of many verb-initial languages of the Americas, e.g., Mayan and other Meso- American languages. \ftx \dt 27/Apr/1998 \key Hord_np \shd Noun phrase (Det, Num, Gen, Mod, RC, N) \txt Noun phrase elements include determiners (4.1.3), numerals (Hgr_mod_num_sys), genitives (possessors) (Hn_poss), modifiers (i.e. attributive adjectives, Hgr_mod_adj_sem), relative clauses (Hcplx_relcl), and the head noun. The head noun is the noun that is modified by all the other elements, e.g. 'dogs', in 'those three big black dogs that are always barking at me.' It is the noun that refers to the same entity that the whole phrase refers to. Some languages also have noun classifiers that optionally or obligatorily occur in noun phrases (4.1.6). \ftx \dt 27/Apr/1998 \key Hord_q_part \shd Question particles \txt Question particles are discussed in Hprag_ndecl_q. In this section, simply exemplify these elements, especially noting their positions with respect to other clause elements. \ftx \dt 27/Apr/1998 \key Hord_q_wd \shd Question words \txt Question words are discussed in Hprag_ndecl_q_qwd. In this section, simply exemplify these elements, especially noting their positions with respect to other clause elements. \ftx \dt 27/Apr/1998 \key Hord_sum \shd Summary \txt Very few languages conform 100% to the general expectations provided by these classic studies. And, as with morphological typology, there is no quantitative method for determining how close a particular language is to its 'ideal' type. In this section of the grammar the fieldworker should summarize the constituent orders of the language in comparison to Greenberg's universals. The language can then be characterized impressionistically as a 'very consistent', 'fairly consistent', or 'inconsistent' language of constituent order type X. For example, English is a fairly consistent AVP language, exhibiting inconsistency only in the placement of its descriptive modifiers in the NP, and in allowing both pre- and post-nominal genitives ('John's house' alongside 'the house of John'). Japanese is a very consistent APV language, Yagua is an inconsistent VPA language, etc. \ftx \txt It is important to note that particular orders within non-main-clause units are not evidence for any main clause order. For example, a language may have the basic order of AVP in main clauses, but every other unit may have orders consistent for APV languages (e.g. post-positions, postverbal auxiliaries, prenominal modifiers, etc.). This is not evidence that the language has APV word order. It may indicate that APV is a diachronically older order for the language, but it is not an argument for a particular synchronic order. Greenberg's universals are simply statistical correlations based on a sample of about 30 languages. They are not predictions of what one will find in any given language. Languages that deviate from Greenberg's ideal types do not 'violate' Greenberg's universals. They are simply inconsistent with the ideal type. Since the majority of languages of the world are inconsistent in this sense, it may be more appropriate to dub a perfectly consistent language as a violation of expectations. Sometimes it may be appropriate to call a language a PV type language, even though the basic constituent order in main clauses is VP, and vice versa. This point of view ascribes no particular importance to main clause constituent order -- it is simply one property among many. If a language has all the characteristics of a VP language except it has PV basic order in main clauses, it is a fairly consistent 'VP-type' language. It only deviates in one respect. \ftx \txt For a chart of Greenberg Universals, jump to: \cf Hord_sum_chart \ftx \dt 27/Apr/1998 \key Hord_sum_chart \shd Greenberg Universals Chart \txt The following chart summarizes expectations for VP and PV languages based on Greenberg's universals. The reader is encouraged to consult Appendix 1 of Greenberg (1963) for a more inclusive summary: \ftx \fln Table #4.2: Summary of Greenberg's Universals (from appendix 2 of \ftx Greenberg 1963) \ftx Greenberg's \ftx Universal Parameter Correlation \ftx \ftx #1 Main Clauses V-P P-V \ftx #3,4 Adpositions Prepositions Postpositions \ftx #2 Genitive (possessor) \ftx and head Noun N-G G-N \ftx #17 Head Noun & Modifier N-M M-N \ftx #24 Relative Clauses \ftx and head Noun N-RelCL RelCL-N \ftx #22 Comparatives Adj-Mkr-Std Std-Mkr-Adj \ftx #16 Inflected Auxiliaries Aux-V V-Aux \ftx #9 Question particles Sent. initial Sent. final \ftx #12 Question words Sent. initial Sent. initial \ftx or elsewhere \ftx #27 Affixes Prefixes Suffixes \ftx \key Hord_vp_v-aux \shd Verb phrase (Verb, Aux) \txt Auxiliaries are verbs in that they satisfy the morphosyntactic definition of verbs (whatever that may be for the language), e.g. they occur in the position of a verb and they carry the inflectional information (subject/object 'agreement' and TAM marking) normally associated with verbs. However, they are auxiliary in that they do not embody the major conceptual relation, state or activity expressed by the clause. They are often semantically 'empty' (e.g. 'do' in English 'he does go to school'), or they convey TAM-like information, e.g. 'will', 'should', 'can', 'might', 'must', 'hæfta', 'yústa' etc. in English. \ftx \txt Auxiliaries normally derive from full verbs. The most likely verbs to become auxiliaries are stative verbs such as be, stand and sit. The next most likely sources for auxiliaries are simple verbs of motion such as go and come. Finally, complement-taking verbs such as say, finish, start, permit, make, force, and want also often become auxiliaries. \ftx \txt If a language has a word that displays the necessary inflectional information in the clause but is distinct from the verb that embodies the main lexical content of the clause, then this word can be called an auxiliary. In a few languages such forms do not seem to come from verbs at all, and the way in which they express the inflectional information is not very much like the way verbs do, i.e. they may exhibit irregular or completely distinct inflectional paradigms (Steele 1978). \ftx \dt 27/Apr/1998 \key Hord_vp_v-aux-adv \shd Verb phrase (Verb, Aux, adverbs) \ftx \ftx \key Hpnom \shd Predicate nominals and related constructions \ftx \txt This topic contains the following subsections. Highlight and jump to the topic of interest: \ftx \shd2 Introductions \cf Hpnom_intro Introduction--Definitions \cf Hpnom_intro_other Other types \cf Hpnom_intro_chart Chart of types \cf Hpnom_intro_consd Practical considerations \cf Hpnom_intro_ex Examples of various types \cf Hpnom_intro_cmp Cross linguistic comparison \shd2 Predicate adjectives \cf Hpnom_p-adj Predicate adjectives \cf Hpnom_p-adj_ncontr contrast with nouns \cf Hpnom_p-adj_other other possibilities \shd2 Predicate nominals \cf Hpnom_p-nom_cop Predicate nominals--Copulas \cf Hpnom_p-nom_cop_op Copula as operation \cf Hpnom_p-nom_cop_part Copula as particle \cf Hpnom_p-nom_cop_pron Copula as pronoun \cf Hpnom_p-nom_cop_tns Copula in non-present tenses \cf Hpnom_p-nom_vcop Copula as verb \cf Hpnom_p-nom_vcop_char Characteristics of copula verbs \cf Hpnom_p-nom_vcop2 example from Mandarin \cf Hpnom_p-nom_vcop3 example from Japanese \cf Hpnom_p-nom_vcop4 example from Korean Hpnom_p-nom_vcop5 verb initial languages \cf Hpnom_p-nom_incl/eq inclusion vs equation \cf Hpnom_p-nom_sum Summary of typology \shd2 Summary of Pred Nom and EPL relationships \cf Hpnom_sum Summary of Pred Nom and EPL relationships \cf Hpnom_sum_chart Charts \shd2 Other topics \cf Hpnom_exst Existentials \cf Hpnom_exst_vscop Existentials vs Copulas \cf Hpnom_p-loc Predicate locatives \cf Hpnom_poss Possessive constructions \dt 12/May/1998 \key Hpnom_exst \shd Existentials \txt Existential constructions typically require a locational or temporal adjunct, e.g. under the bed in the sentence: \ftx \fln (24) There is a cat under the bed. \ftx \txt Many languages treat such sentences as 'There is a God' morphosyntactically as existentials, even though there is no locational adjunct expressed. However, such 'pure' existential sentences are not at all common in everyday discourse. Most languages use an intransitive verb form to express this sort of idea, e.g. 'A God exists.' \ftx \txt Existentials typically serve a PRESENTATIVE function, i.e. to introduce participants onto the discourse stage. Hence NOM is almost always indefinite. Existential constructions in English tend to sound odd with definite NOM elements: \ftx \fln (25) ??There are the lions in Africa. \ftx \txt Usually there is no, or reduced, evidence of grammatical relations in existential constructions, e.g. case marking, verb agreement, etc. This is true in colloquial English: \ftx \fln (26) a. There's bears in the forest. \ftx b. There's ants in the syrup. \ftx c. There's lots of women in linguistics. \ftx \txt Though English teachers may shudder at these examples of verb agreement 'errors', such sentences are extremely common and natural in spoken English. This indicates that the existential 'be' is becoming a defective verb (see ???), a very common property of existential constructions universally. In languages without a tradition of prescriptive grammar, speakers are freer to respond to this sort of functional naturalness. \ftx \txt Existential constructions commonly share features of predicate nominals, e.g the copular morpheme, as in English. Following are some examples from Estonian: \ftx \fln (27) laual on klaas piima 'There is a glass of milk on the table.' \ftx table be glass milk \ftx \txt To continue with a discussion of existentials versus copulas, highlight and jump to the topic below: \cf Hpnom_exst_vscop \ftx \dt 27/Apr/1998 \key Hpnom_exst_vscop \shd Existentials vs Copulas \txt Some languages don't use the copular morpheme in existentials even though they may have a perfectly good copula. The Mandarin sentence below cannot mean 'the book is on the table.' (remember, the copula in Mandarin is shì). \ftx \ftx (28) you shu yi-ben zài zhuozi shàng \ftx EXIST book one-CL at table on \ftx 'There is a book on the table.' \ftx \ftx o of you and e of yi-ben have hacheks. i of yi-ben, u of shu \ftx and o of zhuo-zi have overbars. \ftx \txt In Mandarin the existential particle 'you' is clearly distinct from the copula 'shì'. \ftx \txt Existentials often have special negation strategies, e.g. a verb meaning 'to lack' as in Turkish and Russian: \ftx \fln Turkish positive existential: \ftx (29) a. kösede bir kahve var 'There is a book on the corner.' \ftx on:corner a book EXIST \ftx \ftx Negative existential: \ftx b. kösede bir kahve yok 'There isn't a book on the corner.' \ftx on:corner a book LACK \ftx \fln Russian positive existential: \ftx (30) a. yest kniga na stólu 'There is a book on the table.' \ftx EXIST book on table:LOC \ftx \ftx Negative existential: \ftx b. nyet kniga na stólu 'There isn't a book on the table.' \ftx NEG:EXIST book on table:LOC \ftx \dt 27/Apr/1998 \key Hpnom_intro \shd Introduction--Definitions \txt Every language has constructions in which the main semantic content of the predication is embodied in something other than a lexical verb. Such constructions are sometimes collectively referred to as PREDICATE NOMINALS. However, in this manual we will use the term predicate nominal in a more specific sense, reserving it for those clauses in which the semantic content of the predication is embodied in a noun. This definition distinguishes predicate nominals from similar constructions such as PREDICATE ADJECTIVES, PREDICATE LOCATIVES and others discussed below. These predicate types cohere as a group only in that they tend to be instantiated by constructions that lack a SEMANTICALLY RICH lexical verb. By semantically rich, we mean a verb that itself expresses the major semantic content of the predication. Verbs like 'be' and 'do' in English are not (normally) semantically rich in that they must be accompanied by some other semantically rich lexical item, either a noun (for 'be') or a verb (for 'do') in order to form a predication. These verbs are sometimes termed SEMANTICALLY EMPTY. \ftx \txt To continue this discussion, highlight and jump to the topic below: \cf Hpnom_intro_other \ftx \dt 27/Apr/1998 \key Hpnom_intro_chart \shd Introduction--Chart of types \ftx \ftx \ftx Table #6.1: Predicate types charted according to the likelihood that \ftx they lack a semantically rich lexical verb \ftx \ftx Most likely Not very likely to \ftx to lack a semantically lack a semantically \ftx rich verb. rich verb, but still may. \ftx \ftx Predicate > Predicate > Predicate > Possessive > Locomotion \ftx nominals locatives, adjectives clauses | Predicates \ftx | existentials | \ftx | | \ftx |------- Subject matter for this section ------| \ftx \txt To continue this discussion, highlight and jump to the topic below: \cf Hpnom_intro_consd \ftx \key Hpnom_intro_cmp \shd Introduction--Cross linguistic comparison \txt In a cross-linguistic typological survey (Clark 1978), about 40 languages were compared according to how they treat E, P and L constructions. Clark lumps them all of these under the heading of locational constructions because they all typically have a LOCATION word (LOC) and a NOMINAL (NOM) whose location is specified by the LOC. The following examples illustrate how NOM and LOC are distributed in E, P and L constructions in English: \ftx \ftx E: There is a bee in your bonnet. \ftx NOM LOC \ftx P: Lucretia has nineteen cats. \ftx LOC NOM \ftx L: The cat is under the bed. \ftx NOM LOC \ftx \txt It may seem odd that the possessor in a possessive clause is considered to be a location, but when you think about it, that's what possession is: when you possess something it is literally or figuratively located 'on', 'at' or 'with' you. As illustrated in Hpnom_poss, many (if not most) languages pay attention grammatically to this cognitive similarity in that possessors are often treated the same as locations. \ftx \txt However, word order usually distinguishes possessive clauses from locational clauses, even if nothing else does. The following statistical tendencies on word order in E, P and L constructions are from the findings of Clark (1978): \ftx \ftx E: LOC before NOM (27 of 40 languages) \ftx P: LOC before NOM (35 of 40 languages) \ftx L: NOM before LOC (33 of 40 languages) \ftx \txt From these figures we see that the LOC is much more likely to precede the NOM in possessive constructions than in locationals. \ftx \dt 01/May/1998 \key Hpnom_intro_consd \shd Introduction--Practical considerations \txt In this section of the manual we will somewhat arbitrarily draw the line at possessive constructions. In the corresponding section of any particular grammar, it may be appropriate to include more or fewer of these construction types in a chapter on predicate nominals and related constructions. \ftx \txt First we will elaborate on the difference between a `semantically rich lexical verb' and a semantically empty, or GRAMMATICAL VERB. For example, when one says 'he is a teacher', the predicate is 'be a teacher'. The main semantic content of this predication is embodied in the form teacher. The verb 'be' in English simply serves to mark the construction as a predicate nominal clause and to carry the tense/aspect and person/number information required of predications in English. \ftx \txt Even constructions listed in table #6.1 that employ verbs as auxiliaries or as main predicators tend to use verbs that are BLEACHED semantically and are formally unusual. For example, in English the first three predicate types from the left (including existentials) employ the semantically empty and formally very irregular verb 'be'. Similarly, possessive constructions employ the irregular verb 'have'. In some languages all of these construction types share significant morphosyntactic characteristics. In others, each type is represented by quite different morphosyntax, but in most cases there is enough homogeneity to cause us to want to treat them all within a single section of the grammar. In any particular language, however, some construction types treated in this section may more appropriately be discussed in other sections. For example, English possessive constructions, like 'I have a book', may be treated with transitive verbal predicates. \ftx \txt To continue this discussion, highlight and jump to the topic below: \cf Hpnom_intro_ex \ftx \dt 27/Apr/1998 \key Hpnom_intro_ex \shd Introduction--Examples of various types \txt PREDICATE NOMINALS, as defined here, are those constructions used to express the notions of proper inclusion or equation (see Hpnom_p-nom_incl/eq): \ftx \fln Proper inclusion: Frank is a teacher. \fln Equation: Frank is the teacher. \ftx \txt PREDICATE ADJECTIVES express a quality of an item. Sometimes these clause types are referred to as ATTRIBUTIVE clauses: \ftx \fln Predicate adjective: John is tall. \ftx My car is green. \ftx \txt EXISTENTIALS, LOCATIONAL, and POSSESSIVE clauses often share morphosyntactic properties with predicate nominals and predicate adjectives. For example, they tend to lack a semantically rich lexical verb. \ftx \txt Existential (E) constructions predicate the existence of some entity, usually in some specified location: \ftx \ftx There is a bee in your bonnet. \ftx There is a book on the table. \ftx \fln Locational (L) (or predicate locative) constructions predicate location: \ftx \ftx The gift is in the horse's mouth. \ftx The book is on the table. \ftx \fln Possessive (P) constructions predicate possession: \ftx \ftx Sally has nineteen cats. \ftx The table has a book on it \ftx The book is John's. \ftx \txt Note that this is not the same as a possessive noun phrase, e.g. Sally's cats. Do not treat possessive phrases in this section of the grammar. \ftx \txt Note that in English all of these constructions, except some possive constructions, employ the verb 'be'. It is no accident that in predicate locative and existential constructions the verb is the same one used in predicate nominal and predicate adjective constructions. Possessive clauses in English are rather exotic in that they often employ a special transitive verb, 'have'. In the following sections we will define and exemplify these construction types and will show how they interrelate in about four languages. \ftx \txt To continue this discussion, highlight and jump to the topic below: \cf Hpnom_intro_cmp \ftx \dt 27/Apr/1998 \key Hpnom_intro_other \shd Introduction--Other types \txt Once we have defined the subject matter for this section of the grammar as `predicate types likely to lack a semantically rich lexical verb', we must include at least two predicate types that do not intuitively, at least to English speakers, fall into the same class as predicate nominals. These are EXISTENTIAL constructions, and POSSESSIVE constructions (described below). In fact, predicate types can be arranged along a continuum based on their likelihood to not employ a semantically rich lexical verb. The semantics of such constructions are summarized in the next help screen (Table 6.1) \ftx \txt To continue this discussion, highlight and jump to the topic below: \cf Hpnom_intro_chart \ftx \dt 27/Apr/1998 \key Hpnom_p-adj \shd Predicate adjectives \txt PREDICATE ADJECTIVES (e.g. 'he is tall') are seldom distinct structurally from predicate nominals. Treat them separately only if they exhibit some distinctive formal property or properties. See Hgr_mod_adj_sem for possible distinguishing characteristics of adjectives as opposed to nouns. \ftx \txt The morphosyntax of predicate adjectives is usually identical or similar to that of predicate nominals: \ftx \ftx a. No copula. NP ADJ \ftx b. Copula is a verb: NP V ADJ \ftx c. Copula is a pronoun: NP PRO ADJ \ftx d. Copula is an invariant particle: NP COP ADJ \ftx e. Copula is a derivational operation: [NP]v ADJ \ftx \fln (23) English: a. Rick is a pacifist. PREDICATE NOMINAL \ftx b. Rick is patient. PREDICATE ADJECTIVE \ftx \txt To continue this discussion, highlight and jump to the topic below: \cf Hpnom_p-adj_ncontr \ftx \dt 27/Apr/1998 \key Hpnom_p-adj_ncontr \shd Predicate adjectives--contrast with nouns \txt Sometimes predicate adjectives use a different copula than predicate nominals do: \ftx \fln (20) Spanish: a. Ofelia es profesora. 'Ofelia is a teacher.' \ftx *Ofelia está profesora. \ftx \ftx b. Ofelia está enferma. 'Ofelia is sick.' \ftx \ftx c. Ofelia es enferma. 'Ofelia is a sick person.' \ftx \txt As in many languages, there are few, if any, formal properties that distinguish nouns from adjectives in Spanish. Lexical items that embody properties, e.g. red, sick, large etc. can function exactly as nouns, i.e. 'la roja'-- 'the red one', 'la enferma'-- 'the sick one', 'la grande'-- 'the large one', etc. Both nouns and adjectives normally inflect for gender and number. Students of Spanish learn that the language has two copular verbs. In order to decide which verb to use in a given instance, the rule of thumb is that if the property being predicated is permanent, then 'ser' is used. If the property is temporary, 'estar' is used. This rule probably works 75% of the time, and so is sufficient for most beginning students. However, advanced students also must memorize many exceptions. In fact the two copular verbs contrast in that 'ser' is used for predicate nominals and 'estar' for predicate adjectives and other states, such as locations: \ftx \fln (21) El castillo está en el cerro. 'The castle is (estar) on the hill.' \ftx \ftx *El castillo es en el cerro. ser \ftx \txt For locations (as well as many other situations), 'estar', the copula for predicate adjectives, must be used even though the state can only be understood as permanent. This is because the relationship between subject and predicate in a locational construction is not one of class membership or identity. In most cases properties imparted by adjectives are temporary, whereas the relations of class membership or identity (the semantic 'definition' of a predicate nominal clause) are more permanent, hence the rule of thumb is of some use. \ftx \txt To continue this discussion, highlight and jump to the topic below: \cf Hpnom_p-adj_other \ftx \dt 27/Apr/1998 \key Hpnom_p-adj_other \shd Predicate adjectives--other possibilities \txt Other kinds of phrases can be formed in which the main semantic content of the predication is not embodied in a verb. These usually follow the pattern of predicate adjectives if the latter are at all different from predicate nominals (see Hpnom_p-loc for more information on predicate locatives): \ftx \fln Predicate locative: He is in the dog house. \ftx The cat is under the bed. \ftx This package is from your mother. \ftx \fln Predicate benefactive: This letter is for Melvin. \ftx Trix is for kids. \ftx \fln Predicate accompaniment: Mary was with child. \ftx You were with me that day. \ftx \fln Predicate ?? This award is for outstanding achievement. \ftx My stick is to teach you a lesson. \ftx \txt In some grammars all of these are just lumped together under the label 'predicate nominals'. \ftx \dt 27/Apr/1998 \key Hpnom_p-loc \shd Predicate locatives \txt Some languages, such as English, use the copular verb or morpheme in locational clauses. This is also true in Estonian: \ftx \ftx (22) raamat on laual 'The book is on the table.' \ftx book be:3SG table \ftx \txt There is a secondary type of locational construction in English that uses the verb 'have': 'The table has a book on it.' Notice the correlation between possessor and location - - in English location is equivalent to an inanimate possessor. Hence this secondary means of forming predicate locatives is based on the model of possessive constructions. For many other languages, in particular Russian and Estonian, it's the other way around -- possessive constructions are built on the model of locationals, but with an animate location. \ftx \txt Some languages use a special locative word. This word is often translated as 'be at': \ftx \ftx (23) Mandarin: shu zài zhuo-zi shàng \ftx book be:at table on \ftx 'The book is on the table.' \ftx \ftx u of shu and o of zhuo-zi have overbars. \ftx \txt In Mandarin, as in many languages, the locative word in a locational construction is the same as a preposition in a locative noun phrase. \ftx \dt 27/Apr/1998 \key Hpnom_p-nom_cop \shd Predicate nominals--Copulas \txt Predicate nominal constructions often employ a COPULA. For our purposes a copula is any morpheme (affix, particle or verb) that joins, or 'couples', two nominal elements in a predicate nominal construction. It marks the clause as a predicate nominal and often carries the tense/aspect and other information necessary for predications in the language. The particle 'si' that stands between the two NPs in the Cebuano example in the previous record is not a copula. Rather it is a case marker associated with the noun Juan and occurs in all types of clauses, not just predicate nominals. Also, it does not inflect for tense/aspect etc. In a great many languages a copula will only occur in past tense and in other ways semantically marked predicate nominal clauses. This is true of Russian: \ftx \ftx (3) Ivan bîl uchítyely 'John was a teacher.' \ftx John be:MASC teacher \ftx \txt 'Present tense' (i.e. the unmarked, neutral tense/aspect) predicate nominals are likely to consist simply of two juxtaposed noun phrases. In this section we will describe four types of copular forms that are known to exist in the world's languages. Any given language may employ several or all of these under different circumstances. It may also exhibit copulas that stand 'in between' some of the definitions given here, or it may exhibit a previously unattested copula type. \ftx \dt 27/Apr/1998 \key Hpnom_p-nom_cop_op \shd Predicate nominals--Copula as operation \txt In a few languages, a predicate nominal is formed by taking the predicate noun and applying a derivational operation that forms a verb from that noun. The predicate nominal then becomes a verb grammatically, as evidenced by its position in the sentence and the inflectional information it conveys: \ftx \fln (16) Bella Coola and Kuman on 'handout': \ftx \dt 27/Apr/1998 \key Hpnom_p-nom_cop_part \shd Predicate nominals--Copula as particle \txt Some languages use a special invariant particle to join two nominals in a predicate nominal construction. This particle may derive from a verb or a pronoun, but if it is invariant, i.e. if it remains the same regardless of the person/number/gender of the subject, or the tense/aspect of the clause, then it should be called a particle. For example: \ftx \fln (14) Supyìrè: wuu ne laklibii 'We are students.' \ftx we COP students \ftx \fln (15) Hausa, Gâ, Fante Chemehuevi \ftx \dt 27/Apr/1998 \key Hpnom_p-nom_cop_pron \shd Predicate nominals--Copula as pronoun \txt For many languages the form that joins two nominals in a predicate nominal construction is a pronoun. Normally the copular pronoun corresponds to the subject nominal: \ftx \fln (11) Hebrew: ha-ish hu av-í 'The man is my father.' \ftx DEF-man he father-my \ftx \ftx MORE EXAMPLES \ftx \fln (12) Arabic: EXAMPLES (see Pam) \ftx \fln (13) EXAMPLES FROM MORE LANGUAGES \ftx \dt 27/Apr/1998 \key Hpnom_p-nom_cop_tns \shd Predicate nominals--Copula in non-present tenses \txt Most languages that don't use a copula in simple, present tense predicate nominals, do use a copular verb or morpheme in certain TAM categories. Past and future tenses are common environments in which to find copular verbs or morphemes (17b and c): \ftx \fln Yagua: \ftx (17) a. Máchituru ráy 'I am a teacher.' \ftx teacher 1SG \ftx \ftx b. ra-vyicha-núú-yanu máchituru 'I used to be a teacher.' \ftx 1SG-be-CONT-PAST3 teacher \ftx \ftx c. rà-à vicha máchituru 'I'm gonna be a teacher.' \ftx 1SG-FUT be teacher \ftx \txt Note that in Yagua the copular form is a verb, as evidenced by the fact that it takes verbal aspect (continuative, ex. 17b), tense (distant past, ex. 17b) and mode (future/irrealis, ex. 17c). It also inflects for person, and occurs in sentence-initial position (ex. 17b). \ftx \ftx Russian also employs a copular form in non-past tenses: \ftx \ftx Russian: \fln (18) on bîl uchítyely 'He was a teacher'. \ftx \txt The copula in Russian inflects for gender only, even though verbs inflect for tense and person: \ftx \fln (19) aná bîl-á nyányê 'She was a nurse.' \ftx 3SG:FEM be-FEM nurse:FEM \ftx \txt Hence, it appears that the copula in Russian is a defective verb (see ???). It stands to reason that some copular forms should stand 'in between' verbs and invariant particles since copulas often derive historically from verbs. A defective verb is simply a verb that has lost some of its verbal properties, i.e. it is on its way to becoming an invariant particle but just hasn't gotten there yet. We will see that at least in existential constructions, the number distinction for the English verb 'be' is neutralizing. In this sense, then, 'be' is becoming defective in this environment. \ftx \dt 27/Apr/1998 \key Hpnom_p-nom_incl/eq \shd Predicate nominals--inclusion vs equation \txt PROPER INCLUSION is when a specific entity is asserted to be among the class of items specified in the predicate nominal, e.g. 'He is a teacher'. This sentence might be paraphrased 'he is a member of the class of items designated by the term "teacher"'. Usually the subject of a predicate nominal clause indicating proper inclusion is specific (he), and the predicate nominal is non-specific (a teacher). \ftx \txt EQUATIVE clauses are those which assert that a particular entity (the subject of the clause) is identical to the entity specified in the predicate nominal, e.g. 'he is my father'. Sometimes it is difficult or impossible to determine which nominal is the predicate and which is the subject in equative clauses. Most languages make no grammatical distinction between proper inclusion and equative clauses, though they may. \ftx \txt The most common type of predicate nominal is one in which two noun phrases are juxtaposed. We will term this type 'NP NP juxtaposition': \ftx \fln Cebuano (Austronesian, Philippines) \ftx \ftx (1) magyuyuta si Juan John is a farmer. \ftx farmer ABS John \ftx \fln Russian (Slavic) \ftx \ftx (2) Ivan uchítyely John is a teacher. \ftx John teacher \ftx \dt 27/Apr/1998 \key Hpnom_p-nom_sum \shd Predicate nominals--Summary of typology \txt The following simple formulas summarize the predicate nominal types discussed in this section. These 'formulas' are not meant to represent constituent order: \ftx \ftx a. No copula. NP NP \ftx b. Copula is a verb: NP V NP \ftx c. Copula is a pronoun: NP PRO NP \ftx d. Copula is an invariant particle: NP COP NP \ftx e. Copula is a derivational operation: [NP]v NP \ftx \txt The most common system is to not use a copula in the simplest predicate nominal constructions, i.e. present tense, and to use one of the above copula types in other tenses, aspects and/or modes. \ftx \dt 27/Apr/1998 \key Hpnom_p-nom_vcop \shd Predicate nominals--Copula as verb \txt For some languages the form that joins the two nominals in a predicate nominal construction has many or all the morphosyntactic properties that characterize verbs in that language. For instance, the copula may inflect for tense, aspect and/or mode, and it may occur in the normal verbal position in the sentence, i.e., sentence-initially, sentence-finally or sentence-medially. Semantically, however, copular verbs are 'empty'. That is, they carry little or no semantic content other than whatever is involved in converting a noun phrase into a predicate. A good example of a copular verb is the English verb 'be': \ftx \ftx (4) a. Marty is a sports fan. b. They are Oregonians. \ftx c. She was my favorite teacher. d. You are fine students. \ftx \txt Although 'be' is very irregular, it exhibits all the essential properties of verbs in English: 1) it varies for person (He is, you are), 2) it varies for tense (I am, I was), and 3) it occurs in sentence medial position. \ftx \txt To continue this discussion with examples, highlight and jump to the topic below: \cf Hpnom_p-nom_vcop2 \ftx \txt For a discussion of the characteristics of copular verbs, jump to: \cf Hpnom_p-nom_vcop_char \ftx \dt 27/Apr/1998 \key Hpnom_p-nom_vcop2 \shd Predicate nominals--Copula as verb--example from Mandarin \ftx \txt Mandarin is another verb-medial language that employs a copular verb (examples from Lynn Yang): \ftx \ftx (5) Wo de jie-jie shì yi-ge lao-shi \ftx 1SG GEN older:sister be one-CL teacher \ftx 'My older sister is a teacher.' \ftx \ftx (6) Wo de jie-jie yi-qián shì yi-ge lao-shi \ftx 1SG GEN older:sister before be one-CL teacher \ftx 'My older sister used to be a teacher.' \ftx \txt Character notes: o of Wo, both e of jie-jie, first i of yi- qián and a of lao have hacheks. i of yi-ge and i of lao-shi have overbars. \ftx \txt In Mandarin the morphosyntactic properties of verbs are very limited. However, to the extent that verbs can be identified grammatically, the copula 'shì' in Mandarin can be considered a verb, i.e. it appears in sentence medial position, and it is not restricted to particular tense aspects. \ftx \txt To continue this discussion, highlight and jump to the topic below: \cf Hpnom_p-nom_vcop3 \ftx \dt 27/Apr/1998 \key Hpnom_p-nom_vcop3 \shd Predicate nominals--Copula as verb--example from Japanese \ftx \txt For verb-final languages, the copular verb normally comes at the end of the sentence: \ftx \fln Japanese (from Mitsuyo #Hamada): \ftx \ftx (7) imooto-wa sensei desu \ftx younger:sister-TOP teacher be:PRES \ftx 'My younger sister is a teacher.' \ftx \ftx (8) imooto-wa sensei deshita \ftx younger:sister-TOP teacher be:PAST \ftx 'My younger sister was a teacher.' \ftx \fln The copular element has all the properties of verbs in Japanese. \ftx \txt To continue this discussion, highlight and jump to the topic below: \cf Hpnom_p-nom_vcop4 \ftx \dt 21/Aug/1997 \key Hpnom_p-nom_vcop4 \shd Predicate nominals--Copula as verb--example from Korean \ftx \fln KOREAN (Examples courtesy of Insun Park): \ftx \ftx (9) na-îi nuna-nîn sênsænnim-i-ta \ftx 1SG-GEN elder:sister-TOP teacher-BE-IND \ftx 'My elder sister is a teacher.' \ftx \ftx (10) na-îi nuna-nîn sênsænnim-i-êt-ta \ftx 1SG-GEN elder:sister-TOP teacher-BE-PAST-IND \ftx 'My elder sister was a teacher.' \ftx \txt To continue this discussion, highlight and jump to the topic below: \cf Hpnom_p-nom_vcop5 \ftx \key Hpnom_p-nom_vcop5 \shd Predicate nominals--Copula as verb--verb initial languages \ftx \txt If there is a copular verb in verb initial languages, it normally comes at the beginning of the sentence, but, in fact, verb initial languages seldom employ copular verbs. \ftx \txt To continue this discussion, highlight and jump to the topic below: \cf Hpnom_p-nom_vcop_char \ftx \dt 27/Apr/1998 \key Hpnom_p-nom_vcop_char \shd Predicate nominals--Characteristics of copula verbs \txt There are several properties that tend to characterize copular verbs universally: \ftx \txt 1) They tend to be very irregular. That is, they often exhibit unusual conjugational patterns as compared to the more 'normal' verbs in the language. Paradigms tend to be suppletive and/or defective (see ???). One reason for this fact is that copular constructions are so frequent in discourse. The most frequently used constructions are those that are most likely to develop and retain irregularities. This is because through much use conjugational forms become well-installed in memory as individual units rather than as variants of a central form predictable by a systematic rule. Those individual units then proceed in the natural process of linguistic variation and change independently of changes in the whole system. \ftx \txt 2) Copular verbs tend to derive from very stative verbs, e.g. 'stand', 'sit', 'live', 'exist', etc. Sometimes they derive from simple verbs of motion such as 'go' or 'come'. \ftx \txt 3) Copular verbs tend to function as auxiliaries in other constructions (see Hord_vp_v-aux on auxiliaries). In fact, when a language develops auxiliaries, the first verbs to be used as auxiliaries are the copular verbs. Second are the verbs of motion (Foley and Van Valin 1984). \ftx \dt 27/Apr/1998 \key Hpnom_poss \shd Possessive constructions \txt Languages usually employ existential and/or locational clause structures to express the notion of possession. Occasionally possessive constructions use a special verb like 'to have'. This verb often derives from such sources as the verb for 'hold' or 'carry'. The more common situation, however, is for the possessive construction to use a copular verb or particle. Estonian uses the copular morpheme. To say 'the child has milk' you say literally 'milk is at the child.': \ftx \fln Estonian: \ftx (31) a. lapsel on piima 'The child has milk.' \ftx child:LOC be:3SG milk (Lit: 'milk is at the child'.) \ftx \ftx b. mul on tikku 'I have a match.' \ftx 1SG:LOC be:3SG match (Lit: 'a match is at me') \ftx \txt This is reminiscent of the colloquial English expression: 'Got any money on you?' \ftx \txt Turkish uses the verb meaning 'exist' that also occurs in the existential constructions. To say 'the child has a father', you say literally 'the child's father exists.': \ftx \fln (32) Turkish: cocugun babasi var 'The child has a father.' \ftx child:GEN father exist \ftx \txt Mandarin is similar to Turkish in this respect. To say 'he has a book' you say something like 'to him exists a book.': \ftx \fln (33) Mandarin: ta you yi-ben shu 'He/she has a book.' \ftx 3SG exist one-CL book \ftx \txt Character notes: o of you and e of yi-ben have hacheks. i of yi-ben and u of chu have overbars. \ftx \fln You can also say 'to him is a book' in Mandarin, but this is less natural: \ftx \fln (34) Mandarin: ?ta shì shu 'He/she has/is a book.' \ftx 3SG COP book \ftx \fln Character notes: u has overbar. Check tone on ta. \ftx \txt In French the same verb that is used in existentials in an impersonal sense is also used in possessive clauses. Russian allows the same form for possessive clauses as in predicate locatives (like Estonian). To say 'I have a book' you say 'a book exists to me.': \ftx \fln (35) Russian: u menya (yest) kniga 'I have a book.' \ftx to me:GEN EXIST book \ftx \dt 27/Apr/1998 \key Hpnom_sum \shd Summary of Pred Nom and EPL relationships \txt E, L and P constructions, while apparently serving logically distinct functions, are conceptually quite similar: they all embody a stative (i.e. non-eventive) situation in which the location or existence of one item (NOM) is specified with respect to some other item (LOC). According to Clark, the main functional difference is the relative animacy and definiteness of the two elements (referred to informally as NOM and LOC). As we might expect, however, the difference probably really lies in the notion of topicality. This is a notion which cannot be adequately identified on intuitive evidence alone. Rather, topicality can only be determined through rigorous investigation of texts. Definiteness and animacy are highly correlated with topicality, since human beings tend to select animate and definite entities as topics. Therefore it makes intuitive sense that in a pair of sentences like: \ftx \ftx (36) a. Tom has a book. \ftx b. The book is Tom's. \ftx \txt the phrase 'the book' is more topical in the second sentence than the corresponding phrase in the first. But this intuition derives from the common association between definiteness and topicality, not because definiteness in any way defines topicality. See Hdisc_cnt_ref for further discussion of the notion of discourse topicality. \ftx \txt To continue this discussion, highlight and jump to the topic below: \cf Hpnom_sum_chart \ftx \dt 27/Apr/1998 \key Hpnom_sum_chart \shd Summary of Pred Nom and EPL relationships-- Charts \txt The following table summarizes the morphological characteristic of predicate nominal, existential, locational and possessive constructions in six languages mentioned in the discussion: \ftx \ftx Language PRED NOM E P L \ftx \ftx English be be be/have have \ftx French est a a/est est \ftx Mandarin shì you you zài \ftx Turkish -im(etc. var var var \ftx Russian 0/bïl yest yest 0/bïl \ftx Estonian on on on on \ftx \ftx o of you has hachek. \ftx \txt The following are simple 'formulas' that may help to conceptualize the similarities and differences between these various construction types: \ftx \ftx Pred nom: NOM NOM \ftx +def \ftx \ftx E: NOM LOC \ftx -def -anim \ftx +def \ftx \ftx P: NOM LOC \ftx +def +anim \ftx +def \ftx \ftx L: NOM LOC \ftx +def -anim \ftx +def \ftx \dt 27/Apr/1998 \key Hprag \shd Pragmatically-marked Structures \txt This topic contains all of the following subsections. Highlight and jump to the topic of interest: \cf Hprag_stat Pragmatic Statuses \cf Hprag_iden Identifiability \cf Hprag_foc Focus \cf Hprag_top Topic \cf Hprag_cntr Contrast \shd2 Pragmatic prominence \cf Hprag_prom Pragmatic Prominence \cf Hprag_prom_ord Constituent Order \cf Hprag_prom_ord_ex **incomplete \cf Hprag_prom_ord_dis Dislocation \cf Hprag_prom_ord_dis_other Other kinds of movement \cf Hprag_prom_ord_dis_rule Rules \cf Hprag_prom_ord_dis_ex Examples \cf Hprag_prom_ord_dis_part Use of particles \cf Hprag_prom_form Formatives \cf Hprag_prom_form_ex Examples \cf Hprag_prom_form_ovl Overlay systems \cf Hprag_prom_clft Cleft Constructions \cf Hprag_prom_clft_ex_eng English examples \cf Hprag_prom_clft_ex_other Examples from other languages \shd2 Negation \cf Hprag_neg Negation \cf Hprag_neg_lex Lexical Negation \cf Hprag_neg_mor Morphological Negation \cf Hprag_neg_anal Analytic Negation \cf Hprag_neg_anal_part Analytic Negation by particle \cf Hprag_neg_anal_part_ex1 Examples I \cf Hprag_neg_anal_part_ex2 Examples II \cf Hprag_neg_anal_part_ex3 Examples III \cf Hprag_neg_anal_v-init Analytic by initial finite verb \cf Hprag_neg_anal_v-fin Analytic by final finite verb \cf Hprag_neg_s-mod Secondary Modifications \cf Hprag_neg_s-mod_ord Word order \cf Hprag_neg_s-mod_tone Tone \cf Hprag_neg_s-mod_t/a Tense/Aspect \cf Hprag_neg_s-mod_infl Inflection \cf Hprag_neg_s-mod_case Case \cf Hprag_neg_n-cl Non-Clausal Negation \cf Hprag_neg_n-cl_drv Derivational \cf Hprag_neg_n-cl_quan Quantifiers \cf Hprag_neg_n-cl_sco Scope \shd2 Non-declarative speech acts \cf Hprag_ndecl Non-Declarative Speech Acts \cf Hprag_ndecl_theory Theory \cf Hprag_ndecl_q Interrogatives \cf Hprag_ndecl_q_ynq Yes/No Questions \cf Hprag_ndecl_q_ynq_inton Intonation \cf Hprag_ndecl_q_ynq_other Other Phonological Markers \cf Hprag_ndecl_q_ynq_ord Word order \cf Hprag_ndecl_q_ynq_spec Special Particles \cf Hprag_ndecl_q_ynq_tag Tag questions \cf Hprag_ndecl_q_ynq_func Functions of Y/N Qs \cf Hprag_ndecl_q_qwd Question Word Questions \cf Hprag_ndecl_q_qwd_func Functions of Q words \cf Hprag_ndecl_q_qwd_vo Q words in VO languages \cf Hprag_ndecl_q_qwd_ov Q words in OV languages \cf Hprag_ndecl_q_qwd_case Q words and case markers \cf Hprag_ndecl_imp Imperatives \dt 29/May/1998 \key Hprag_cntr \shd Contrast \fln A prototypical contrastive sentence presupposes that \ftx \ftx a) a particular event E (taken loosely to mean any state of \ftx affairs described by a proposition) occurred, \ftx \ftx b) there is a group of entities that might have had a role, \ftx R, in E, and \ftx \ftx c) the hearer 'incorrectly' (in the eyes of the speaker) \ftx believes that one of the entities did in fact have the role R. \ftx \fln The contrastive sentence then asserts \ftx \ftx a) the 'correct' identity of the entity involved, according \ftx to the perception of the speaker, and \ftx \ftx b) the proposition that the entity the hearer thought had \ftx the role R in fact did not. \ftx \txt So for example, the English sentence 'SALLY made the salad' (with 'contrastive stress' on Sally) implies that \ftx \ftx a) there was a group of people, perhaps just Sally and \ftx Harry, that might have made the salad, \ftx \ftx b) the speaker has reason to believe that the hearer \ftx incorrectly thinks Harry made the salad. \ftx \fln By uttering this sentence, then, the speaker asserts that \ftx \ftx a) Sally was the person who made the salad, and \ftx \ftx b) Harry did not make the salad. Not every instance of \ftx contrast will have all of these characteristics, but this is \ftx the prototype. \ftx \txt Typically languages will use STRESSED PRONOUNS (sometimes these are the only type of free pronominal devices in the language), and some kind of cleft construction to signal contrast. \ftx \fln References: Chafe (1976), Dik (1981), D. Payne (1987) \ftx \dt 27/Apr/1998 \key Hprag_foc \shd Focus \txt The following is a brief overview of ways in which the term 'focus' (and various expansions of that term) have been used in the recent linguistic literature. This typology is adapted from Chafe 1976, Watters 1979 and Dik 1981. \ftx \fln There are three general approaches to the term focus. These are: \ftx \ftx 1. 'Focus' is a term applied to some morphosyntactic \ftx operation or category whose function has not been adequately \ftx analyzed. \ftx \ftx 2. 'Focus' is a term applied to one element of every \ftx sentence. In this approach, focus can pretty much be equated \ftx with 'new information' or 'asserted information'. \ftx \ftx 3. 'Focus' describes a condition of some pragmatically \ftx marked sentences. Other sentences can be 'focus-neutral' or \ftx 'unfocussed.' \ftx \shd2 First approach \ftx \txt The first approach to the term 'focus' will not be discussed at length here. It is evident in such locutions as 'word order varies for focussing purposes'. What this probably means is the writer does not understand the functions of the various word orders in the language being described. Also, since all languages possess structures whose function can only be determined in relation to a theory of discourse, there is a strong tendency in treatments that make no provision for discourse function to label such structures as conveying 'focus' or 'emphasis' or 'stylistic variation' without explicit definition. These terms become the wastebasket for linguistic structure that is not describable within the bounds of a sentence. \ftx \shd2 Second approach \ftx \txt The second approach to focus stems from the work of the FUNCTIONAL SENTENCE PERSPECTIVE linguists of the PRAGUE SCHOOL (e.g. Mathesius 1939). Under this system, every sentence has two parts; the part that refers to what the hearer is presumed to already know, or have in mind, and the part that adds some new information. Some sentences may consist entirely of new material. Although the early Prague School linguists did not use the term 'focus', they are to be credited with the concept of focus as the part of the sentence that conveys new information. Other terms that are applied to this notion are RHEME, ASSERTION and NEW INFORMATION. \ftx \txt The heuristic for determining which part of a sentence is focussed in this conceptualization is to imagine the sentence as an answer to an information question (see Hprag_ndecl_q_qwd on information questions). The focus is the part of the answer that fills in the information requested in the prompting question: \ftx \fln What happened? Billy pushed Johnny off the porch. (Whole sentence) \ftx \ftx \fln What did Billy do? He pushed Johnny off the porch. (Predicate focus) \ftx \ftx \fln Who pushed Billy off the porch? Johnny pushed Billy off the porch. \ftx (Subject focus) \ftx \fln Who did Billy push off the porch? He pushed Johnny off the porch. \ftx (Object focus) \ftx \fln Where did Billy push Johnny? He pushed him off the porch. \ftx (Location focus) \ftx \shd2 Third approach \ftx \txt The third conception of the term 'focus' is the view that takes focus to be a special pragmatic status that is not evident in all sentences. Sometimes this conception of focus is termed MARKED FOCUS. Sentences that are 'focussed', or have a 'focussed constituent' in this sense are PRAGMATICALLY MARKED. That is, they deviate in their pragmatic nuances from the standard, most common sentence type in the language. Many authors (e.g. Chafe 1976, Givón 1979) use the term CONTRAST to describe this pragmatic function (see Hprag_cntr). It must be understood that even within this general approach to 'focus', there are many different types of focus. Any given author may use the term 'focus' to refer to any one of the focus types described here. Similarly, any given language may be sensitive to only one, a few or none of these focus types. One shouldn't necessarily hope to find all focus types grammaticalized in every language. On the other hand, when describing the structures that a language uses to convey focus-like functions, one should always be as explicit as possible as to which kind (or kinds) of focus the language appears to be sensitive to. \ftx \shd2 Scope of Focus \ftx \txt The major distinction in the typology of marked focus falls under the heading of 'scope of focus'. The scope of focus of a sentence is either the truth value of the entire sentence or a constituent of the sentence: \ftx \fln Scope of focus: Entire sentence = Truth Value Focus (TVF) \ftx A particular constituent = Constituent Focus (CF) \ftx \txt Truth value focus counters the assumed presupposition that the truth-value of the entire clause is in question. Bahasa Indonesian grammaticalizes TVF with the existential particle 'ada' (data from Dik 1981): \ftx \fln (12) a. Ali pergi ke pasar Ali went to the market \ftx Ali go to market \ftx \ftx b. Ali ada pergi ke pasar Ali DID go to the market. \ftx Ali EXIST go to market \ftx \txt Example 12a is a focus-neutral sentence in Indonesian, while 12b is a sentence in which the truth value is focussed. Note that in the English translations, the same function is accomplished with the semantically empty auxiliary verb 'do' and a non-finite main verb. Presumably, 12b would be uttered in a situation where the speaker had reason to believe the hearer believed that Ali did not go to the market. That is, 12b is an assertion in CONTRAST to the presupposition of its negative. Sometimes TVF is called POLAR FOCUS. \ftx \txt If the scope of focus for a particular sentence is a constituent of the sentence (CF), then it can be any one of the following focus types: \ftx \ftx a) Assertive focus. S believes H has no knowledge of the \ftx information: \ftx \fln (13) They brought me a bowl of this thick, green, mushy stuff. \ftx \ftx a) Counter-presuppositional focus. This focus type comes \ftx closest to contrastive focus in the tradition of Chafe 1976 \ftx and Givón 1979(CHECK THE GIVON REFERENCE). T. Payne 1985 \ftx calls this 'exclusive contrast': \ftx \fln (14) Sally and Robert came over last night, but SHE got drunk. \ftx (Presupposition: You thought Robert would, but he didn't) \ftx \ftx a) Exhaustive listing focus: That information which S \ftx asserts is unique in that the rest of the sentence is true \ftx only with respect to it and false with respect to all other \ftx possible information: \ftx \fln (15) I drank only Pepsi at the party \ftx \txt Aghem, a Bantoid language of Cameroon, employs a complex system including constituent order and particles to convey all of these focus types, and a few others (Waters 1979). See Hprag_prom_ord_ex for a brief presentation of the Aghem data. \ftx \dt 27/Apr/1998 \key Hprag_iden \shd Identifiability \txt Two pragmatic statuses that play a significant role in the grammars of most languages are IDENTIFIABILITY and REFERENTIALITY. Certain noun phrases refer to entities that the speaker judges should be identifiable by the hearer. The particle 'the' is one means of conveying identifiablity in English: \ftx \fln (1) The Duke of Wimple trod on the Princess' toe. \ftx \txt The use of 'the' in this example instructs the hearer that there is a unique Duke of Wimple and Princess that the speaker is referring to. Furthermore, the speaker assumes that the hearer is capable of identifying which Duke of Wimple and which princess is meant. That is, the participants in question are identifiable given the information the speaker assumes the hearer has available to him or her. If the particle 'a' were used in place of 'the', the effect would be that there is not a unique referent for each of these terms. That is, there may be many Dukes of Wimple and many princesses that the hearer might identify and it is unknown or it just doesn't matter which specific ones were involved in the action. Traditionally the term DEFINITE has been used to describe the status we will refer to as identifiable. \ftx \shd2 Proper Names \ftx \txt Noun phrases can be IDENTIFIED (i.e. made identifiable) in several ways. The use of a proper noun normally implies that the speaker assumes the hearer can identify the referent: \ftx \fln (2) George embraced Saddam. \ftx \txt In a normal conversation this use of proper names implies that the speaker assumes that the hearer can identify the unique individuals described. There is no need to say 'There was this guy named George . . . ', or 'Do you remember that guy named George we met at the party last weekend? . . .' to establish the identity of the participants we are referring to. Similarly, upon hearing a sentence like 2, a hearer will assume the speaker is referring to some identifiable referent, and will quickly attach the name to a referent if at all possible. If a plausible referent is not identified, the hearer is likely to protest: 'Hey wait a minute. Who's George?' \ftx \txt Often a noun phrase is identified by its association with some other already identified noun phrase. For example: \ftx \fln (3) George's wife embraced Saddam. \ftx \txt In this sentence the referent of the noun 'wife' is established via its association with the proper name 'George'. Since 'George' is identifiable, and since presumably George has only one wife, then his wife should also be identifiable. So, NPs that are grammatically possessed by identifiable NPs are also identifiable. \ftx \shd2 Implicit in Situation \ftx \txt Identifiability is not necessarily explicit. Identifiability in real language is always significant only in relation to the communication situation. That is, something is treated as identifiable if its referent is explicit enough for the speaker's current purposes. For example, consider the following sentence: \ftx \fln (4) I got mad at Hosni for writing on the living room wall. \ftx \txt Here the phrase 'the living room wall' is identifiable even though most living rooms would have more than one wall. It is just not relevant for the speaker's purpose in this case to distinguish which living room wall is being referred to. Similarly, even 'George's wife' in 3 may not in itself identify a specific message world entity (i.e. in a situation in which George is known to have more than one wife). However, a sentence such as this would still be acceptable if either a) it just didn't matter which wife were involved, or b) the particular wife were identified in terms of the context, e.g. George took only one of his wives to visit Saddam therefore it could only plausibly be that wife that embraced Saddam. \ftx \shd Referentiality \ftx \shd2 Objective Referentiality \ftx \txt Referentiality is similar, but not identical, to identifiability. Here I will briefly contrast two approaches to the notion of referentiality (Givón 1979, Du Bois 1980). The first approach to referentiality I will term OBJECTIVE REFERENTIALITY. The second is DISCOURSE REFERENTIALITY. \ftx \txt An entity is objectively referential if it really exists as a bounded, individuated entity in the message world. Sometimes referentiality in this sense is referred to as SPECIFICITY. This definition excludes the following: \ftx \fln (5) Generics: All men are ridiculous. \ftx Non-specifics: Someday I'd like to buy a cabin by the seashore. \ftx \fln Analogous examples with objectively referential participants would be: \ftx \fln (6) Those men are ridiculous. \ftx Someday I'd like to buy your cabin by the seashore. \ftx \txt Note that objective referentiality is not the same as identifiability. A generic referent can be identifiable in the sense that the speaker assumes the hearer can identify the genera (e.g. all men in example 5), though there is no specific individual being referred to. This fact is reflected in English grammar in that the particle 'the' can mark generic noun phrases: \ftx \fln (7) The elephant is a huge mammal. \ftx \txt Here the speaker instructs the hearer to identify the generic class referred to by 'elephant' but not necessarily to single out any individual (objectively referential) elephant. \ftx \txt Similarly, non-identifiable entities need not be non-referential. For example: \ftx \fln (8) I'm looking for a good book. \ftx \txt This sentence is ambiguous in English. It could mean that I'm looking for any book that happens to be good, or it could mean that I have a specific good book in mind but I just don't assume that you as hearer can identify that book. In either case the book is treated as non-identifiable (as conveyed via the particle 'a'). In the first case it is non-referential (or non-specific), whereas in the second case it is objectively referential (or specific). \ftx \txt Spanish grammaticalizes the referentiality distinction for human direct objects. Referential human direct objects take the preposition 'a' (example 9a), whereas non-referential human direct objects take no preposition (example 9b): \ftx \fln (9) a. Estoy buscando a una empleada. 'I'm looking for a \ftx be:1SG look:for REF one housekeeper (specific) housekeeper.' \ftx \ftx b. Estoy buscando una empleada. 'I'm looking for a (any) \ftx be:1SG look:for one housekeeper housekeeper.' \ftx \shd2 Discourse Referentiality \ftx \txt In contrast to objective referentiality, discourse referentiality has to do with continuing importance over a portion of text (Du Bois 1980). In general this is a more restrictive concept than is objective referentiality. That is, it is common to have objectively referential entities that are not discourse referential, but it is more difficult to conceive of discourse referential entities that are not also objectively referential. For example, any prop in a story might be objectively referential, as in the following: \ftx \fln (10) She came in through the bathroom window. \ftx \txt In this sentence 'the bathroom window' is treated as objectively existing in the scene established in the discourse. In this sense it is objectively referential. However, if the window is never mentioned again, it would not be discourse referential, in terms of Du Bois (1980), because it would not have continuing presence on the discourse stage. Many languages have been shown to be more sensitive to this notion of referentiality than to the notion of objective referentiality. For example, in Papago, items that are introduced into the discourse for the first time appear before the verb if they are `destined' to figure prominently in the succeeding text (discourse referential), but appear after the verb if they are only transitory (Doris Payne 1987). This is independent of the objective referentiality of the items. \ftx \txt Wright and Givón (1987) have shown that the demonstrative 'this' in spoken English is, among other things, an indicator of discourse referentiality. In spoken narratives, items introduced with this are much more likely to persist, i.e. be mentioned repeatedly, than are items introduced with either 'the' or 'a': \ftx \fln (11) I was just sitting there minding my own business when this guy walks up. \ftx \txt In 11 the speaker is very likely to continue talking about the referent of the expression 'this guy'. In this sense this is a marker of discourse referentiality. Other terms that have been used for this concept are DEPLOYABILITY (Jaggar 1985), MANIPULABILITY (Hopper and Thompson 1984) and IMPORTANCE (Givón 1990). However, the important fact to remember is that natural languages tend to be more sensitive to this status, whatever it may be called, than to objective referentiality as defined within classic philosophy. An entity is discourse referential if it can be referred to in subsequent text as the 'same'. \ftx \dt 27/Apr/1998 \key Hprag_ndecl \shd Non-declarative Speech acts \txt Languages typically have various morphosyntactic devices that express whether a sentence is asserting information, requesting information (interrogatives) or commanding some action (imperatives). Those morphosyntactic devices are often modal in character (see Hv_tam_mode on the linguistic definition of mode). For example, questions and commands are irrealis in mode, therefore morphology associated with irrealis assertions often appear in questions and commands. However, declarative-interrogative- imperative does not describe a modal parameter per se. In this section we will look at various means that languages employ to express the speech act value of a clause. \ftx \txt To continue this discussion, highlight and jump to the topics below: \cf Hprag_ndecl_theory Theory \cf Hprag_ndecl_q Interrogatives \cf Hprag_ndecl_imp Imperatives \ftx \dt 27/Apr/1998 \key Hprag_ndecl_imp \shd Imperatives \txt IMPERATIVES are verb forms or construction types that are used to directly command the hearer to perform some actions, e.g. 'eat this!' (Sischo 1939). Usually imperatives are understood to refer to second person subjects. Because it is so common, so expected for the intended subject of an imperative sentence to be the hearer, reference to the subject is not necessary and so the subject is often omitted. Imperatives typically allow fewer TAM contrasts than other construction types. This is because it is simply pragmatically impossible to command someone to perform acts with certain TAM operations, e.g. *Ate that!, *be having a baby! etc. In the following paragraphs we will discuss and exemplify certain formal properties of imperative constructions. \ftx \shd2 Imperatives and verb forms \ftx \txt Imperatives sometimes take special verb forms. In Greenlandic Eskimo, the distinction between declarative and imperative clauses is signalled by the morphological distinction between '-v' and '-gi': \ftx \ftx Greenlandic: \fln (116) a. iga-v-o-t 'You are cooking (something).' \ftx cook-DECL-INTRNS-2 \ftx \ftx b. iga-gi-t 'Cook! (something)' \ftx cook-IMP-2 \ftx \shd2 Imperatives and negation \ftx \txt Imperatives sometimes take special negation. In Greenlandic, the negative operator 'na-' is used in imperatives (117a) and in dependent clauses (117b): \ftx \fln (117) a. Attor-na-gu 'Do not disturb this.' \ftx disturb-NEG-IMP:3SG \ftx \ftx b. Attor-na-gu iser-p-o-q \ftx disturb-NEG-INF:3SG enter-DECL-INTRNS-3SG \ftx 'Without disturbing him, he came in.' \ftx \txt This '-na' is distinct from the negative marker used in independent clauses. Note also that the third person singular non-finite verb suffix '-gu' is the same as the imperative marker when the object is an identifiable third person argument. \ftx \shd2 Imperatives and mode \ftx \txt Imperatives are often associated with other irrealis modes. For example, in modern Israeli Hebrew there is a specific verb form for imperatives (ex. 118a). However, the future can also be understood as an imperative (exx. 118b and 118c). The only way of forming a negative imperative is with the future form of the verb plus a special negative marker: \ftx \ftx Modern Israeli Hebrew: \fln (118) a. Shev 'Sit down!' \ftx sit(IMP) \ftx \ftx b. Teshev 'Sit down!' OR \ftx sit(2SG.FUT.INDIC) 'You will sit down' \ftx \ftx c. Hu lo' yoshev 'He is not sitting' \ftx he not sit(MASC.SG.PRES.INDIC) \ftx \ftx d. *Lo' shev \ftx \ftx e. Lo' teshev 'You will not sit down' \ftx not sit(2SG.FUT.INDIC) \ftx \ftx f. 'Al teshev 'Do not sit down!' \ftx not sit(2SG.FUT.INDIC) \ftx \fln In Yagua the imperative and the future are exactly the same: \ftx \fln (119) a. Y-â-maasa 'Sit down' or \ftx 2SG-IRR-sit 'You will sit down.' \ftx \ftx b. vuryâ-â-murray 'Let's sing' or \ftx 1PL-IRR-sing 'We will sing.' \ftx \fln Imperatives and case \ftx \txt Sometimes imperatives affect case marking. For example, in Finnish P arguments normally occur in the morphological case traditionally termed 'accusative' (120a). However, in imperatives the P argument occurs in the 'nominative' case (120b): \ftx \ftx Finnish: \fln (120) a. Maija soi kala-n 'Maija ate fish.' \ftx Maija:NOM ate fish-ACC \ftx \ftx b. Syö kala 'Eat fish!' \ftx eat fish:NOM \ftx \dt 27/Apr/1998 \key Hprag_ndecl_q \shd Interrogatives \txt In this section we will discuss sentence types that are overtly interrogative. In the section on pragmatics there will be more complete coverage of the means of making requests, eliciting information, indicating wishes, etc. \ftx \txt Languages always have some grammaticalized means of specifying that a particular utterance is to be understood as a request for information rather than an assertion. Such grammatical structures we will term INTERROGATIVE clauses. In traditional English grammar the term interrogative is described as a 'mode', along with declarative and imperative. This sense of the term mode overlaps yet conflicts with our definition of mode as the degree to which a situation is presented as being 'actual'. Certainly interrogative clauses tend to be irrealis in mode, along with many other clause types. However, interrogatives do not themselves constitute a point on the modal scale. \ftx \txt Within the class of interrogative clauses, languages typically distinguish two subtypes: those for which the information requested is a simple affirmation or disaffirmation (yes or no), and those for which the requested information is a more elaborate locution, either a phrase, a proposition or an entire discourse. In the following two subsections these two general types of interrogative clauses will be discussed. \ftx \txt To continue this discussion, highlight and jump to the topics below: \cf Hprag_ndecl_q_ynq Yes/No Questions \cf Hprag_ndecl_q_qwd Question Word Questions \ftx \dt 27/Apr/1998 \key Hprag_ndecl_q_qwd \shd Question word (information, content) questions \txt Questions that require a more elaborate response than simply an affirmation or disaffirmation are called QUESTION WORD QUESTIONS, CONTENT QUESTIONS, INFORMATION QUESTIONS or WH-QUESTIONS. The last term reflects the fact that in written English, the question words all contain a w and an h. Even though this mnemonic device may be helpful to speakers of English, we will not use this term in this manual. Rather, we will employ the term 'question-word questions'. \ftx \txt All languages have a set of special words that occur in question word questions. These words are often similar or identical to a set of pronouns used elsewhere in the language, e.g. the relative pronouns or pronouns used to refer to non-specific, non-identified entities. For example, the set of question words in English is practically identical to the set of relative pronouns: \ftx \fln English: \ftx Question word Relative pronoun Meaning \ftx \ftx who who human, subject \ftx whom (archaic) whom human, non-subject \ftx what -- non-human \ftx where where location \ftx why why reason \ftx how -- manner \ftx when when time \ftx which which generic \ftx \txt In some dialects of English it is possible to use 'what' and 'how' as relative pronouns, e.g. 'the house what I saw', 'the way how you did it'. In 'standard' English, however, the generic complementizer 'that' would be more typical in these circumstances. \ftx \fln Often question words are similar to indefinite pronouns: \ftx \fln (107) Tamang: khaima khaima klang-pa 'Sometimes he plays' \ftx when when play-INDEF \ftx \txt To continue this discussion, highlight and jump to the topics below: \cf Hprag_ndecl_q_qwd_func Functions of Q words \cf Hprag_ndecl_q_qwd_vo Q words in VO languages \cf Hprag_ndecl_q_qwd_ov Q words in OV languages \cf Hprag_ndecl_q_qwd_case Q words and case markers \ftx \dt 27/Apr/1998 \key Hprag_ndecl_q_qwd_case \shd Question words and case markers \txt Question words can usually take case markers and/or adpositions. When a question word in an oblique role is fronted (sometimes EXTRACTED), the adposition may remain with the 'gap' (115a) or it may go along with the question word (115b): \ftx \fln (115) a. What did you eat with 0? \ftx b. With what did you eat 0? \ftx \fln 'Pied-piping' is an informal term for the phenomenon illustrated in 115b. \ftx \dt 27/Apr/1998 \key Hprag_ndecl_q_qwd_func \shd Functions of question words \txt Question words accomplish two tasks: (1) they mark the clause as a question, and (2) they indicate what information is being requested. For example, 108b through e are English questions formed from the declarative sentence in 108a: \ftx \fln (108) a. Zebedee threw stones at the herring \ftx b. Who threw stones at the herring? \ftx c. What did Zeb throw 0 at the herring? \ftx d. What did Zeb throw stones at 0? \ftx e. What did Zeb do to the herring? \ftx \txt The presence of the special question word at the beginning of the sentence marks the sentence as a question. The actual question word chosen, plus a 'gap' somewhere in the sentence (108c and d) or the pro-verb 'do' (108e), specify what information the speaker is requesting the hearer to fill in. \ftx \txt To continue this discussion, highlight and jump to the topic below: \cf Hprag_ndecl_q_qwd_vo \ftx \dt 27/Apr/1998 \key Hprag_ndecl_q_qwd_ov \shd Question words in OV languages \txt In OV languages the question word sometimes remains in the 'normal' position (in situ), rather than moving to the front: \ftx \ftx Japanese (Maynard 1987): \fln (111) Zentai doko itteta da? \ftx in:the:world where have:been COP \ftx `Where in the world have you been?' \ftx \ftx Tibetan: \fln (112) yoqöö qhare see-pê-ree? \ftx servant what eat-PAST-INTER \ftx 'What did the servant eat?' \ftx \txt However, in most OV languages the question word can either remain in situ or it can move to the front: \ftx \fln (113) Wappo: a. may ce chici hak'she? 'Whom does that bear like?' \ftx who that bear like \ftx \ftx b. ce chici may hak'she? 'Whom does that bear like \ftx \fln (114) a. ita mi¿ yok'-okh hak'she? \ftx where you sit-INF like \ftx 'Where would you like to sit?' \ftx \ftx b. mi¿ yok'-okh hak'she ita? \ftx you sit-INF like where \ftx 'Where would you like to sit?' \ftx \txt To continue this discussion, highlight and jump to the topic below: \cf Hprag_ndecl_q_qwd_case \ftx \dt 27/Apr/1998 \key Hprag_ndecl_q_qwd_vo \shd Question words in VO languages \txt In VO languages, such as English, it is typical for the question word to appear at the beginning of the clause. This fact was observed by Greenberg (1966) in his universals 11 and 12. The following are some examples of clause-initial question words in a non-Indo-European language: \ftx \fln (109) Zapotec: a. tu biiya-lu? 'Whom did you see?' \ftx who saw-you \ftx \ftx b. zhi bi'ni-lu? 'What did you do?' \ftx \ftx c. tu najii Betu? 'Who loves Betu?' \ftx \txt Occasionally, even in VO languages, the question word will remain in the same position the requested information would have occurred in if the sentence were a declarative: \ftx \fln (110) Mandarin: a. shei kan ni? 'Who saw you?' \ftx who see you \ftx \ftx b. ni kan shei? 'Whom did you see?' \ftx \ftx c. ta dao nali qu? 'Where did he go?' \ftx 3SG to where go \ftx \txt To continue this discussion, highlight and jump to the topic below: \cf Hprag_ndecl_q_qwd_ov \ftx \dt 27/Apr/1998 \key Hprag_ndecl_q_ynq \shd Yes/No Questions \txt We will use the term YES/NO QUESTIONS to refer to interrogative clauses for which the expected answer is either 'yes' or 'no'. The following paragraphs discuss the various ways languages are known to form yes/no questions: \ftx \txt To continue this discussion, highlight and jump to the topics below: \cf Hprag_ndecl_q_ynq_inton Intonation \cf Hprag_ndecl_q_ynq_other Other Phonological Markers \cf Hprag_ndecl_q_ynq_ord Word order \cf Hprag_ndecl_q_ynq_spec Special Particles \cf Hprag_ndecl_q_ynq_tag Tag questions \cf Hprag_ndecl_q_ynq_func Functions of Y/N Qs \ftx \dt 27/Apr/1998 \key Hprag_ndecl_q_ynq_func \shd Functions of Yes/No Questions \txt So far we have dealt with the various ways in which languages form yes/no questions. In most languages, the morphosyntax of yes/no questions is employed in several different ways in discourse. In this section we will briefly survey some of the ways in which yes/no questions are known to function. Since many of these are present at least marginally in English, we will illustrate these functions primarily from English. It should be kept in mind, however, that some of these functions that are only marginal in English (e.g., intensification) are much more well-installed as discourse devices in other languages. Furthermore, there may be other creative uses made of yes/no question structures that have yet to be attested. \ftx \ftx 1. To solicit information. This is the basic use of yes/no \ftx questions, e.g.: \ftx \fln (98) 'Is it time for class?' \ftx \ftx 2. To request action. This is quite different from soliciting \ftx information: \ftx \fln (99) 'Could you close the window?' \ftx \txt Spanish illustrates the difference between usages 1 and 2 in that there exist two distinct lexical verbs to describe the two senses of 'asking': \ftx \fln (100) preguntar 'to ask for information' \ftx pedir 'to request a thing or some action' \ftx \fln (101) Me preguntó qué hora fue. 'He asked me what time it was.' \ftx *Me pidió qué hora fue. \ftx \fln (102) Me pidió un Bolívar. 'He asked me for a Bolívar.' \ftx *Me preguntó un Bolívar. \ftx \fln (103) Me pidió escribir una carta. 'He asked me to write a letter.' \ftx *Me preguntó escribir una carta. \ftx \ftx 3. For rhetorical effect. Rhetorical questions expect no answer: \ftx \fln (104) 'Is the Pope Catholic?' \ftx \ftx 4. Confirmation of information already possessed by the speaker: \ftx \fln (105) 'You're going, aren't you?' \ftx 'Aren't you going?' \ftx \ftx 5. Intensification: \ftx \fln (106) 'Did he ever yell!' \ftx \txt Although this sentence type does not have question intonation, it does exhibit the syntax of a yes/no question. \ftx \dt 27/Apr/1998 \key Hprag_ndecl_q_ynq_inton \shd Yes/No Questions--Intonation \txt Intonation. Yes/no questions universally tend to involve distinctive intonation patterns. The intonation pattern employed in yes/no questions is usually rising, as in English, but is sometimes falling, as in Russian. Question intonation can either be the only indication that a clause is a question or it can accompany any of the other strategies listed below. \ftx \txt To continue this discussion, highlight and jump to the topic below: \cf Hprag_ndecl_q_ynq_other \ftx \dt 27/Apr/1998 \key Hprag_ndecl_q_ynq_ord \shd Yes/No Questions--word order \txt Many languages, especially languages that are of the VO constituent order type, employ distinctive constituent orders in yes/no questions. Usually this distinctive order involves an 'inversion' (or reversal) of the order of subject and verb. This is very common in Austronesian and European languages. For example: \ftx \ftx Malay: \fln (81) ASSERTION: bapak datangkah nanti 'Father will come later.' \ftx father come:FUT later \ftx \fln (82) QUESTION: datangkah bapak nanti 'Will father come later?' \ftx come:FUT father later \ftx \txt English employs a somewhat exotic inversion system in yes/no questions. Instead of reversing the order of subject and main verb, English reverses the order of subject and auxiliary verb (83a and b). If the corresponding assertion contains no auxiliary, the 'dummy' auxiliary 'do' is inserted (83c): \ftx \fln (83) a. Will he arrive on time? \ftx \ftx b. Can they bite corn nuts? \ftx \ftx c. Do you want to subsume these clause types? \ftx \txt In American English, simple subject-verb inversion occurs in predicate nominal, existential and locational sentences (84a, b and c). In British English this extends to possessive constructions (84d): \ftx \fln (84) a. Is he a ringmaster? \ftx \ftx b. Are there cats under your flowerpots? \ftx \ftx c. Were you in the butterscotch pudding? \ftx \ftx d. Have you a match? (Chiefly British) \ftx \txt To continue this discussion, highlight and jump to the topic below: \cf Hprag_ndecl_q_ynq_spec \dt 27/Apr/1998 \key Hprag_ndecl_q_ynq_other \shd Yes/No Questions--other phonological markers \fln Non-intonational phonological markers, e.g. glottalization, voicing. \ftx \txt To continue this discussion, highlight and jump to the topic below: \cf Hprag_ndecl_q_ynq_ord \ftx \key Hprag_ndecl_q_ynq_spec \shd Yes/No Questions--special particles \txt Other than intonation, the most common means of forming a yes/no question universally is with an interrogative particle. This strategy is most common with OV languages, but does occur in VO languages as well. The question particle (QP) can be cliticized to the first constituent in the clause (either before or after), or occur at end of the clause. Often the question particle can be omitted, leaving only intonation and the pragmatics of the situation to distinguish the clause as a question: \ftx \ftx Tagalog: \fln (85) ASSERTION: mabait si Pilar. 'Pilar is kind.' \ftx \fln (86) QUESTION: mabait (ba) si Pilar? 'Is Pilar kind?' \ftx \ftx Latin: \fln (87) erat-ne te-cum 'Was he with you?' \ftx he:was-QP you-with \ftx \ftx Zapotec: \fln (88) (nee) nuu bisoze-lu 'Is your father there?' \ftx QP is father-your \ftx \ftx Yagua: \fln (89) Jidyeetu-viy júnaa-chara? 'Does your daughter cry?' \ftx your:daughter-QP cry-HABIT \ftx \fln (90) Sa-ya-viy Quityo-mu-jû? 'Did he go to Iquitos?' \ftx 3SG-go-QP Iquitos-LOC-DIR \ftx \ftx Mandarin: \fln (91) ta xihuan chi pingguo ma 'Does she like to eat apples?' \ftx 3SG like eat apple QP \ftx \ftx Wappo: \fln (92) eephi mansana pa¿ukh hak'she¿ he¿ 'Does she like to eat \ftx 3SG apple eat like QP apples?' \ftx \txt In Canadian and some other varieties of English a question particle is an alternative to subject-AUX inversion: \ftx \fln (93) You want to feed my sled dogs, eh? \ftx \txt To continue this discussion, highlight and jump to the topic below: \cf Hprag_ndecl_q_ynq_tag \ftx \dt 27/Apr/1998 \key Hprag_ndecl_q_ynq_tag \shd Yes/No Questions--Tag questions \txt A TAG QUESTION is a yes/no question consisting of a declarative sentence plus a 'tag' that requests confirmation or disconfirmation of the declarative sentence. Usually tag questions are a secondary yes/no question device. That is to say, in languages that employ tag questions there is always some other, more fully grammaticalized means of forming yes/no questions. However, the tag is often the historical source for question particles (see above). Spoken English uses tag questions in particular pragmatic environments. For example: \ftx \fln (94) English: She's leaving, isn't she? \ftx She's leaving, right? \ftx \txt These questions seem to imply that the speaker expected an affirmative answer. The basic yes/no question strategy does not carry this pragmatic expectation, e.g. Is she leaving? \ftx \fln The following are some additional examples of tag questions: \ftx \fln (95) Russian: ti yevo slushil, pravda? 'You heard him, didn't you?' \ftx you him heard true \ftx \fln (96) Lamani: u jan-wa cha, koni ka? 'He's going, isn't he?' \ftx he goes-he PRES NEG QP \ftx \fln A tag question is sometimes a reduced form of a conjoined alternative clause: \ftx \fln MANDARIN (Li and Thompson 1981:546): \ftx \fln (97) a. nimen shi jiu dianzhong kai mén de, duì bu duì? \ftx 2PL be nine o'clock open door NOM right not right \ftx `You opened at nine o'clock, right?' \ftx \txt Character notes: Hachek over i of first word u of jiu and a of dianzhong. Bar over o of dianzhong. \ftx \ftx b. women qu chi shuiguo, hao bu hao? \ftx 1PL go eat fruit good not good \ftx 'Let's go eat some fruit, OK?' \ftx \txt Character notes: Hachek over o of women, i and o of shuiguo and a of both hao. Bar over i of chi. \ftx \txt To continue this discussion, highlight and jump to the topic below: \cf Hprag_ndecl_q_ynq_func \ftx \dt 27/Apr/1998 \key Hprag_ndecl_theory \shd Non-declarative Speech acts--theory \txt The term 'declarative' in traditional grammar refers to sentences that simply assert information. Often the term 'declarative mode' will be found in the literature. In this manual, and in the scientific study of language in general declarative is not a mode. In the tradition of speech act theory, the term ASSERTION most closely approximates the traditional notion of declarative. \ftx \txt Declarative clauses are usually the normal, unmarked clause type. If there are special markings for speech act types, declarative is usually expressed via a zero marker. Tibetan is one exception to this generalization. In Tibetan, both declarative and interrogative clauses receive a special marker: \ftx \fln (77) yoqöö môômôô see-pê-ree DECLARATIVE \ftx servant dumplings eat-PAST-DECL \ftx 'The servant ate dumplings.' \ftx \fln (78) yoqöö môômôô see²-qi-ree DECLARATIVE \ftx servant dumplings eat-FUT-DECL \ftx 'The servant will eat dumplings.' \ftx \fln (79) yoqöö môômôô see-pê-repee INTERROGATIVE \ftx servant dumplings eat-PAST-QP \ftx 'Did the servant eat dumplings?' \ftx \fln (80) yoqöö qhare see-pê-ree? INTERROGATIVE \ftx servant what eat-PAST-INTER \ftx 'What did the servant eat?' \ftx \dt 27/Apr/1998 \key Hprag_neg \shd Negation \txt This topic contains the following subsections. Highlight and jump to the topic of interest: \cf Hprag_neg_lex Lexical Negation \cf Hprag_neg_mor Morphological Negation \cf Hprag_neg_anal Analytic Negation \cf Hprag_neg_s-mod Secondary Modifications \cf Hprag_neg_n-cl Non-Clausal Negation \dt 27/Apr/1998 \key Hprag_neg_anal \shd Analytic Negation \txt This topic contains the following subsections. Highlight and jump to the topic of interest: \cf Hprag_neg_anal_part Analytic Negation by particle \cf Hprag_neg_anal_part_ex1 Examples I \cf Hprag_neg_anal_part_ex2 Examples II \cf Hprag_neg_anal_part_ex3 Examples III \cf Hprag_neg_anal_v-init Analytic by initial finite verb \cf Hprag_neg_anal_v-fin Analytic by final finite verb \dt 27/Apr/1998 \key Hprag_neg_anal_part \shd Negation--Analytic by particle \txt There are two kinds of analytic negation: negative particles and finite negative verbs. Negative particles are normally associated with the main verb of the clause. However, they may also be sentence level clitics. Negative particles can be invariant, as the English 'not' and its allomorph '-nt', or Russian 'ne': \ftx \fln (52) a. on ne igraet 'He doesn't play.' \ftx he NEG play \ftx \ftx b. ne igraj 'Don't play!' \ftx NEG play:IMP \ftx \ftx c. on ne durak 'He is not a fool.' \ftx he NEG fool \ftx \txt In many languages, however, the negative particle varies depending on the tense, aspect, mode or other factors. A typical distinction in negative particles is between plain negatives and negatives of existence. \ftx \txt To continue this discussion, highlight and jump to the topic below: \cf Hprag_neg_anal_part_ex1 \ftx \dt 27/Apr/1998 \key Hprag_neg_anal_part_ex1 \shd Negation--Analytic by particle--Examples I \txt In Tagalog, and most other Austronesian languages, there are two ways of saying 'no'. In Tagalog the particles are 'wala' and 'hindi'. 'Wala' is the negative of existence. It is the appropriate negative response to a yes/no question relating to the existence or presence of some item: \ftx \fln (53) a. Mayroon ka bang pera? 'Do you have any money.' \ftx \ftx b. Wala 'None.' \ftx *Hindi \ftx \txt 'Hindi', on the other hand, is the standard means of responding negatively to non-existential propositions: \ftx \fln (54) a. Pupunta ka ba sa sayawan? 'Are you going to the dance?' \ftx \ftx b. Hindi 'No.' \ftx *Wala \ftx \txt In addition to being the negative responses to questions, 'hindi' and 'wala' are also the particles used to form negative sentences. Not surprisingly the difference between the two particles is that 'wala' negates existential propositions (ex. 55a, b) while 'hindi' negates other sorts of propositions (ex. 56a, b): \ftx \fln (55) a. Wala akong pera 'I don't have any money.' \ftx \ftx b. Wala akong alam 'I don't know anything.' \ftx (Lit: 'I lack knowledge') \ftx \fln (56) a. Hindi ako papasok sa eskwela. 'I'm not going to school.' \ftx \ftx b. Hindi ko alam 'I don't know.' \ftx \txt To continue this discussion, highlight and jump to the topic below: \cf Hprag_neg_anal_part_ex2 \ftx \dt 27/Apr/1998 \key Hprag_neg_anal_part_ex2 \shd Negation--Analytic by particle--Examples II \txt Mandarin has at least three negative particles. The most common particle is 'bu' (ex. 57a). The existential negative is 'mei' (ex. 57b), and the negative particle used in imperatives is 'bie' (ex. 57c): \ftx \fln (57) a. Ta bu he jiu 'He doesn't/didn't drink wine.' \ftx 3SG NEG drink wine \ftx \ftx b. Ta mei you gege 'She doesn't have an older brother.' \ftx 3SG NEG exist older:brother \ftx \ftx c. Bie zou 'Don't go!' \ftx NEG:IMP go \ftx \txt To continue this discussion, highlight and jump to the topic below: \cf Hprag_neg_anal_part_ex3 \ftx \dt 27/Apr/1998 \key Hprag_neg_anal_part_ex3 \shd Negation--Analytic by particle--Examples III \txt Many languages, among them Iraqi Arabic, employ one invariant negative particle in verbal predicates (predicates headed by a verb) as in 58a, and another in verbless predicates (e.g. predicate nominals, locationals, existentials, etc) as in 58b, c, and d: \ftx \ftx Iraqi Arabic: \fln (58) a. ¿êli ma: ra:h lidda:¿ire 'Ali didn't go to the office.' \ftx Ali NEG went to:office \ftx \ftx b. ¿ubu:jê mu: muha:mi 'My father is not a lawyer.' \ftx father:my NEG lawyer \ftx \ftx c. hadê ssati mu: rêmli 'This beach is not sandy.' \ftx this beach NEG sandy \ftx \ftx d. lwêktu:b mu: ¿ili 'The letter is not for me.' \ftx the:letter NEG for:me \ftx \txt To continue this discussion, highlight and jump to the topic below: \cf Hprag_neg_anal_v-init \ftx \dt 27/Apr/1998 \key Hprag_neg_anal_v-fin \shd Negation--Analytic by final finite verb \txt Tongan and Squamish are both verb-initial languages. The following examples are from verb final languages that employ a finite negative verb as a primary negation device: \ftx \ftx Diegeño (Yuman): \fln (61) ¿nya:-c ¿-a¿m-x ¿-ma:w 'I didn't go.' \ftx I-SUBJ 1SG-go-IRR 1SG-NEG \ftx \txt In Diegeño both finite and complement verbs take person inflection. However, in this example it is clear that the verb meaning 'go' is subordinate because it is marked with the irrealis suffix '-x'. Also the negative stem 'ma:w' occurs in sentence final position, as expected for main verbs. \ftx \ftx Evenki (Tungus, Siberia): \fln (62) a. Bi dukuwu:n-ma duku-ca:-w 'I wrote a letter.' \ftx 1SG letter-ACC write-PAST-1SG \ftx \ftx b. Bi dukuwu:n-ma ê-cê:-w duku-ra \ftx 1SG letter-ACC NEG-PAST-1SG write-PPLE \ftx 'I didn't write a letter.' \ftx \txt Although Evenki is a verb-final language, as illustrated in example 62a, the negative verb does not occur in final position, at least when the sentence being negated has an overtly expressed direct object. Nevertheless, since the stem that expresses the negative sense, 'ê', is inflected like a verb, and the other verb is inflected like a participle, it would still be appropriate to term this strategy a finite negative verb. \ftx \dt 27/Apr/1998 \key Hprag_neg_anal_v-init \shd Negation--Analytic by initial finite verb \txt The second type of analytic negation involves a finite negative verb and a complement clause (see Hcplx_cplcl). Morphosyntactic properties of negative verbs are exactly those of finite verbs in general for the language. For example, the negative verb will take finite verbal inflectional morphology and will occur in the normal position of a verb. The affirmative verb, i.e. the verb that expresses the main semantic content of the clause, is treated like a complement verb. That is, it may be introduced by a complementizer or take non-finite or irrealis verbal morphology. This negation strategy occurs primarily in verb-initial or verb-final languages. Following are some examples from various languages that employ a finite negative verb: \ftx \ftx Tongan (Polynesian): \fln (59) a. Na'e-alu 'a Siale 'Charlie went.' \ftx COMPL-go ABS Charlie \ftx \ftx b. Na'e-'ikai [ ke 'alu 'a Siale ] 'Charlie didn't go.' \ftx COMPL-NEG IRR go ABS Charlie \ftx \txt The portion of sentence 59b enclosed in brackets is a complement clause. It is marked as irrealis (a non-finite category in Tongan) and the negative stem ''ikai' takes inflectional morphology common to main verbs. It also occurs in the sentence position characteristic of main verbs. \ftx \ftx Squamish (Salish): \fln (60) Ha'u-c-0-ap qaly-c'ic'a'p' 'You (pl) didn't work.' \ftx NEG-AUX-PAST-2PL COMP-work \ftx \txt In this sentence the stem that carries the negative sense, 'ha'u', takes all the verbal inflection, i.e. tense and person marking. It also occurs in the normal position for main verbs. The verb meaning 'work', on the other hand is clearly subordinated by the complementizer 'qaly-'. \ftx \txt To continue this discussion, highlight and jump to the topic below: \cf Hprag_neg_anal_v-fin \ftx \dt 27/Apr/1998 \key Hprag_neg_intro \shd Negation--Introduction \txt A NEGATIVE sentence is one which asserts that some event, situation or state of affairs does not hold. In this section we will discuss and exemplify various ways in which languages treat these special pragmatically marked clauses. \ftx \txt The most common negative strategies in any language are those used to make an entire clause negative. These we will describe as CLAUSAL NEGATION, e.g. 'I didn't do it'. Other types of negation are associated with particular constituents of clauses, e.g. 'I have no bananas'. This will be referred to as CONSTITUENT NEGATION. Although the semantic effect of constituent negation can be very similar or identical to that of clausal negation, constituent negation is always less common as a grammatical device than clausal negation. In these sections we will primarily discuss clausal negation. Towards the end we will deal briefly with constituent negation. \ftx \dt 27/Apr/1998 \key Hprag_neg_lex \shd Negation--Lexical \txt As might be expected, lexical negation describes a situation in which the concept of negation is part and parcel of the lexical semantics of a particular verb. For example, the verb 'lack' in English can be thought of as the lexical negative of 'have'. However, it is sometimes difficult to isolate a particular verb as the lexical negative of some other verb. For example, is 'stand' the lexical negative of 'sit', or are these just two distinct verbs? \ftx \dt 27/Apr/1998 \key Hprag_neg_mor \shd Negation--Morphological \txt Morphemes that encode clausal negation are normally, if not always, associated with the verb. Many languages, e.g. Farsi as illustrated in 50a and b, employ a simple verbal affix: \ftx \fln (50) a . na-xar-am 'I didn't buy.' \ftx NEG-buy-1SG \ftx \ftx b. na-mi-xar-am 'I'm not buying.' \ftx NEG-PRES-buy-1SG \ftx \txt As in the case of Farsi, it is very common for negative affixes to be reflexes of older negative particles (see Hprag_neg_anal_part -- analytic negation). \ftx \txt Negation is often tied up with other verbal inflections. For example, Nanai (Tungus) uses special tense markers in negative clauses. Note also that the stem vowel is lengthened in the negative: \ftx \fln (51) a. xola-j-si 'You are reading.' \ftx read-PRES-2SG \ftx \ftx b. xola-xa-si 'You were reading.' \ftx read-PAST-2SG \ftx \ftx c. xola:-si-si 'You aren't reading.' \ftx read-NEG:PRES-2SG \ftx \ftx d. xola:-ci-si 'You weren't reading.' \ftx read-NEG:PAST-2SG \ftx \dt 27/Apr/1998 \key Hprag_neg_n-cl \shd Non-clausal Negation \txt So far we have discussed ways in which languages express negative assertions instantiated in propositions. In this section we will discuss various ways in which specific clause constituents can be negated. We will begin with derivational negation, and will continue with negative quantifiers. Finally we will discuss briefly the notion of negative scope. \ftx \txt To continue this discussion, highlight and jump to one of the topics below: \cf Hprag_neg_n-cl_drv Derivational \cf Hprag_neg_n-cl_quan Quantifiers \cf Hprag_neg_n-cl_sco Scope \ftx \dt 27/Apr/1998 \key Hprag_neg_n-cl_drv \shd Non-clausal Negation--derivational \txt Occasionally languages will allow a stem to be transformed into its 'opposite' by use of some derivational morphology. This can be termed DERIVATIONAL NEGATION. English uses the prefixes 'un-' and perhaps 'non-' for this purpose: \ftx \fln (69) unhappy non-smoker \ftx unselfish non-past tense \ftx unreasonable non-entity etc. \ftx \txt In English 'un-' is largely restricted to adjectival stems and 'non-' to adjectival or nominal stems. Furthermore, neither of these is fully productive. In other languages, however, derivational negation can be more prevalent. For example, in Panare, many verbs stems are built on a root plus a negative suffix '-(i)ka'. The resulting stem then embodies a concept which in some loose sense can be understood as the opposite of the concept embodied by the original root. For example: \ftx \fln (70) a. t-ama-yaj chu 'I threw it out.' \ftx 1:3-throw:out-PAST 1SG \ftx \ftx b. t-ama-ika-yaj chu 'I kept it/stored it/placed it.' \ftx 1:3-throw:out-NEG-PAST 1SG \ftx \fln (71 ) a. y-otawë-yaj 'It got dark.' \ftx 3-get:dark-PAST \ftx \ftx b. y-otawë-ika-yaj kën 'He/she woke up.' \ftx 3-get:dark-NEG-PAST 3SG \ftx \txt This derivational operation is related to, but quite distinct from, standard inflectional negation. Standard negation in Panare is expressed via a post-verbal particle 'ka' (counter-expectation) or 'pï' (consistent with expectations). \ftx \txt To continue this discussion, highlight and jump to the topic below: \cf Hprag_neg_n-cl_quan \ftx \dt 27/Apr/1998 \key Hprag_neg_n-cl_quan \shd Non-clausal Negation--quantifiers \txt Many languages employ quantifiers that are either inherently negative (e.g. English 'none', 'nothing') or are negated independently of clausal negation (e.g. 'not many'). Most languages allow or require negative quantifiers to be accompanied by clausal negation. For example, in Russian the form 'nikto-- 'nobody' when referring to the subject of a clause must be accompanied by clausal negation: \ftx \fln (72) a. nikto nje prisel 'Nobody came.' \ftx nobody NEG came \ftx \ftx b . *nikto prisel \ftx \fln Standard English is exotic in disallowing this use of a 'double negative': \ftx \fln (73) a. Nobody came. b. Nobody didn't come. \ftx (means 'everybody came', not \ftx 'nobody came' as in Russian.) \ftx \txt To continue this discussion, highlight and jump to the topic below: \cf Hprag_neg_n-cl_sco \ftx \dt 27/Apr/1998 \key Hprag_neg_n-cl_sco \shd Non-clausal Negation--scope \txt Sometimes constituent negation and clausal negation interact to cause variations in NEGATIVE SCOPE. Scope refers to the variable portions of a clause that can be negated. Clausal negation has scope over the entire clause. Constituent negation has scope over a particular constituent of the clause. The fieldworker should be mindful of variations in negative scope, though in early stages of fieldwork it may be very difficult to discern such variation. Examples of variations in negative scope will be provided from English: \ftx \fln (74) a. Not many people like Vonnegut. SCOPE: Subject quantifier only. \ftx \ftx b. Many people do not like Vonnegut. SCOPE = Entire clause \ftx \fln (75) a. I deliberately didn't bump into her. SCOPE = Entire clause \ftx \ftx b. I didn't deliberately bump into her. SCOPE = Adverb only \ftx \fln (76) a. I won't force you to marry Zelda. SCOPE = Entire sentence \ftx \ftx b. I will force you not to marry Zelda. SCOPE = Complement clause \ftx \fln References: Horn, Laurence R. (1978) UHL. J. Payne (1985) Shopen vol. I. \ftx \dt 27/Apr/1998 \key Hprag_neg_s-mod \shd Negation--Secondary modifications \txt Secondary modifications. So far we have discussed the primary devices that languages use to express a negative proposition. In addition to these primary devices, there are sometimes secondary devices that accompany the primary devices. To date none of the following devices have been found to be the only indication of negation in a sentence. \ftx \txt To continue this discussion, highlight and jump to one of the topics below: \cf Hprag_neg_s-mod_ord Word order \cf Hprag_neg_s-mod_tone Tone \cf Hprag_neg_s-mod_t/a Tense/Aspect \cf Hprag_neg_s-mod_infl Inflection \cf Hprag_neg_s-mod_case Case \ftx \dt 27/Apr/1998 \key Hprag_neg_s-mod_case \shd Negation--Secondary modifications--case \txt In a few languages, special case marking patterns occur in negative clauses. For example, in Russian with certain transitive verbs the object occurs in the accusative case in affirmative sentences (68a) and in the genitive case in negative sentences (68b): \ftx \fln (68) a. on zabudjet tot vecer 'He will forget that evening.' \ftx he forget:FUT that:ACC evening:ACC \ftx \ftx b. on nje zabudjet togo vecera \ftx he NEG forget:FUT that:GEN evening:GEN \ftx 'He will not forget that evening.' \ftx \dt 27/Apr/1998 \key Hprag_neg_s-mod_infl \shd Negation--Secondary modifications-- inflection \txt A few languages employ special person/number or tense/aspect markers on verbs in negative sentences. These normally are reflexes of older forms of the language. It is typically the case that negative clauses (along with other non-basic clause types) retain older morphosyntactic patterns. For example, in Kawaiisu in past tense affirmative sentences the suffix is '-kïdiine' (67). In negative sentences, the suffix is '-keeneeneene' (67b): \ftx \fln (67) a. ta¿nipuzi-a pïkee-kïdiine momo¿o-na 'Man saw woman.' \ftx man-SUBJ see-PAST woman-OBJ \ftx \ftx b. ta¿nipuzi-a yuweatï pïkee-keneeneene momo¿o-na \ftx man-SUBJ NEG see-PAST:3->3 woman-OBJ \ftx 'Man didn't see woman.' \ftx \txt To continue this discussion, highlight and jump to the topic below: \cf Hprag_neg_s-mod_case \ftx \dt 27/Apr/1998 \key Hprag_neg_s-mod_ord \shd Negation--Secondary modifications--word order \txt Many VO languages employ a special word order in negative clauses. For example, Kru uses SVO order in affirmative clauses (63a) and SOV order in negative clauses (63b): \ftx \ftx Kru (Niger-Congo, Ivory Coast) \fln (63) a. ó të kó 'He bought rice.' \ftx he:COMPL buy rice \ftx \ftx b. ó se kò të 'He didn't buy rice.' \ftx he:COMPL NEG rice buy \ftx \txt To continue this discussion, highlight and jump to the topic below: \cf Hprag_neg_s-mod_tone \ftx \dt 27/Apr/1998 \key Hprag_neg_s-mod_tone \shd Negation--Secondary modifications--tone \txt Many African languages employ a distinct tone on the verb or auxiliary for negative clauses. For example, the incompletive auxiliary in Igbo carries low tone in affirmative sentences (64a) but high tone in negative sentences (64b): \ftx \ftx Igbo (Kwa, Niger-Congo) \fln (64) a. ò nà àsá akwà 'She is doing the wash.' \ftx she INCOMPL do wash \ftx \ftx b. ò ná-ghí àsá akwà 'She has not done the wash.' \ftx she INCOMPL-NEG do wash \ftx \txt To continue this discussion, highlight and jump to the topic below: \cf Hprag_neg_s-mod_t/a \ftx \dt 27/Apr/1998 \key Hprag_neg_s-mod_t/a \shd Negation--Secondary modifications-- tense/aspect \txt Neutralization of tense-aspect distinctions. Sometimes there are fewer tense-aspect distinctions in the negative than in the affirmative. For example, Komi exhibits a present-future distinction in the affirmative (65a and b), but no such distinction in the negative (65c): \ftx \fln (65) a. gizö 'He writes.' \ftx \ftx b. gizas 'He will write.' \ftx \ftx c. oz giz 'He doesn't write,' or \ftx 'He won't write.' \ftx \txt Similarly, Bembe allows two future tense markers, 'ká' and 'kà', in affirmative sentences (66a and b) but only 'kà' in negative sentences (66c): \ftx \ftx Bembe (Bantu, Niger-Congo, Zambia) \fln (66) a. n-kà-boomba 'I'm about to work.' \ftx 1SG-FUT:1-work \ftx \ftx b. n-ká-boomba 'I will work (later).' \ftx 1SG-FUT:2-work \ftx \ftx c. n-shi-kà-boomba 'I won't work.' \ftx 1SG-NEG-FUT:1-work \ftx \ftx d. *n-shi-ká-boomba \ftx 1SG-NEG-FUT:2-work \ftx \txt To continue this discussion, highlight and jump to the topic below: \cf Hprag_neg_s-mod_infl \ftx \dt 27/Apr/1998 \key Hprag_prom \shd The morphosyntax of pragmatic prominence (e.g. focus, contrast and 'topicalization') \txt As mentioned earlier, languages can employ intonation, constituent order, special formatives (bound morphemes or particles) or cleft constructions to convey the special pragmatic statuses that commonly fall under the rubric of 'focus', 'contrast' or 'topicalization'. As mentioned above, I am using the non-technical term 'pragmatic prominence' as a cover for all of these three domains (and perhaps others). \ftx \txt The use of intonation is fairly self evident -- speakers ascribe pragmatic prominence to (or 'highlight') parts of their sentences by pronouncing them more loudly or with a higher pitch. A few languages (e.g. Russian) are said to employ lower pitch (but still higher volume) for prominent information. We will have nothing further to say about intonation here. In the following sections we will provide examples of constituent order, formatives and cleft constructions. \ftx \txt To continue this discussion, highlight and jump to one of the topic below: \cf Hprag_prom_ord Constituent Order \cf Hprag_prom_form Formatives \cf Hprag_prom_clft Cleft Constructions \ftx \dt 27/Apr/1998 \key Hprag_prom_clft \shd Cleft constructions \txt A CLEFT CONSTRUCTION is a type of predicate-nominal consisting of a noun phrase (NPi) and a relative clause whose relativized NP is coreferential with NPi (see Hcplx_relcl -- Relative Clauses). NPi is commonly referred to as the 'clefted constituent', and is normally found to the left of the rest of the sentence, though it may appear in other positions. Cleft constructions can be formulated as follows: \ftx \fln (37) NPi (COP) [ ... NPi ... ]Srel \ftx \txt The form that Srel takes depends on what relativization strategies the language employs, i.e. it could be a nominalization, a participial clause or a more prototypical relative clause (see Hcplx_relcl). Similarly the presence or absence of COP depends on the general structure of predicate nominal constructions in the language. As stated above, the presence of COP is a clear indication that one is dealing with a cleft construction. If COP is absent, the construction may still be a cleft if the language allows predicate nominal constructions with no copular element (see Hpnom_intro). In such a case, a cleft is distinguished from plain left-dislocation in that the clause that follows the leftward NP is a relative clause or other type of participant nominalization. In a very few languages, namely those that both allow predicate nominals with no copula and relative clauses with no relativizer or other special morphology, some structures may be indeterminate as to whether they are best thought of as left-dislocation or clefting. Dera (Chadic) is apparently such a language: \ftx \fln (38) wuni wun kapa kurei 'THOSE ONES plant corn.' \ftx they ones plant corn or 'Those are the ones who plant corn.' \ftx \txt To continue this discussion, highlight and jump to the topics below: \cf Hprag_prom_clft_ex_eng English examples \cf Hprag_prom_clft_ex_other Examples from other languages \ftx \dt 27/Apr/1998 \key Hprag_prom_clft_ex_eng \shd Cleft constructions--English examples \fln Some examples of clefts in English include the following: \ftx \fln (39) a. Home is [where the heart is 0 ]S. (c.f. 'The heart is at home.') \ftx NPi COP REL NPi \ftx \ftx b. Lucretia is [whom I love 0 ]S. (c.f. 'I love Lucretia') \ftx NPi COP REL NPi \ftx \txt English has at least two types of cleft constructions. These have traditionally been termed clefts and PSUEDO-CLEFTS: \ftx \fln (40) Clefts: \ftx a. It is Lucretia who grimaced. \ftx b. It's the Duke whom Lucretia distains. \ftx c. It's the Duke who trod on poor Lucretia's watermelon. \ftx \fln (41) Psuedo-Clefts (NP COP Srel) \ftx a. Lucretia is the one who grimaced. \ftx b. The Duke is the one whom Lucretia distains. \ftx c. The Duke is the one who trod on poor Lucretia's watermelon. \ftx d. Home is where the heart is. \ftx \fln (42) Psuedo-Clefts (Srel COP NP) \ftx a. What happened was you blew a heater hose. (c.f. 'that you \ftx blew a heater hose \ftx happened') \ftx b. What John ate was beans. \ftx c. The one who grimaced was Yassar. \ftx d. That which we have seen with our own eyes is what we are \ftx reporting to you. (Both NPs in the predicate nominal are \ftx relative clauses) \ftx \txt In fact, by our definition of cleft as a predicate nominal consisting of a noun phrase and a relative clause that relativizes that noun phrase, all of these construction types are clefts. 'It-clefts' (ex. 40) and 'the one clefts' (ex. 41a, b and c) simply represent two different means that English employs to avoid having to use a headless relative clause (see Hcplx_relcl_hdless). The most 'natural' (from a universal point of view) form of a cleft in English would involve a headless relative clause (41d, 42a, b and the following): \ftx \fln (43) ?Lucretia is who grimaced. \ftx ?The Duke is whom Lucretia distains. \ftx ?The Duke is who trod on poor Lucretia's turnip. \ftx \txt Clefts in many languages exhibit this pattern (NP COP Headless-RC, see below). However, headless relative clauses in English are generally avoided, at least in written and otherwise planned speech. For this reason a 'dummy' element is employed, either as the subject (it in 40) or as the head (the one in 42a, b and c). \ftx \txt Prince (1978) provides an interesting and insightful analysis of the functions of various kinds of cleft constructions in English. This work could serve as a model for other studies of the functions of pragmatically marked structures in other languages. However, one must always be aware that similar structures from one language to the next may or may not have similar functions. Therefore the fieldworker should be careful not to project Prince's findings on data from another language without adequate empirical evidence from the language itself. \ftx \txt To continue this discussion, highlight and jump to the topic below: \cf Hprag_prom_clft_ex_other \ftx \dt 27/Apr/1998 \key Hprag_prom_clft_ex_other \shd Cleft constructions--Examples from other languages \txt The following are examples of various types of clefts in other languages (data from Harries-Delisle 1978): \ftx \ftx HEADLESS RELATIVE CLAUSE: \fln (44) Mandarin: Yohàn kàn-jièn de shìh ge nán rén \ftx John saw REL be CL male person \ftx 'Who John saw was a man.' (not a woman) \ftx \fln (45) Indonesian: bukan saya yang beladjar bahasa indónésia \ftx NEG I REL study language Indonesia \ftx 'I am not who is studying Indonesian.' \ftx \ftx PARTICIPIAL PHRASE: \fln (46) German: Der segelt das ist mein Bruder. \ftx the sail:PP? that be my brother \ftx 'The sailing (one), that is my brother.' \ftx \ftx NOMINALIZATION: \fln (47) Amharic: êssu naw yamattaw \ftx 3SG be NOM:came \ftx 'He is who came'. (He is the 'comer') \ftx \txt The following are examples from Malayalam (from Andrews 1985:84-85). Example 48 is an unclefted sentence. Examples 49a through 49d are clefts formed with various constituents. Note that in Malayalam, the clefted constituent does not have to appear strictly to the left of the sentence (though it may): \ftx \fln (48) Kutti innale ammakkê aanaye kotuttu \ftx child:NOM yesterday mother:DAT elephant:DAT gave \ftx 'The child gave an elephant to the mother yesterday.' \ftx \fln (49) a. Kuttiy-aanê innale ammakkê aanaye kotutt-atê \ftx child:NOM-is yesterday mother:DAT elephant:DAT gave-it \ftx 'The child is (he who) gave an elephant to the mother yesterday.' \ftx \ftx b. Kutti innale ammakk-aanê aanaye kotutt-atê \ftx child:NOM yesterday mother:DAT-is elephant:DAT gave-it \ftx 'It is the mother that the child gave an elephant to yesterday.' \ftx \ftx c. Kutti innaley-aanê ammakkê aanaye kotutt-atê \ftx child:NOM yesterday-is mother:DAT elephant:DAT gave-it \ftx 'It is yesterday that the child gave an elephant to the mother.' \ftx \ftx d. Kutti innale ammakkê aanayey-aan kotutt-atê \ftx child:NOM yesterday mother:DAT elephant:DAT-is gave-it \ftx 'The child gave an elephant to the mother yesterday.' \ftx \dt 27/Apr/1998 \key Hprag_prom_form \shd Formatives \txt Some languages employ morphemes or particles to ascribe special pragmatic status to noun phrases in sentences. There is a continuum between morphological case markers (see Hgrel_case) and markers of pragmatic status. This continuum can be roughly divided as follows: \ftx \ftx Pragmatic status markers English articles, \ftx Aghem focus particles, etc. \ftx Overlay systems Japanese and Korean 'topic marking' \ftx Case Markers Latin, Eskimo, Russian, Quechua, etc. \ftx \txt It must be kept in mind throughout this discussion that these structure types really represent a continuous scale. It is in principle very difficult to tease apart grammatical relations, semantic roles and pragmatic status since they all influence one another to a great extent. However, generalizations can be made concerning the commonest, or most prototypical, functions of certain structure types. Prototypically, case markers are those grammatical devices that most directly indicate grammatical relations, i.e. grammaticalized semantic roles and pragmatic statuses (see Qintro_what_hbehav_rout2 -- Grammaticalization, and Hgrel_intro -- Grammatical Relations). The 'articles' of English are examples of pragmatic status markers. Typically, pragmatic status markers partially correlate with grammatical relations. For example, noun phrases that have the grammatical relation of subject are also likely to have the pragmatic status of identifiable (or definite) in English, while objects are more likely to be non-identifiable (or indefinite). If this statistical correlation becomes a 100% generalization, then the pragmatic status marker 'the' can be said to have become a subject case marker (see Shibatani 1990 for discussion of how pragmatic categories can become grammaticalized as grammatical relations). \ftx \txt To continue this discussion, highlight and jump to the topics below: \cf Hprag_prom_form_ex Examples \cf Hprag_prom_form_ovl Overlay systems \ftx \dt 27/Apr/1998 \key Hprag_prom_form_ex \shd Formatives--Examples \txt Aghem uses verb morphology and a focus particle to convey various pragmatic nuances. For example, there is a special form of the auxiliary verb that is used for TVF clauses in the perfective aspect. Example 32a illustrates a neutral perfective aspect clause, while 32b illustrates a TVF perfective aspect clause: \ftx \fln (32) a. éná¿ mò fúo kï-be â fïn-ghó \ftx Inah AUX give fufu to friends \ftx 'Inah gave fufu to his friends.' \ftx \ftx b. éná¿ má'á fúo kï-be â fïn-ghó \ftx Inah AUX:FOC give fufu to friends \ftx 'Inah DID give fufu to his friends.' \ftx \txt There is also a 'focus particle' 'nò' in Aghem that appears to the right of a focussed constituent. Sometimes the choice of whether to use word order or the particle 'nò' to accomplish a particular focus task appears to be completely free (see Hprag_prom_ord_ex for examples of word order as a focussing device in Aghem): \ftx \fln (33) a. fú kï mô ñïn nò á kï-'bé 'The rat RAN (i.e. did not walk) \ftx rat SM AUX run FOC in compound in the compound.' \ftx \ftx b. fú kï mô ñïn á kï-'bé nò 'The rat ran in the COMPOUND \ftx rat SM AUX run in compound FOC (not in the house).' \ftx \txt In Akan (Schachter 1985:37) there is a 'focus' particle 'na' (example 34) and a 'contrastive' particle 'de' (ex. 35): \ftx \fln (34) Kwame na ôbeye adwuma no. \ftx Kwame FOC he:will:do work the \ftx 'It's Kwame who will do the work.' \ftx \fln (35) Kwame de ôbekô, na Kofi de ôbetena ha \ftx Kwame CONTR he:will:go and Kofi CONTR he:will:stay here \ftx 'KWAME will go, but KOFI will stay here.' \ftx \txt To continue this discussion, highlight and jump to the topic below: \cf Hprag_prom_form_ovl \ftx \dt 27/Apr/1998 \key Hprag_prom_form_ovl \shd Formatives--Overlay systems \txt Overlay systems for marking pragmatic status of nominal elements are a combination of morphological case marking systems and pragmatic status marking systems. The essence of an overlay system is that one or more basic case markers are replaced ('overlaid') by the pragmatic status markers when a nominal element is singled out for special pragmatic treatment. Both Japanese and Korean have overlay systems for marking 'topic' (defined in a language-specific sense). The topic marker in Japanese is 'wa'. It can overlay either the subject marker 'ga' (example 36b, the object marker 'o' 36c or other nominal case markers: \ftx \fln (36) a. Unmarked: taroo ga hon o katta 'Taro bought a book.' \ftx Taro SUBJ book OBJ bought \ftx \ftx b. taroo wa hon o katta 'As for Taro, he bought \ftx TOP a book.' \ftx \ftx c. hon wa taroo ga katta 'As for the book, Taro \ftx TOP bought it.' \ftx \txt In linguistic articles, English translations of Japanese sentences with 'wa' marked nominals typically employ the 'as for X . . ' left-dislocation construction. In fact Japanese 'wa' has various functions, and is still a matter of some controversy (see, e.g., Hinds, Maynard and Iwasaki 1987). \ftx \dt 27/Apr/1998 \key Hprag_prom_ord \shd Constituent order--General \txt The first step in determining what constituent orders are used to convey special pragmatic statuses is to determine the basic constituent order of the language (see Hord_cl_case). This is because pragmatically marked constituent orders are always determined in relation to the basic constituent order. If the language has no basic order, then order probably does not convey special pragmatic statuses. Rather, it probably conveys something more mundane, like discourse referentiality or identifiability (see Hprag_iden -- Identifiability and Discourse Referentiality). \ftx \txt There are two general clause-internal positions where pragmatically prominent information may be placed: (1) Sentence initial position and (2) adjacent to the verb. For example, since English is an AVP language, sentence-initial position and immediately pre-verbal position are candidates for pragmatically prominent P arguments. English appropriates only sentence-initial position for this purpose: \ftx \fln (17) a. Beans I like. \ftx b. *I beans like. \ftx \txt Since sentence-initial is the normal position for A arguments in English, that position does not ascribe special prominence to A arguments. Similarly, sentence-final position does not ascribe prominence to P arguments. Other logically possible positions are simply not utilized in English: \ftx \fln (18) a. *Like I beans \ftx b. *Like beans I \ftx \txt The fact that English allows a very small set of possible constituent orders indicates that English is a CONFIGURATIONAL language (Hale 1983). Languages that allow all orders of constituents are sometimes termed NON-CONFIGURATIONAL languages. \ftx \shd2 Aghem example \ftx \txt If a position adjacent to the verb is a pragmatically prominent position, it may be immediately before the verb (IBV) or immediately after the verb (IAV), or both. Aghem employs both positions for various kinds of prominence. Like most languages of West Africa, Aghem exhibits the basic constituent order A AUX V P. The function of IAV position in Aghem is to convey focus as asserted or new information (characterization #2 of the typology of focus outlined in Hprag_foc): \ftx \fln (19) Question: 'Who ran?' \ftx \ftx Answer: à mò ñïn éná¿ 'INAH ran.' \ftx it AUX run Inah \ftx \fln (20) Question: fïl á mò zï kwò 'What did the friends eat?' \ftx friend SM AUX eat what \ftx \ftx Answer: fïl á mò zï kï-be 'The friends ate FUFU.' \ftx friend SM AUX eat fufu \ftx \fln (21) Question: fïl á mò zï ghe be-'kó 'Where did the friends eat fufu?' \ftx where \ftx \ftx Answer: fïl á mò zï án 'sóm be-'kó 'The friends ate fufu on \ftx on farm the farm.' \ftx \fln (22) Question: fïl á mò zï zïn be-'kó 'When did the friends eat fufu?' \ftx when \ftx \ftx Answer: fïl á mò zò á'zóô be-'kó 'The friends ate fufu \ftx yesterday yesterday.' \ftx \txt In all of these examples, including the questions, the focussed constituent comes in the IAV position. Intransitive subjects normally precede the verb, but in 19 the focussed subject follows. Similarly, in the rest of the examples, the element that corresponds to the question word in the question appears in IAV position. \ftx \txt The position immediately before the verb (IBV) is employed for contrastive (or counter-presuppositional) focus in Aghem (under characterization #3 of the typology of focus in Hprag_foc): \ftx \fln (23) fïl á mò be-'kï án 'sóm zï in the farm (not the house).' \ftx á'zóô yesterday (not two days ago).' \ftx án 'wó with hands (not spoons).' \ftx 'The friends ate fufu . . . \ftx \txt To continue this discussion, highlight and jump to the topic below: \cf Hprag_prom_ord_dis \ftx \dt 27/Apr/1998 \key Hprag_prom_ord_dis \shd Constituent order--Dislocation \txt DISLOCATION (left and right) refers to the placing of a clause element outside the syntactic boundaries of the clause. Sometimes dislocation is referred to as EXTRAPOSITION. Left-dislocation is sometimes referred to as PREPOSING and right-dislocation as POST-POSING. The term TOPICALIZATION refers to left-dislocation in the tradition of generative grammar and other autonomous approaches to syntax. Right dislocation is sometimes referred to as AFTERTHOUGHT TOPICALIZATION. All of these terminologies assume that the leftward nominal in left-dislocation occupies a constituent structure position that stands outside the sentence but is still adjoined to the sentence at a higher level. In the generative tradition, that position is often referred to as the TOPIC position: \ftx \fln (24) S' --> TOPIC S \ftx \fln Various elements can be moved out of S into the TOPIC position, e.g. \ftx \fln (25) a. My father, he likes Beethoven. \ftx b. Beethoven, now I enjoy his music. \ftx \txt This notion of topic is strictly structural. Whatever functional (i.e. communicative) properties may be associated with topicalization constructions in the generative tradition are tangential to their structural status. \ftx \txt To continue this discussion, highlight and jump to the topic below: \cf Hprag_prom_ord_dis_other Other kinds of movement \cf Hprag_prom_ord_dis_rule Rules \cf Hprag_prom_ord_dis_ex Examples \cf Hprag_prom_ord_dis_part Use of particles \ftx \dt 27/Apr/1998 \key Hprag_prom_ord_dis_ex \shd Constituent order--Examples \ftx \fln (28) Fronting: Beans I like. \ftx Left dislocation: Beans, I like them. \ftx As for beans, I think they're great. \ftx Apposition: Beans. Why do we always have leftovers? \ftx Cleft: Beans are what I like. \ftx What I like is beans. \ftx \fln To continue this discussion, highlight and jump to the topic below: \cf Hprag_prom_ord_dis_part \ftx \key Hprag_prom_ord_dis_other \shd Constituent order--Other kinds of movement \txt Apparently all languages employ left dislocation as a grammaticalized construction. Some also employ right dislocation. It may be difficult to distinguish left- dislocation from (1) APPOSITION of a free noun phrase to the sentence, (2) FRONTING of an element within the clause, and (3) CLEFTING (see Hprag_prom_clft). The corresponding difficulty may also entail for right-dislocation. However, the following discussion will be couched exclusively in terms of left dislocation. The issue is whether the element to the left of the main predication is grammatically a part of the predication or not. That is, there are three possible grammatical statuses of a pragmatically prominent noun phrase that is in sentence-initial position. These statuses can be schematized as follows: \ftx \fln (26) a. [NP] [S] Apposition \ftx b. [NP S]S' Left dislocation \ftx c. [NP ... ]S Fronting (if NP is not initial in the neutral \ftx constituent order) \ftx \txt In other words, noun phrases that are placed in sentence-initial position for pragmatic highlighting can be grammatically separate from the following sentence (26a), grammatically adjoined to the sentence but not an integral part of it (26b), or an integral part of the sentence (26c). \ftx \txt NOTE: Foley and Van Valin (1985) use the term topicalization for 26b and left-dislocation for 26c. Here we do not adopt this terminology for two reasons: 1) In the interests of maintaining a truly universal perspective, we avoid using functional terms, such as topicalization, to describe morphosyntactic devices. This is because structures that are analogous from one language to the next may not have analogous functions. 2) To consider 26c to be left-dislocation entails that the leftward NP is external to the clause. From the field linguist's point of view there is no independent evidence for this entailment. In fact, it would seem to imply that there are four levels of grammatical integration, one of which is unattested in English: \ftx \ftx Fronted but clause-internal: (unattested in English) \ftx left-dislocation Beans I like. \ftx Topicalization (As for) beans, I like them. \ftx Apposition Beans. What a great lunch. \ftx \txt A survey of the literature reveals no other sources that use Foley and Van Valin's terminology. \ftx \txt In addition to these grammatical statuses, a leftward NP may also be clefted. The grammatical structure of a cleft construction may be schematized as follows: \ftx \fln (27) [NPi] COP [ ... NPi ...]S \ftx \fln Cleft constructions are discussed in more detail in: \cf Hprag_prom_clft \ftx \txt To continue this discussion, highlight and jump to the topic below: \cf Hprag_prom_ord_dis_rule \ftx \dt 27/Apr/1998 \key Hprag_prom_ord_dis_part \shd Constituent order--Use of particles \txt Many languages employ particles to set left dislocated elements off from the rest of the sentence. Such constructions have much in common with clefts, discussed in Hprag_prom_clft. Here we will provide a way to tell the difference. If the particle that sets left-dislocated NPs off from the rest of the sentence is not 'be', then the construction can still be thought of as left-dislocation. If, however, the particle is a form of the verb or invariant particle that appears in predicate nominal constructions (i.e. the form that means 'be'), then it is a cleft. \ftx \txt The special particles that many languages employ in left-dislocation constructions often derive historically from older copular forms. This fact illustrates that the distinction between clefting and left-dislocation is continuous rather than absolute. For expository purposes, however, it is convenient to draw the line at the point where the particle that sets off the left dislocated element ceases to function as a copula in predicate nominal constructions. The following are examples of languages that employ special particles to set off left dislocated noun phrases: \ftx \fln (29) Tagalog: Ang babae ay humiram ng pera sa bangko. \ftx ABS woman LD A:borrow OBL money OBL bank \ftx 'The woman, she borrowed money from a bank.' \ftx \txt In this language the particle 'ay' functions like comma intonation does in other languages. It does not occur in predicate nominal constructions (unless the subject of the predicate nominal is left-dislocated). There is also a fronting construction that does not employ 'ay', and a distinct cleft construction (see Hprag_prom_clft). \ftx \fln (30) Malagasy: izahay no tia anao. 'WE love you.' \ftx we LD love you \ftx \fln (31) Akan: kòfí nà ówó Engìrési. 'KOFI is in England.' \ftx Kofi LD be:in England \ftx \txt In general, the functions of these pragmatically marked structures progress in markedness from fronting to clefting (see Hprag_prom_clft), i.e. fronting is more marked than the neutral constituent order, left-dislocation is more marked than fronting and clefting is more marked than left-dislocation. However, the specific functions of these construction types for any given language need to be investigated individually. \ftx \dt 27/Apr/1998 \key Hprag_prom_ord_dis_rule \shd Constituent order--Rules relating to dislocation \txt The following rules of thumb will help determine what kind of construction one is dealing with (these rules are ordered): \ftx \ftx 1. If the construction normally falls under a single \ftx intonation contour, i.e. there is typically no pause or \ftx 'comma' intonation after the initial NP, AND there is no \ftx special particle between the initial NP and the rest of the \ftx clause AND there is no reference to the initial nominal \ftx within the clause (other than grammatical agreement), it is \ftx fronting. \ftx \ftx 2. If the initial NP is recapitulated within S by a free \ftx referring form (i.e. anything besides grammatical agreement) \ftx AND a pause or a special particle (other than the copula) \ftx can naturally intervene between the initial NP and S, then \ftx it is probably left dislocation. \ftx \ftx 3. If the initial NP has no role in S, and/or adverbial \ftx elements can intervene between the initial NP and S, then it \ftx is probably apposition (sometimes referred to as \ftx JUXTAPOSITION). \ftx \ftx 4. If the element that intervenes between the initial NP and \ftx the rest of the clause is a form of the copula AND/OR the \ftx main predication has the form of a relative clause, then it \ftx is a cleft. \ftx \txt To continue this discussion, highlight and jump to the topic below: \cf Hprag_prom_ord_dis_ex \ftx \key Hprag_prom_ord_ex \shd \dt 21/Aug/1997 \key Hprag_stat \shd Pragmatic statuses--Introduction \txt Pragmatics has to do with the way a message is structured so as to take into account the context, including the hearer's presumed 'mental state'. Whereas semantic roles are features of the content of the discourse (see Hgr_v_cl_sem), PRAGMATIC STATUSES relate the content to the context. Labels that have been used to describe pragmatic statuses include: GIVEN, NEW, PRESUPPOSED, FOCUS, TOPIC, DEFINITE and REFERENTIAL. These terms will be described in the following subsections. But first we will sketch the conceptual background to these pragmatic notions. \ftx \txt Human beings have a limited ability to attend to stimuli. When communicating with other people, we as speakers need to constantly 1) assess our audience's present mental state, e.g. what they already know, what they are currently thinking about, what they are interested in, etc., and 2) we must construct our message so as to revise that mental state in the direction we want it to go. For example, we may highlight items that we want someone to pay attention to right now, and which we sense he or she is not already paying attention to. Also, we may spend little communicative energy on information which we sense the hearer is already thinking about, or has available in short term memory. Some languages have morphosyntactic devices that respond more or less directly to these kinds of pragmatic notions. \ftx \txt It should be pointed out that grammatical relations are one major means of conveying pragmatic information about nominal elements in discourse (see chapter #7). For example, subjects universally tend to be definite, given and already available in memory. Direct objects are either given or new in about equal proportions. Obliques tend to encode new information and/or information that is not central to the ongoing development of the discourse (Givón 1983b). Also, the pragmatic status of a nominal as given or new is influenced by many factors, including semantic roles. So, for example, people are more likely to choose AGENTS as the main topics of their discourses (conversely, one might say that discourse topics tend to be highly agentive). This is because there is a natural human tendency to discuss things that exercise power and control rather than things which don't. \ftx \txt In addition to the routine pragmatic statuses accorded to nominal elements in clauses by grammatical relations, languages typically express special (i.e. marked or unusual) pragmatic statuses via special morphosyntactic devices. These are commonly referred to as 'focus' or 'emphatic' devices. However, the fieldworker should not use these terms unless they are defined very explicitly. These are probably the most overused and misused terms in the science of linguistics. Instead, we will use the non-technical term 'pragmatic prominence' as a cover for all marked or unusual pragmatic statuses. There are two reasons for this lumping: (1) there is little standardization of terminology within this domain (e.g. the terms 'focus' and 'topic' are opposites in some characterizations and synonyms in others!), and (2) a particular device may function in different ways in different languages. The devices described in Hprag_prom, however, are united in that they typically ascribe some sort of unusual pragmatic status to a sentence element. Which particular status that is varies from language to language. \ftx \txt Probably the most common way of ascribing pragmatic prominence to particular pieces of information is INTONATION. We draw special attention to parts of the sentences we utter by pronouncing those parts more loudly, and/or at a higher pitch. Other common means of ascribing pragmatic prominence are word order, morphosyntactic operators (morphemes or particles) and various kinds of CLEFT constructions. These devices are described and exemplified in: \ftx \cf Hprag_prom_ord Constituent Order \cf Hprag_prom_form Formatives \cf Hprag_prom_clft Cleft Constructions \ftx \dt 27/Apr/1998 \key Hprag_top \shd Topic \txt Like the term 'focus', the term 'topic' has been characterized according to several broad approaches: \ftx \ftx 1. The topic is a dislocated sentence constituent (see \ftx Hprag_prom_ord_dis on left and right dislocation). \ftx Sometimes such elements are termed 'topicalized', and the \ftx pragmatically marked structures that encode them \ftx 'topicalization'. \ftx \ftx 2. The topic is a sentence-level notion that can be \ftx paraphrased 'what the sentence is about'. Every (or almost \ftx every) sentence has a topic in this sense (Reinhart 1982). \ftx \ftx 3. The topic is a discourse-level notion that can be \ftx paraphrased 'what the discourse is about'. Not every \ftx sentence in a discourse may mention the topic in this sense. \ftx \ftx 4. The topic is 'the [concepetual or referential] frame \ftx within which the rest of the predication holds' (Li and \ftx Thompson 1976). \ftx \ftx 5. Topicality is a scalar discourse notion. Every nominal \ftx participant is topical to a certain degree. Topicality is \ftx measured in terms of how often a participant is mentioned \ftx over a span of text (Givón 1983). \ftx \txt The notion of topicalization as left- and right-dislocation is most prominent in approaches that assume that syntax is autonomous of communicative function (see Hprag_prom_ord_dis). \ftx \txt The notion of topic as a sentence level notion probably stems from the work of the Prague school linguists (see above). Like the term focus, topic was not used by these early linguists. Nevertheless, they came up with the concept that part of every (or almost every) sentence is old, given or known information. This part of the sentence was called the THEME by the Prague school linguists. It was defined in contrast to the RHEME, i.e. that part of the sentence that conveys new or asserted information. This conceptual distinction is what eventually evolved into the topic/focus distinction. \ftx \fln REFERENCES: \ftx \dt 27/Apr/1998 \key Htopics_outline \shd Linguistic Field Manual: first level contents \txt (Some of these topics are marked as incomplete (usually with double asterisks **.) This only means that the cross-referencing is not complete. Many of the topics are quite exhaustive. For more complete cross-referencing, go to the associated topic in the Grammar Outline database and jump from there.) \cf Heth Demographic and Ethnographic Information \cf Hmor Morphological Typology \cf Hgr Grammatical Categories \cf Hord Constituent Order Typology \cf Hn Noun and Noun Phrase Operations \cf Hpnom Predicate Nominals and Related Constructions \cf Hgrel Grammatical Relations \cf Hval Valence-changing Operations \cf Hv Verb and Verb Phrase Operations \cf Hprag Pragmatically-marked Structures \cf Hcplx Complex Constructions \cf Hdisc The Language as a Vehicle for Discourse \dt 29/May/1998 \key Hv \shd Verb and verb-phrase operations \txt This topic contains all of the following subsections. Highlight and jump to the topic of interest: \shd2 Nominalization \cf Hv_nz nominalization \cf Hv_nz_actn Action Nominalization \cf Hv_nz_part Participant Nominalization \cf Hv_other Other Deverbalizing Processes \shd2 Compounding \cf Hv_cmp Compounding \cf Hv_cmp_n Noun Incorporation \cf Hv_cmp_v Complex Verbs \cf Hv_cmp_v_cmp/ser Compound vs serial verbs \shd2 Tense/Aspect/Mode \cf Hv_tam Tense/Aspect/Mode \cf Hv_tam_tns Tense \cf Hv_tam_asp Aspect \cf Hv_tam_asp_perfv Perfective Aspect \cf Hv_tam_asp_imprf Imperfective Aspect \cf Hv_tam_asp_perf Perfect Aspect \cf Hv_tam_asp_cplv Completive Aspect \cf Hv_tam_asp_incpt Inceptive Aspect \cf Hv_tam_asp_plu Pluperfect Aspect \cf Hv_tam_asp_cntv Continuative/Progressive Aspect \cf Hv_tam_asp_punc Punctual Aspect \cf Hv_tam_asp_iter Iterative Aspect \cf Hv_tam_asp_hab Habitual Aspect \cf Hv_tam_mode Mode \cf Hv_tam_mode_subj Subjunctive Mode \cf Hv_tam_mode_opt Optative Mode \cf Hv_tam_mode_potn Potential Mode \cf Hv_tam_mode_hypo Hypothetical Mode \cf Hv_tam_mode_cnd Conditional Mode \cf Hv_tam_mode_deon Deontic Mode \cf Hv_tam_mode_epist Epistemic Mode \cf Hv_tam_msyn Morphosyntax \cf Hv_tam_ncode Nominal Coding \cf Hv_tam_adv Temporal Adverbs \cf Hv_tam_auxv Auxiliary Verbs \shd2 Location/Direction \cf Hv_loc Location/Direction \cf Hv_loc_cult Cultural influence \shd2 Agreement \cf Hv_agr Agreement \cf Hv_evid Evidentiality \cf Hv_evid_vsval Evidentiality vs Validation \cf Hv_evid_ex1 Examples I \cf Hv_evid_modl Modal vs Validational marking \shd2 Miscellaneous \cf Hv_misc Miscellaneous \cf Hv_misc_lxtim Lexical Time Reference \cf Hv_misc_dist Distributive \cf Hv_misc_env Environmental \cf Hv_misc_env_ex Athabaskan example \dt 12/May/1998 \key Hval \shd Valence changing operations \txt This topic contains the following subsections. Highlight and jump to the topic of interest: \shd2 Valence and predicate calculus \cf Hval_p-calc Valence and Predicate Calculus \cf Hval_p-calc_intro General Discussion \cf Hval_p-calc_as Valence as Predicate Calculus \shd2 Valence increasing operations \cf Hval_incr Valence Increasing Operations \cf Hval_incr_caus Causatives \cf Hval_incr_appl Applicatives \cf Hval_incr_dat_int Dative of interest \cf Hval_incr_dat_shift Dative Shift \shd2 Valence decreasing operations \cf Hval_decr Valence Decreasing Operations \cf Hval_decr_refl Reflexives \cf Hval_decr_refl_other Other Coreference \cf Hval_decr_recip Reciprocals \cf Hval_decr_pass Passives \cf Hval_decr_mid Middle Constructions \cf Hval_decr_anti-p Antipassives \cf Hval_decr_demot Object Demotion \cf Hval_decr_ncorp Object incorporation \dt 12/May/1998 \key Hval_decr \shd Valence decreasing operations \txt Languages can have morphological, lexical and periphrastic/analytic means of reducing the valence of a verb. The most common morphological valence decreasing operations are reflexives/ reciprocals, passives, and antipassives. These will be discussed in the following three subsections. Here we will introduce valence decreasing operations with an illustration from Panare, a Carib language of Venezuela. \ftx \txt Carib languages are famous for valence decreasing operations. In fact, in Panare it seems that most intransitive verbs in the lexicon are derived from transitives, though there are some semantically determined exceptions. Mayan languages also employ extensive detransitivization. \ftx \dt 27/Apr/1998 \key Hval_decr_anti-p \shd Antipassives \txt Like passive, ANTIPASSIVE is a valence-decreasing operation. That is, it downplays the conceptual centrality of one participant by downgrading the syntactic status of the verbal argument that refers to that participant. Unlike passives, however, antipassives downplay the centrality of a PATIENT or 'P' argument rather than an AGENT or 'A' argument. For example: \ftx \ftx Eskimo: \fln (77) TRANSITIVE: Yero-m keme-q nerre-llru-a. 'Yero ate the meat.' \ftx Y. -ERG meat-ABS eat-PAST-3SG/3SG \ftx \fln (78) ANTIPASSIVE: \ftx Yero-q (kemer-meng) nerre-llru-u-q. 'Yero ate (meat).' \ftx Y. -ABS meat-INST eat-PAST-INTRNS-3SG \ftx \fln Prototypical antipassives have the following formal characteristics: \ftx \ftx 1) The P argument occurs in the INSTRUMENTAL or some other \ftx oblique case. \ftx \ftx 2) The verb or verb phrase contains some overt marker of \ftx intransitivity (e.g. it may take an explicit marker of \ftx intransitivity, inflect like an intransitive verb, etc. \ftx depending on the formal characteristics of intransitive \ftx verbs in that language.) \ftx \ftx 3) The 'A' appears in the ABSOLUTIVE case. \ftx \txt The best examples of antipassives are found in morphologically ergative languages, i.e. those that have a morphologically defined absolutive case. In non-ergative languages, object demotion or omission serves essentially the same function as antipassive does in ergative languages. The crucial difference, if it is necessary to draw a distinction between object demotion/omission and antipassive (e.g. if a given language has both), is that in antipassives the verb takes some specific marker of antipassivization or intransitivity, whereas in object demotion /omission no such verbal marker occurs. \ftx \dt 27/Apr/1998 \key Hval_decr_demot \shd Object demotion \txt Like antipassive, OBJECT DEMOTION is an operation that downplays the centrality of a P argument. In fact, some linguists (e.g. Heath 1976) have treated object demotion and object omission as types of antipassivization. Object demotion sometimes indicates 'less involvement' of the P in the event expressed by the verb. For example: \ftx \ftx Bzhedukh (NW Caucasian language): \fln (79) TRANSITIVE: c¿aalya-m c¿êgwo-êr ya-zwoa \ftx boy-ERG field-ABS 3SG-plows \ftx 'The boy plows the field.' \ftx \fln (80) OBJECT DEMOTION: c¿aalya-r c¿êgwo-êm ya-zwoa \ftx boy-ABS field-OBL 3SG-plows \ftx 'The boy is trying to plow the field.' \ftx \txt In 79 the interpretation is that the field is in fact being plowed, whereas in 80 the field may or may not actually be affected by the action of the boy. Hence it can be said that the P is 'less involved' with the activity of plowing in example 80 than in example 79. The only reason we would probably not want to call 80 an antipassive is that the verb does not contain any overt representation of detransitivization. \ftx \ftx Adyghe (NW Caucasian): \fln (81) TRANSITIVE: JeRedzakw¿e-m s¿ale-r jïwïsïjas \ftx teacher-ERG boy-ABS admonished \ftx 'The teacher (successfully) admonished the boy.' \ftx (i.e. the boy's behavior changed as a result.) \ftx \fln (82) OBJECT DEMOTION: JeRedzakw¿e-r s¿ale-m jïwïsïjas \ftx teacher-ABS boy-ERG/OBL admonished \ftx 'The teacher (unsuccessfully) admonished the boy.' \ftx (i.e. the boy's behavior did not change as a result.) \ftx \txt These examples are very parallel to those in 79 and 80 above. Again the P is less involved in, or less affected by the action of the verb in 82 than in 81. \ftx \fln Object demotion also occurs in non-ergative languages. For example: \ftx \ftx English: \fln (83) TRANSITIVE: The hunter shot the deer. \ftx \fln (84) OBJECT DEMOTION: The hunter shot at the deer. \ftx \txt Note that, like the Caucasian languages cited above, the object demotion construction in English tends to convey a situation in which the P participant is less involved in or less affected by the action of the verb. \ftx \dt 27/Apr/1998 \key Hval_decr_mid \shd Middle constructions \txt The term middle or middle voice has been used in a variety of ways in different language traditions. What all of these 'middle' constructions have in common is that they involve detransitivization. The motivation for the term is that these constructions are neither passive nor active -- they are in between or 'middle'. We will consider a functional middle construction to be one that expresses a semantically transitive situation in terms of a process undergone by the PATIENT, rather than as an action carried out by an AGENT. Middle constructions are not to be confused with 'medial clauses' or 'medial verbs' (see Hcplx_med_medcl). The verb 'break' in English is a good example of a middle verb. Sometimes verbs of this class are called LABILE VERBS. We will use the term 'middle verb' simply to capture the functional similarity between intransitive constructions formed with these verbs and morphological 'middle' constructions in other languages. Used transitively break is a standard transitive verb. When used intransitively, however, the PATIENT rather than the AGENT is the subject, and the situation is expressed as a process (in terms of Chafe 1970) rather than as an action, e.g. \ftx \fln (70) TRANSITIVE: The workers broke the vase. \ftx \fln (71) MIDDLE VOICE: The vase broke (*by the workers). \ftx \txt This property distinguishes verbs such as 'break' from other verbs that can be either transitive or intransitive (probably the majority of verbs in English) e.g. \ftx \fln (72) TRANSITIVE: I hit the vase. \ftx \fln (73) 'MIDDLE': *The vase hit. \ftx \txt In English, middle constructions are lexical, i.e. whether a verb can head a middle voice clause is an inherent lexical property of verbs. In many languages, however, middle constructions are overtly signalled: \ftx \txt The only difference between a functional passive and a functional middle construction is that a passive treats the situation as an action carried out by an agent but just downplays the identity of the agent. A middle construction, on the other hand, treats the situation as a process, i.e. it completely ignores the role of the agent. Because passive and middle functions are so similar, many languages use the same morphology to express both. In Greek, for example, middle and passive constructions are the same in all tense/aspects except aorist: \ftx \ftx Koiné Greek \fln (74) Loose (present) \ftx Active lúo 'I let (someone) loose' \ftx Passive lúomai 'I am let loose (by someone)' \ftx Middle lúomai 'I become loose'/'I let myself loose' \ftx \ftx Loose (aorist) \ftx Active élusa 'I let (someone) loose' \ftx Passive elúTen 'I was let loose (by someone)' \ftx Middle elusámen 'I became loose'/'I let myself loose' \ftx \txt Mayan and Cariban languages, however, consistently treat middle constructions as distinct from passives: \ftx \ftx K'iche' Mayan (England 1988:74): \fln (75) Hit \ftx Active ch'ay 'hit' \ftx Passive xch'aay 'be hit (by someone)' \ftx Middle xch'aayik 'become hit' \ftx \ftx Panare (Cariban): \fln (76) Keep \ftx Active amaika 'keep' \ftx Passive amaikasa' 'be kept' \ftx Middle samaika 'stay/sit/remain' \ftx \txt Frequently, middle constructions convey the notion that the subject is both the controller and the affected participant. However, this characterization provides no way of distinguishing functional middles from functional reflexives. Indeed, in many languages reflexives and middles are expressed by the same morphosyntax, but not always. In order to consistently distinguish middle and reflexive functions, we must employ the notion of process versus action. A middle construction expresses the scene as a process whereas a reflexive (and a passive) express the situation as an event. \ftx \fln Sometimes morphological middle constructions are called ANTICAUSATIVES. \ftx \dt 27/Apr/1998 \key Hval_decr_ncorp \shd Object incorporation--noun incorporation defined \txt NOUN INCORPORATION is where a core noun phrase (subject or object) of a clause becomes 'attached to' or 'incorporated into' the verb. Incorporation exhibits all the characteristics of compounding discussed in Hv_cmp_n, namely: \ftx \ftx 1) a stress pattern characteristic of words rather than phrases \ftx 2) possibly unusual word order and \ftx 3) meanings that are more specific than the meanings of the \ftx individual parts. \ftx \txt OBJECT INCORPORATION is far more common than subject incorporation. In English both are possible, but not very productive. Incorporated forms are either lexicalized expressions such as 'babysit', or they are severely restricted with respect to their syntactic possibilities, e.g. 'fox-hunt' can only be used in the progressive 'we went fox hunting', not *'I fox hunted all morning', or *'I fox hunt for a living'. However, just about all transitive verb that describe customary activities can incorporate a direct object in the progressive aspect, e.g. 'girl-watching', 'car-washing' etc. Occasionally one will hear an example of subject incorporation, e.g. 'this medicine is doctor recommended'. This construction is, like non-lexicalized examples of object incorporation, highly restricted, occurring only in the passive voice, c.f. *'He went doctor recommending last week', or *'She doctor recommended aspirin for my headache'. In this case English reflects the universal tendency for object incorporation to be more prominent than subject incorporation. \ftx \txt Formally, object incorporation is a valence decreasing operation, since the object ceases to function as an independent noun phrase and becomes part of a formally intransitive verb. Object incorporation is common in Amerindian and Siberian languages. For example: \ftx \ftx Chukchee (Siberia): \fln (85) TRANSITIVE: Tumg-e na-ntêwat-ên kupre-n. \ftx friends-ERG 3SG-set-TRANS net-ABS \ftx 'The friends set the net.' \ftx \fln (86) INCORPORATION: Tumg-êt kupra-ntêwat-g'at. \ftx friends-NOM net-set-INTRANS \ftx 'The friends set nets.' \ftx \txt The functions of object incorporation tend to overlap considerably with those of other operations that downgrade the syntactic prominence of a P argument. Here is an interesting example from English: \ftx \fln (87) You may pay the amount to the service desk within the next 30 days or \ftx you will be pay deducted. \ftx \txt References: Sapir (1911), Georgia Green (1981), Mariane Mithun (1984), Jerrold Sadock (1986). \ftx \dt 27/Apr/1998 \key Hval_decr_panare \shd Valence decreasing operations--Panare \ftx \key Hval_decr_pass \shd Passives--prototypical \txt The definition of a PASSIVE clause to be employed here is based on a prototype (Comrie 1981, Givón 1984:164, Shibatani 1985). A prototypical passive clause is characterized both morphosyntactically and in terms of its discourse function. Morphosyntactically a passive is a semantically transitive (two core argument) clause for which the following three properties hold: \ftx \ftx 1) the AGENT (or most AGENT-like argument) is either omitted (not \ftx 'zero-pronominalized', see Hval_p-calc_zero) or demoted to \ftx an oblique role. \ftx \ftx 2) the other core argument (the 'P') possesses all \ftx properties of subjects relevant for the language as a whole. \ftx \ftx 3) the verb possesses any and all language-specific formal \ftx properties of intransitive verbs. \ftx \txt In terms of discourse function a prototypical passive is used in contexts where the A is relatively low in topicality with respect to the P. It is crucial to note that this is not a criterial definition. Rather it defines a prototype against which passive-like constructions can be compared. A construction may exhibit many or few of the morphosyntactic properties. Similarly, a passive-like construction may sometimes be used in contexts where a passive would be unexpected, given the above characterization. However, it is the case that constructions that possess the morphosyntactic properties of passives also generally exhibit the discourse functional property mentioned above. Givón (1982b, 1984:164 and 1990) provides a relatively comprehensive typology of various passive-like phenomena according to a definition similar to the one given here. In the following subsections we will discuss personal and impersonal passives. Under personal passives we will provide examples of lexical, morphological and analytic passives. \ftx \txt PERSONAL PASSIVES are constructions for which some specific agent is implied, but either is not expressed, or is expressed in an oblique role. Personal passives can be lexical, morphological or periphrastic/analytic. Examples of each type are provided in the following sections. \ftx \txt A LEXICAL PASSIVE is any clause headed by a verb that is inherently passive in character. To be inherently passive, the verb must embody a cognitive model that includes the presence of a causing AGENT, but the PATIENT must be the grammatical subject. A verb such as 'break' in English is not a lexical passive because when used intransitively it does not automatically embody a scene in which some AGENT acts upon some PATIENT, e.g. 'The window broke'. The verb 'bááryî' in Yagua, on the other hand, does specifically assert that the subject was the object of killing on the part of a conscious AGENT: Yagua: \ftx \fln (60) Sa-bááryî-máá. 'He was killed in battle.' \ftx 3SG-be:killed:in:battle-PERF \ftx \txt If we imagine an English verb such as 'murder' falling into the same class as 'break' in English we would have some idea of the sense of this Yagua verb, i.e. 'he murdered' would mean 'he was murdered.' Lexical passives are apparently quite rare. \ftx \txt MORPHOLOGICAL PASSIVES are very common. They often employ the same or similar morphology as does perfect aspect (see Hval_p-calc_src). Passive morphemes are also sometimes derived from copulas or affixes/particles that form nominalizations on the PATIENT of a verb. For example: \ftx \fln Kera (Afroasiatic, Chadic): \ftx (61) TRANSITIVE: Hùlúm gà-ng hàrgá-ng gìdè hiúw-a. \ftx man:DEF put-PAST goat-DEF womb pen-LOC \ftx 'The man put the goat in the pen.' \ftx \ftx (62) PASSIVE: Hàrgá-ng dè-gà-gè gìdè hiúw-a (kás hùlúm-a). \ftx goat-DEF PASS-put-REDUP womb pen-LOC hand man-LOC \ftx 'The goat was put in the pen (by the man).' \ftx \fln UTE (Agent of passive may not be expresed): \ftx (63) TRANSITIVE: Ta'wóci tûpüyci tïráabi-kya \ftx man rock throw-PAST \ftx 'The man threw the rock.' \ftx \ftx (64) PASSIVE : Tûpüyci tïráabi-ta-xa. \ftx rock throw-PASS-PAST \ftx 'The rock was thrown,' or 'Someone threw the rock.' \ftx \ftx \txt English has analytic passives. In English passives a copular verb plus the 'past participle' (a PATIENT nominalization) of the active verb is used: \ftx \fln (65) The city was destroy-ed (by the enemy). \ftx COP -NOM \ftx \ftx \txt The function of IMPERSONAL PASSIVES is essentially the same as that of basic passives: they downplay the centrality of an AGENT. In a sentence like the German 'Yesterday there was dancing', the identity of the participants in the dance is not central to the speaker's communicative goal; only the fact that dancing took place. English does not have a grammaticalized impersonal passive. The following English examples are close functional approximations of impersonal passive constructions in languages that have them. However, the English examples are based on other sentence patterns, namely a 'normal' active verb with a third person plural subject for 66a and an existential construction in 66b: \ftx \fln (66) a. They say that by 2000 there will be no more trains in America. \ftx \ftx b. There will be dancing in the streets. \ftx \txt Some other languages, e.g. German and Lithuanian, employ passive morphology when the AGENT is unspecified: \ftx \ftx German: \fln (67) Es wird hier ge-tanzt. 'Dancing takes place here.' \ftx it be here PASS-dance \ftx \fln Lithuanian (indicates uncertainty, doubt, etc.): \ftx \fln (68) Jo cia per griovî sokta. \ftx 3SG:GEN here over ditch jumped \ftx '(Evidently) he jumped over the ditch here.' \ftx (lit: By him here the ditch was jumped over.) \ftx \fln Spanish uses reflexive morphology in one kind of impersonal passive: \ftx \fln (69) Spanish: Se cae mucho acá. 'They fall a lot here.' \ftx \txt We know of no languages that employ specific morphology just for impersonal passives. This is not particularly surprising, as the same statement is almost true for personal passives as well. As mentioned above, both morphological and analytic personal passives tend to employ formal structures, either bound morphemes or free words as the case may be, whose 'basic' function is one of the following: \ftx \ftx 1) perfect aspect markers \ftx 2) copulas \ftx 3) PATIENT nominalizers. \ftx \dt 27/Apr/1998 \key Hval_decr_recip \shd Reflexives/reciprocals \txt A RECIPROCAL clause is very similar conceptually to a reflexive. For this reason, reciprocals and reflexives are often expressed identically. A prototypical reciprocal clause is one in which two participants equally act upon each other, i.e. both are equally AGENT and PATIENT. For example English 'they saw each other.' These are conceptually similar to reflexives in that both indicate that AGENT and PATIENT are coreferential, though for different reasons. \ftx \txt LEXICAL RECIPROCALS are verbs for which reciprocity is a built-in component of their semantics. Some lexically reciprocal verbs in English are 'kiss', 'meet', and 'shake hands', e.g. 'John and Mary kissed' usually means 'John and Mary kissed each other'. If some other situation is to be communicated, the object must be explicitly mentioned, e.g. 'John and Mary kissed Grandma.' \ftx \txt Many languages that have morphological reflexives also have MORPHOLOGICAL RECIPROCALS. These languages typically express reflexives and reciprocals with the same morphological operators. Here we will provide examples from Spanish and Yagua. \ftx \ftx Spanish: \fln (50) Juán se-quemó. 'Juan burned himself.' \ftx J. REFL-burn:3SG:PAST \ftx \fln (51) Juán y María se-conocieron en Lima. \ftx J. and M. REFL-meet:3PL:PAST in Lima \ftx 'Juan and Maria met (each other) in Lima.' \ftx \fln (52) Juán y María se-quemaron. \ftx J. and M. REFL-burn:3PL:PAST \ftx 'Juan and Maria burned themselves.' \ftx or 'Juan and Maria burned each other.' \ftx \txt Often such constructions are technically ambiguous, e.g. examples 51 and 52 above. However, there are some ways of resolving the ambiguity. When the subject is singular, the reflexive reading is demanded (e.g. ex 50). However, when the subject is plural, both reflexive and reciprocal readings are possible. In such cases, the context disambiguates. So example 51 would probably not mean 'Juan and María met themselves', as this represents a pragmatically bizarre interpretation, whereas 529 is more ambiguous out of context. \ftx \txt For more examples of morphological reciprocals, highlight and jump to: \cf Hval_decr_recip_mor_ex1 \ftx \shd2 Analytic Reciprocals \ftx \txt In English, reflexives and reciprocals are both analytic, but are not isomorphic. Reflexives use the reflexive pronouns, whereas reciprocals use the special anaphoric operation 'each other': \ftx \fln (55) Reciprocal: John and Mary saw each other. \ftx \dt 27/Apr/1998 \key Hval_decr_recip_mor_ex1 \shd Morphological Reciprocals--Examples I \txt Yagua is another language in which morphological reflexives and reciprocals are isomorphic (i.e. they have the same morphosyntactic form). In Yagua the reflexive/reciprocal enclitic is '-yu': \ftx \fln (53) Suunumívachiyu. 'He painted himself.' \ftx sa-junumívay-sìy-yù \ftx 3SG-paint-PAST1-REFL \ftx \fln (54) Ruuvañúúyanúyu. 'They were killing each other.' \ftx riy-juvay-núúy-janú-yù \ftx 3PL-kill-CONT-PAST3-REFL \ftx \dt 27/Apr/1998 \key Hval_decr_refl \shd Reflexives/reciprocals \ftx \txt A prototypical REFLEXIVE construction is one in which subject and object are the same entity, e.g. English 'she saw herself'. Reflexive operations reduce the semantic valence of a transitive clause by specifying that there are not two separate entities involved. Rather, one entity fulfills two semantic roles and/or grammatical relations. As with many functional operations, reflexives can be coded lexically, morphologically or analytically. \ftx \txt A LEXICAL REFLEXIVE is one which is tied to the lexical meaning of a particular verb. For example, the English verbs to 'get dressed', 'wash up', 'put on', 'shave', etc. all normally imply that the object is coreferential with the subject. e.g. \ftx \fln (43) Edward shaved, washed and got dressed. \ftx \txt This sentence implies that Edward shaved himself, washed himself and dressed himself. If some other object is intended, it must be explicitly mentioned, e.g. \ftx \fln (44) Edward washed Claire. \ftx \txt A MORPHOLOGICAL REFLEXIVE is expressed by one of the morphological processes discussed in Hmor_mproc. English has no morphological reflexives. The most well-known examples of morphological reflexives are probably those of Romance languages. However, the writing systems of these languages tend to obscure the fact that the reflexive morphemes are actually bound clitics rather than free words. \ftx \txt For examples of morphological reflexives, highlight and jump to: \cf Hval_decr_refl_mor_ex1 \ftx \shd2 Analytic reflexives \ftx \txt English has analytic reflexives. These are signalled by the 'reflexive pronouns', i.e. myself, yourself, himself, herself, ourselves, yourselves, themselves and itself, e.g. \ftx \fln (49) McGovern burned himself. \ftx \txt This is an analytic reflexive because the presence of the reflexive operation is conveyed via a lexical word that is distinct from the verb. From a purely syntactic point of view, the analytic reflexive operation of English is not a valence decreasing device. This is because, there are still two syntactic arguments - 'McGovern' and 'himself'. We may want to say, however, that this clause is 'semantically intransitive' because the two syntactic arguments refer to a single entity in the message world. \ftx \dt 27/Apr/1998 \key Hval_decr_refl_mor_ex1 \shd Morphological reflexive--Examples I \txt In Spanish a reflexive is formed from a transitive verb by the addition of the pro-clitic 'se': \ftx \fln (45) NON-REFLEXIVE: \ftx Juán quemó la cena. 'John burned dinner.' \ftx \fln (46) REFLEXIVE: \ftx Juán se-quemó. 'John burned himself.' \ftx \txt Note that even verbs that are lexical reflexives in English, must take the reflexive prefix to be understood as reflexive in Spanish: \ftx \fln (47) a. Juán lavó el carro. 'John washed the car.' \ftx Juán se-lavó. 'John washed (himself).' \ftx *Juán lavó. \ftx \ftx b. Juán afeitó el tigre. 'John shaved the tiger.' \ftx Juán se-afeitó. 'John shaved (himself).' \ftx *Juán afeito. \ftx \ftx c. Juán vistió al niño. 'John dressed the boy.' \ftx Juán se-vistió. 'John got dressed.' \ftx *Juan vistió. \ftx \txt Russian offers additional examples of morphological reflexives. It also illustrates the fact that 'valence decreasing' is a somewhat unified functional notion and that reflexive is only one interpretation of a possible range of valence decreasing functions that may be encoded by a single form. In Russian, a reflexive is formed by the addition of a suffix '-sja': \ftx \fln (48) NON-REFLEXIVE: \ftx a. Ivan umïvátj djetj-oj `John washes the children.' \ftx John wash child-PL:ACC \ftx \fln REFLEXIVE: \ftx b. Ivan umïvát-sja `John washes (himself).' \ftx John wash-REFL \ftx \dt 27/Apr/1998 \key Hval_decr_refl_other \shd Other Coreference \txt In some languages, especially those that have morphological reflexives, the reflexive/reciprocal morphology also occurs in noun phrases to indicate coreference between the possessor of the noun and an argument of the verb. Eskimo provides a ready example of this phenomena: \ftx \fln (56) EXAMPLE NEEDED \ftx \ftx \fln Yagua provides an example of a related phenomenon: \ftx \fln (57) Suumutyô jïïta naandaanúyu 'Heri son answered heri \ftx sa-jumutyô jïïta naana-daa-nú-yù \ftx 3SG-answer JIITA 3DL-little-person-REFL \ftx \txt In this example, the reflexive marker '-yu' indicates coreference between the object of the clause and the possessor of the subject noun phrase. If the son belongs to some other person, the regular 3rd person object enclitic is used. There is no direct analog to this phenomenon in English. In the sentence: \ftx \fln (58) Her own son answered her. \ftx \txt In this sentence the reflexive marker 'own' goes with the possessor. In Yagua it goes with the object. That is, it would be as if in English we could say: \ftx \fln (59) *Her son answered herself. \ftx \txt This use of reflexive morphology does not decrease the valence of the clause. This is because its primary function is to encode coreference between a possessor and a core clause constituent. It does not reduce the number of core participants in the clause. \ftx \dt 27/Apr/1998 \key Hval_incr \shd Valence increasing operations \txt The following sections will deal with some of the major uses of valence changing operations: \ftx \cf Hval_incr_caus Causatives \cf Hval_incr_appl Applicatives \cf Hval_incr_dat_shift Dative Shift \ftx \dt 27/Apr/1998 \key Hval_incr_appl \shd Applicatives \txt Some languages have operations whereby a verb is marked for the semantic role of a direct object. Such operations are described using various terminology. Here we will refer to them as APPLICATIVES, though you also might hear them called 'advancements' or 'promotions' to direct object. Applicatives are common in Amerindian, Austronesian and Bantu languages. \ftx \txt To continue this discussion, highlight and jump to the topic below: \cf Hval_incr_appl_ex1 \ftx \dt 27/Apr/1998 \key Hval_incr_appl_ex1 \shd Applicatives--Examples I \txt A straightforward example of the occurrence of applicatives comes from Kinyarwanda, a Bantu language of Rwanda: \ftx \ftx Kinyarwanda: \fln (23) a. Umugóre a-ra-kor-er-a umuhuungu igitabo. \ftx woman she-PRES-read-BEN-ASP boy book \ftx 'The woman is reading the boy the book.' \ftx (valence = 3, boy = BENEFACTIVE) \ftx \ftx b. Umwáalimu y-oohere-jé-ho ishuuri igitabo \ftx teacher he-send-ASP-LOC school book \ftx 'The teacher sent the book to the school.' \ftx (valence = 3, school = LOCATIVE) \ftx \txt In example 23a the suffix '-er' indicates that the first object after the verb has the semantic role of BENEFACTIVE. In example 23b, the suffix '-ho' indicates that the first object is a LOCATIVE. There are syntactic tests in Kinyarwanda that show that these elements really are syntactic direct objects of the verb (Kimenyi 1980). The verbal suffix indicates the semantic role of the object. \ftx \txt To continue this discussion, highlight and jump to the topic below: \cf Hval_incr_appl_ex2 \ftx \dt 27/Apr/1998 \key Hval_incr_appl_ex2 \shd Applicatives--Examples II \txt In Yagua, '-ta' indicates that a locative or instrumental participant is in direct object position. \ftx \fln (24) a. Sa-duu rá-viimú 'He blows into it.' \ftx 3SG-blow INAN-into \ftx (valence = 1) \ftx \ftx b. Sa-duu-tá-ra 'He blows it.' \ftx 3SG-blow-TA-INAN:OBJ \ftx (valence = 2) \ftx \ftx c. sî-îchití-rya javanu quiichi-tya \ftx 3SG-poke-INAN:OBJ meat knife-INST \ftx 'He poked the meat with the/a knife'. \ftx \ftx d. sî-îchití-tya-ra quiichiy \ftx 3SG-poke-TA-INAN:OBJ knife \ftx 'He poked (something) with the knife'. \ftx \txt Note that in 24c, the postposition that marks a nominal as having the semantic role of INSTRUMENT is the same form as the applicative verbal suffix ('-tya' and '-rya' are phonologically conditioned allomorphs of '-tá' and '-ra' respectively). \ftx \txt To continue this discussion, highlight and jump to the topic below: \cf Hval_incr_appl_ex3 \ftx \dt 27/Apr/1998 \key Hval_incr_appl_ex3 \shd Applicatives--Examples III \txt In Nomatsiguenga, a Pre-Andine Maipuran Arawakan language of the Eastern Peruvian foothills, a much more complicated system is found. In this language there are at least nine applicative suffixes that express a variety of semantic roles. In example 30 '-te' indicates that the direct object has the semantic role of ablative (a location that is the goal of some locationally directed action). That the form Ariberto is in fact the syntactic direct object of 26 is confirmed by the facts that a) it occurs directly after the verb, and b) it is preceded by the 3SG direct object enclitic '-ri'. The same arguments hold for the direct objects of verbs with the other applicative suffixes illustrated in 27 through 34: \ftx \ftx Nomatsiguenga: \fln (25) Pablo i-niake-ro inato 'Paul saw mother' \ftx Paul he-see-her mother \ftx \fln (26) Pablo i-pë-ri Ariberito kireki 'Paul gave Albert money' \ftx he-give-him Albert money \ftx \fln (27) i-pëna-ben-t-i-na-ro mabesi naro \ftx he-pay-WRT-E-TENSE-me-it pineapple I \ftx 'He paid for the pineapple for me' \ftx \fln (28) i-pë-mi-ri tiapa obiro 'He gave you the chicken' \ftx he-give-you-him chicken you \ftx \fln (29) a. n-areeka siointi-kë 'I arrived at Shointi's' \ftx I-arrive -at \ftx \ftx b. n-areeka-ri siointi 'I arrived at Shointi's'. \ftx I-arrive-him \ftx \fln (30) LOCATIVE-1 \ftx Pablo i-hoka-te-ta-be-ka-ri Ariberito i-gotsirote \ftx he-throw-toward-E-frust-REFL-him Albert his-knife \ftx 'Paul threw his knife toward Albert' \ftx \fln (31) LOCATIVE-2 \ftx Pablo i-kenga-mo-ta-h-i-ri Ariberito \ftx he-narrate-in:presence:of-E-frust-REFL-him Albert \ftx 'Paul narrated it in Albert's presence' \ftx \fln (32) INSTRUMENT \ftx ora pi-nets-an-ti-ma-ri hitatsia negativo \ftx that you-look:at-INST-FUT-FUT:REFL-him name negative \ftx 'Look at it (the sun during an eclipse) with that which \ftx is called a negative'. \ftx \fln (33) ASSOCIATIVE \ftx Juan i-komota-ka-ke-ri Pablo otsegoha \ftx he-dam:stream-ASSOC-PAST-him Paul river:branch \ftx 'John dammed the river branch with Paul'. \ftx \fln (34) PURPOSE \ftx a. Pablo i-ata-si-ke-ri Ariberito. \ftx he-go-PURP-PAST-him Albert \ftx 'Paul went with Albert in mind (e.g., to see him)' \ftx \ftx b. ni-ganta-si-t-ë-ri hompiki \ftx I-send-PURP-E-TENSE-him pills \ftx 'I sent him for pills' \ftx \fln (35) REASON \ftx a. Pablo i-kisa-biri-ke-ri Juan \ftx he-be:angry-REASON-PAST-him John \ftx 'Paul was angry on account of John'. \ftx \ftx b. Pablo i-atage-biri-ke-ri Juan \ftx he-go-REASON-PAST-him \ftx 'John was the reason for Paul's going'. \ftx \fln (36) BENEFACTIVE \ftx Pablo i-pë-ne-ri Ariberito tiapa singi \ftx he-give-BEN-him Albert chicken corn \ftx 'Paul gave the chickens corn for Albert' \ftx \fln (37) INCLUDED (with reference to) \ftx a. Pablo i-samë-ko-ke-ro i-gisere \ftx he-sleep-INC-PAST-it his-comb \ftx 'Paul went to sleep with reference to the comb' \ftx (he was making it and dropped it) \ftx \ftx b. Pabli i-komoto-ko-ke-ri pabati otsegoha \ftx he-dam:stream-INC-PAST-him father river:branch \ftx 'Paul dammed the river branch with reference to father' \ftx (c.f. 35 above) \ftx \fln (38) WITH REFERENCE TO \ftx a. Pablo i-pëna-ben-ta-h-i-ri yaniri kireki \ftx he-pay-WRT-E-FRUST-REFL-him howler:monkey money \ftx 'Paul paid money for the howler monkey' \ftx \ftx b. pi-ngaki-ben-kima yaniri \ftx you-stay:awake-WRT howler:monkey \ftx 'Stay awake with reference to the howler monkey \ftx (because of him)' \ftx \txt To continue this discussion, highlight and jump to the topic below: \cf Hval_incr_appl_vscaus \ftx \dt 27/Apr/1998 \key Hval_incr_appl_vscaus \shd Applicatives versus Causatives \txt In some languages the instrumental applicative can also be construed as a causative. For example, in Kinyarwanda, the causative and the applicative are the same morpheme, 'iis'. The functional basis for this isomorphism is apparent in the following pair of examples (Kimenyi 1980:164): \ftx \fln (39) a. Umugabo a-ra-andik-iis-a umugabo íbárúwa \ftx man 3SG-PRES-write-CAUSE-ASP man letter \ftx `The man is making the man write a letter.' \ftx \ftx b. Umugabo a-ra-andik-iis-a íkárámu íbárúwa \ftx man 3SG-PRES-write-APPL-ASP pen letter \ftx `The man is writing a letter with a pen.' \ftx \txt The only real difference between these two sentences is the animacy of the `causee'. In both cases the controller acts on something to accomplish some action. In 39a the thing he acts on is another human, whereas in 39b the thing he acts on is a pen. Other languages in which the same kind of isomorphism obtains are Yagua (see ex. 24 above), Malay and Djirbal (Croft 1991:242). In many other languages the causative and instrumental applicatives are different morphemes. Nevertheless, the fact that they are often formally similar underscores the conceptual similarity between these apparently distinct functional types. \ftx \dt 27/Apr/1998 \key Hval_incr_caus \shd Causatives \txt Linguists and philosophers have always been interested in causation. Causative constructions (or causatives) are the linguistic instantiations of the conceptual notion of causation. Causatives can be divided into three types: lexical, morphological and paraphrastic/ analytic. The morphological causative is one kind of 'valence increasing' operation. An example of each of the three possible answers to theis question would be: \ftx \ftx a) lexical kill \ftx b) morphological die+cause \ftx c) analytic/periphrastic cause to die \ftx \txt For definitions, highlight and jump to: \cf Hval_incr_caus_def \ftx \txt For a discussion of Transitivity, highlight and jump to: \cf Hval_incr_caus_tr \ftx \txt To continue this discussion, highlight and jump to: \cf Hval_incr_caus_lex \ftx \dt 27/Apr/1998 \key Hval_incr_caus_anal \shd Analytic Causatives \txt Most causatives in English are analytic in that they involve the addition of a separate causative verb, e.g. 'make', 'cause', 'force', 'compel' etc. \ftx \fln (12) He made me do it. \ftx Gloucester caused Lucretia to die. \ftx Williams forced his barber to relinquish his position. \ftx Marie compelled Tyler to dance with her. \ftx \txt Analytic causatives are not normally considered to be valence increasing operations, even though semantically they can be interpreted as such. Rather, in most cases they can be considered to consist of a matrix clause (CAUSE) whose sentential complement refers to the caused event (see Hcplx_cplcl on complement clauses). We will continue to include analytic causatives in this section because of the interesting functional generalizations that can be made across the three causative types. \ftx \txt To continue this discussion, highlight and jump to the topic below: \cf Hval_incr_caus_c/e_1 \ftx \dt 27/Apr/1998 \key Hval_incr_caus_c/e_cod_dist1 \shd Coding principles--Structural distance--I \txt For languages that have more than one formal kind of causative the 'smaller' one (i.e. the one where cause and effect are most closely united formally) will be used for more direct causation, while the 'larger' one (i.e. the one where morphosyntactic size of the causative construction is greater) will be used for less direct causation. \ftx \txt This principle is illustrated by Haiman (1983) in terms of an 'iconicity pyramid'. In this pyramid the pinnacle is the construction in which cause and effect are embodied by a single lexical form. This is what we would term a 'lexical causative'. Morphological and finally analytic causatives are found at lower levels of the pyramid. These are construction types in which the cause is increasingly more distant morphosyntactically from the effect. This increase in morphosyntactic distance is correlated to an increase in conceptual distance at the lower levels of the pyramid. \ftx \fln (13) X (lexical causative) More direct causation \ftx Y+Z (morphological causative) | | | \ftx Y#Z (analytic causative) Less direct causation \ftx \txt Note that this pyramid makes no claims as to the semantics of lexical vs. analytic causatives in two different languages, but only in those languages that have more than one causative to express essentially the same idea. \ftx \txt Longer linguistic distance (according to Haiman's pyramid) is always correlated with greater conceptual distance. For this reason analytic causatives often 'require' an animate causee. The greater conceptual distance implied by the longer analytic causatives conveys that the causer does not have direct physical control over the causee. Rather the causee retains some degree of control over the caused event. Such control is incongruous with an inanimate causee. \ftx \txt In English there is no grammatical constraint that the causee in an analytic causative be animate. However, if the causee is not animate (or for some other reason has no control over the caused event), the analytic causative sounds strange: \ftx \fln (14) I caused the tree to fall. \ftx the chicken to die. \ftx the cup to rise to my lips. \ftx \txt All of these seem to imply magical powers because of the conceptual distance between cause and effect. This is not the case with corresponding lexical causatives. \ftx \fln (15) I felled the tree. \ftx I killed the chicken. \ftx I raised the cup to my lips. \ftx \txt These imply a close connection between cause and effect, e.g. direct physical contact and complete control of the causer over the causee. \ftx \txt Amharic illustrates this principle even within one type of causative. Amharic has two morphological causatives, one signalled by the prefix 'a-' and the other by the prefix 'as-'. The shorter of these is used for direct causation, while the longer one, 'as-', is always used for indirect causation: \ftx \fln (16) a. Abbat lëgun sëga a-bälla \ftx father boy meat CAUSE-eat \ftx 'The father fed the boy the meat.' (direct physical control) \ftx \ftx b. Abbat lëgun sëga as-bälla \ftx father boy meat CAUSE-eat \ftx 'The father forced the boy to eat the meat.' \ftx (indirect control, e.g. by threat) \ftx \fln Korean illustrates this principle in morphological and analytic causatives: \ftx \fln (17) a. ip-hi-ta 'to dress someone' \ftx b. ip-key ha-ta 'persuade to get dressed' \ftx \fln (18) a. ket-I-ta 'force to walk' \ftx b. ket-key ha-ta 'enable to walk' \ftx \txt In each of these examples the first sentence involves direct, physical action on the part of the causer, while the second sentence involves more removed, less direct causation. These examples also illustrate that indirect causation often has additional semantic overtones, e.g. 'enable,' 'permit', 'persuade', 'tell' etc. \ftx \txt Finally, Miztec also illustrates the principle that morphosyntactic distance correlates with conceptual distance in morphological and analytic causatives: \ftx \fln (19) a. s-kée 'Feed him.' \ftx CAUSE-eat(potential) (= put food in his mouth.) \ftx \ftx b. sá:à hà nà kee 'Make him eat.' \ftx CAUSE NOM OPT eat (= prepare food for him to eat.) \ftx \ftx \shd2 Finiteness and Case \ftx \txt The more distant the cause from the effect in time, space or participants, the more finite the verb that expresses the effect will be. \ftx \txt If the causee retains a high degree of control over the caused event, it will appear in a case normally associated with AGENTS, e.g. the nominative or ergative case. If it retains little or no control (i.e. it is completely manipulated by the causer), it will appear in a case normally associated with patients, e.g. the accusative or absolutive case. When the causee retains some degree of freedom of action, it appears in the nominative case in English: \ftx \fln (20) a. I asked that he leave. (request, causee retains right to say no.) \ftx NOM \ftx \ftx b. I asked him to leave. (command, less likelihood that causee \ftx ACC has the option to say no.) \ftx \ftx c. I made him leave. (No control retained by causee. Causee \ftx ACC appears in accusative case, and comple- \ftx ment lacks 'to'. Effect very closely \ftx integrated to the predicate of cause.) \ftx \txt In Hungarian, the causee appears in the accusative case when the causee retains no control over the event (99), but in the instrumental case when it retains some control (99): \ftx \fln Hungarian: \ftx (21) En köhogtettem a gyerek-et. \ftx I caused:to:cough the child-ACC \ftx 'I made the child cough,' (e.g. by slapping him/her on the back) \ftx \ftx (22) En köhogtettem a gyerek-kel. \ftx I caused:to:cough the child-INST \ftx 'I got the child to cough,' (e.g. by asking him/her to do so.) \ftx \fln References: Comrie (1981:ch. 8). Givón (1984). Haiman (1983). \ftx \dt 27/Apr/1998 \key Hval_incr_caus_c/e_cod_fin \shd \key Hval_incr_caus_c/e_1 \shd Cause & Effect \txt There is a relationship between structural integration and conceptual integration between cause and effect. Conceptual integration refers to how integrated or 'close' the cause and effect are in the message world. Structural integration refers to how integrated the element expressing the cause and the element expressing the effect are in the causative construction. Conceptual integration is commonly described in terms of the distinction between DIRECT CAUSATION and INDIRECT CAUSATION. \ftx \txt Direct causation is where the causer is directly, instantly and probably physically responsible for the effect. For example, the verb 'kill' in English is a lexical causative that conveys direct causation, whereas 'cause to die' is an analytic causative that expresses indirect causation. A sentence like 'Jesse killed the gunfighter' is likely to describe a situation in which Jesse is directly and physically responsible for the gunfighter's death. The sentence 'Jesse caused the gunfighter to die', on the other hand, might describe a situation in which the act that resulted in the gunfighter's death is removed physically and/or temporally from the act of his dying, e.g. Jesse may have tampered with the gunfighter's gun, or distracted him during a gunfight. \ftx \txt The relationship between structural integration and conceptual integration between cause and effect is instantiated in at least three different ways in the known languages of the world. \ftx \ftx 1. Structural distance: The amount of phonological bulk \ftx necessary to encode the causative operation is iconically \ftx related to the amount of conceptual distance between the \ftx cause and the effect (Haiman 1983). \ftx \ftx 2. Finite vs non-finite verb forms: If cause and effect are \ftx the same in terms of tense/aspect/modality/evidentiality and \ftx or location, the verb that expresses the effect will not \ftx need to be marked for TAM. \ftx \ftx 3. Morphological case of the causee: If the causee retains a \ftx high degree of control over the caused event, it will appear \ftx in a case normally associated with AGENTS, e.g. the \ftx nominative or ergative case. If it retains little or no \ftx control (i.e. it is completely manipulated by the causer), \ftx it will appear in a case normally associated with patients, \ftx e.g. the accusative or absolutive case. \ftx \txt These three 'coding principles of causatives' will be discussed and exemplified below. \ftx \txt To continue, highlight and jump to: \cf Hval_incr_caus_c/e_cod_dist1 \ftx \dt 27/Apr/1998 \key Hval_incr_caus_def \shd Causatives--Definitions \txt Definition: A causative is a linguistic expression that contains in semantic/logical structure a predicate of cause, one argument of which is a predicate expressing an effect. A causative sentence can be symbolized as: \ftx \ftx CAUSE(x, P) = x causes P \ftx \txt One possible English instantiation of this predicate calculus statement is the following: \ftx \fln CAUSE(Montezuma, Eat(Cortez, possum)) = Montezuma caused Cortez to eat possum. \ftx \fln The definitions of the predicates in a causative sentence are: \ftx \fln PREDICATE OF CAUSE = The predicate that contains the notion of causation, \ftx e.g. CAUSE(x, P). Sometimes the predicate of cause is \ftx referred to as a MATRIX predicate (or matrix clause) \ftx because the predicate of effect is embedded within \ftx the predicate of cause (see Hcplx_cplcl). \ftx \fln PREDICATE OF EFFECT = The predicate that expresses the effect of the \ftx causative situation, e.g. EAT(Cortez, possum). \ftx Sometimes it is said that the predicate of effect \ftx is EMBEDDED in the predicate of cause. \ftx \fln The definitions of the core arguments on a causative sentence are: \ftx \fln CAUSEE = subject of caused event. Sometimes referred to as \ftx the COERCED ENDPOINT (Croft 1991:241). \ftx \fln CAUSER = subject of the predicate of cause and so normally \ftx also of the causative situation. Sometimes referred \ftx to as the AGENT OF CAUSE. \ftx \dt 27/Apr/1998 \key Hval_incr_caus_lex \shd Lexical Causatives \txt There are at least three subtypes of what we will term lexical causatives. The unifying factor behind all of these types is the fact that in each case the notion of cause is wrapped up in the lexical definition of the verb itself. It is not added by any additional operator: \ftx \ftx 1. No change in verb: non-causative: The vase broke. \ftx causative: MacBeth broke the vase. \ftx \ftx 2. Some idiosyncratic change in verb: \ftx non-causative: The tree fell. (verb = 'to fall') \ftx causative: Bunyan felled the tree. (verb = 'to fell') \ftx \ftx 3. Different verb: non-cause: Stephanie ate beans. \ftx cause: Gilligan fed Stephanie beans. \ftx \ftx non-cause: Lucretia died. \ftx cause: Gloucester killed Lucretia. \ftx \ftx see/show, etc. \ftx \fln All of the above are examples of what we will term lexical causatives. \ftx \txt To continue this discussion, highlight and jump to the topic below: \cf Hval_incr_caus_mor \ftx \dt 27/Apr/1998 \key Hval_incr_caus_mor \shd Morphological Causatives \txt Morphological causatives involve a productive change in the form of the verb. The verb 'to fell' in English does not qualify as a morphological causative because it is not derived by a rule that can be applied to many other verbs in the language. \ftx \fln Turkish (Indo-European): \ftx \ftx (3) ol = die / dir = cause / oldür = kill \ftx \txt Most morphological causatives express at least causation and permission. Georgian exhibits one such construction (fr. Comrie 1978:164): \ftx \ftx (4) Mama shvil-s ceril-s a-cer-ineb-s. \ftx father son-DAT letter-ACC PREF-write-CAUSE-3SG \ftx 'Father makes/helps/lets his son write the letter.' \ftx \txt To see some examples, highlight and jump to the topic below: \cf Hval_incr_caus_mor_yup \ftx \txt To continue this discussion, highlight and jump to the topic below: \cf Hval_incr_caus_anal \ftx \dt 27/Apr/1998 \key Hval_incr_caus_mor_pun \shd Morphological Causatives--Punjabi \txt The Panjabi morphological causative borders on being lexical because the rule that derives a causative verb from a non-causative is very opaque (i.e. hard to formulate explicitly): \ftx \fln Panjabi (Indo-European): \ftx \fln (11) NON-CAUSATIVE: k'ad \ftx a. o-ne k'aNa k'ad-a 'He ate food.' \ftx 3SG-ERG food eat-PAST:SGM \ftx \ftx CAUSATIVE: k'lay \ftx b. timi-ne o-nu k'aNa k'lay-a \ftx woman-ERG 3SG-DAT food eat:CAUSE-PAST:3SGM \ftx 'The woman made him eat food.' \ftx \ftx NON-CAUSATIVE: dore \ftx c. Ram dore-a 'Ram ran.' \ftx Ram run-PAST:SGM \ftx \ftx CAUSATIVE: dêray \ftx d. munci-ne Ram-nu dGray-a \ftx teacher-ERG Ram-DAT run:CAUSE-PAST:3SG \ftx 'The teacher made Ram run.' \ftx \txt There are no more examples. To continue this discussion, close each jump window until you have returned to the following record: \cf Hval_incr_caus_mor \ftx \dt 27/Apr/1998 \key Hval_incr_caus_mor_que \shd Morphological Causatives--Quechua \txt Quechua is another language that allows morphological causatives of inherently transitive stems (ex. 99). However, even in Quechua it is more common to use a periphrastic causative when the caused event is transitive: \ftx \fln INTRANSITIVE ROOT (sleep): \ftx \fln (9) a. noqa puñu-: 'I sleep.' (Non causative) \ftx 1SG sleep-1SG \ftx \ftx b. noqa-ta puñu-chi-ma-n 'It makes me sleep.' \ftx 1SG-ACC sleep-CAUSE-1SG-3SG \ftx \fln TRANSITIVE ROOT (hit): \ftx \fln (10) a. Qam noqa-ta maqa-ma-nki \ftx 2SG 1SG-ACC hit-1SG-2SG \ftx `You hit me.' (non-causative) \ftx \ftx b. Pay qam-wan noqa-ta maqa-chi-ma-n \ftx 3SG 2SG-COM 1SG-ACC hit-CAUSE-1SG-3SG \ftx `He makes you hit me.' \ftx \txt Example 10b illustrates a very common pattern with morphological causatives of lexically transitive verbs: the CAUSEE goes into an oblique case, while the P of the verb root remains in the accusative case. Another possibility is for the causative of a transitive clause to allow two accusatives. \ftx \txt To continue with an example from Punjabi, highlight and jump to: \cf Hval_incr_caus_mor_pun \ftx \dt 27/Apr/1998 \key Hval_incr_caus_mor_yup \shd Morphological Causatives--Eskimo \txt Many morphological causatives are restricted to use with inherently intransitive stems. The following examples from Yup'ik Eskimo illustrate a typical range of functions often associated with morphological `causative' operators that are restricted to intransitive roots (Reed, et. al 1977:177): \ftx \ftx Root Stem \ftx \ftx tuqu- `die' tuqute- `kill' \ftx tai- `come' taite- `bring' \ftx uita- `stay' uitate- `let stay / leave alone' \ftx tatame- `be startled' tamate- `startle' \ftx ane- `go out' ante- `put outside' \ftx itr- `go in' iterte- `put in/insert' \ftx atrar- `go down' atrarte- `take down' \ftx mayur- `go up' mayurte- `put up' \ftx \txt However, Yup'ik also has other causative operators that function with transitive or intranstive roots: \ftx \fln (5) INTRANSITIVE ROOT (go up): \ftx Qetunra-ni tage-vkar-aa \ftx son-ABS:POSS go:up-CAUSE-3SG>3SG \ftx 'He makes/lets his own son go up.' \ftx \fln (6) TRANSITIVE ROOT (eat): \ftx Arnam irnia-mi-nun neqerrlu-ut nere-vkar-ai. \ftx woman-ERG child-POSS-OBL dryfish-ABS:PL eat-CAUSE-3SG>3PL \ftx 'The woman makes/lets her child eat the dryfish.' \ftx \fln (7) INTRANSITIVE ROOT (go): \ftx Ayag-cess-gu. \ftx go-CAUSE-IMP:SG>3SG \ftx `Make/let him go.' \ftx \fln (8) TRANSITIVE ROOT (dry) \ftx Nukalpia-m aana-mi-nun kenir-cet-aa kemek \ftx young:man-ERG mother-POSS-OBL dry-CAUSE-3SG>3SG meat:ABS \ftx `The young man made/let his own mother dry the meat.' \ftx \txt To see an example from Quechua, highlight and jump to the topic below: \cf Hval_incr_caus_mor_que \ftx \dt 27/Apr/1998 \key Hval_incr_caus_tr \shd Causatives--Transitivity \txt Causative sentences can be constructed on the basis of intransitive or transitive caused events. Causative predicates always involve one more argument than the caused predicate. Hence if the caused event is intransitive, the causative is transitive. If the caused event is transitive, the causative is di-transitive, etc. For example: \ftx \ftx Intransitive caused event: Cortez made Montezuma laugh. \ftx \ftx Transitive caused event: Montezuma made Cortez eat possum. \ftx \txt Other concepts that add one participant to a scene are also sometimes encoded by a valence increasing operation. Sometimes these operations are identical to the operation, or one of the operations, used to encode causatives. For example: \ftx \ftx Believe: Montezuma believes Cortez ate possum. BELIEVE(m, P) \ftx Say: " says " " " SAY(m, P) \ftx -------------------------------------------------- \ftx Want: " wants " to eat possum. WANT(m, P) \ftx Ask: " asked " " " " ASK(m, P) \ftx Permission: " let " eat possum. LET(m, P) \ftx (failure to prevent) \ftx \txt The last three often use exactly the same morphosyntax as 'pure' causatives, especially when they are expressed morphologically. \ftx \txt Causatives can be expressed lexically, morphologically or analytically/periphrastically. \ftx \dt 27/Apr/1998 \key Hval_incr_dat_int \shd Dative of interest \txt Some languages allow a participant that is associated with the event in some grammatically unspecified way to be referred to with a dative pronoun. Spanish is one well-known example: \ftx \fln (41) Se me quemó la cena. \ftx REFL 1SG burn:3SG:PAST DEF:FEM:SG dinner \ftx 'Dinner burned on me.' (valence = 2) \ftx \txt This sentence might be translated 'dinner burned with respect to me,' or 'dinner burned for me'. With transitive verbs, the dative of interest can indicate that the participant referred to with a dative pronoun is the possessor of the direct object: \ftx \fln (42) Le cortó el pelo. 'She cut hair (with respect to/on/for) him.' \ftx (i.e. 'She cut his hair.') \ftx \txt This last construction is sometimes called POSSESSOR ASCENSION. However, this terminology (from Relational Grammar) assumes that the dative participant is at some deep level a syntactic possessor of the direct object, as in the English translation equivalent. However, there is no particular reason to make this assumption, especially in light of the fact that Spanish has a fully productive 'dative of interest' construction type. \ftx \txt Dative of interest constructions are distinct from applicatives and dative-shift constructions in that the argument that is added to the proposition is instantiated as a 'dative' participant, i.e. as the third argument in a tri-valent construction. With applicatives and dative-shift constructions, the additional argument appears as a direct object. \ftx \txt To continue this discussion, highlight and jump to: \cf Hval_incr_dat_shift \dt 27/Apr/1998 \key Hval_incr_dat_shift \shd Dative shift and other valence increasing operations. \txt Many languages have two alternative morphosyntactic means of expressing a tri-valent proposition. Tri-valent propositions normally involve an AGENT, a PATIENT (usually an item conveyed from one person to another), and a RECIPIENT. Some verbs that commonly embody tri-valent propositions are; 'show', 'give', 'send'. For each of these verbs the RECIPIENT (or perhaps EXPERIENCER in the case of 'show') occurs sometimes in the dative case, marked by the preposition 'to', and sometimes with no case marker. The construction in which the RECIPIENT does not take a preposition is termed a dative shift construction: \ftx \fln (40) Normal: a. Prudence gave her greatcoat to the curator. \ftx Dative shift: b. Prudence gave the curator her greatcoat. \ftx \txt We consider dative-shift to be a valence increasing operation because it is a means of bringing participants with peripheral semantic roles, e.g. RECIPIENT and BENEFACTIVE, onto 'center-stage' in addition to whatever participants may already be on stage. If there are other non-subject arguments in the sentence, they acquire status as the 'second object'. In this position they may or may not retain all morphosyntactic properties of direct objects. \ftx \txt There are two rather subtle differences between applicative and dative-shift constructions. These are: 1) applicatives involve some marking on the verb whereas dative-shift constructions do not, and 2) applicative constructions normally allow INSTRUMENTS to become direct objects, whereas dative-shift constructions typically allow only RECIPIENTs and BENEFACTIVEs. If a language allows all of these semantic roles to become direct object by a productive operation, there may be no reason to distinguish applicative from dative-shift. This appears to be the case in Nomatsiguenga, cited above. \ftx \txt To continue this discussion, highlight and jump to the topic below: \cf Hval_incr_dat_int \ftx \dt 27/Apr/1998 \key Hval_p-calc \shd Valence and predicate calculus \txt VALENCE can be thought of as a semantic notion, a syntactic notion or a combination of the two. SEMANTIC VALENCE refers to the number of participants that are 'on stage' (see Qintro_embod_play_A) in the scene embodied by the verb. For example, the verb 'eat' in English has a semantic valence of two, since for any given event of eating there must be at least an eater and an eaten thing. In terms of predicate calculus, the concept EAT is a relation between two variables, x and y, where x is a thing that eats and y is a thing that undergoes eating. This semantic relationship would be represented in predicate calculus notation as EAT(x,y) (see below). \ftx \txt GRAMMATICAL VALENCE (or SYNTACTIC VALENCE) refers to the number of arguments present in any given clause. An ARGUMENT of a verb is a nominal element (including possibly zero, if this is a referential operator in the language) that bears a direct grammatical relation to the verb (see chapter #7). So, for example, a given instance of the verb eat in English may have a syntactic valence of 1 or 2. In a sentence like 'Have you eaten yet?' there is no direct object, hence the only argument of the verb is the eater. Similarly, in 'she ate away at the bone' there is only one argument of the verb, since 'bone' is an oblique adjunct, i.e. it does not bear a direct grammatical relation to the verb. \ftx \txt To continue this discussion, highlight and jump to the topic below: \cf Hval_p-calc_zero \ftx \dt 27/Apr/1998 \key Hval_p-calc_as \shd Valence as predicate calculus \txt It may be useful to think of valence in terms of PREDICATE CALCULUS. At various points in the discussion of valence increasing and decreasing operations, concepts and notation from predicate calculus will be employed. Also, it is common to find this notation used in the linguistic literature. It is a good idea for anyone involved in linguistic analysis to be comfortable with these concepts -- it is essential for anyone planning to undertake a graduate program in linguistics. \ftx \fln TERM = designation for a thing, as opposed to a property of \ftx things or a relation between things (see variable and \ftx constant below). \ftx \fln VARIABLE = a term that has no specific reference, e.g. 'x', \ftx 'y', etc. \ftx \fln CONSTANT = a term that does have specific reference in the \ftx message world, e.g. 'Yankee stadium', 'Socrates', \ftx 'Montezuma', 'the unicorn in my garden', 'the number 3' \ftx etc. \ftx \fln ARGUMENT = a place within a function to be filled by a term \ftx or a proposition. \ftx \fln PREDICATE/FUNCTION = a property that can be applied to a term or a \ftx relation between terms, e.g. 'be human', 'be mortal', \ftx 'die', 'eat', 'laugh' etc. \ftx \fln PROPOSITIONAL FUNCTION = A predicate applied to a variable or set of \ftx variables, e.g. 'x is human', 'some unspecified \ftx thing ate some unspecified thing', 'some unspecified \ftx person laughed', etc. \ftx \fln PROPOSITION = a predicate applied to a constant or set of \ftx constants, e.g. 'Socrates is human', 'Montezuma laughed', \ftx 'Bill kissed Mary', etc. \ftx \txt f(x) is pronounced 'function of x'. This notation refers to some property of items that might instantiate the variable x. Another way of thinking of this use of the term 'function' is that it refers to a relation between x and itself at various stages (if one prefers to think of all functions as 'mappings' from one item to another). \ftx \txt To continue this discussion, highlight and jump to the topic below: \cf Hval_p-calc_as_ex \ftx \dt 27/Apr/1998 \key Hval_p-calc_as_ex \shd Valence as predicate calculus--Examples \txt Usually capital letters designate predicates, e.g. LAUGH is a predicate. It refers to the particular property of variables (or relation between various stages of a single variable) designated by the English lexeme 'laugh'. P(x) (pronounced 'predicate of x') is a kind of f(x), where P is a predicate. P(x) is a propositional function (rather than a proposition) because it is applied to a variable 'x' rather than a constant. LAUGH(x) is one of the many possible functions (properties) of x. \ftx \fln LAUGH(Montezuma) is a proposition. It is one possible instantiation of the \ftx propositional function LAUGH(x). \ftx \txt 'Montezuma laughed' is one English instantiation of the proposition LAUGH(Montezuma). \ftx \txt So far we have been concerned with 'one place' functions. Often functions have more than one argument. \ftx \txt f(x, y) is pronounced 'a function from x to y'. Sometimes this can be restated as 'a relation between x and y'. \ftx \fln EAT(x, y) is one possible relation between two entities x and y. By \ftx convention we normally express the initiating argument first. \ftx We will call this convention the LINEARITY CONVENTION. \ftx So EAT(x, y) can be thought of as referring to the relationship \ftx of 'eating' that holds between some as yet unspecified eater (x) \ftx and some as yet unspecified eaten thing (y). \ftx \fln EAT(Cortez, possum) is one proposition that instantiates the propositional \ftx function EAT(x, y). \ftx \txt 'Cortez ate possum' is one English instantiation of the proposition EAT(Cortez, Possum). ('The possum ate Cortez' is not an instantiation of this propositional function because of the linearity convention). \ftx \txt Propositions can also be arguments of functions. This is why argument and term are not the same thing. Let capital letters (P, Q, R etc.) be abbreviations for propositions. These can fill argument slots in other propositional functions, e.g.: \ftx \ftx If P = LAUGH(Montezuma), then TRUE(P) = TRUE(LAUGH(Montezuma)) = \ftx the English expression 'it is true that Montezuma laughed'. \ftx \txt The following illustrates a two-place proposition filling an argument slot of another propositional function: \ftx \ftx P = KISS(x, y) = Xavier kisses Yolanda. \ftx \ftx P' = WANT(x, P) = Xavier wants to kiss Yolanda. \ftx \ftx Alternatively: WANT(Xavier, (KISS(Xavier, Yolanda))) \ftx \dt 27/Apr/1998 \key Hval_p-calc_ex \shd Valence and predicate calculus--Examples \ftx \key Hval_p-calc_p-drop \shd Valence and predicate calculus--Pro-drop languages \txt In many languages zero anaphora is much more prevalent than it is in English. In such languages (sometimes called pro-drop languages) it may be difficult to distinguish constructions with reduced syntactic valence from those with zero pronouns. In the extreme case of languages with no morphological means of coding grammatical relations, and few restrictions on zero anaphora, the only way to decide is to examine the discourse context. But then, for such languages (e.g. Chinese, Thai), it is largely a moot point whether a particular construction constitutes reduced valence or not. The concept of syntactic valence is valuable insofar as it leads to an understanding of alternative arrangements of grammatical relations (e.g. alternative case marking patterns, verbal affixation or particles). If the language provides no such alternatives, then syntactic valence is not much of an issue. \ftx \txt To continue this discussion, highlight and jump to the topic below: \cf Hval_p-calc_tr \ftx \dt 27/Apr/1998 \key Hval_p-calc_role \shd Valence and predicate calculus--Participant roles \txt As always, in this discussion we will consider semantic properties to be properties of the conceptual representation of things and events in the message world, and syntactic properties to be properties of linguistic elements in sentences. So for us the semantic valence of a verb V refers to the number of core participants in the cognitive model embodied by V. Syntactic valence, then, is the number of verbal arguments in a clause. Languages typically have various ways of adjusting, i.e. increasing or decreasing, the syntactic valence of clauses. The semantic (i.e. conceptual) effect of increasing syntactic valence can be characterized most generally as upgrading a peripheral participant to center stage, whereas the effect of decreasing valence is to downgrade a normally center stage participant to peripheral status. Furthermore, the participants brought onto or taken off of center stage can be CONTROLLERS, i.e. agents or agent-like participants, AFFECTED or patient-like participants or they may be recipients or benefactees. Thus we can identify a typology of valence changing operations as follows: \ftx \shd2 Valence increasing devices: \ftx Those that add a controlling \ftx argument: CAUSATIVES \ftx \ftx Those that add an affected \ftx argument: APPLICATIVES \ftx \ftx Those that add a \ftx recipient/benefactee: DATIVE SHIFT \ftx \shd2 Valence decreasing devices: \ftx Those that subtract a controlling SUBJECT OMISSION \ftx argument: PASSIVES \ftx \ftx Those that subtract an affected OBJECT OMISSION \ftx argument: ANTIPASSIVES \ftx OBJECT DEMOTION \ftx \ftx Those that merge controlling and \ftx affected arguments: REFLEXIVES \ftx RECIPROCALS \ftx MIDDLES \ftx \txt The following subsections systematically describe and exemplify valence-related devices according to this typology: \cf Hval_incr \cf Hval_decr \ftx \txt To continue this discussion, highlight and jump to the topic below: \cf Hval_p-calc_src \ftx \dt 27/Apr/1998 \key Hval_p-calc_src \shd Valence and predicate calculus--Sources \txt Valence changing operators tend to derive from free verb roots that, at an earlier stage of the language, formed analytic constructions. Occasionally, however, valence changing operators derive from inflectional operators. Morphological reflexives are one example of valence changing operators that often are best categorized as inflectional morphology (i.e. languages often have a reflexive operator that enters into the verb coding paradigm for person and number - see Hval_decr_recip_prot). Furthermore, there is a distinct tendency for passive voice and perfect aspect markers to be related synchronically and/or etymologically. This isomorphism is well-motivated by the cognitive notion of viewpoint (Delancey, 1982, 1990). Both passive and perfect aspect clauses share the functional domain of 'endpoint orientation'. For passives the endpoint is the resultant state of an action directed from an AGENT to a PATIENT. For perfect aspect, the orientation is likewise a state that has resulted from some previous action, though the orientation is not necessarily on the effect of that action upon some individuated PATIENT (see Hv_tam_asp_perf). For this reason passive and perfect aspect morphemes are sometimes isomorphic. \ftx \dt 27/Apr/1998 \key Hval_p-calc_tr \shd Valence and predicate calculus--Transitivity \txt The notion of valence is closely aligned with the traditional idea of TRANSITIVITY, i.e. a TRANSITIVE verb is one that describes a relation between two participants such that one of the participants acts towards or upon the other. An INTRANSITIVE verb is one that describes a property, state or situation involving only one participant. Sometimes intransitive verbs, e.g. 'run' in 'I run', are called 'uni-valent', i.e. they have a semantic valence of one. Similarly, transitive verbs such as kill in 'he killed a bear' are called 'di-valent'. Bi-transitive (or tri-valent) verbs are those that have three core participants, e.g. 'give' in 'He gave Mary a book.' Although recent studies (notably Hopper and Thompson 1980) have taken the term 'transitivity' to mean the degree to which an event 'carries over' from an active, volitional AGENT to a PATIENT, still it is common to find the term used in the traditional way. \ftx \txt Unfortunately, in the past linguists have not always been careful to distinguish semantic transitivity from grammatical transitivity. So, for example, there are some who would say 'eat' is always a transitive verb. These linguists use the term transitive in the sense we use the term 'semantically transitive'. Others would say 'eat' is sometimes transitive and sometimes intransitive. These linguists are most likely referring to syntactic transitivity. Still others would say that there are two related verbs 'eat' in the lexicon of English, one of which is transitive and the other intransitive. \ftx \txt To continue this discussion, highlight and jump to the topic below: \cf Hval_p-calc_role \ftx \dt 27/Apr/1998 \key Hval_p-calc_zero \shd Valence and predicate calculus--types of zero \txt Note that there is an important difference between the omission of a verbal argument in a sentence like 'John already ate' and a zero pronoun as in 'John came in and 0 sat down.' In the first, the lack of a direct object is due to the unimportance of the identity of the eaten thing. This assertion is easily demonstrated by observation of English speakers in action -- in the overwhelming majority of instances when a verb with a semantic valence of two occurs with no reference to the second argument, the situation is one in which the identity of the item which fills that second argument role has not been established, and need not be established in order for the speaker to achieve his or her communicative goal. On the other hand, the 'zero-pronoun' in the example 'John came in and 0 sat down' functions in exactly the opposite kind of situation, namely when the identity of the referent is so well and recently established that confusion with some other entity is impossible. One would hardly ask the question 'who sat down?', or even entertain the possibility that it was anyone other than 'John' after someone utters the above sentence. On the other hand, one could very naturally ask the question 'what did he eat?' after someone says 'John already ate.' \ftx \txt To continue this discussion, highlight and jump to the topic below: \cf Hval_p-calc_p-drop \ftx \dt 27/Apr/1998 \key Hv_agr \shd Agreement \txt [NOTE: If this is a particularly complex area in the language, you may want to just describe the verb agreement paradigm for main, declarative, perfective aspect clauses here and provide pointers to where the other paradigms are described. Then be sure to describe the other paradigms in the sections that pertain to their use, e.g. describe the paradigm for subjunctive mode in the section on mode, and the paradigm for negatives in the section on negatives, etc.] \ftx \txt VERB AGREEMENT or VERB CODING (sometimes called CONCORD, especially among linguists who specialize in African languages) is where certain features of verbal arguments are obligatorily represented on the verb, regardless of whether those arguments are also referred to by free noun phrases or pronouns. Arguments represented by verb agreement are said to have a GRAMMATICAL RELATION to the verb (see Hgrel_val). Sometimes verb agreement is an ANAPHORIC device (e.g. Spanish), and sometimes not (e.g. English). We say that verb agreement is anaphoric in Spanish because it can constitute the only reference to an entity in the clause (see above and Hord_vp). For example, the word 'hablo' is a fully grammatical sentence in Spanish meaning 'I speak'. The '-o' suffix in itself constitutes an adequate reference to the subject argument. Hence we say that '-o' (and verb agreement generally in Spanish) is anaphoric. Sometimes languages which have anaphoric verb agreement are called PRO-DROP languages. \ftx \txt In English, on the other hand, a verb form like 'am' is not a fully grammatical sentence even though it does 'agree with' a first person singular subject. In English, verb agreement must be accompanied by a free form reference to the subject participant, e.g. 'I am'. Therefore verb agreement in English is non-anaphoric. Sometimes non-anaphoric verb agreement is referred to as GRAMMATICAL AGREEMENT. \ftx \txt Verb agreement (whether anaphoric or not) can be indicated by any of the morphological processes mentioned in Hmor_mproc, e.g. prefixing, suffixing, stem changes, etc. Verb agreement, both anaphoric and grammatical, almost always arises from a diachronic process extending from free pronouns, through anaphoric clitics, to verb agreement. For this reason, agreement markers are often similar in form to free pronouns. \ftx \dt 27/Apr/1998 \key Hv_cmp \shd Compounding \txt This topic contains the following subsections. Highlight and jump to the topic of interest: \cf Hv_cmp_n Noun Incorporation \cf Hv_cmp_v Complex Verbs \dt 27/Apr/1998 \key Hv_cmp_n \shd Compounding--noun incorporation \txt NOUN INCORPORATION is where a core noun phrase (subject or object) of a clause becomes 'attached' to the verb. Incorporation may exhibit all the characteristics of compounding discussed in Hn_cmp, namely: \ftx \ftx 1) a stress pattern characteristic of words rather than phrases \ftx 2) unusual word order and \ftx 3) meanings that are more specific than the meanings of the individual \ftx parts. \ftx \txt The syntax and semantics of object incorporation are discussed in: \cf Hval_decr_ncorp \ftx \dt 27/Apr/1998 \key Hv_cmp_v \shd Compounding--complex verbs \txt Sometimes verb roots combine to form more complex stems. Verbs of motion commonly enter into such compounds. For example, in Yagua the verb 'jasúmiy' means 'to go up'. This verb can be compounded with just about any other verb in the language to indicate action accomplished in a rising direction: \ftx \fln (23) Sa-súúy-asúmiy 'He shouts rising.' \ftx 3SG-shout-rise \ftx \txt Verbs that freely enter into such compounds often lose their verbal character and ultimately become derivational suffixes. For example, in Quechua at the time of the Inca empire there was a complement taking verb '-kacha' that meant `to go about doing'. In modern Huallaga Quechua (Weber 1989:150) there is a verb suffix '-ykacha' that means `iterative' or `vacilating aimlessly'. The 'y' component is the old infinitive marker that would have appeared on the complement verb: \ftx \fln (24) kuyu-ykacha:-chi-shun \ftx move-ITER-CAUSE-3PL:IMPER \ftx `Let's make it move (back and forth).' \ftx \txt To continue this discussion, highlight and jump to the topic below: \cf Hv_cmp_v_cmp/ser \ftx \dt 27/Apr/1998 \key Hv_cmp_v_cmp/ser \shd Compounding--compound versus serial verbs \txt At times it can be difficult or even impossible to objectively distinguish compound verbs from serial verbs (see Hcplx_sv), verbs plus a complement (Hcplx_cplcl), or verbs with derivational suffixes. In general the following rules of thumb may be of some use in selecting appropriate terminology: \ftx \ftx 1) If one of the roots is not a contemporary verb in its own \ftx right, then you have a verb plus a derivational affix. \ftx \ftx 2) If both roots can function as independent verbs, and if \ftx anything can come in between the two roots (e.g., \ftx inflectional morphology or object nominals) then you have \ftx serial verbs or complementation. See Hcplx_sv and \ftx Hcplx_cplcl for further information. \ftx \ftx 3) If the form of one of the roots is substantially distinct \ftx from its form as a independent verb, AND if the meaning of \ftx the verb is 'bleached', i.e. less specific than the meaning \ftx of the same root as an independent form, then you may \ftx consider calling it a derivational affix. \ftx \txt These rules of thumb would identify the Yagua example to be verb-verb compounding, since 'jasúmiy' is still a viable verb in the language; but would identify Quechua '-ykacha' to be a derivational suffix since '-kacha' no longer functions as an independent verb. \ftx \txt Of course, since some derivational affixes arise by continuous diachronic development from verbs, there is no absolute dividing line between compounding and affixation. However, in most cases the fieldworker will be able to make a reasonable judgement based on the above suggestions. \ftx \fln References: Sapir (1911), Green (1981), Mithun (1984, 1986), Sadock (1986). \ftx \dt 27/Apr/1998 \key Hv_evid \shd Evidentiality--Introduction \txt EVIDENTIALITY has to do with how languages convey relative certainty of truth. It has been called 'the linguistic coding of epistemology' (Chafe and Nichols 1986). EPISTEMOLOGY is the science of knowing. There are some things people are sure of, either because they have reliable evidence for them, or because they have unquestioning faith that they are true. There are other things people are less sure of, and some things they think are only within the realm of possibility. Languages typically provide morphosyntactic devices for expressing these various attitudes toward knowledge. Often, speakers present things as unquestionably true, e.g., 'it's raining.' Sometimes speakers of English use an adverb to express something about the reliability of what they say, or the probability of its truth: 'It's probably raining', or 'Maybe it's raining.' Inference from indirect evidence may be expressed with a modal auxiliary: 'It must be raining.' Or the specific kind of evidence on which an inference is based may be indicated with a separate verb: 'It sounds like it's raining.' The perception that a piece of information does not match the prototypical meaning of a verbal expression may be expressed formulaically: 'It's sort of raining.' Or an adverb may suggest that some information is different from what was expected: 'Actually, it's raining' (Chafe and Nichols 1986). Other languages may make evidential distinctions in the verb morphology. In such languages the evidential system is almost always linked to the tense/aspect/mode (TAM) system. \ftx \txt To continue this discussion, highlight and jump to the topic below: \cf Hv_evid_vsval \ftx \dt 27/Apr/1998 \key Hv_evid_ex1 \shd Evidentiality--Examples I \txt The most common type of evidential marker in language seems to be a 'hearsay' particle. Yup'ik Eskimo provides a straightforward example (Reed, et al 1976): \ftx \fln (53) Tua-llu-gguq nunaa-t uku-t uita-lri-it \ftx Then-and-HSY village-ABS DEM-ABS be-PAST-3 \ftx 'And then there was this village, they say.' \ftx \txt Without '-gguq' this sentence would imply that the speaker has direct experience with the village described. Yup'ik also possesses an 'inferential' evidential enclitic that contrasts with '-gguq'-- 'hearsay' and '-0'-- 'direct'. The inferential enclitic is '-ggem': \ftx \fln (54) Ak'a-ggem ayag-llru-uq \ftx already-INFER leave-PAST-3 \ftx 'It seems he already left.' \ftx \txt Example 54 would be used in a situation where the speaker did not personally see the person leave, nor was told about his leaving by someone else. Instead the speaker infers that the subject has left, e.g. from noticing that he is no longer present. \ftx \txt In Eskimo the evidential particles are 'second position enclitics' (see Hgr_n_prn/clt_def_clt_ex). It is often the case that evidentials and validationals operate on the sentence, rather than verb (phrase) level. \ftx \ftx \txt Huallaga Quechua has three enclitics that are clearly evidential. These enclitics follow the clausal element that conveys new, or asserted, information (Weber 1986:419ff). It stands to reason that the evidentials should be associated with the new information in a clause, since one is more likely to question the source and status of new information than of given information. These enclitics are '-mi'-- 'direct evidence', '-shi'-- 'hearsay' and '-chi'-- 'inference': \ftx \fln (55) Qam-pis maqa-ma-shka-nki a. -mi \ftx you-also hit-1-PERF-2 b. -shi \ftx c. -chi \ftx `You also hit me.' \ftx \ftx a. I saw/felt you hit me and I was conscious. \ftx b. I was drunk, and someone informed me that you hit me.' \ftx c. A group of people beat me up, and I think you might have been \ftx one of them. \ftx \txt In the future or other irrealis contexts, the evidentials express more validational, modal or rhetorical force: \ftx \fln (56) Noqa a. -mi chaya -:-man aywa-r-qa \ftx b. -shi \ftx c. -chi \ftx 1SG arrive-1-COND go-ADV-TOP \ftx `I would (-mi)/could (-shi)/might (-chi) arrive, if I were to go.' \ftx \txt To continue this discussion, highlight and jump to the topic below: \cf Hv_evid_modl \ftx \dt 27/Apr/1998 \key Hv_evid_modl \shd Evidentiality--Modal versus Validational marking \txt The difference between modal and validational marking is sometimes subtle. As usual, there is no particular reason to expect that a language will have an operation that is distinctly one or the other, though it might. Here we will try to provide a means of distinguishing the two. Validational marking has to do with how firmly the speaker believes a statement to be true. Modal operations have to do with how likely a proposition is to be realized. In English we use matrix verbs to indicate both validational and modal concepts, e.g. 'I think', 'I believe', 'I know' are validational whereas 'I must', 'I should', 'I might' are modal. Other matrix verbs are evidential, e.g. 'they say,' 'it seems' and 'I see'. Like TAM marking, evidential and validational marking can be coded as verbal affixes, clause combining (as in English) or periphrastically. \ftx \txt Panare illustrates the interaction of evidentiality with tense and aspect. The immediate past perfective suffix, '-yaj', tends to convey first hand evidentiality. In this respect it contrasts with the non-specific aspect markers: \ftx \fln (57) a. Tiyaj kën Kamánapana \ftx të-yaj kën kamána-pana \ftx go-PPERF1 AN:INVIS Camana-toward \ftx 'He left for Camana (and I saw him go)'. \ftx \ftx b. Yutën kën Kamánayaka \ftx y-u-të-n kën kamána-yaka \ftx 3-INTRNS-go-NON:SPEC AN:INVIS Camana-to \ftx 'He went to Camana (at some unspecified time; \ftx I may not know when because I didn't see him go.') \ftx \txt Also, there are two perfect aspect suffixes. One of these, '-sa'', typically conveys first hand evidentiality whereas the other, '-jpë', conveys inferential evidentiality: \ftx \fln (58) a. wëtësa' këj kën 'He has left (I saw him go.' \ftx w-të-sa' këj kën \ftx INTRNS-go-PERF1 AN:PROX AN:INVIS \ftx \ftx b. wëtëjpë këj kën 'He must have left (e.g. because \ftx w-të-jpë këj kën all his clothes are gone).' \ftx INTRNS-go-PERF2 AN:PROX AN:INVIS \ftx \dt 27/Apr/1998 \key Hv_evid_vsval \shd Evidentiality versus Validation \txt Some linguists (e.g. Weber 1986) make a distinction between EVIDENTIAL FORCE and VALIDATIONAL FORCE. In this view evidential marking is strictly limited to indicating the source of the information contained in the clause, whereas validational marking indicates the degree of commitment the speaker makes as to the truth of the assertion. Naturally, these two parameters are very similar, since one is likely to be strongly convinced of the truth of information gained from direct experience and less convinced of the truth of information gained indirectly. Nevertheless, Weber shows that, at least for Quechua, even if the speaker is absolutely convinced of the truth of a proposition he has not directly witnessed, he may not use the direct evidence evidential. For example, in Quechua I cannot without hedging say 'my mother's Grandfather's name was Henderson' unless I have personally met my mother's grandfather. It should be clear also that evidentiality is closely tied to tense, aspect and mode. We are more likely to be sure of past events than future events, the completion of perfective events than of events still in process, realis assertions than irrealis assertions. As with location and direction marking, evidentiality and validationality are often difficult to tease apart from the TAM system, and there may be complex diachronic and synchronic relationships. \ftx \txt To continue this discussion, highlight and jump to the topic below: \cf Hv_evid_ex1 \ftx \dt 27/Apr/1998 \key Hv_intro \shd Other verb and verb-phrase operations-- Introduction \txt In this chapter we will discuss a collection of major operations likely to be grammaticalized in verbs or verb phrases, but not covered in other chapters. The first three, nominalization, deverbalization and compounding are most likely to be derivational whereas the other five 1) tense/aspect/mode (TAM), 2) evidentiality, 3) location/direction, 4) nondeclaratives and 5) negation are much more likely to be inflectional. Many of these operations are likely to be indistinct from each other in any given language. However, because there is a long tradition of describing them separately, and because they are, at least to the Indo-European mind, logically distinct, it will be convenient to discuss them individually. It should be kept in mind, however, that in most cases there is significant semantic and morphosyntactic overlap among these groups. \ftx \dt 27/Apr/1998 \key Hv_loc \shd Location/direction \txt Even as tense 'grounds' a situation in time, location and directional marking ground situations in space. Spatial grounding has not been given as much prominence as tense in the literature because Greek and Latin did not have verbal operations expressing spatial grounding. However, lots of other languages do grammaticalize spatial grounding. In fact for some languages, spatial deixis is more central to the verbal system than temporal deixis. Directional formatives are often related etymologically to the basic verbs of motion ('go', 'come', and perhaps 'arrive', 'return', 'depart', 'go up' and 'go down'). Some languages have only one basic verb of motion, and rely on directional formatives to distinguish whether the motion is away from a point of reference ('go') or towards a point of reference ('come'). This is true of Lahu and many other Tibeto-Burman languages (Matisoff 1973). \ftx \txt To continue this discussion, highlight and jump to the topic below: \cf Hv_loc_cult \ftx \dt 27/Apr/1998 \key Hv_loc_cult \shd Location/direction--cultural influence \txt The system of verbal locational marking is often sensitive to the culture and/or environment of the people who speak the language. For example, many of the Quechuan languages, spoken in the Andes Mountains of South America, have verbal locational suffixes indicating action 'uphill', 'downhill' and 'at the same altitude'. Yagua, a language spoken along the rivers of lowland Peru has similar suffixes that mean action done 'upriver', 'downriver' or 'overland.' Many Papuan languages (Papua New Guinea and Irian Jaya) have elaborate systems of locational marking on verbs. The following is an example from Orya (Tor-Lake Plain stock, Irian Jaya. Example courtesy of Phil Fields): \ftx \fln (49) esek-gul-bla-in-hal-za \ftx slide-NOM:SG>ACC:F-DAT:MASC-down-away-to:here \ftx `Slide it down and away to me.' \ftx \txt In Yagua, the suffixes '-nuvïï', '-nuvaa' and a few others designate that the action expressed by the verb they attach to occurs relative to a particular locational scene and trajectory of movement (T. Payne 1989): \ftx \fln (50) a. Naani-ipeni-yââ-nuvïï 'They dance all over on \ftx 3DL-dance-DIST-on:arrival1 arrival (current scene).' \ftx \ftx b. Naani-núú-ñuvee 'They look on arrival (new scene).' \ftx 3DL-look-on:arrival2 \ftx \txt Both '-nuvïï' and '-nuvaa' indicate that the action expressed by the verb occurs upon arrival on some scene. The opposition between the two is determined by whether that scene is the currently activated one or if it implies the activation of a new scene. Other similar locational suffixes in Yagua include: '-rïï'-- 'passing by', '-ja'-- 'moving horizontally, across water or land', '-jasúmiy'-- 'moving upwards', '-siy'-- 'departing'. \ftx \txt In Otomí (Otomanguean, Mexico), the verb phrase contains an auxiliary element that inflects for person and number of the subject, for aspect, and for whether the action is away from a designated deictic center (exocentric) or towards a designated deictic center (centric). The deictic center is usually, but not necessarily, the location of the speaker at the time of speaking (examples courtesy of Henrietta Andrews): \ftx \fln (51) ¿bü x-tí tzon nìr ngû 'When you arrive at your house \ftx when FUT1-2:EXO arrive your house (over there) . . . ' \ftx \fln (52) ngû g-rí ¿úni 'As you give (it) (here) . . .' \ftx as FUT-2:CENT give \ftx \fln References: Matisoff 1973, T. Payne 1984, DeLancey 1986. \ftx \dt 27/Apr/1998 \key Hv_misc \shd Miscellaneous \txt This topic contains the following subsections. Highlight and jump to the topic of interest: \cf Hv_misc_lxtim Lexical Time Reference \cf Hv_misc_dist Distributive \cf Hv_misc_env Environmental \dt 27/Apr/1998 \key Hv_misc_dist \shd Miscellaneous--Distributive \txt Distributive, i.e. 'all over the place', 'with a back and forth motion'. \ftx \key Hv_misc_env \shd Miscellaneous--Environmental \txt Environmental, e.g. 'at night', 'over water' (on motion verbs), etc. Athabaskan languages are particularly rich in what we can only describe as 'miscellaneous' verbal operators. They are miscellaneous not only because they express semantic notions not embodied in the verbal morphology of languages familiar to most linguists, but also because their functions are variable and difficult to describe with a single inclusive statement. The presence of, or choice between such morphemes typically is dependent on a complex of factors including verb semantics, verb stem shape and syntactic valence. Such operations do exist in other language families of the world as well. Furthermore, most (if not all) verb morphology exhibits a certain degree of randomness and variability; it's just that Athabaskan languages seem to have taken this characteristic to an extreme. \ftx \txt To continue this discussion, highlight and jump to the topic below: \cf Hv_misc_env_ex \ftx \dt 27/Apr/1998 \key Hv_misc_env_ex \shd Miscellaneous--Athabaskan example \txt A brief presentation of one Athabaskan system is in order here. The following data on Koyukon come exclusively from Thompson (1989). \ftx \txt Some verb morphemes or verb phrase particles may have no clear or productive semantic effect. They may simply be required for certain roughly defined classes of verbs, e.g. verbs of motion, manipulation, transitive verbs, etc. Athabaskan languages are famous for their 'verb classifiers'. Koyukon has four verb classifier morphemes: '0-' , 'l~-', 'di-' and 'li-'. In the following examples the choice of classifier is not predictable from the semantics or phonological shape of the verb stems: \ftx \fln (59) na-ghonh 'You are making them.' \ftx 2S:SUBJ-make:PL [0 classifier] \ftx \fln (60) ni-l~-tsee 'You are making it.' \ftx 2S:SUBJ-CL-make:SG \ftx \fln (61) di-bits 'It is wide.' \ftx CL-wide \ftx \fln (62) li-ts'ul~ 'It is clean.' \ftx CL-clean \ftx \txt Sometimes a given verb root can occur with more than one of these operators, in which case the operator 'derives' a verb from one sub-class to another. In Koyukon there is a tendency for the 'l~-' classifier to be used with transitive verbs, though examples such as 62 above illustrate that 'l~-' cannot be considered a marker of transitive verbs. Nevertheless, any intransitive verb can be made transitive by changing any classifier to 'l~-': \ftx \fln (63) a. atsah 'He/she is crying.' \ftx cry [0 classifier] \ftx \ftx b. ni-l~-tsah 'You are making him/her cry.' \ftx 2S:SUBJ-CL-cry \ftx \txt The verb classifiers of Koyukon (and Athabaskan languages generally) operate within many of the functional systems described in earlier sections, such as causation (63b), passive (ex. 64b and 65b), and applicative (ex. 66b): \ftx \ftx ACTIVE: \fln (64) a. y-ee-to-ts'iyh 'He/she will pinch him/her once.' \ftx 3S:DO-once-FUT-pinch [0 classifier] \ftx \ftx PASSIVE: \ftx b. ee-to-di-ts'iyh 'He/she will be pinched once.' \ftx once-FUT-CL-pinch \ftx ACTIVE: \fln (65) a. n-ee-to-l~-dzis 'He/she will hit you once.' \ftx 2S:DO-once-FUT-CL-hit \ftx \ftx PASSIVE: \ftx b. ee-ta-gh-ee-l-dzis 'You will be hit once.' \ftx once-FUT-PRG-2S:SUBJ-CL-hit [li- classifier] \ftx \ftx NORMAL TRANSITIVE: \fln (66) a. li-tl~-baats 'I boiled it.' \ftx PERF-1SG:SUBJ:CL-boil \ftx \ftx \ftx 'SELF BENEFACTIVE' \ftx b. daa-l-gi-baats 'I boiled it for myself.' \ftx THM-PERF-1SG:SUBJ:CL-boil \ftx \txt These examples also illustrate what have been called the 'theme' morphemes of Athabaskan. These are morphemes that are simply required for certain stems. For any such miscellaneous operations, argue for why you have not treated them as TAM or location/direction marking. \ftx \dt 27/Apr/1998 \key Hv_misc_lxtim \shd Miscellaneous--Lexical time reference \txt Some typical miscellaneous verb or verb phrase operations include: Lexical time reference (as opposed to tense), e.g. yesterday, tomorrow. Koyukon 'ee-'-- 'once'. Yagua '-jásiy'-- 'earlier today', '-jay'-- 'yesterday'. These have been called degrees of 'tense' (Payne and Payne 1990), but really are more like miscellaneous derivational morphemes, because: \ftx \ftx 1) they are not required at all by the verb system. Our feeling \ftx (though no statistical study has been done) is that -jay is about \ftx as common as the time adverbial yesterday in English. \ftx 2) The information these suffixes express is very specific -- \ftx i.e. their meanings are more characteristic of lexical items \ftx than of grammatical morphemes. \ftx \dt 27/Apr/1998 \key Hv_nz \shd Nominalization \txt Nominalizations are words that function as nouns but are based on roots of some other grammatical category. In this section we will concentrate on nominalizations based on verb roots. Some languages make extensive use of verb-based nominalization. Any of the dependent clause types discussed in chapter #11 (relative, complement and adverbial clauses) can be formally instantiated by verb-based nominalizations. Therefore extensive discussion of the uses of certain kinds of nominalizations is found in chapter #11. In the present section I will provide a simple typology of verb-based nominalizations. \ftx \txt Nominalizations can be divided functionally into two types: action nominalizations and participant nominalizations (Comrie and Thompson 1985). As with other verbal operations, nominalization can be lexical, morphological or analytic. In this section we will try to give examples of all three formal strategies used to accomplish action nominalization and various kinds of participant nominalization. \ftx \txt To continue this discussion, highlight and jump to the topics below: \cf Hv_nz_actn Action Nominalization \cf Hv_nz_part Participant Nominalization \dt 27/Apr/1998 \key Hv_nz_actn \shd Action nominalization \ftx \txt An action nominalization refers to the action, usually in the abstract, expressed by the verb root on which it is based. English is particularly rich in action nominalization strategies. For example, the root walk is basically a verb: \ftx \fln (1) I walk to school. \ftx \txt However, there are various ways to form an action nominalization from this verb--see the succeeding sections for examples. \ftx \shd2 Action nominalization--lexical \ftx \txt The simplest way to nominalize lexically is with a 'zero' operator. As in other sections of this manual, zero derivation can be considered to be a lexical process: \ftx \ftx (2) Let's go for a walk. \ftx \txt In this example, the root 'walk' is being used as a noun that refers to a specific instance of walking. Occasionally, a verb plus a preposition can form a lexical nominalization: \ftx \ftx (3) That was a significant breakthrough. (< break through) \ftx He has a hangup. (< hang up) \ftx She gave him a talking to. (< talk to) \ftx She posed for a pin-up. (< pin up) \ftx They gave her a makeover. (< make over) \ftx \shd2 Action nominalization--morphological \ftx \txt In addition to a lexical strategy formed with a zero operator, action nominalizations can be formed morphologically in English: \ftx \ftx (4) Walking is good for you. \ftx I'm looking for employment. (< employ) \ftx He worked in construction. (< construct) \ftx That's a new procedure. (< procede) \ftx The process wore me out. (< procede) \ftx It was a display of abundance (< abound) \ftx Prosperity is not an end in itself. (< prosper) \ftx Economic growth is down. (< grow) \ftx Flight 24 is now boarding at gate 3c. (< fly) \ftx \txt Note that all of these morphological strategies are lexically restricted. For example, even a strategy as common as '-ing' cannot be used with all verbs: \ftx \ftx (5) *Employing is good for you. \ftx *I like the water's spewing. \ftx \txt Other strategies, e.g., '-ess', '-th', '-ure' and '-ight', are extremely non-productive as verb-based nominalizers, hence they are close to being lexical nominalization processes. \ftx \shd2 Action nominalization--analytic \ftx \txt In isolating languages, nominalization may be analytic. For example, Mandarin Chinese employs the particle 'de' to form many kinds of nominalizations, including action nominalizations (examples from Li and Thompson 1981): \ftx \ftx (6) women hézuò de wèntí hen jiandan \ftx we cooperate NOM problem very simple \ftx `The problem concerning our cooperation is very simple.' \ftx \txt Character notes: Hachek over o of women, e of hen and first a of last word. Line over 2nd a of last word. \ftx \fln Further references: Comrie and Thompson (1985). \ftx \dt 27/Apr/1998 \key Hv_nz_part \shd Participant nominalizations \ftx \shd2 Agent nominalizations \ftx \txt A nominalization that refers to the agent of the nominalized verb is an agent nominalization. One function of the suffix -er/-or in English is as a morphological agent nominalizer: \ftx \fln (7) Employer (< employ) = a person who employs someone. \ftx Tax collector (< tax collect) = a person who collects taxes. \ftx \shd2 Lexical agent nominalization is marginal in English: \ftx \fln (8) a drip of water (< it drips??) \ftx a pickpocket (< he picks pockets) \ftx a scarecrow (< it scares crows) \ftx It was a flop (< it flopped) \ftx \txt In English, agent nominalizations generally function to refer to entities according to characteristic activities that they are involved in, rather than specific events. However, in many other languages, agent nominalizations are used to refer to specific activities: \ftx \ftx Yagua: \fln (9) dapüüñu 'hunter/one who is hunting'. \ftx dapüüy-nu \ftx hunt-NOM \ftx \txt Mandarin Chinese employs the particle 'de' to form agent nominalizations (examples from Li and Thompson 1981): \ftx \fln (10) zhòng shuiguo de hen nán guòhuó \ftx grow fruit NOM very difficult make:living \ftx 'It is difficult for fruit-growers to make a living.' \ftx \fln Character notes: Hachek over i and o of 2nd word and e of 4th word. \ftx \fln (11) mài qìche de dàbàn dou shì hao rén. \ftx sell car NOM majority all be good person \ftx `Car sellers are mostly good people.' \ftx \fln Character notes: Hachek over a of hao. Line over e of 2nd word and o of dou. \ftx \shd2 Patient nominalizations \ftx \txt A nominalization that refers to the patient of the nominalized verb is a patient nominalization. Sometimes patient nominalizations are called PAST PARTICIPLES. The '-ee' suffix in English comes from the French past participle and is a morphological patient nominalizer: \ftx \fln (12) He is a new employee. (< employ) \ftx He is a Vietnam returnee. (< return) \ftx a retiree \ftx a divorcee \ftx an escapee \ftx \txt Like agent nominalization, lexical patient nominalization is marginal in English, functioning mostly with the modifiers `good' and `bad': \ftx \fln (13) This book is a good read. \ftx That's a bad buy. \ftx \txt Panare has two patient nominalizers. One is a true 'past participle' in that it refers to an entity according to some event that that entity was involved in in the past: \ftx \fln (14) yïpetyu'masa' 'The hit one.' \ftx y-petyu'ma-sa' \ftx TRNS-hit-PPART \ftx \txt There is also a 'future participle' that refers to an entity according to some event it is 'destined' to be involved in in the future: \ftx \fln (15) ej-ke mën y-onpa-se'ña wï-ch-ireemë-në-to'. \ftx be-NEG INAN:DIST TRN-eat-FUT:PART INTR-DETRANS-feed-INCL-PURP \ftx `There is no food for us to eat.' \ftx \fln (16) Mo ka n-aj y-apanawa-se'ña? \ftx EXIST QM 3SG-AUX TRN-rub-FUT:PART \ftx `Do you have something to rub on it?' \ftx \txt Consistent with its morphological typology, Mandarin Chinese employs an analytic strategy to form patient nominalizations. This strategy is also used to form other types of nominalizations. Only the context determines which kind of nominalization is meant: \ftx \fln (17) zhèi zhong zhíwù keyi dang-zuò chi de \ftx this type plant can take-be eat NOM \ftx `(One) can take this type of plant as food.' (chi de = 'food') \ftx \txt Character notes: Hachek over o of 2nd word and e and i of 4th word. Line over a of dang and i of chi. \ftx \fln (18) wo mài de shì Zhongguó huò \ftx I sell NOM be China product \ftx `What I sell is Chinese merchandise.' \ftx \shd2 Instrument nominalizations \ftx \txt Instrument nominalizations are often the same as agent nominalizations. This is true for English where the '-er' suffix is used for both: \ftx \fln (19) Coffee grinder, can opener, etc. \ftx \txt Spanish employs a productive compounding strategy to form instrument nominalizations of transitive verbs: \ftx \fln (20) para-brisa-s `windshield' \ftx stop-wind-PL \ftx \ftx abre-lata-s `can opener.' \ftx open-can-PL \ftx \ftx saca-punta-s `pencil sharpener.' \ftx take:out-point-PL \ftx \ftx \shd2 Location nominalization \ftx \txt Many languages have strategies that form nominalizations that refer in a general way to some entity associated with the verb root. Often these nominalizations refer to a location where the activity described by the verb tends to occur: \ftx \fln Eskimo -vik: \ftx (21) cali-vik Workshop. \ftx work-NOM \ftx \ftx ner-vik Restaurant, eating place. \ftx eat-NOM \ftx \ftx kumarr-vik Fireplace \ftx burn-NOM \ftx \ftx mis-vik Landing strip \ftx alight-NOM \ftx \fln Yagua -jo: \ftx (22) músá-jo Port (place where one descends \ftx descend-NOM to the river to get water) \ftx \ftx jasúmiy-jo Ladder, stairway \ftx ascend-NOM \ftx \shd2 Product nominalizations \ftx \txt English has lexical and morphological means of forming nouns that refer to the product of a verb root: \ftx \fln Lexical: It's only a scratch. He has a growth on his neck. \ftx \txt Morphological (stress shift to first syllable): 'This donut is a réject'. 'He is a recent cónvert'. 'We had to buy a pérmit'. \ftx \shd2 Manner nominalizations \ftx \fln 'He has a mean slice'. 'I can't hit his curve'. \ftx \dt 27/Apr/1998 \key Hv_other \shd Other deverbalizing processes \txt Every language has ways of changing the grammatical category of a lexeme, either to make it more or less verb-like. Such devices typically are sensitive to the demands of discourse, i.e., a device which renders a form more like a verb is used when the concept that form refers to constitutes an EVENT in the discourse (see Hopper and Thompson 1980, 1984). On the other hand, devices which render a form less verb-like are used when the form refers to a concept that is less event-like, e.g., a process or a state rather than an action, a description rather than an event, background rather than foreground, etc. The following diagram represents two contiuous scales that tend to parallel each other. Any given language may formally instantiate more or fewer of the categories listed on the right, but however many members of this scale a language instantiates, their functions will be related to one another according to the left-hand scale: \ftx \ftx Function | Formal instantiation \ftx | \ftx Transitive, active, | Active transitive verb \ftx kinetic event | Active intransitive verb \ftx Process | Non-active intransitive verb \ftx State | Stative verb \ftx | Adjective \ftx Abstract object | Mass noun \ftx Amorphous object | \ftx Concrete object | Count noun \ftx \dt 27/Apr/1998 \key Hv_tam \shd Tense/aspect/mode--introduction \txt It is not uncommon for the tense/aspect/mode system of a language to interact in a significant way with other seemingly distinct subsystems such as nominal case marking or verb agreement. For example, many languages use a nominative/accusative system of case marking and/or verb agreement in present tense and non-perfective aspects, but an ergative/absolutive system in clauses with perfective aspect or past tense (see Hgrel_splt for discussion of these terms and for examples). Beyond the sentence level syntax of the language, TAM marking is often deployed in interesting ways in the discourse (see Hdisc_cnt_act on action continuity). For example, so-called 'present tense' is often used to make a narrative describing past events more vivid: 'Then he says to me . . .' \ftx \txt Functionally, TENSE, ASPECT and MODE (TAM for short) are operations that anchor or GROUND the information contained in a clause according to its sequential, temporal or epistemological orientation. Tense is associated with the sequence of events in real time, aspect with the internal temporal 'structure' of a situation, while mode relates the situation to reality, i.e., how probable or actual a particular situation is. \ftx \shd2 Interrelations \ftx \txt Tense, aspect and mode are seldom distinct categories of verbal inflection. Rather there are usually one or more 'areas' of the verbal word or verb phrase where operations that impart tense, aspect and modal flavor are encoded. In fact, it may be that we only think of tense, aspect and mode as distinct because they are somewhat distinct categories in the classical languages and in Indo-European generally. Operators that occur in the TAM area of the verb or verb phrase are likely to have indistinct semantic ranges; their meanings may seem to vary depending on the verb they are attached to, or the case marking or other characteristics of the core NPs (subject and object) in the clause. TAM formatives may seem to combine tense, aspect, modal, evidential (section 4.2.7) and/or locational/directional concepts (section 4.2.3). Some languages pay more attention to tense (e.g. English), others to aspect (e.g. Austronesian and African generally), others to mode (Eskimo), and still others to location and direction (many American, Australian and Papuan languages). Furthermore, some verb stems may not allow certain operations while favoring others. Finally, certain combinations of TAM operators may co-occur with greater than chance frequency, whereas other logically possible combinations may seldom or never occur. Groups (usually pairs) of operators that commonly cluster together are sometimes referred to as HYPERMORPHEMES. \ftx \txt Beware of too quickly and confidently giving a TAM operator a gloss. Explanation consists of much more than supplying labels. For example, in elicitation it may be common for a particular TAM operator to occur in responses to English past-tense prompts. It is tempting to gloss such an operator with the label 'past tense' without investigating its semantics in relation to other TAM operators in the system. Many languages have been analyzed as having a tense system when in fact aspect is the more relevant parameter. \ftx \txt In the following sections we will discuss several concepts that have traditionally been described under the terminology of tense, aspect and mode. \ftx \shd2 Summary \txt Tense 'grounds' a situation with reference to a temporal deictic center, usually the time of utterance. Aspect describes the internal temporal structure of a situation. Mode describes varying degrees of likelihood that a situation is actual (like adjusting the brightness on a television screen). \ftx \txt This topic continues in the following subsections. Highlight and jump to the topic of interest: \cf Hv_tam_tns Tense \cf Hv_tam_asp Aspect \cf Hv_tam_mode Mode \cf Hv_tam_msyn Morphosyntax \cf Hv_tam_ncode Nominal Coding \cf Hv_tam_adv Temporal Adverbs \cf Hv_tam_auxv Auxiliary Verbs \dt 27/Apr/1998 \key Hv_tam_adv \shd Tense/aspect/mode and temporal adverbs \txt Languages often code temporal notions periphrastically. Temporal adverbs are periphrastic indicators of temporal relations. These do not constitute part of the tense or aspect system. For example, \ftx \fln (44) 'I see the doctor tomorrow.' ('tense': future) \ftx 'I see the doctor every day.' ('aspect': habitual) \ftx \txt It is obvious to us that 'tomorrow' and 'every day' are not tense markers in English, but that may be only because we have another marker whose function is clearly to indicate tense (-ed). In other languages, this may not be so clear. Indonesian, for example, along with many other languages of Insular and Mainland Southeast Asia, codes very few aspectual and no tense distinctions. The only way of making temporal grounding explicit is via adverbials or extra-clausal periphrastic devices. \ftx \dt 27/Apr/1998 \key Hv_tam_asp \shd Tense/aspect/mode--Aspect \txt ASPECT describes the internal temporal structure of an event or state -- the temporal 'shape' of a situation. The following diagrams and English examples may help conceptualize some of the more common aspectual distinctions languages make. Keep in mind, however, that no language necessarily grammaticalizes any of these aspects, and that the aspectual operations grammaticalized in any given language may not line up exactly with these notions. In particular, English does not grammaticalize many aspectual notions. This does not mean that English sentences 'have no aspect'. It just means that there are few well-oiled grammatical means of expressing aspectual variation. For example, completion and inception are not fully grammaticalized in English. They are expressed by analytic structures involving a matrix verb and a participial complement: 'I finished working', 'I started working.' \ftx \txt NOTE: Recall that signs of grammaticalization include: 1) formal shift and 2) semantic bleaching. The 'markers' of English inceptive and completive aspects are the complement taking verbs 'begin' and 'finish'. They neither take on special form, nor convey any idiosyncratic semantic overtones when occurring in inceptive or completive clauses. This is in contrast to the verb 'have', for example, that does have special semantics and special formal properties when occurring as a marker of perfect aspect or debitative mode: \ftx \fln (67) PERFECT ASPECT: They've fallen. \fln (68) DEBITATIVE MODE: They hæftê go now. \ftx \txt Note that in the lexical use of the verb 'have', the contractions illustrated in these sentences are much more unusual: \ftx \fln (69) LEXICAL USE: They have two dollars. \ftx ??They've two dollars. \ftx ??They hæftê dollars./??They hæf two dollars. \ftx \txt This is a formal indication of grammaticalization. The semantic indication that grammaticalization has taken place with the aspectual and modal use of 'have' is that the meanings of 67 and 68 are quite different from the standard lexical use of 'have' to mean something like 'possess'. 'Finish' and 'begin' as complement taking verbs, on the other hand, convey pretty much the same concepts as they do when the complement is not a clause: \ftx \fln (70) LEXICAL USE: He finished the bookcase. \ftx He began Mozart's sonata in G. \ftx \txt Of course, grammaticalization is a process rather than an event. Therefore, any given operator may be more or less grammaticalized. Verbs that have been grammaticalized as aspect markers in English include 'have' and 'go'. 'Will' is an archaic lexical verb (meaning 'want') that is now used almost exclusively as a future tense marker. 'Ought' is related to the lexical verb 'owe', and 'can' is related to the lexical verb 'know'. But these modals have diverged so significantly from their lexical sources that it is difficult for native speakers to discern any formal or functional connection. Interestingly, in both of these cases the aspect marker reflects the more conservative, older, form. \ftx \txt This topic contains the following subsections. Highlight and jump to the topic of interest: \cf Hv_tam_asp_perfv Perfective Aspect \cf Hv_tam_asp_imprf Imperfective Aspect \cf Hv_tam_asp_perf Perfect Aspect \cf Hv_tam_asp_cplv Completive Aspect [INCOMPLETE] \cf Hv_tam_asp_incpt Inceptive Aspect [INCOMPLETE] \cf Hv_tam_asp_plu Pluperfect Aspect \cf Hv_tam_asp_cntv Continuative/Progressive Aspect \cf Hv_tam_asp_punc Punctual Aspect \cf Hv_tam_asp_iter Iterative Aspect \cf Hv_tam_asp_hab Habitual Aspect \ftx \dt 27/Apr/1998 \key Hv_tam_asp_cntv \shd Tense/aspect/mode--Continuative/progressive aspect \txt CONTINUATIVE or PROGRESSIVE aspect implies an ongoing, dynamic process. This is opposed to STATIVE aspect which implies no change over time. Continuous or progressive aspect is distinguished from HABITUAL in that continuative or progressive refers to actual events, whereas habitual expresses the general truth that some event characteristically takes place from time to time. Habitual does not refer to any specific events. \ftx \txt The auxiliary 'be' is a grammaticalized marker of progressive aspect in English. \ftx \ftx >-------------------------> 'He is writing letters.' \ftx \dt 27/Apr/1998 \key Hv_tam_asp_cplv \shd Tense/aspect/mode--Completive aspect -- INCOMPLETE \ftx \key Hv_tam_asp_hab \shd Tense/aspect/mode--Habitual aspect \txt As mentioned above under continuative/progressive, HABITUAL aspect conveys an assertion that a certain type of event, such as Rudyard walking to school, regularly takes place (i.e. is instantiated by actual events) from time to time. It does not imply that an instance of the event is taking place 'now'. \ftx \ftx <-------------------------> 'He drinks.' \ftx \txt The simple 'present tense' verb forms in English do not indicate present tense, as defined here, for dynamic verbs. That is, the present tense in English does not anchor dynamic events (events that involve change over time) as occurring at the same time as the time of speaking ('now'). A sentence such as 'he walks to school' either means a) habitual ('he walks to school every day'), b) 'historical present', ('So he gets out of bed, gets dressed and has breakfast. Then he walks to school, see?') which actually anchors the event at some point in the past, or c) 'future' ('Tomorrow he walks to school; I refuse to take him anymore.'). Since 'walk' describes a dynamic event, in order to anchor it to the moment of speaking, a progressive form is required: 'he is walking to school'. 'Static' or STATIVE situations do not involve change over time. Verbs of sensation and mental state typically describe static situations, e.g.: 'she knows the answer', 'he sees a bear', 'I wonder what happened to Jane?' etc. For such verbs, the present tense form actually does anchor the event at the time of utterance. These also, in their normal senses, do not occur in the progressive aspect: '?she is remembering his name.' \ftx \dt 27/Apr/1998 \key Hv_tam_asp_imprf \shd Tense/aspect/mode--Imperfective aspect \txt In IMPERFECTIVE aspect the situation is viewed from 'inside', as an on going process. Habitual and progressive aspects are subtypes of imperfective. See below and Comrie (1978b) for more details. \ftx \ftx <------------------------> 'He writes letters.' \ftx \key Hv_tam_asp_incpt \shd Tense/aspect/mode--Inceptive aspect -- INCOMPLETE \ftx \key Hv_tam_asp_iter \shd Tense/aspect/mode--Iterative aspect \txt ITERATIVE aspect is where a punctual event takes place several times in succession. \ftx \ftx >-x-x-x-x-x-x-x-x-x-x-x-x-> 'He was coughing.' \ftx \txt With inherently punctual verbs like 'cough', the progressive implies iterativity. With 'non-punctual' (durative) verbs like 'run' the progressive implies continuity. The progressive cannot be used naturally with stative verbs like 'know', e.g. *'Rudyard was knowing the answer.' \ftx \dt 27/Apr/1998 \key Hv_tam_asp_perf \shd Tense/aspect/mode--Perfect aspect \txt PERFECT aspect normally describes a currently relevant state brought about by the situation (normally an event) refered to by the verb. \ftx \ftx --------|X 'He has come from Aqaba.' \ftx \txt 'He has come' implies 'he is here now', whereas 'he came' does not. He may be here now or he may have come and left again. There is no direct implication of current relevance inherent in the simple past tense of English. Perfect aspect is not the same as completive. 'He has finished working' (perfect completive) implies 'he is not working now', whereas the perfective completive, 'he finished working', does not carry this implication. For example, it is pragmatically acceptable to say 'He finished working at 12:00 and began again at 2:00.' In the perfect, however, this sounds strange: '??He has finished working at 12:00 and began again at 2:00.' The similarity between the terms perfect and perfective is unfortunate, as these refer to very different aspectual categories. Nevertheless, these terms are very standard in the literature, and the field worker needs to be careful not to confuse them. \ftx \dt 27/Apr/1998 \key Hv_tam_asp_perfv \shd Tense/aspect/mode--Perfective aspect \ftx \fln [NOTE: In the diagrams in the following sections, the following notation \ftx is used: \ftx \ftx > = unbounded time \ftx | = a temporal boundary \ftx ¦ = completion \ftx + = inception \ftx x = a punctual event, i.e. an event that occurs instantaneously and \ftx therefore has no internal temporal structure] \ftx \txt In PERFECTIVE aspect the situation is viewed in its entirety, independent of tense. The terms 'preterite' and 'aorist' usually refer to past tense plus perfective aspect. The main events of a narrative are normally recounted in perfective aspect, whereas collateral, explanatory, descriptive material occurs in various non-perfective aspects (e.g. imperfective, progressive, habitual): \ftx \ftx |--------| 'He wrote a letter.' \ftx \txt The sentence 'he wrote' in English could be perfective, habitual, iterative or almost any other aspect. The verb form simply encodes tense, leaving the aspectual distinctions to be disambiguated by the context. For example, 'he wrote a letter' is not likely to be habitual or iterative, therefore perfective aspect is much more likely, though there is no specific marker of perfective aspect in the clause. On the other hand, 'he wrote letters' is probably iterative (though it could be habitual) even though the verb form is the same as that in the previous example. Hence we say that English does not grammaticalize aspect. This is different from saying that these sentences in English 'have no aspect'. \ftx \dt 27/Apr/1998 \key Hv_tam_asp_plu \shd Tense/aspect/mode--Pluperfect aspect \txt PLUPERFECT, like 'preterite', refers to a combination of an aspect and a tense. Pluperfect combines perfect aspect and past tense. The effect of this combination is to shift the deictic center (DC) from 'now' to some point in the past. That is, the state that results from an event is presented as occurring at some point in the past: \ftx \ftx --------|DC ---------(now) 'I had entered a congested zone.' \ftx \dt 27/Apr/1998 \key Hv_tam_asp_punc \shd Tense/aspect/mode--Punctual aspect \txt PUNCTUAL events are those which have no internal temporal structure because they occur in an instant in time. Sometimes this aspect is referred to as INSTANTANEOUS. \ftx \ftx x 'He sneezed.' \ftx \dt 27/Apr/1998 \key Hv_tam_auxv \shd Tense/aspect/mode and auxillary verbs \txt Another periphrastic means of conveying TAM notions is predicate combining, i.e., the use of auxiliary or matrix verbs (see Hord_adpp_vsdepv): \ftx \fln (45) 'He has gone' (perfect aspect \ftx 'I am going' (progressive aspect) \ftx \txt Mode is not a verbal category in English. Rather it is expressed by auxiliaries (predicate combining): \ftx \fln (46) I should walk to school. \ftx can \ftx must \ftx might \ftx hæftê \ftx etc. \ftx \txt Aspect is not a verbal category in English either. Again, to the extent that it is grammaticalized at all, it is expressed by predicate combining: \ftx \fln (47) I have come. \ftx I was coming. \ftx I used to walk to school. \ftx I finished working. \ftx \txt Furthermore, 'tense' as a verbal category in English really has only two values: '-ed'-- 'past' (plus several allomorphs) and '0'-- 'non-past'. The future is coded by predicate combining (as are modals): \ftx \fln (48) PAST: I walked to school. \ftx NONPAST: I walk to school. (habitual, future or historical present) \ftx FUTURE: I will walk to school. \ftx \txt References: DeLancey (1982), Hopper (1979, 1982), Hopper and Thompson (1980), Comrie (1978b). Wallace (1982). \ftx \dt 27/Apr/1998 \key Hv_tam_mode \shd Tense/aspect/mode--Mode \txt MODE describes the degree of 'reality', or 'actuality' of a situation. It sometimes describes the speaker's estimation of the relevance of the situation to him or her self. The terms mode, mood and modality are used interchangeably. The highest level distinction in modal operations is between REALIS and IRREALIS, though like most conceptual distinctions these terms describe a continuum. A prototypical realis mode strongly asserts that a specific event or state of affairs has actually happened, or actually holds true. A prototypical irrealis mode makes no such assertion whatsoever. Note that irrealis mode does not necessarily assert that an event did not take place or will not take place. It simply makes no claims with respect to the actuality of the event or situation described. Negative clauses do assert that events or situations do not hold, but these are subject to the same realis-irrealis continuum as are affirmative clauses. For example, I can assert the reality of the statement 'he did not clean the kitchen' just as weakly or strongly as I can assert the reality of its affirmative counterpart (see Hprag_neg_intro on negative assertions). \ftx \txt Mode interacts significantly with aspect and tense. For example, habitual aspect clauses are less realis than perfective aspect clauses since habitual aspect describes an event type that is instantiated from time to time by actual events. Similarly mode interacts with the referentiality and definiteness of noun phrases. For example, entities under a highly realis mode assertion are more likely to be referential than those under an irrealis assertion: \ftx \fln (26) Rudyard ate the cheerios that were in the cupboard. \ftx \fln However, a less realis mode with a specific referential object sounds odd: \ftx \fln (27) ??Rudyard always eats the cheerios that were in the cupboard. \ftx \fln A less realis mode with a generic (non-referential) object sounds better: \ftx \fln (28) Rudyard always eats cheerios for breakfast. \ftx \txt Irrealis mode can refer to an event/situation which is presented as occurring in a contingent world. For example: \ftx \fln (29) If you eat Wheaties, you'll be like the big boys. \ftx \txt In this sentence the condition, 'if you eat Wheaties', is irrealis. Irrealis mode makes no assertion that the event/state IS true. INTERROGATIVE MODE and IMPERATIVE MODE are likely to be irrealis, since they do not assert that X did happen, but order it to come about, or question whether it will or did come about. So if a language grammaticalizes the notion of irrealis, chances are interrogative and/or imperative clauses will fall into the irrealis category. Some terms that have been used for various kinds of assertions that are close to the irrealis end of the realis-irrealis continuum are: Subjunctive, optative (wishes), potential (might, ability to), hypothetical/imaginary, conditional (if), probability, debitative (should, must, have to). \ftx \txt This topic contains the following subsections. Highlight and jump to the topic of interest: \cf Hv_tam_mode_subj Subjunctive Mode \cf Hv_tam_mode_opt Optative Mode \cf Hv_tam_mode_potn Potential Mode \cf Hv_tam_mode_hypo Hypothetical Mode \cf Hv_tam_mode_cnd Conditional Mode \cf Hv_tam_mode_deon Deontic Mode \cf Hv_tam_mode_epist Epistemic Mode \ftx \dt 27/Apr/1998 \key Hv_tam_mode_cnd \shd Tense/aspect/mode--Conditional Mode \fln The following are examples of CONDITIONAL mode from English: \ftx \fln (34) If you eat your cheerios, you will be strong. \ftx If you come home before six, we can go to the movie. \ftx \key Hv_tam_mode_deon \shd Tense/aspect/mode--Deontic Mode \fln The following are examples of DEONTIC (OBLIGATION) mode from English: \ftx \fln (35) I have to earn a million dollars this year. \ftx We should send out a Christmas letter. \ftx There ought to be a law. \ftx They must have dinner with us. \ftx \txt Sometimes what I have called 'potential' mode is treated as part of the deontic continuum. \ftx \dt 27/Apr/1998 \key Hv_tam_mode_epist \shd Tense/aspect/mode--Epistemic Mode \txt The following is an example of EPISTEMIC (PROBABILITY) mode from English: \ftx \fln (36) They must have left already. \ftx \txt Note that 'must' in English has both an EPISTEMIC (ex. 36) and a DEONTIC (ex. 35) sense. \ftx \dt 27/Apr/1998 \key Hv_tam_mode_hypo \shd Tense/aspect/mode--Hypothetical Mode \fln The following are examples of HYPOTHETICAL mode from English: \ftx \fln (33) Let's suppose that I had a million dollars... \ftx Now if it were possible to earn a million dollars as a \ftx college professor . . . \ftx If you had eaten your cheerios as a child, you would be \ftx doing better in school today. \ftx \key Hv_tam_mode_opt \shd Tense/aspect/mode--Optative Mode \fln The following is an example of OPTATIVE mode in English: \ftx \fln (31) I wish I had a million dollars. \ftx I want to earn a million dollars. \ftx \key Hv_tam_mode_potn \shd Tense/aspect/mode--Potential Mode \fln The following is an example of POTENTIAL mode from English: \ftx \fln (32) I might earn a million dollars. \ftx I can/am able to earn a million dollars. \ftx \key Hv_tam_mode_subj \shd Tense/aspect/mode--Subjunctive Mode \fln The following is an example of SUBJUNCTIVE mode in Spanish: \ftx \fln (30) Si no hubiera sido por Anita, mi reloj sería perdido. \ftx if NEG have:SUBJ been for Anita my watch would:be lost \ftx 'If it had not been for Anita, my watch would be lost'. \ftx \key Hv_tam_msyn \shd Tense/aspect/mode--Morphosyntax \txt Morphosyntax of TAM: TAM is most often associated with the verbal word (for polysynthetic languages) or verb phrase (for more isolating languages). It can be expressed lexically, morphologically or analytically: \ftx \fln (37) English: \ftx is > was Past: Lexical (suppletion) \ftx walk > walked Past: Morphological \ftx see > will see Future: Analytic \ftx \fln (38) Spanish: \ftx ir > fue 'go'>'went' Tense: Lexical \ftx hablar > habló 'speak'>'spoke' Tense: Morphological \ftx habló > hablaba 'spoke'>'was speaking' Aspect: Morphological \ftx fue > fuera 'was'>'would be' Mode: Morphological \ftx haber hablado 'have spoken' Aspect: Analytic \ftx \fln (39) Mandarin: \ftx Zéi pâo le Aspect: Analytic \ftx thief run PERFECTIVE \ftx 'The thief ran away.' \ftx \dt 27/Apr/1998 \key Hv_tam_ncode \shd Tense/aspect/mode and nominal coding \txt Nominal coding often affects the aspect of a clause (less often the tense or mode). For example, in English the difference between a generic and a specific direct object can convey the difference between habitual and perfective aspect: \ftx \fln (40) I built houses (past tense / habitual aspect) \ftx I built a house (past tense / perfective aspect) \ftx \txt In Finnish, the difference between an accusative and a partitive direct object conveys the distinction between perfective and progressive aspect: \ftx \fln (41) Finnish: \ftx a. han luki kirjan 'He read the book.' \ftx he read book:ACC (past perfective) \ftx \ftx b. han luki kirjaa 'He was reading the book.' \ftx he read book:PART. (past progressive) \ftx \txt For this reason, case markers can be mistaken for TAM markers. For example, in Guaymí the ergative case marker '-gwe' only occurs in perfective aspect clauses: \ftx \fln (42) a. Dorí-gwe ti dëma-íni 'Doris greeted me.' \ftx Doris-ERG 1SG greet-PERF \ftx \fln (43) a. Dori ti dëma-e 'Doris greets me.' \ftx Doris 1SG greet-PRES \ftx \txt The original grammatical description of Guaymí analyzed '-gwe' as a 'tense marker' since it only occurred in what the author analyzed as past tense. \ftx \dt 27/Apr/1998 \key Hv_tam_tns \shd Tense/aspect/mode--Tense \txt TENSE is the relation of the time of an event described in a clause to some specific reference point in time, usually the moment the clause is uttered. If we think of time as a line, with 'now' represented by a point moving from left to right, we can conceptualize tense in terms of the following diagram: \ftx \ftx now \ftx ---------------------------->------------------------------ \ftx \txt Languages divide up this conceptual notion for purposes of grammatical marking in many different ways. One common tense system is past, present and future: \ftx \ftx past |present| future \ftx -----------------------+--->---+-------------------- \ftx \txt Probably more common, however, are two-way distinctions, either past/non-past or future/non-future: \ftx \ftx past | non-past \ftx --------------------------|>------------------------ \ftx non-future | future \ftx --------------------------->|----------------------- \ftx \txt Another possibility that is rumored to occur is a two way distinction between present and non-present. In such a system 'past' and 'future' actions would be coded with a single form: \ftx \ftx not-now | now | not-now \ftx -----------------------+-->--+---------------------- \ftx \txt Some languages make many divisions in the past. The following diagram illustrates the tense system of Yagua: \ftx \ftx Distant |1 year |1 month |1 week |today/ |'now' |future \ftx past |ago |ago |ago |yes'day| | \ftx ------------------------------------------->----------> \ftx \txt Bembe is a language that is reported to have grammaticalized two degrees of proximity in the future: \ftx \fln (25) a. n-kà-boomba 'I'm about to work.' \ftx 1SG-FUT:1-work \ftx \ftx b. n-ká-boomba 'I will work (later).' \ftx 1SG-FUT:2-work \ftx \txt Note that the near future prefix has low tone, while the distant future prefix has high tone. No language has been shown to have more than two degrees of proximity in the future. If you find an exception to this, it is worth a paper describing it. \ftx \dt 27/Apr/1998 \key Qcplx \shd Complex constructions \fln For a complete listing of the Help topics covered, jump to: \cf Hcplx Complex constructions \ftx \txt The following topics have questions available: \cf Qcplx_intro Introduction \cf Qcplx_sv Serial Verbs \cf Qcplx_cplcl Complement Clauses \cf Qcplx_advcl Adverbial Clauses \cf Qcplx_advcl_time Time \cf Qcplx_advcl_loc Location \cf Qcplx_advcl_man Manner \cf Qcplx_advcl_purp Purpose \cf Qcplx_advcl_rsn Reason \cf Qcplx_advcl_circ Circumstantial \cf Qcplx_advcl_sim Simultaneous \cf Qcplx_advcl_cnd Conditional \cf Qcplx_advcl_cncs Concessive \cf Qcplx_advcl_sbst Substitutive \cf Qcplx_advcl_add Additive \cf Qcplx_advcl_abs Absolutive \cf Qcplx_med Medial Clauses \cf Qcplx_med_chn Clause Chaining \cf Qcplx_med_medcl Medial Clauses \cf Qcplx_med_sw Switch Reference Systems \cf Qcplx_relcl Relative Clauses \cf Qcplx_relcl_np Noun Phrase Encoding \cf Qcplx_relcl_elem Clause Elements Capable of being Relativized \cf Qcplx_coord Coordination \cf Qcplx_coord_np Strategies -- Parallel to Noun Phrases \cf Qcplx_coord_zero Strategies -- Zero \cf Qcplx_coord_cnj Strategies -- Coordinating Conjunction \cf Qcplx_coord_other Other Coordination Strategies \cf Qcplx_coord_log Logical Relations \cf Qcplx_other Other Complex Constructions \dt 25/Aug/1997 \key Qcplx_advcl \shd Adverbial clauses \fln For Help jump to: \cf Hcplx_advcl \ftx \q How are adverbial clauses formed? \ftx \q What kinds of adverbial clauses are there? e.g. time, manner, purpose, reason, consequence, sequence, conditional. \ftx \q Can adverbial clauses occur in more than one place in a sentence? \ftx \q If so are there any differences in meaning associated with the various allowable positions for any given adverbial clause type? \ftx \q Among the conditionals, are there any sub-divisions? e.g. contrafactual (If I had done it differently, that wouldn't have happened), hypothetical (If I were you, I'd do it differently). \ftx \q What restrictions are there on the TAM marking of conditional clauses? \ftx \dt 27/Apr/1998 \key Qcplx_advcl_abs \shd Adverbial clauses--Absolutive \fln For Help jump to: \cf Hcplx_advcl_abs \ftx \q How are absolutive adverbial clauses formed in the language? \ftx \dt 20/Aug/1997 \key Qcplx_advcl_add \shd Adverbial clauses--Additive \fln For Help jump to: \cf Hcplx_advcl_add \ftx \q How are additive adverbial clauses formed in the language? \ftx \dt 20/Aug/1997 \key Qcplx_advcl_circ \shd Adverbial clauses--Circumstantial \fln For Help jump to: \cf Hcplx_advcl_circ \ftx \q How are circumstantial adverbial clauses formed in your language? \ftx \dt 20/Aug/1997 \key Qcplx_advcl_cncs \shd Adverbial clauses--Concessive \fln For Help jump to: \cf Hcplx_advcl_cncs \ftx \q How are concessive adverbial clauses formed in the language? \ftx \dt 20/Aug/1997 \key Qcplx_advcl_cnd \shd Adverbial clauses--Conditional \fln For Help jump to: \cf Hcplx_advcl_cnd \ftx \q How are conditional adverbial clauses formed in your language? \ftx \dt 20/Aug/1997 \key Qcplx_advcl_loc \shd Adverbial clauses--Location \fln For Help jump to: \cf Hcplx_advcl_loc \ftx \q How are adverbial clauses of location formed in your language? \ftx \dt 20/Aug/1997 \key Qcplx_advcl_man \shd Adverbial clauses--Manner \fln For Help jump to: \cf Hcplx_advcl_man \ftx \q How are adverbial clauses of manner formed in your language? \ftx \dt 20/Aug/1997 \key Qcplx_advcl_purp \shd Adverbial clauses--Purpose \fln For Help jump to: \cf Hcplx_advcl_purp \ftx \q How are adverbial clauses of purpose formed in your language? \ftx \dt 20/Aug/1997 \key Qcplx_advcl_rsn \shd Adverbial clauses--Reason \fln For Help jump to: \cf Hcplx_advcl_rsn \ftx \q How are adverbial clauses of reason formed in your language? \ftx \dt 20/Aug/1997 \key Qcplx_advcl_sbst \shd Adverbial clauses--Substitutive \fln For Help jump to: \cf Hcplx_advcl_sbst \ftx \q How are substitutive adverbial clauses formed in your language? \ftx \dt 20/Aug/1997 \key Qcplx_advcl_sim \shd Adverbial clauses--Simultaneous \fln For Help jump to: \cf Hcplx_advcl_sim \ftx \q How are simultaneous adverbial clauses formed in your language? \ftx \dt 20/Aug/1997 \key Qcplx_advcl_time \shd Adverbial clauses--Time \fln For Help jump to: \cf Hcplx_advcl_time \ftx \q How are adverbial clauses of time formed in your language? \ftx \dt 20/Aug/1997 \key Qcplx_coord \shd Coordination \fln For Help jump to: \cf Hcplx_coord \ftx \q How are the following kinds of clausal coordination typically coded? \q a. Conjunction (a and b)/(neither a nor b) \q b. Disjunction (a or b) \q c. Exclusion (a and not b) \ftx \dt 27/Apr/1998 \key Qcplx_coord_cnj \shd Coordination strategies--Coordinating conjunction I \fln For Help jump to: \cf Hcplx_coord_cnj_1 \cf Hcplx_coord_cnj_2 \ftx \q Does your language use coordinating conjunctions for joining clauses? \q Is the coordinating conjunction also used with instrumental and comitative elements in your language? \dt 27/Apr/1998 \key Qcplx_coord_cnj_2 \shd Coordination strategies--Coordinating conjunction II \fln For Help jump to: \cf Hcplx_coord_cnj_2 \ftx \q Is the coordinating conjunction also used with instrumental and comitative elements in your language? \ftx \dt 27/Apr/1998 \key Qcplx_coord_log \shd Coordination--Logical relations \fln For Help jump to: \cf Hcplx_coord_log \ftx \q How is disjunction expressed in your language? \ftx \dt 20/Aug/1997 \key Qcplx_coord_np \shd Coordination strategies--Parallel to noun phrases \fln For Help jump to: \cf Hcplx_coord_np \ftx \q Does the strategy used by your language for the coordination of clauses apply also to the conjoining of noun phrases? \ftx \dt 27/Apr/1998 \key Qcplx_coord_other \shd Other Coordination strategies \fln For Help jump to: \cf Hcplx_coord_other \ftx \q Is there more than one coordinating conjunction used to link clauses in your language? \q What others may occur and how are they used? \ftx \dt 27/Apr/1998 \key Qcplx_coord_zero \shd Coordination strategies--Zero \fln For Help jump to: \cf Hcplx_coord_zero \ftx \q Does your language use simple juxtaposition as a strategy for joining clauses? \ftx \dt 27/Apr/1998 \key Qcplx_cplcl \shd Complement clauses \fln For Help jump to: \cf Hcplx_cplcl \ftx \q What kinds of complement clauses does the language have? \ftx \q Are particular complement types common for particular classes of complement-taking verbs? \q Does the language allow subject and object complements, or just object complements? \ftx \dt 27/Apr/1998 \key Qcplx_cplcl_cplx \shd Complement clauses--the Complexity Continuum \fln For Help jump to: \cf Hcplx_cplcl_cplx \ftx \fln For a note on the Complexity Continuum, jump to the help screen. \ftx \dt 20/Aug/1997 \key Qcplx_cplcl_indq \shd Complement clauses--Indirect questions \fln For Help jump to: \cf Hcplx_cplcl_indq \ftx \q How are indirect questions handled in your language? \ftx \dt 20/Aug/1997 \key Qcplx_cplcl_mrgcpl \shd Complement clauses--Merged complements \fln For Help jump to: \cf Hcplx_cplcl_mrgcpl \ftx \q Is there evidence for merged complements in your language? How may they be described? \ftx \dt 27/Apr/1998 \key Qcplx_cplcl_prop \shd Complement clauses--Propositional complements \fln For Help jump to: \cf Hcplx_cplcl_prop \ftx \q How may propositional complements be characterized in your language? \ftx \dt 20/Aug/1997 \key Qcplx_cplcl_s/o \shd Subject & Object Complement clauses \fln For Help jump to: \cf Hcplx_cplcl_s/o \ftx \q Does your language show evidence for both subject and object complement clauses? \ftx \dt 27/Apr/1998 \key Qcplx_intro \shd Complex constructions--Introduction \fln For Help jump to: \cf Hcplx_intro \ftx \q What kinds of complex multi-verb constructions are found in your language? \ftx \dt 27/Apr/1998 \key Qcplx_med \shd Medial clauses, clause chaining and switch reference \fln For Help jump to: \cf Hcplx_med \q Does the language have any grammaticalized device that explicitly indicates whether a participant in one clause is the same or different than some participant in another clause? \ftx \fln If so, answer the following questions: \ftx \q a) What direction does the dependency go? i.e. does a marker signal coreferentiality with a yet to be mentioned participant, or an already mentioned participant? (Maybe both, depending on other factors). \ftx \q b) What can 'antecede' one of these markers? i.e. is coreferentiality always with respect to a 'subject' participant, or can non-subject AGENTS, or nominals of other grammatical relations also antecede a coreference form? \ftx \q c) On what categories of elements can these markers go?, e.g. verbs, nouns, conjunctions, etc. \ftx \q Can one clause be inflected for the person/number of the subject of some other clause? \ftx \q Do the markers of interclausal coreference also carry other information, e.g. tense/aspect or semantic relations between clauses? \ftx \q How extensive is this phenomenon? \ftx \dt 27/Apr/1998 \key Qcplx_med_chn \shd Clause chaining \fln For Help jump to: \cf Hcplx_med_chn \ftx \q Is clause chaining a feature of your language? \ftx \dt 20/Aug/1997 \key Qcplx_med_medcl \shd Medial clauses \fln For Help jump to: \cf Hcplx_med_medcl \ftx \q What is the structure of 'medial' clauses, if they are a distint feature of your language? \ftx \dt 27/Apr/1998 \key Qcplx_med_sw \shd Switch reference systems \fln For Help jump to: \cf Hcplx_med_sw \ftx \q Are verbs in your language marked to show whether succeeding subjects in clause chains are the same or different? \ftx \dt 27/Apr/1998 \key Qcplx_other \shd Other Complex Constructions \ftx \q Are there other kinds of dependent clauses? (e.g. a compendium of the ways in which various kinds of nominalization accomplish functions normally associated with dependent clauses). \ftx \dt 27/Apr/1998 \key Qcplx_relcl \shd Relative clauses--Introduction \fln For Help jump to: \cf Hcplx_relcl \ftx \q What kind of relative clauses does the language have? \q a. Prenominal \q b. Postnominal \q c. Internally headed \q d. Headless \q e. Corelative \ftx \q What positions on the following relativizability hierarchy can be relativized? \ftx \ftx Subject > Direct object > Indirect object > Oblique > Possessor \ftx \q What strategy is used for each position? \ftx \dt 27/Apr/1998 \key Qcplx_relcl_elem \shd Clause elements capable of being relativized \fln For Help jump to: \cf Hcplx_relcl_elem \ftx \q Which elements in a clause may be relativized in your language? \ftx \dt 20/Aug/1997 \key Qcplx_relcl_hdless \shd Relative clauses--Headless examples \fln For Help jump to: \cf Hcplx_relcl_hdless \ftx \q Are there examples of headless relative clauses in your language? \ftx \dt 20/Aug/1997 \key Qcplx_relcl_inthd \shd Relative clauses--Internally headed examples \fln For Help jump to: \cf Hcplx_relcl_inthd \ftx \q Does your language utilize internally headed relative clauses? \ftx \dt 20/Aug/1997 \key Qcplx_relcl_np \shd Noun phrase encoding in relative clauses, verb medial clauses \fln For Help jump to: \cf Hcplx_relcl_np \ftx \q How do relative clauses encode the pertinent noun phrase so that its role is recoverable? \ftx \dt 27/Apr/1998 \key Qcplx_relcl_np_2 \shd Noun phrase encoding in relative clauses, verb initial languages \fln For Help jump to: \cf Hcplx_relcl_np_2 \ftx \q Are there examples of relative clauses in your language such that the role of the noun phrase is not recoverable from formal features? \ftx \dt 27/Apr/1998 \key Qcplx_relcl_np_3 \shd Noun phrase encoding in relative clauses--Pronoun retention \fln For Help jump to: \cf Hcplx_relcl_np_3 \ftx \q Does your language use pronoun retention in relative clauses to encode the noun phrase and avoid ambiguity? \ftx \dt 27/Apr/1998 \key Qcplx_relcl_np_4 \shd Noun phrase encoding in relative clauses--Relative pronouns \fln For Help jump to: \cf Hcplx_relcl_np_4 \ftx \q Does your language use relative pronouns in relative clauses to avoid ambiguity? \ftx \dt 27/Apr/1998 \key Qcplx_relcl_post-n \shd Relative clauses--Post-nominal examples \fln For Help jump to: \cf Hcplx_relcl_post-n \ftx \q Does your language utilize post-nominal relative clauses? \ftx \dt 20/Aug/1997 \key Qcplx_relcl_pre-n \shd Relative clauses--Pre-nominal examples \fln For Help jump to: \cf Hcplx_relcl_pre-n \ftx \q Are there any occurrences of pre-nominal relative clauses in your language? \ftx \dt 27/Apr/1998 \key Qcplx_relcl_typ \shd Relative clauses--typological parameters \fln For Help jump to: \cf Hcplx_relcl_typ \ftx \q What are the relevant typological parameters for grouping relative clauses in your language? \ftx \dt 27/Apr/1998 \key Qcplx_relcl_vscplcl \shd Headless Relative clauses versus complement clauses \fln For Help jump to: \cf Hcplx_relcl_vscplcl \ftx \q Does your language show a contrast between headless relative clauses and complement clauses? \ftx \dt 20/Aug/1997 \key Qcplx_sv \shd Serial verbs \fln For Help jump to: \cf Hcplx_sv \ftx \q Does the language have serial verbs (or 'co-verbs' in the East Asian tradition)? \ftx \q Which verbs are most likely to occur in serial constructions? \ftx \q Are there any that are losing their semantic content and becoming more like auxiliaries, adpositions, or TAM markers when they occur in serial constructions? \ftx \dt 27/Apr/1998 \key Qcplx_sv_ex \shd Serial verbs--Examples \fln For Help jump to: \cf Hcplx_sv_ex \ftx \txt For examples of serial verb constructions in various languages, jump to the help screen. \ftx \dt 27/Apr/1998 \key Qcplx_sv_extr \shd Serial verbs--Extraction \fln For Help jump to: \cf Hcplx_sv_extr \ftx \q If your language has serial verbs, does it show evidence of interaction with an extraction process? \ftx \dt 27/Apr/1998 \key Qcplx_sv_subj \shd Serial verbs--Subjects \fln For Help jump to: \cf Hcplx_sv_subj \ftx \q If your language has serial verb constructions do they affect the marking of subject? \ftx \dt 27/Apr/1998 \key Qcplx_sv_tam \shd Serial verbs--Tense/aspect/mode \fln For Help jump to: \cf Hcplx_sv_tam \ftx \txt For an example from Yoruba of the way these principles affect tense/aspect/mode in serial verbs, jump to the help screen. \ftx \dt 20/Aug/1997 \key Qcplx_sv_vsaux \shd Serial verbs versus Auxillaries \fln For Help jump to: \cf Hcplx_sv_vsaux \ftx \q Is there a clear separation between serial verbs and auxillaries in your language? \ftx \dt 27/Apr/1998 \key Qcplx_sv_vschn \shd Serial verbs versus Clause chaining \fln For Help jump to: \cf Hcplx_sv_vschn \ftx \q Are serial verb constructionss different from clause chaining? \ftx \dt 20/Aug/1997 \key Qdisc \shd ***Conclusions: the language as a vehicle for discourse \fln The following is a list of section headings in this chapter: \ftx \ftx H12.0.A Discourse analysis and linguistic analysis \ftx H12.0.B Discourse analysis versus interpretation \ftx H12.1 Continuity (cohesion) and discontinuity \ftx H12.1.A Topic Continuity \ftx H12.1.B. Action Continuity \ftx H12.1.C. Thematic Continuity \ftx H12.1.1 Topic (referential) continuity \ftx H12.1.1.A Topic continuity--Referential devices \ftx H12.1.1.B Topic continuity--Other referential devices \ftx H12.1.1.C Topic continuity--Inherent versus context imparted factors \ftx H12.1.1.D Indices of Topic continuity \ftx H12.1.2.A Thematic continuity--Hierarchy of speech \ftx H12.1.2.B Thematic continuity--Inference \ftx H12.1.2.C Thematic continuity--Examples \ftx H12.1.2.D Thematic continuity--Structural frameworks \ftx H12.1.3. Action continuity \ftx H12.1.3.A Action continuity--Procedural Text \ftx H12.1.4 Episodic Prominence \ftx H12.1.4.1 Climax/peak \ftx H12.1.4.2.A Intensification--Rhetorical questions \ftx H12.1.4.2.B Intensification--Negation \ftx H12.2 Genres \ftx H12.2.1.A Conversation I \ftx H12.2.1.B Conversation II \ftx H12.2.2 Narrative \ftx H12.2.2.1 Personal experience \ftx H12.2.2.2 Historical \ftx H12.2.2.3 Folk stories \ftx H12.2.2.4 Mythology \ftx H12.2.3 Hortatory \ftx H12.2.4 Procedural \ftx H12.2.5 Expository \ftx H12.2.6 Descriptive \ftx H12.2.7 Ritual speech \ftx H12.3.1 Idiomatic expressions / proverbs \ftx H12.3.2 Sound symbolism \ftx H12.3.3 Typological findings \ftx H12.3.4 Bibliography or references cited \ftx \txt By highlighting any of the above topics and using the jump feature you should be able to jump directly to the help notes on any of these items. Be sure your search path is set correctly, then if you highlight the section number or heading including the initial H you should go directly to the Help file. If you highlight it excluding the initial H, it should take you to that specific question. It doesn't matter how much of the rest of the section title you highlight, provided you start with either the initial H or the first digit of the section number. \ftx \dt 27/Apr/1998 \key Qdisc_bib \shd Bibliography or references cited \fln For Help jump to: \cf Hdisc_bib \ftx \q What relevant reference material have you used in studying the language and in answering these questions? \ftx \dt 27/Apr/1998 \key Qdisc_cnt \shd Continuity (cohesion) and discontinuity \fln For Help jump to: \cf Hdisc_cnt \ftx \q How does your language maintain cohesion and continuity in discourse? \ftx \dt 20/Aug/1997 \key Qdisc_cnt_act \shd Action continuity \fln For Help jump to: \cf Hdisc_cnt_act \ftx \q What kinds of devices does your language use to maintain action continuity? \q How is tense/aspect marking deployed in discourse? (Answer will probably vary according to genre. See Hdisc_genr). \q What morphosyntactic devices are used to signal the 'events' in a narrative discourse? \q What about the 'non-events', i.e. collateral descriptive material? \ftx \dt 27/Apr/1998 \key Qdisc_cnt_act_proc \shd Action continuity--Procedural Text \fln For Help jump to: \cf Hdisc_cnt_act_proc \ftx \q How might the structure of procedural or other varieties of text in your language differ from that of narrative? \ftx \dt 27/Apr/1998 \key Qdisc_cnt_prom \shd Episodic Prominence \fln For Help jump to: \cf Hdisc_cnt_prom \ftx \q What devices are used to ascribe special prominence to portions of text? \ftx \q Can you isolate the kinds of prominence that the language is sensitive to? \ftx \dt 20/Aug/1997 \key Qdisc_cnt_prom_clim \shd Climax/peak \fln For Help jump to: \cf Hdisc_cnt_prom_clim \ftx \q Are there special morphosyntactic devices characteristically used at the climax or peak of a narrative? \ftx \q Is there a recognizable peak in other genres? \ftx \dt 27/Apr/1998 \key Qdisc_cnt_prom_intns \shd Intensification \fln For Help jump to: \cf Hdisc_cnt_prom_intns \ftx \q Are rhetorical questions used as 'highlighting' devices in discourse? Give examples. \ftx \q Is negation used as a 'highlighting' device in discourse? Give examples. \ftx \dt 27/Apr/1998 \key Qdisc_cnt_ref \shd Topic (referential) Continuity \fln For Help jump to: \cf Hdisc_cnt_ref \q What devices does your language use to maintain topic continuity? \q What are the discourse functions of the various referential devices? i.e., which code highly continuous referents, and which code highly discontinuous referents? \ftx \shd2 Related questions: \q How are referents introduced into narrative and/or conversational discourse? \q Are referents introduced differently depending on whether or not they are 'destined' to figure prominently in the following text? (I.e. Does the language clearly distinguish introductions of 'discourse manipulable' referents?) \q Are there different coding devices used to introduce referents that have some honorific status? \ftx \dt 27/Apr/1998 \key Qdisc_cnt_ref_ind \shd Indices of Topic continuity \fln For Help jump to: \cf Hdisc_cnt_ref_ind \ftx \fln For a suggested list of topic continuity indices, jump to the help screen. \ftx \dt 20/Aug/1997 \key Qdisc_cnt_ref_inher \shd Topic continuity--Inherent versus context imparted factors \fln For Help jump to: \cf Hdisc_cnt_ref_inher \ftx \q Are there other relevant features relating to referential continuity in discourse in your language? \ftx \dt 27/Apr/1998 \key Qdisc_cnt_ref_other \shd Topic continuity--Other referential devices \fln For Help jump to: \cf Hdisc_cnt_ref_other \ftx \q What other devices does the language use to keep track of referents, such as indicating relative rank or importance? \ftx \dt 27/Apr/1998 \key Qdisc_cnt_ref_refd \shd Topic continuity--Referential devices \fln For Help jump to: \cf Hdisc_cnt_ref_refd \ftx \q How are new referents introduced into a discourse, or maintained once they have been introduced? \ftx \dt 27/Apr/1998 \key Qdisc_cnt_thm \shd Thematic Continuity \fln For Help jump to: \cf Hdisc_cnt_thm \ftx \q What devices does your language use to maintain thematic continuity? \ftx \dt 20/Aug/1997 \key Qdisc_cnt_thm_ex \shd Thematic continuity--Examples \fln For Help jump to: \cf Hdisc_cnt_thm_ex \ftx \txt For examples of the way inference is used to show structure in text, jump to the help screen. \ftx \dt 27/Apr/1998 \key Qdisc_cnt_thm_hier \shd Thematic continuity--Hierarchy of speech \fln For Help jump to: \cf Hdisc_cnt_thm \ftx \q How does your language handle the mismatch between the linear production of speech and the hierarchical nature of discourse? \q Are there particular morphosyntactic devices which permit the recovery of discourse structure? \ftx \dt 27/Apr/1998 \key Qdisc_cnt_thm_infr \shd Thematic continuity--Inference \fln For Help jump to: \cf Hdisc_cnt_thm_infr \ftx \q How important is inference in understanding the hierarchical structure of text in your language? \ftx \dt 27/Apr/1998 \key Qdisc_cnt_thm_strc \shd Thematic continuity--Structural frameworks \fln For Help jump to: \cf Hdisc_cnt_thm_strc \ftx \q Have you given consideration to the various possible frameworks, or the semantic principles, which might be used for describing the hierarchical structure of text in your language? \ftx \dt 27/Apr/1998 \key Qdisc_genr \shd Genres \fln For Help jump to: \cf Hdisc_genr \ftx \q What discourse genres are demonstrably distinct in this language? Exemplify and discuss the significant characteristics of each. \ftx \dt 27/Apr/1998 \key Qdisc_genr_cnvs \shd Conversation \fln For Help jump to: \cf Hdisc_genr_cnvs \ftx \q What are the characteristic features of conversational exchanges in the language? \q What devices does your language use to indicate that a speaker wants to continue speaking, or, conversely, that he has finished and wants to turn control over to someone else? \ftx \dt 27/Apr/1998 \key Qdisc_genr_desc \shd Descriptive \fln For Help jump to: \cf Hdisc_genr_desc \ftx \q Have you collected samples of descriptive texts? \txt For some suggested possibilities, jump to the help screen. \ftx \dt 27/Apr/1998 \key Qdisc_genr_exp \shd Expository \fln For Help jump to: \cf Hdisc_genr_exp \ftx \q Have you collected samples of expository texts? \txt For some suggested possibilities, jump to the help screen. \ftx \dt 27/Apr/1998 \key Qdisc_genr_hort \shd Hortatory \fln For Help jump to: \cf Hdisc_genr_hort \ftx \q Have you collected samples of hortatory texts? \txt For some suggested possibilities, jump to the help screen. \ftx \dt 27/Apr/1998 \key Qdisc_genr_nar \shd Narrative \fln For Help jump to: \cf Hdisc_genr_nar \ftx \q What are the characteristic features of narrative in the language? \ftx \dt 20/Aug/1997 \key Qdisc_genr_nar_flk \shd Folk stories \fln For Help jump to: \cf Hdisc_genr_nar_flk \ftx \q Have you collected examples of folk story narrative texts? \txt For some suggested possibilities, jump to the help screen. \ftx \dt 27/Apr/1998 \key Qdisc_genr_nar_hist \shd Historical \fln For Help jump to: \cf Hdisc_genr_nar_hist \ftx \q Have you collected samples of historical narrative texts? \txt For some suggested possibilities, jump to the help screen. \ftx \dt 27/Apr/1998 \key Qdisc_genr_nar_myth \shd Mythology \fln For Help jump to: \cf Hdisc_genr_nar_myth \ftx \q Have you collected samples of mythological narrative texts? \txt For some suggested possibilities, jump to the help screen. \ftx \dt 27/Apr/1998 \key Qdisc_genr_nar_pers \shd Personal experience \fln For Help jump to: \cf Hdisc_genr_nar_pers \ftx \q Have you collected samples of personal experience narrative texts? \txt For some suggested possibilities and how to elicit them, jump to the help screen. \ftx \dt 27/Apr/1998 \key Qdisc_genr_proc \shd Procedural \fln For Help jump to: \cf Hdisc_genr_proc \ftx \q Have you collected samples of procedural texts? For some suggested possibilities, jump to the help screen. \ftx \dt 27/Apr/1998 \key Qdisc_genr_rit \shd Ritual speech \fln For Help jump to: \cf Hdisc_genr_rit \ftx \q Have yu collected samples of ritual speech? For some suggested possibilities, jump to the help screen. \ftx \dt 27/Apr/1998 \key Qdisc_idm \shd Idiomatic expressions / proverbs \fln For Help jump to: \cf Hdisc_idm \ftx \q What kind of idiomatic expressions, metaphors, and proverbs are found in the language? \ftx \dt 27/Apr/1998 \key Qdisc_intro_intrp \shd Discourse analysis versus interpretation \fln For Help jump to: \cf Hdisc_intro_intrp \ftx \txt For a discussion of the difference between analysis of discourse and interpretation of discourse, jump to the help screen. \ftx \dt 20/Aug/1997 \key Qdisc_intro_ling \shd Discourse analysis and linguistic analysis \fln For Help jump to: \cf Hdisc_intro_ling \ftx \txt For a discussion of the relationship between language and discourse analysis, jump to the help screen. \ftx \dt 20/Aug/1997 \key Qdisc_misc \shd Miscellaneous and Conclusions \fln For Help jump to: \cf Hdisc_misc \ftx \q What else is particularly interesting about discourse in this language? \ftx \dt 20/Aug/1997 \key Qdisc_sound \shd Sound symbolism \q Does the language make extensive and productive use of sound symbolism? \q What are some common ideophones? \q How is the phonological system of ideophones and sound symbolism in general different from that of the rest of the language? \q How is the morphology different? How is the syntax different? \ftx \dt 20/Aug/1997 \key Qdisc_typ \shd Typological findings \q What are the features of this language that are particularly interesting? \q What typological surprises does it present? \q How does this work contribute to our understanding of the notion `possible human language'? \q What directions for further research do you recommend and/or plan to undertake yourself? \dt 27/Apr/1998 \key Qeth \shd ***Demographic and ethnographic information \txt This chapter presents some basic questions to help you answer the important sociolinguistic questions about the people and the language you are studying. \ftx \fln The following is a list of section headings in this chapter. \ftx \ftx H1.1.A The name of the language--External \ftx H1.1.B The name of the language--Internal \ftx H1.1.C The name of the language--Origins \ftx H1.2.A Ethnography--Economic activity \ftx H1.2.B Ethnography--Ecosystem \ftx H1.2.C Ethnography--Material culture \ftx H1.2.D Ethnography--Cosmology \ftx H1.3.A Demography--Locale \ftx H1.3.B Demography--Other groups \ftx H1.3.C Demography--Interaction \ftx H1.4.A Genetic affiliation--Family \ftx H1.4.B Genetic affiliation--Relatives \ftx H1.4.C Genetic affiliation--Published sources \ftx H1.5 Previous research \ftx H1.6 Sociolinguistic situation \ftx H1.6.1.A Multilingualism--Percentage \ftx H1.6.1.B Multilingualism--Languages \ftx H1.6.2 Contexts of use \ftx H1.6.3.A Viability--Use by children \ftx H1.6.3.B Viability--Pressures \ftx H1.6.3.C Viability--Semi-speakers \ftx H1.6.4 Loan words \ftx H1.7 Dialects \ftx \txt By highlighting any of the section numbers or topics and using the jump feature you should be able to jump directly to the help notes on any of these items. Be sure your search path is set correctly, then if you highlight the section number or heading including the initial H you should go directly to the Help file. If you highlight it excluding the initial H, it should take you to that specific question. It doesn't matter how much of the rest of the title you include in the highlight, provided you start with either the initial H or the first digit of the section number. \ftx \dt 27/Apr/1998 \key Qeth_affil_fam \shd Genetic affiliation--Family \fln For Help jump to: \cf Heth_affil_fam \ftx \q What language family does this language belong to? \ftx \dt 20/Aug/1997 \key Qeth_affil_rel \shd Genetic affiliation--Relatives \fln For Help jump to: \cf Heth_affil_rel \ftx \q What are its closest relatives? \ftx \dt 20/Aug/1997 \key Qeth_affil_src \shd Genetic affiliation--Published sources \fln For Help jump to: \cf Heth_affil_src \ftx \q What published and unpublished linguistic work has been done in this language and/or its close relatives? \ftx \dt 27/Apr/1998 \key Qeth_dem_intct \shd Demography--Interaction \fln For Help jump to: \cf Heth_dem_intct \ftx \q What is the nature of the interaction with these language groups? \ftx \dt 20/Aug/1997 \key Qeth_dem_loc \shd Demography--Locale \fln For Help jump to: \cf Heth_dem_loc \ftx \q Where is the language spoken, and how are the people distributed in this area? \ftx \dt 27/Apr/1998 \key Qeth_dem_other \shd Demography--Other groups \fln For Help jump to: \cf Heth_dem_other \ftx \q Are there other language groups inhabiting the same area? \ftx \dt 20/Aug/1997 \key Qeth_dlect \shd Dialects \fln For Help jump to: \cf Heth_dlect \ftx \q Is there significant dialect variation? \q What kinds of differences distinguish the dialects? Give examples. \q What dialect is represented in the sketch? \ftx \dt 20/Aug/1997 \key Qeth_eth_cosm \shd Ethnography--Cosmology \fln For Help jump to: \cf Heth_eth_cosm \ftx \q What is their cosmology like? \ftx \dt 20/Aug/1997 \key Qeth_eth_cult \shd Ethnography--Material culture \fln For Help jump to: \cf Heth_eth_cult \ftx \q What is their material culture like? \ftx \dt 20/Aug/1997 \key Qeth_eth_econ \shd Ethnography--Economic activity \fln For Help jump to: \cf Heth_eth_econ \ftx \q What is the dominant economic activity of the people? \ftx \dt 20/Aug/1997 \key Qeth_eth_ecosys \shd Ethnography--Ecosystem \fln For Help jump to: \cf Heth_eth_ecosys \ftx \q What is their ecosystem like? \ftx \dt 20/Aug/1997 \key Qeth_nam_ext \shd The name of the language--External \fln For Help jump to: \cf Heth_nam_ext \ftx \q What is the language known as to outsiders? \ftx \dt 20/Aug/1997 \key Qeth_nam_int \shd The name of the language--Internal \fln For Help jump to: \cf Heth_nam_int \ftx \q What term do the people use to distinguish themselves from other language groups? \ftx \dt 27/Apr/1998 \key Qeth_nam_orig \shd The name of the language--Origins \fln For Help jump to: \cf Heth_nam_orig \ftx \q What is the origin of these terms (if known)? \ftx \dt 20/Aug/1997 \key Qeth_prevres \shd Previous research \fln For Help jump to: \cf Heth_prevres \ftx \q What other research has been carried out with respect to this language? \ftx \dt 20/Aug/1997 \key Qeth_socl \shd Sociolinguistic situation \fln For Help jump to: \cf Heth_socl \txt The following sections will lead you through a number of questions pertinent to the sociolinguistic situation.*** \ftx \dt 27/Apr/1998 \key Qeth_socl_cntxt \shd Contexts of use \fln For Help jump to: \cf Heth_socl_cntxt \ftx \q In what contexts are multilingual individuals likely to use the language described in this sketch? \ftx \dt 27/Apr/1998 \key Qeth_socl_loan \shd Loan words \fln For Help jump to: \cf Heth_socl_loan \ftx \q Does the lexicon of this language contain many words from other languages? \q If so, in what semantic domains do these tend to occur? Give examples. \ftx \dt 27/Apr/1998 \key Qeth_socl_mult_lang \shd Multilingualism--Languages \fln For Help jump to: \cf Heth_socl_mult_lang \ftx \q What language(s) are people multilingual in, and to what degree? \ftx \dt 20/Aug/1997 \key Qeth_socl_mult_perc \shd Multilingualism--Percentage \fln For Help jump to: \cf Heth_socl_mult_perc \ftx \q What percentage of the people are monolingual? \ftx \dt 20/Aug/1997 \key Qeth_socl_via_pres \shd Viability--Pressures \fln For Help jump to: \cf Heth_socl_via_pres \ftx \q What pressures are there on young people to a) learn another language, and b) reject their own language? \q How strong are these pressures in your estimation? \ftx \dt 27/Apr/1998 \key Qeth_socl_via_semi \shd Viability--Semi-speakers \fln For Help jump to: \cf Heth_socl_via_semi \ftx \q Are there considerable numbers of people who are judged by the community to be imperfect speakers of the language? \ftx \dt 27/Apr/1998 \key Qeth_socl_via_use \shd Viability--Use by children \fln For Help jump to: \cf Heth_socl_via_use \ftx \q Are children learning the language as their first language? \q If so, how long do they remain monolingual? \ftx \dt 20/Aug/1997 \key Qgr \shd ***Grammatical Categories \fln The following is a list of section headings in this chapter. \ftx \ftx H3.0.A Minor Grammatical categories \ftx H3.1.A Nouns--General \ftx H3.1.B Nouns--Morphosyntactic criteria \ftx H3.1.C Nouns--Distributional characteristics \ftx H3.1.D Nouns--Structural characteristics \ftx H3.1.E Nouns--Changing Grammatical Category, an Example \ftx H3.1.F Nouns--Other structural characteristics of the example \ftx H3.1.G Nouns--Solutions to the example \ftx H3.1.H Nouns--Discussion of Solution 1 \ftx H3.1.I Nouns--Discussion of Solution 2 \ftx H3.1.J Nouns--Discussion of Solution 3 \ftx H3.1.K Nouns--Recommendation \ftx H3.1.1.A Types of nouns \ftx H3.1.1.B Mass and Count Nouns \ftx H3.1.1.1 Proper names \ftx H3.1.1.2 Count nouns \ftx H3.1.1.3 Mass nouns \ftx H3.1.1.4 etc. \ftx H3.1.2 The structure of the noun or noun-phrase \ftx H3.1.3.A Pronouns and/or anaphoric clitics \ftx H3.1.3.B Pronouns and/or anaphoric clitics \ftx H3.1.3.C Example of anaphora versus agreement \ftx H3.1.3.D Definitions--Pronouns \ftx H3.1.3.E Definitions--Anaphoric clitics \ftx H3.1.3.F Anaphoric clitics in Romance languages \ftx H3.1.3.G Relevant Distinctions \ftx H3.1.3.H Pronoun/Anaphoric clitic distinctions--Person \ftx H3.1.3.I Pronoun/Anaphoric clitic distinctions--Number \ftx H3.1.3.J Pronoun/Anaphoric clitic distinctions--Gender \ftx H3.1.3.K Pronoun/Anaphoric clitic distinctions--Grammatical relations \ftx H3.1.3.L Pronoun/Anaphoric clitic distinctions--Semantic roles \ftx H3.1.3.M Pronoun/Anaphoric clitic distinctions--Definiteness/specificity \ftx H3.1.3.N Pronoun/Anaphoric clitic distinctions--Honorifics \ftx H3.2 Verbs \ftx H3.2.1.A Verb classes-general \ftx H3.2.1.B Verb classes-Semantic roles \ftx H3.2.1.C Verb classes-common Semantic roles \ftx H3.2.1.D Verb classes-common Semantic roles--Agent in English \ftx H3.2.1.E Verb classes-common Semantic roles--Agent in Guaymí \ftx H3.2.1.F Verb classes-common Semantic roles--Force \ftx H3.2.1.G Verb classes-common Semantic roles--Instrument \ftx H3.2.1.H Verb classes-common Semantic roles--Experiencer \ftx H3.2.1.I Verb classes-common Semantic roles--Recipient \ftx H3.2.1.J Verb classes-common Semantic roles-Patient \ftx H3.2.1.K Verbs--Encoding Semantic Roles \ftx H3.2.1.L Verb classes-Semantic roles versus grammatical relations \ftx H3.2.1.M Verb classes-Universal semantic roles \ftx H3.2.1.N Verb classes-Examples from Guaymí \ftx H3.2.1.O Verb classes-Examples from Gujarati \ftx H3.2.1.P Verb classes-Examples from Tagalog \ftx H3.2.1.2 Weather verbs \ftx H3.2.1.3 States \ftx H3.2.1.4 Involuntary processes \ftx H3.2.1.5 Bodily functions \ftx H3.2.1.6.A Motion \ftx H3.2.1.6.B Locomotion \ftx H3.2.1.6.C Motion--verbal operators \ftx H3.2.1.7 Position \ftx H3.2.1.8 Actions \ftx H3.2.1.9 Action processes \ftx H3.2.1.10 Factives \ftx H3.2.1.11 Cognition \ftx H3.2.1.12 Sensation \ftx H3.2.1.13 Emotion \ftx H3.2.1.14 Utterance \ftx H3.2.1.15 Manipulation \ftx H3.2.2.A Verb or verb-phrase structure \ftx H3.2.2.B Verb or verb-phrase structure \ftx H3.2.2.C Verb or verb-phrase structure \ftx H3.2.2.D Verb or verb-phrase structure \ftx H3.2.2.E Questions to answer for all verbal operations \ftx H3.3 Modifiers \ftx H3.3.1.A Descriptive adjectives--form \ftx H3.3.1.B Descriptive adjectives--agreement \ftx H3.3.1.C Descriptive adjectives--semantic classes \ftx H3.3.1.1 Size \ftx H3.3.1.2 Shape \ftx H3.3.1.3 Color \ftx H3.3.1.4 Value (good/bad) \ftx H3.3.1.5 Texture \ftx H3.3.1.6 Functionality \ftx H3.3.1.7 Other \ftx H3.3.2 Non-numeral quantifiers \ftx H3.3.3.A Numerals--system used \ftx H3.3.3.B Numerals--extent \ftx H3.3.3.C Numerals--agreement \ftx H3.4.A Adverbs--General \ftx H3.4.B Adverbs versus complement-taking verbs \ftx H3.4.1 Manner \ftx H3.4.2 Time \ftx H3.4.3 Direction/location \ftx H3.4.4 Evidential/epistemic \ftx \txt By highlighting any of the above topics and using the jump feature you should be able to jump directly to the help notes on any of these items. Be sure your search path is set correctly, then if you highlight the section number or heading including the initial H you should go directly to the Help file. If you highlight it excluding the initial H, it should take you to that specific question. It doesn't matter how much of the rest of the section title you include in the highlight, provided you start with either the initial H or the first digit of the section number. \ftx \dt 27/Apr/1998 \key Qgrel \shd ***Grammatical relations \fln The following is a list of section headings in this chapter: \ftx \ftx H7.0.1.A Grammatical relations--Introduction \ftx H7.0.1.B Grammatical relations--pragmatic status \ftx H7.0.1.C Grammatical relations--functional approach \ftx H7.0.1.D Grammatical relations--less prototypical cases \ftx H7.0.1.E Grammatical relations--core categories \ftx H7.0.2.A Valency \ftx H7.0.2.B Encoding grammatical relations \ftx H7.0.3.A System type--nominative/accusative \ftx H7.0.3.B System type--ergative/absolutive \ftx H7.0.3.C System type--examples I \ftx H7.0.3.D System type--other possibilities \ftx H7.0.3.E System type--examples II \ftx H7.0.3.F System type--examples III \ftx H7.0.3.G System type--constituent order \ftx H7.0.3.H System type--other possibilities \ftx H7.0.3.I System type--rationale I \ftx H7.0.3.J System type--rationale II \ftx H7.0.3.K System type--ergativity characterized \ftx H7.0.3.L System type--ergativity in English \ftx H7.0.4 Split systems \ftx H7.0.4.A Split systems--intransitive \ftx H7.0.4.A.1 Split systems--intransitive examples I \ftx H7.0.4.A.2 Split systems--intransitive examples II \ftx H7.0.4.A.3 Split systems--intransitive--other possibilities \ftx H7.0.4.A.4 Split systems--intransitive examples III \ftx H7.0.4.A.5 Split systems--intransitive--Fluid-S languages \ftx H7.0.4.B.1 Split systems--transitive--agent/patient distinction \ftx H7.0.4.B.2 Split systems--transitive--agentive hierarchy \ftx H7.0.4.B.3 Split systems--transitive--agreement and pronouns \ftx H7.0.4.B.4 Split systems--transitive--topicality \ftx H7.0.4.B.5 Split systems--transitive--examples I \ftx H7.0.4.B.6 Split systems--noun phrases in transitive clauses \ftx H7.0.4.B.7 Split systems--transitive--solutions \ftx H7.0.4.B.8 Split systems--transitive--examples II \ftx H7.0.4.B.9 Split systems--transitive--examples III \ftx H7.0.4.B.10 Split systems--transitive--examples IV \ftx H7.0.4.B.11 Split systems--transitive--examples V \ftx H7.0.4.B.12 Split systems--transitive--examples VI \ftx H7.0.4.B.13 Split systems--transitive--examples VII \ftx H7.0.4.B.14 Split systems--transitive--summary \ftx H7.0.4.C Split systems--tense/aspect \ftx H7.0.4.C.1 Split systems--tense/aspect--Examples I \ftx H7.0.4.C.2 Split systems--tense/aspect--Explanation \ftx H7.0.4.C.3 Split systems--tense/aspect--Examples II \ftx H7.0.4.C.4 Split systems--tense/aspect--Examples III \ftx H7.0.4.D Split systems--Summary \ftx H7.0.5 Syntactic ergativity \ftx H7.0.5.A Syntactic ergativity--the process I \ftx H7.0.5.B Syntactic ergativity--the process II \ftx H7.0.5.C Syntactic ergativity--Summary \ftx H7.0.5.D Syntactic ergativity--Example I \ftx H7.0.5.E Syntactic ergativity--conjunction reduction \ftx H7.0.5.F Syntactic ergativity--Example II \ftx H7.1 Case marking \ftx H7.2 Verb coding \ftx H7.3 Constituent order \ftx H7.4 Agentivity/Empathy Hierarchies \ftx H7.5 Summary \ftx \txt By highlighting any of the above topics and using the jump feature you should be able to jump directly to the help notes on any of these items. Be sure your search path is set correctly, then if you highlight the section number or heading including the initial H you should go directly to the Help file. If you highlight it excluding the initial H, it should take you to that specific question. It doesn't matter how much of the rest of the section title you highlight, provided you start with either the initial H or the first digit of the section number. \ftx \dt 27/Apr/1998 \key Qgrel_ag/emp \shd Agentivity/Empathy Hierarchies \ftx \q Are there situations in which only pragmatic knowledge distinguishes A from P in transitive clauses? Describe those situations and give examples. \q How prominent are these situations and how does the language deal with them? \ftx \dt 27/Apr/1998 \key Qgrel_case \shd Case marking \fln For Help jump to: \cf Hgrel_case \ftx \q Does the language have morphological cases? If so, how do the cases function? \q Are they primarily sensitive to semantic roles, pragmatic roles or grammatical relations? \ftx \q What kind of system governs morphological case marking? Nominative/accusative, ergative/absolutive or a split system? Give examples (answer may vary depending on tense/aspect or clause type, e.g. dependent vs. independent, etc.). Include pronominal forms as well as case markers on full noun phrases. \ftx \dt 27/Apr/1998 \key Qgrel_erg \shd Syntactic ergativity \fln For Help jump to: \cf Hgrel_erg \ftx \q Does your language show evidence of 'syntactic' ergativity? \ftx \dt 20/Aug/1997 \key Qgrel_erg_cnj \shd Syntactic ergativity--conjunction reduction \fln For Help jump to: \cf Hgrel_erg_cnj \ftx \q Does your language show evidence of conjunction reduction? \ftx \dt 20/Aug/1997 \key Qgrel_erg_ex1 \shd Syntactic ergativity--Example I \fln For Help jump to: \cf Hgrel_erg_ex1 \ftx \fln For an example from Tagalog, jump to the help screen. \ftx \dt 20/Aug/1997 \key Qgrel_erg_ex2 \shd Syntactic ergativity--Example II \fln For Help jump to: \cf Hgrel_erg_ex2 \ftx \fln For an example from Yup'ik Eskimo, jump to the help screen. \ftx \dt 20/Aug/1997 \key Qgrel_erg_sum \shd Syntactic ergativity--Summary \fln For Help jump to: \cf Hgrel_erg_sum \ftx \txt For a summary of syntactic processes and a chart of the possibilities, jump to the help screen. \ftx \dt 27/Apr/1998 \key Qgrel_erg_1 \shd Syntactic ergativity--the process I \fln For Help jump to: \cf Hgrel_erg_1 \ftx \txt For further discussion of the syntactic process for a nominative/accusative system, jump to the help screen. \ftx \dt 20/Aug/1997 \key Qgrel_erg_2 \shd Syntactic ergativity--the process II \fln For Help jump to: \cf Hgrel_erg_2 \ftx \txt For discussion of the syntactic process for a ergative/absolutive system, jump to the help screen. \ftx \dt 27/Apr/1998 \key Qgrel_intro \shd Grammatical relations--Introduction \fln For Help jump to: \cf Hgrel_intro \ftx \q Is the notion of (a) 'subject' and (b) direct object viable and useful in this language? \q If so, how is it defined? That is, what are the distinctive morphosyntactic properties of subject/direct object nominals? \ftx \shd2 Pragmatic status \fln For Help jump to: \cf Hgrel_intro_prag \ftx \q Are there special nominal markers for indicating the semantic role of core arguments? (e.g. dative case for experiencers, locative case for cognizers, etc. \ftx \shd2 Functional approach \fln For Help jump to: \cf Hgrel_intro_func \ftx \q What is the relationship between pragmatic/semantic roles and grammatical relations? \ftx \shd2 Less prototypical cases \fln For Help jump to: \cf Hgrel_intro_lprot \ftx \q What happens when a noun phrase is less than a prototypical agent or topic? \ftx \shd2 Core categories \fln For Help jump to: \cf Hgrel_intro_core \ftx \q Are there special verbal forms that identify the semantic roles of the core nominal arguments in a clause? \ftx \dt 27/Apr/1998 \key Qgrel_ord \shd Constituent order \ftx \q Does constituent order enter into the determination of grammatical relations? \ftx \dt 27/Apr/1998 \key Qgrel_splt \shd Split systems \fln For Help jump to: \cf Hgrel_splt \ftx \q Is there a split system for organizing grammatical relations? \q If so, what determines the split? (i.e. under what circumstances does the nominative/accusative system occur and under what circumstances does the ergative/absolutive system occur?) \q a. Does the system for pronouns operate on the same basis as that of full NPs? \q b. Are there different GR marking systems depending on the predicate type? \q c. Are there different GR marking systems depending on the tense and/or aspect of the clause? \ftx \dt 27/Apr/1998 \key Qgrel_splt_intr \shd Split systems--intransitive \fln For Help jump to: \cf Hgrel_splt_intr \ftx \q Do some verbs inflect differently if the subject acts with volition (on purpose), versus accidentally? \ftx \shd2 Other possibilities \fln For Help jump to: \cf Hgrel_splt_intr_other \ftx \q Does your language use Split-S marking of intransitive verbs? \ftx \shd2 Fluid-S languages \fln For Help jump to: \cf Hgrel_splt_intr_fluid \ftx \q Does your language exhibit any of the characteristics of Fluid-S languages? \ftx \dt 27/Apr/1998 \key Qgrel_splt_sum \shd Split systems--Summary \fln For Help jump to: \cf Hgrel_splt_sum \ftx \txt For a summary of split systems and the way they organize grammatical relations, jump to the help screen. \ftx \dt 20/Aug/1997 \key Qgrel_splt_tr_agr \shd Split systems--transitive--agreement and pronouns \fln For Help jump to: \cf Hgrel_splt_tr_agr \ftx \txt For a discussion of the relationship of the agentivity hierarchy and the use of verb agreement and pronouns, jump to the help screen. \ftx \dt 27/Apr/1998 \key Qgrel_splt_tr_ag/emp \shd Split systems--transitive--agentive hierarchy \fln For Help jump to: \cf Hgrel_splt_tr_ag/emp \ftx \fln For a discussion of the agentive hierarchy, jump to the help screen. \ftx \dt 20/Aug/1997 \key Qgrel_splt_tr_a/p \shd Split systems--transitive--agent/patient distinction \fln For Help jump to: \cf Hgrel_splt_tr_a/p \ftx \q How does your language use pragmatic inference to distinguish between Agent and Patient? \ftx \dt 27/Apr/1998 \key Qgrel_splt_tr_np \shd Split systems--noun phrases in transitive clauses \fln For Help jump to: \cf Hgrel_splt_tr_np \ftx \q How are noun phrases marked for agentivity in transitive clauses in the language? \ftx \dt 27/Apr/1998 \key Qgrel_splt_tr_top \shd Split systems--transitive--topicality \fln For Help jump to: \cf Hgrel_splt_tr_top \ftx \q Does your language use agreement markers or pronouns to encode topicality? \ftx \dt 27/Apr/1998 \key Qgrel_splt_t/a \shd Split systems--tense/aspect \fln For Help jump to: \cf Hgrel_splt_t/a \ftx \q Does the way grammatical relations are marked in the language depend on particular tense/aspect categories? \ftx \dt 27/Apr/1998 \key Qgrel_splt_t/a_expl \shd Split systems--tense/aspect--Explanation \fln For Help jump to: \cf Hgrel_splt_t/a_expl \ftx \txt For a proposed explanation of the previous example from a functional point of view, jump to the help screen. \ftx \dt 20/Aug/1997 \key Qgrel_splt_t/a_ex1 \shd Split systems--tense/aspect--Examples I \fln For Help jump to: \cf Hgrel_splt_t/a_ex1 \ftx \fln For an example from Georgian, jump to the help screen. \ftx \dt 20/Aug/1997 \key Qgrel_splt_t/a_ex2 \shd Split systems--tense/aspect--Examples II \fln For Help jump to: \cf Hgrel_splt_t/a_ex2 \ftx \fln For an example from Newari, jump to the help screen. \ftx \dt 20/Aug/1997 \key Qgrel_splt_t/a_ex3 \shd Split systems--tense/aspect--Examples III \fln For Help jump to: \cf Hgrel_splt_t/a_ex3 \ftx \fln For an example from Guaymí, jump to the help screen. \ftx \dt 20/Aug/1997 \key Qgrel_sum \shd Summary \fln For Help jump to: \cf Hgrel_sum \ftx \txt For a chart summarizing the definitions of various terms, jump to the help screen. \ftx \dt 20/Aug/1997 \key Qgrel_typ_erg/abs \shd System type--ergative/absolutive \fln For Help jump to: \cf Hgrel_typ_erg/abs \ftx \q Is the system of grammatical relations in basic clauses organized on an ergative/absolutive, or some other system than nominative/accusative? (Be sure to provide evidence from case marking, and/or verb coding, etc. depending on the characteristics of the language.) \ftx \dt 27/Apr/1998 \key Qgrel_typ_nom/acc \shd System type--nominative/accusative \fln For Help jump to: \cf Hgrel_typ_nom/acc \ftx \q Is the system of grammatical relations in basic clauses organized on a nominative/accusative system? (Be sure to provide evidence from case marking, and/or verb coding, etc. depending on the characteristics of the language.) \ftx \dt 27/Apr/1998 \key Qgrel_typ_ord \shd System type--constituent order \fln For Help jump to: \cf Hgrel_typ_ord \ftx \q Does your language use constituent order to mark grammatical relations? \ftx \dt 27/Apr/1998 \key Qgrel_typ_ord_other \shd System type--other possibilities \fln For Help jump to: \cf Hgrel_typ_ord_other \ftx \q Is the system of grammatical relations in basic clauses organized on the basis of some other system than nominative/accusative or ergative/absolutive? (Be sure to provide evidence from case marking, and/or verb coding, etc. depending on the characteristics of the language.) \ftx \dt 27/Apr/1998 \key Qgrel_val \shd Valency \fln For Help jump to: \cf Hgrel_val \ftx \q Exemplify some simple uni-valent ('intransitive') divalent ('transitive') and trivalent ('ditransitive') clauses. Trivalent clauses may not unequivocally exist. \ftx \shd2 Encoding grammatical relations \fln For Help jump to: \cf Hgrel_val_enc \ftx \q How are the grammatical relations A, S, and P coded? \ftx \dt 27/Apr/1998 \key Qgrel_vcode \shd Verb coding \ftx \q Do verbs contain operators that vary with the person and/or number of certain verbal arguments? \ftx \q If so, what kind of system is manifested by the various verb coding paradigms? \ftx \dt 27/Apr/1998 \key Qgr_adv_dir/loc \shd Direction/location \fln For Help jump to: \cf Hgr_adv_dir/loc \ftx \q What adverbs of direction/location are there? \ftx \dt 20/Aug/1997 \key Qgr_adv_ev/epis \shd Evidential/epistemic \fln For Help jump to: \cf Hgr_adv_ev/epis \ftx \q What evidential/epistemic adverbs are there? \ftx \dt 20/Aug/1997 \key Qgr_adv_gen \shd Adverbs--General \fln For Help jump to: \cf Hgr_adv_gen \ftx \q What characterizes a form as being an adverb in this language? \ftx \dt 20/Aug/1997 \key Qgr_adv_man \shd Manner \fln For Help jump to: \cf Hgr_adv_man \ftx \q What adverbs of manner are there? \ftx \dt 20/Aug/1997 \key Qgr_adv_time \shd Time \fln For Help jump to: \cf Hgr_adv_time \ftx \q What adverbs of time are there? \ftx \dt 20/Aug/1997 \key Qgr_adv_vscpl \shd Adverbs versus complement-taking verbs \fln For Help jump to: \cf Hgr_adv_vscpl \ftx \q Are any of these classes of adverbs related to older complement-taking (matrix) verbs? \ftx \dt 27/Apr/1998 \key Qgr_cat \shd Grammatical categories \fln For Help jump to: \cf Hgr_cat \ftx \q What are the formally distinct grammatical categories of this language? \ftx \dt 20/Aug/1997 \key Qgr_cat_min \shd Minor Grammatical categories \ftx \dt 20/Aug/1997 \key Qgr_mod \shd Modifiers \fln For Help jump to: \cf Hgr_mod \ftx \q What kinds of modifiers occur in the language? \ftx \dt 20/Aug/1997 \key Qgr_mod_adj_agr \shd Descriptive adjectives--agreement \ftx \q Do adjectives agree with their heads? (e.g. in number, case and/or noun class) \ftx \dt 27/Apr/1998 \key Qgr_mod_adj_col \shd Color \fln For Help jump to: \cf Hgr_mod_adj_col \ftx \q What adjectives of color are there? \ftx \dt 20/Aug/1997 \key Qgr_mod_adj_form \shd Descriptive adjectives--form \fln For Help jump to: \cf Hgr_mod_adj_form \ftx \q What characterizes a form as being an adjective in this language? \ftx \dt 20/Aug/1997 \key Qgr_mod_adj_func \shd Functionality \fln For Help jump to: \cf Hgr_mod_adj_func \ftx \q What adjectives of functionality are there? \ftx \dt 20/Aug/1997 \key Qgr_mod_adj_other \shd Other \q What other kinds of adjectives are there? \ftx \dt 20/Aug/1997 \key Qgr_mod_adj_sem \shd Descriptive adjectives--semantic classes \fln For Help jump to: \cf Hgr_mod_adj_sem \ftx \q How can you characterize semantically the class of concepts coded by this formal category? \ftx \dt 27/Apr/1998 \key Qgr_mod_adj_shp \shd Shape \fln For Help jump to: \cf Hgr_mod_adj_shp \ftx \q What adjectives of shape are there? \ftx \dt 20/Aug/1997 \key Qgr_mod_adj_size \shd Size \ftx \q What adjectives of size are there? \ftx \key Qgr_mod_adj_txt \shd Texture \fln For Help jump to: \cf Hgr_mod_adj_txt \ftx \q What adjectives are there indicating texture? \ftx \dt 20/Aug/1997 \key Qgr_mod_adj_val \shd Value (good/bad) \fln For Help jump to: \cf Hgr_mod_adj_val \ftx \q What adjectives indicating value are there? \ftx \dt 20/Aug/1997 \key Qgr_mod_non-num \shd Non-numeral quantifiers \ftx \q What non-numeral quantifiers are there? \ftx \key Qgr_mod_num_agr \shd Numerals--agreement \ftx \q Do numerals agree with their head nouns? (e.g. in number, case, and/or noun class) \ftx \dt 27/Apr/1998 \key Qgr_mod_num_ext \shd Numerals--extent \fln For Help jump to: \cf Hgr_mod_num_ext \ftx \q How high can a fluent native speaker count without resorting either to words from another language or to a generic word like many? Exemplify the system up to this point. \ftx \dt 27/Apr/1998 \key Qgr_mod_num_sys \shd Numerals--system used \fln For Help jump to: \cf Hgr_mod_num_sys \ftx \q What kind of system does the language employ for counting? \ftx \dt 20/Aug/1997 \key Qgr_n_dist \shd Nouns--Distributional characteristics \fln For Help jump to: \cf Hgr_n_dist \ftx \q What are the distributional characteristics of nouns? \ftx \dt 20/Aug/1997 \key Qgr_n_ex_chgcat \shd Nouns--Changing Grammatical Category, an Example \fln For Help jump to: \cf Hgr_n_ex_chgcat \ftx \txt If you would like to see an illustrative example of how to determine the categorical status of a derived form check the help file corresponding to this entry. \ftx \dt 27/Apr/1998 \key Qgr_n_ex_other \shd Nouns--Other structural characteristics of the example \fln For Help jump to: \cf Hgr_n_ex_other \ftx \txt If you would like to see some other structural characteristics of nouns, check the help file corresponding to this entry. \ftx \dt 27/Apr/1998 \key Qgr_n_ex_sol \shd Nouns--Solutions to the example \fln For Help jump to: \cf Hgr_n_ex_sol \ftx \txt If you would like to see some possible solutions to the 'walking' example check the help file corresponding to this entry. \ftx \dt 27/Apr/1998 \key Qgr_n_ex_sol_rec \shd Nouns--Recommendation \fln For Help jump to: \cf Hgr_n_ex_sol_rec \ftx \txt For a discussion of the recommended approach to problem cases, see the help file corresponding to this entry. \ftx \dt 27/Apr/1998 \key Qgr_n_ex_sol_1 \shd Nouns--Discussion of Solution 1 \fln For Help jump to: \cf Hgr_n_ex_sol_1 \ftx \txt For a discussion of Solution 1, see the help file corresponding to this entry. \ftx \dt 27/Apr/1998 \key Qgr_n_ex_sol_2 \shd Nouns--Discussion of Solution 2 \fln For Help jump to: \cf Hgr_n_ex_sol_2 \ftx \txt For a discussion of solution 2 see the help file corresponding to this entry. \ftx \dt 27/Apr/1998 \key Qgr_n_ex_sol_3 \shd Nouns--Discussion of Solution 3 \fln For Help jump to: \cf Hgr_n_ex_sol_3 \ftx \txt For a discussion of solution 3, see the help file corresponding to this entry. \ftx \dt 27/Apr/1998 \key Qgr_n_gen \shd Nouns--General \fln For Help jump to: \cf Hgr_n_gen \q How is a NOUN defined in your language? \ftx \dt 20/Aug/1997 \key Qgr_n_msyn \shd Nouns--Morphosyntactic criteria \fln For Help jump to: \cf Hgr_n_msyn \ftx \q Have you considered the morphosyntactic criteria for nounhood in the language? \ftx \dt 27/Apr/1998 \key Qgr_n_n/npstr \shd The structure of the noun or noun-phrase \fln For Help jump to: \cf Hgr_n_n/npstr \ftx \q What is the basic structure of the noun word (for polysynthetic languages) and/or noun phrase (for more isolating languages)? \ftx \dt 27/Apr/1998 \key Qgr_n_prn/clt_def_clt \shd Definitions--Anaphoric clitics \fln For Help jump to: \cf Hgr_n_prn/clt_def_clt \ftx \fln For a definition of Anaphoric clitics, see the corresponding help file. \ftx \dt 20/Aug/1997 \key Qgr_n_prn/clt_def_clt_ex \shd Anaphoric clitics in Romance languages \fln For Help jump to: \cf Hgr_n_prn/clt_def_clt_ex \ftx \txt For an example of the use of anaphoric clitics in Romance languages see the corresponding help file. \ftx \dt 27/Apr/1998 \key Qgr_n_prn/clt_def_pron \shd Definitions--Pronouns \fln For Help jump to: \cf Hgr_n_prn/clt_def_pron \ftx \fln For a definition of pronouns, see the corresponding help file. \ftx \dt 20/Aug/1997 \key Qgr_n_prn/clt_dist \shd Relevant Distinctions \fln For Help jump to: \cf Hgr_n_prn/clt_dist \ftx \q Is the pronoun/anaphoric clitic distinction relevant in the language? \ftx \dt 20/Aug/1997 \key Qgr_n_prn/clt_dist_def \shd Pronoun/Anaphoric clitic distinctions--Definiteness/specificity \ftx \q What distinctions for definiteness/specificity are made in the language? \ftx \key Qgr_n_prn/clt_dist_gen \shd Pronoun/Anaphoric clitic distinctions--Gender \fln For Help jump to: \cf Hgr_n_prn/clt_dist_gen \ftx \q What distinctions for gender are made in the language? \ftx \key Qgr_n_prn/clt_dist_grel \shd Pronoun/Anaphoric clitic distinctions--Grammatical relations \ftx \q What distinctions for grammatical relations are made in the language? \ftx \key Qgr_n_prn/clt_dist_hon \shd Pronoun/Anaphoric clitic distinctions--Honorifics \fln For Help jump to: \cf Hgr_n_prn/clt_dist_hon \ftx \q What distinctions for honorifics are made in the language? \ftx \dt 20/Aug/1997 \key Qgr_n_prn/clt_dist_num \shd Pronoun/Anaphoric clitic distinctions--Number \fln For Help jump to: \cf Hgr_n_prn/clt_dist_num \ftx \q What distinctions for number are made in the language? \ftx \dt 20/Aug/1997 \key Qgr_n_prn/clt_dist_pers \shd Pronoun/Anaphoric clitic distinctions--Person \fln For Help jump to: \cf Hgr_n_prn/clt_dist_pers \ftx \q What distinctions for person are made in the language? \ftx \dt 20/Aug/1997 \key Qgr_n_prn/clt_dist_role \shd Pronoun/Anaphoric clitic distinctions--Semantic roles \fln For Help jump to: \cf Hgr_n_prn/clt_dist_role \ftx \q What distinctions for semantic roles are made in the language? \ftx \dt 20/Aug/1997 \key Qgr_n_prn/clt_intro \shd Pronouns and/or anaphoric clitics \fln For Help jump to: \cf Hgr_n_prn/clt_intro \ftx \q Does the language have free pronouns and/or anaphoric clitics? If so, create a chart of them. \ftx \dt 27/Apr/1998 \key Qgr_n_prn/clt_vsagr \shd Example of anaphora versus agreement \fln For Help jump to: \cf Hgr_n_prn/clt_vsagr \ftx \q Is there a contrast between anaphora and agreement in the language? \ftx \dt 20/Aug/1997 \key Qgr_n_strc \shd Nouns--Structural characteristics \fln For Help jump to: \cf Hgr_n_strc \ftx \q What are the structural characteristics of nouns? \ftx \dt 20/Aug/1997 \key Qgr_n_typ \shd Types of nouns \fln For Help jump to: \cf Hgr_n_typ \ftx \q What are the major formally distinct subcategories of nouns? \ftx \dt 20/Aug/1997 \key Qgr_n_typ_count \shd Count nouns \ftx \q What are the distinctive characteristics of count nouns? \ftx \key Qgr_n_typ_mass \shd Mass nouns \ftx \q What are the distinctive characteristics of mass nouns? \ftx \key Qgr_n_typ_m/cdist \shd Mass and Count Noun distinction \fln For Help jump to: \cf Hgr_n_typ \ftx \q Is there a distinction between mass and count nouns in the language? \ftx \dt 20/Aug/1997 \key Qgr_n_typ_other \shd Other subcategories of noun. \fln For Help jump to: \cf Hgr_n_typ_other \ftx \q What are the distinctive characteristics of other subcategories of nouns? \ftx \dt 27/Apr/1998 \key Qgr_n_typ_prop \shd Proper names \ftx \q What are the distinctive characteristics of proper names? \ftx \key Qgr_v \shd Verbs \fln For Help jump to: \cf Hgr_v \ftx \q What are the morphosyntactic characteristics of verbs? \ftx \dt 20/Aug/1997 \key Qgr_v_cl \shd Verb classes-general \fln For Help jump to: \cf Hgr_v_cl \ftx \q What are the major subclasses of verbs? \ftx \dt 19/Aug/1997 \key Qgr_v_cl_actn \shd Actions \fln For Help jump to: \cf Hgr_v_cl_actn \ftx \q Are there special action verbs? \ftx \dt 19/Aug/1997 \key Qgr_v_cl_actproc \shd Action processes \fln For Help jump to: \cf Hgr_v_cl_actproc \ftx \q Are there special verbs for certain action processes? \ftx \dt 19/Aug/1997 \key Qgr_v_cl_bod \shd Bodily functions \fln For Help jump to: \cf Hgr_v_cl_bod \ftx \q Are there specific verbs to describe bodily functions? \ftx \dt 20/Aug/1997 \key Qgr_v_cl_bod_loc \shd Locomotion \fln For Help jump to: \cf Hgr_v_cl_bod_loc \ftx \q Are there special locomotion verbs as opposed to normal motion verbs? \ftx \dt 20/Aug/1997 \key Qgr_v_cl_bod_mot \shd Motion \fln For Help jump to: \cf Hgr_v_cl_bod_mot \ftx \q Are there special motion verbs? \ftx \dt 20/Aug/1997 \key Qgr_v_cl_bod_mot_op \shd Motion--verbal operators \fln For Help jump to: \cf Hgr_v_cl_bod_mot_op \ftx \txt If you wish to see an example of how motion verbs are produced using verbal operators see the corresponding help screen. \ftx \dt 27/Apr/1998 \key Qgr_v_cl_cogn \shd Cognition \fln For Help jump to: \cf Hgr_v_cl_cogn \ftx \q Are there special verbs of cognition? \ftx \dt 19/Aug/1997 \key Qgr_v_cl_emot \shd Emotion \fln For Help jump to: \cf Hgr_v_cl_emot \ftx \q Are there special verbs of Emotion? \ftx \dt 19/Aug/1997 \key Qgr_v_cl_ex1 \shd Verb classes-Examples from Guaymí \fln For Help jump to: \cf Hgr_v_cl_ex1 \ftx \txt For examples of how semantic roles are encoded in Guaymí see the corresponding help screen. \ftx \dt 27/Apr/1998 \key Qgr_v_cl_ex2 \shd Verb classes-Examples from Gujarati \fln For Help jump to: \cf Hgr_v_cl_ex2 \ftx \txt For examples of ways semantic roles are treated in Gujarati see the corresponding help screen. \ftx \dt 19/Aug/1997 \key Qgr_v_cl_ex3 \shd Verb classes-Examples from Tagalog \fln For Help jump to: \cf Hgr_v_cl_ex3 \ftx \txt For examples of how semantic roles are treated in Tagalog see the corresponding help screen. \ftx \key Qgr_v_cl_fact \shd Factives \fln For Help jump to: \cf Hgr_v_cl_fact \ftx \q Are there special factive verbs? \ftx \dt 19/Aug/1997 \key Qgr_v_cl_manip \shd Manipulation \fln For Help jump to: \cf Hgr_v_cl_manip \ftx \q Are there special verbs of manipulation? \ftx \dt 19/Aug/1997 \key Qgr_v_cl_pos \shd Position \fln For Help jump to: \cf Hgr_v_cl_pos \ftx \q Are there special verbs of position? \ftx \dt 19/Aug/1997 \key Qgr_v_cl_sem \shd Verb classes-Semantic roles \fln For Help jump to: \cf Hgr_v_cl_sem \ftx \q What semantic roles are found in this language? \ftx \dt 19/Aug/1997 \key Qgr_v_cl_sem_com \shd Verb classes-common Semantic roles \fln For Help jump to: \cf Hgr_v_cl_sem_com \ftx \txt For a brief overview of semantic roles found in most languages see the corresponding help screen. \ftx \dt 27/Apr/1998 \key Qgr_v_cl_sem_com_ag \shd Verb classes-common Semantic roles-Agent \fln For Help jump to: \cf Hgr_v_cl_sem_com_ag \ftx \txt For an example of how Agent functions in two languages -- English and Guaymí -- see the corresponding help screen. \ftx \key Qgr_v_cl_sem_com_exp \shd Verb classes-common Semantic roles-Experiencer \fln For Help jump to: \cf Hgr_v_cl_sem_com_exp \ftx \txt For an example of how the semantic role Experiencer is used see the corresponding help screen. \ftx \dt 19/Aug/1997 \key Qgr_v_cl_sem_com_force \shd Verb classes-common Semantic roles-Force \fln For Help jump to: \cf Hgr_v_cl_sem_com_force \ftx \txt For an example of how the semantic role Force is used see the corresponding help screen. \ftx \dt 19/Aug/1997 \key Qgr_v_cl_sem_com_instr \shd Verb classes-common Semantic roles-Instrument \fln For Help jump to: \cf Hgr_v_cl_sem_com_instr \ftx \txt For an example of how the semantic role Instrument is used see the corresponding help screen. \ftx \dt 19/Aug/1997 \key Qgr_v_cl_sem_com_pat \shd Verb classes-common Semantic roles-Patient \fln For Help jump to: \cf Hgr_v_cl_sem_com_pat \ftx \txt For an example of how the semantic role Patient is used see the corresponding help screen. \ftx \dt 19/Aug/1997 \key Qgr_v_cl_sem_com_recip \shd Verb classes-common Semantic roles-Recipient \fln For Help jump to: \cf Hgr_v_cl_sem_com_recip \ftx \txt For an example of how the semantic role Recipient is used see the corresponding help screen. \ftx \dt 19/Aug/1997 \key Qgr_v_cl_sem_encod \shd Verbs--Encoding Semantic Roles \fln For Help jump to: \cf Hgr_v_cl_sem_encod \ftx \q How does the language encode semantic roles? \ftx \dt 19/Aug/1997 \key Qgr_v_cl_sem_univ \shd Verb classes-Universal semantic roles \fln For Help jump to: \cf Hgr_v_cl_sem_univ \ftx \txt For a discussion of Universal semantic roles see the corresponding help screen \ftx \dt 27/Apr/1998 \key Qgr_v_cl_sem_vsrel \shd Verb classes-Semantic roles versus grammatical relations \fln For Help jump to: \cf Hgr_v_cl_sem_vsrel \ftx \q Do the semantic roles and grammatical relations correspond? \ftx \dt 19/Aug/1997 \key Qgr_v_cl_sens \shd Sensation \fln For Help jump to: \cf Hgr_v_cl_sens \ftx \q Are there special verbs of Sensation? \ftx \dt 19/Aug/1997 \key Qgr_v_cl_st \shd States \fln For Help jump to: \cf Hgr_v_cl_st \ftx \q Are there specific stative verbs in the language? \ftx \dt 19/Aug/1997 \key Qgr_v_cl_utt \shd Utterance \fln For Help jump to: \cf Hgr_v_cl_utt \ftx \q Are there special verbs dealing with utterance? \ftx \dt 19/Aug/1997 \key Qgr_v_cl_weath \shd Weather verbs \fln For Help jump to: \cf Hgr_v_cl_weath \ftx \q Are there distinctive weather verbs in the language? \ftx \dt 19/Aug/1997 \key Qgr_v_cl__inv \shd Involuntary processes \fln For Help jump to: \cf Hgr_v_cl__inv \ftx \q Are there specific verbs to describe involuntary processes? \ftx \dt 19/Aug/1997 \key Qgr_v_strc \shd Verb or verb-phrase structure \ftx \q Where do the various verbal operators fall within the verbal word or verb phrase? \ftx \fln For Help jump to: \cf Hgr_v_strc_A \ftx \q Are directional and/or locational notions coded in the verb or verb phrase at all? If so, do they refer to the direction/location of the subject, the object or both? \ftx \fln For Help jump to: \cf Hgr_v_strc_B \ftx \q Is there any overlap or indeterminacy between the functions of directional/locational morphosyntax and tense/aspect/mode marking? \ftx \fln For Help jump to: \cf Hgr_v_strc_C \ftx \q Are the directional/locational markers related to the verbs of motion or position (go, come, stand, ascend, descend)? \ftx \fln For Help jump to: \cf Hgr_v_strc_D \ftx \shd2 Questions to answer for all verbal operations: \ftx \fln For Help jump to: \cf Hgr_v_strc_E \ftx \q a) Is this operation obligatory? That is, does one member of the paradigm have to occur in every finite verb or verb phrase? \q b) Is it productive, i.e. can the operation be specified for all verb stems, and does it have the same meaning with each one? (nothing is fully productive, but some operations are more productive than others). \q c) Is this operation primarily coded morphologically, analytically or lexically? Are there any exceptions to the general case? \q d) Where in the verb phrase or verbal word is this operation likely to be coded? Can it occur in more than one place? \ftx \dt 27/Apr/1998 \key Qmor \shd ***Morphological Typology \fln The following is a list of section headings in this chapter. \ftx \ftx H2.A Synthesis and Fusion--General \ftx H2.B Synthesis and Fusion--the Morpheme \ftx H2.C Synthesis and Fusion--Conceptualization \ftx H2.D Synthesis and Fusion--Definitions \ftx H2.E Synthesis and Fusion--Allomorphs \ftx H2.F Synthesis and Fusion--Where it fits \ftx H2.G Synthesis and Fusion--Morphophonemic rules \ftx H2.1 Morphological Typology \ftx H2.1.1 Synthesis \ftx H2.1.2 Fusion \ftx H2.2.A Inflection and derivation--General \ftx H2.2.B Inflection and derivation--Derivational \ftx H2.2.C Inflection and derivation--Inflectional \ftx H2.3 Roots and stems \ftx H2.4 Morphological processes \ftx H2.5 Head/dependent marking \ftx \txt By highlighting any of the section numbers or topics and using the jump feature you should be able to jump directly to the help notes on any of these items. Be sure your search path is set correctly, then if you highlight the section number or heading including the initial H you should go directly to the Help file. If you highlight it excluding the initial H, it should take you to that specific question. It doesn't matter how much of the heading you include in the highlight, provided you start it either with the H or the first digit of the section number. \ftx \dt 27/Apr/1998 \key Qmor_hd/dep \shd Head / dependent marking \fln For Help jump to: \cf Hmor_hd/dep \ftx \q If the language is at all polysynthetic, is it dominantly 'head-marking', 'dependent-marking' or mixed? \q If the language is of a mixed type, where is it head marking and where is it dependent marking? \ftx \dt 27/Apr/1998 \key Qmor_infl_drv \shd Inflection and derivation--Derivational operations \fln For Help jump to: \cf Hmor_infl_drv \ftx \q What kinds of derivational affix does the language use? \txt For a description of derivational type operations, jump to the help screen. \ftx \dt 27/Apr/1998 \key Qmor_infl_gen \shd Inflection and derivation--General \fln For Help jump to: \cf Hmor_infl_gen \ftx \q Does the language make a distinction between inflectional and derivational affixes? \txt For a comment on the distinction, jump to the help screen. \ftx \dt 27/Apr/1998 \key Qmor_infl_infl \shd Inflection and derivation--Inflectional operations \fln For Help jump to: \cf Hmor_infl_infl \ftx \q What kinds of inflectional affix occur in the language? \txt For a description of inflectional type operations, jump to the help screen. \ftx \dt 27/Apr/1998 \key Qmor_mproc \shd Morphological processes \fln For Help jump to: \cf Hmor_mproc \ftx \q If the language is at all agglutinative, is it dominantly prefixing, suffixing or neither? \ftx \txt Illustrate the major and secondary patterns (including infixation, stem modification, reduplication, suprasegmental modification and suppletion). \ftx \dt 27/Apr/1998 \key Qmor_rt \shd Roots and stems \fln For Help jump to: \cf Hmor_rt \ftx \q Does the language have both the categories root and stem? For a discussion of the concepts root and stem, jump to the help screen. \ftx \dt 27/Apr/1998 \key Qmor_typ \shd Morphological Typology \fln For Help jump to: \cf Hmor_typ \ftx \txt For a discussion of the significance of typology for grammatical analysis, jump to the help screen. \ftx \dt 19/Aug/1997 \key Qmor_typ_fus \shd Fusion \fln For Help jump to: \cf Hmor_typ_fus \ftx \q If the language is at all polysynthetic, is it dominantly agglutinative or fusional? Give examples of its dominant pattern and any secondary patterns. \ftx \dt 27/Apr/1998 \key Qmor_typ_synth \shd Synthesis \fln For Help jump to: \cf Hmor_typ_synth \ftx \q Is the language dominantly isolating or polysynthetic? \ftx \dt 19/Aug/1997 \key Qn \shd ***Noun and noun-phrase operations \fln The following is a list of section headings in this chapter: \ftx \ftx H5.0 Nominal operations \ftx H5.1.A Compounding--Noun-noun compounds \ftx H5.1.B Compounding--Identifying \ftx H5.1.C Compounding--Noun-verb compounds \ftx H5.1.D Compounding--Productivity \ftx H5.1.E Compounding--Frequency of occurrence \ftx H5.2.A Denominalization--I \ftx H5.2.B Denominalization--II \ftx H5.2.C Denominalization--III \ftx H5.3.A Number--Singular/Plural distinctions \ftx H5.3.B Number--Other distinctions \ftx H5.3.C Number--Optional marking \ftx H5.4.A Case--Contrasting with adpositions \ftx H5.4.B Case--Examples \ftx H5.4.C Case--Semantic roles \ftx H5.5.A Articles \ftx H5.5.B Demonstratives--the class \ftx H5.5.C Demonstratives--degrees of distance \ftx H5.5.D Demonstratives--other parameters \ftx H5.5.E Article like particles \ftx H5.6.A Possessors \ftx H5.6.B Possessors--alienable vs inalienable \ftx H5.6.C Possessors--other kinds \ftx H5.6.D Possessors--constituent order \ftx H5.7.A Class (including gender) \ftx H5.7.B Class--other dimensions \ftx H5.7.C Class--relation to other constituents \ftx H5.8 Diminution/augmentation \ftx H5.9 Other operations encoded in the noun phrase \ftx \txt By highlighting any of the above topics and using the jump feature you should be able to jump directly to the help notes on any of these items. Be sure your search path is set correctly, then if you highlight the section number or heading including the initial H you should go directly to the Help file. If you highlight it excluding the initial H, it should take you to that specific question. It doesn't matter how much of the rest of the section title you include in your highlight, provided you start with either the H or the initial digit of the section number. \ftx \dt 27/Apr/1998 \key Qn_art \shd Articles \fln For Help jump to: \cf Hn_art \ftx \q Do noun phrases have articles? \q If so, are they obligatory or optional? \q Are they separate words, or bound morphemes? \ftx \dt 19/Aug/1997 \key Qn_art_dem \shd Demonstratives--the class \fln For Help jump to: \cf Hn_art_dem \ftx \q Is there a distinct class of demonstratives? \ftx \shd2 Degrees of Distance \fln For Help jump to: \cf Hn_art_dem_deg \ftx \q How many degrees of distance are there in the system of demonstratives? \ftx \shd2 Other Parameters \fln For Help jump to: \cf Hn_art_dem_other \ftx \q Are there other distinctions? \ftx \dt 19/Aug/1997 \key Qn_art_like \shd Article like particles \fln For Help jump to: \cf Hn_art_like \ftx \q Are there any other article like particles which occur in the language? \ftx \dt 19/Aug/1997 \key Qn_case \shd Case \fln For Help jump to: \cf Hn_case Case \q Do nouns exhibit morphological case? \q If so, what are the cases? (the functions of the cases will be elaborated in later sections.) \shd2 Semantic roles \fln For Help jump to: \cf Hn_case_role \q What are semantic roles which are manifested by case marking? \ftx \dt 27/Apr/1998 \key Qn_cls \shd Class (including gender) \fln For Help jump to: \cf Hn_cls \ftx \q Is there a noun class system? What are the classes, and how are they manifested in the noun phrase? \ftx \shd2 Other dimensions \fln For Help jump to: \cf Hn_cls_other \ftx \q What dimension of reality is most central to the noun class system? (e.g. animacy, shape, function, etc.).What other dimensions are relevant? \ftx \shd2 Relation to other constituents \fln For Help jump to: \cf Hn_cls_rel Class--relation to other constituents \ftx \q Do the classifiers occur with numerals? adjectives? verbs? What is their function in these contexts? \ftx \dt 27/Apr/1998 \key Qn_cmp \shd Compounding \fln For Help jump to: \cf Hn_cmp \ftx \q Is there compounding? \ftx \shd2 Identifying Compounds \fln For Help jump to: \cf Hn_cmp_iden \ftx \q How do you know it is compounding? \ftx \shd2 Noun-verb Compounds \fln For Help jump to: \cf Hn_cmp_n-v \ftx \q Is there noun-verb compounding that results in a noun? \ftx \shd2 Productivity of Compounds \fln For Help jump to: \cf Hn_cmp_prod \ftx \q Are these processes productive? \ftx \shd2 Frequency of occurrence \ftx \q How common is compounding? \ftx \dt 19/Aug/1997 \key Qn_dim/aug \shd Diminution/augmentation \fln For Help jump to: \cf Hn_dim/aug \ftx \q Does the language employ diminutive and/or augmentative operators in the noun or noun-phrase? \ftx \dt 27/Apr/1998 \key Qn_dnom \shd Denominalization \fln For Help jump to: \cf Hn_dnom \ftx \q Are there any processes (productive or not) that form a verb from a noun? \ftx \shd Adjectives \ftx \q Are there any processes (productive or not) that form an adjective from a noun? \ftx \shd Adverbs \ftx \q Are there any processes (productive or not) that form an adverb from a noun? \ftx \dt 27/Apr/1998 \key Qn_intro \shd Nominal operations \fln For Help jump to: \cf Hn_intro \ftx \q What are the characteristic nominal operations occurring in this language? \ftx \dt 19/Aug/1997 \key Qn_num \shd Number \txt In this very short section, just describe number marking on nouns. The numeral system of the language should be described in: gr_mod_num_sys. \ftx \shd Singular/Plural distinctions \fln For Help jump to: \cf Hn_num_s/pl \ftx \q Is the distinction between singular and non-singular obligatorily signalled for all nouns, some subclasses of nouns, or not at all? \ftx \shd Other distinctions \fln For Help jump to: \cf Hn_num_other \ftx \q What non-singular distinctions are there? \ftx \shd Optional marking \fln For Help jump to: \cf Hn_num_opt \ftx \q If number marking is 'optional', when does it tend to occur, and when does it tend not to occur? \ftx \dt 27/Apr/1998 \key Qn_other \shd Other operations encoded in the noun phrase \fln For Help jump to: \cf Hn_other \ftx \q Are there other operations which are encoded in the noun phrase? \ftx \dt 19/Aug/1997 \key Qn_poss \shd Possessors \fln For Help jump to: \cf Hn_poss \ftx \q How are possessors referred to in the noun phrase? \q Do nouns agree with their possessors, do possessors agree with possessed nouns, neither, or both? \ftx \shd2 Alienable vs inalienable posssession \fln For Help jump to: \cf Hn_poss_al/inal \ftx \q Is there a distinction between alienable and inalienable possession? \ftx \shd2 Other kinds of possession \fln For Help jump to: \cf Hn_poss_other \ftx \q Are there other types of possession? \ftx \shd2 Constituent order \fln For Help jump to: \cf Hn_poss_ord \ftx \q When the possessor is a full noun, where does it usually come with respect to the head of the NP in which it occurs? \ftx \dt 27/Apr/1998 \key Qord \shd ***Constituent Order Typology--Contents \fln The following is a list of section headings in this chapter: \ftx \ftx H4.0.1.A Historical and theoretical background--Clauses \ftx H4.0.1.B Historical and theoretical background--Constituent order \ftx H4.0.1.C Historical and theoretical background--Syntactic correlates of \ftx constituent order \ftx H4.0.1.D Historical and theoretical background--Problems with \ftx Greenberg's work \ftx H4.0.1.E Historical and theoretical background--Concluding remarks \ftx \ftx H4.0.2.A Distribution of constituent order types around the world \ftx --common orders \ftx H4.0.2.B Distribution of constituent order types around the world \ftx --verb initial clauses \ftx H4.1.A Constituent order in main clauses--general case \ftx H4.1.B Constituent order in main clauses--'Rigid' constituent order \ftx languages \ftx H4.1.C Constituent order in main clauses--`Flexible' constituent order \ftx languages \ftx H4.1.D Constituent order in main clauses--Pragmatic factors I \ftx H4.1.E Constituent order in main clauses--Pragmatic factors II \ftx H4.1.F Constituent order in main clauses--Pragmatic factors III \ftx H4.1.G Constituent order in main clauses--Cautionary note \ftx H4.2.A Verb phrase (Verb, Aux) \ftx H4.2.B Verb phrase (Verb, Aux, adverbs) \ftx H4.3 Noun phrase (Det, Num, Gen, Mod, RC, N) \ftx H4.4.A Adpositional phrases (prepositions and postpositions) \ftx H4.4.B Adpositional phrases versus dependent verbs \ftx H4.4.C Adpositional phrases versus nouns \ftx H4.5 Comparatives \ftx H4.6.A Question particles \ftx H4.6.B Question words \ftx H4.7.A Summary \ftx H4.7.B Greenberg Universals Chart \ftx \txt By highlighting any of the above topics and using the jump feature you should be able to jump directly to the help notes on any of these items. Be sure your search path is set correctly, then if you highlight the section number or heading including the initial H you should go directly to the Help file. If you highlight it excluding the initial H, it should take you to that specific question. It doesn't matter how much of the rest of the section title you include in the highlight, provided you start with either the initial H or the first digit of the section number. \ftx \dt 27/Apr/1998 \key Qord_adpp \shd Adpositional phrases (prepositions and postpositions) \fln For Help jump to: \cf Hord_adpp \ftx \q Is the language dominantly prepositional or postpositional? Give examples. \ftx \shd Adpositional phrases versus dependent verbs \fln For Help jump to: \cf Hord_adpp_vsdepv \ftx \q Does your language use both adpositions and dependent verbs? Are they in contrast? \ftx \shd Adpositional phrases versus nouns \fln For Help jump to: \cf Hord_adpp_vsn \ftx \q Does the language use complex adpositions built up fo other adpositions, or of adpositions and nouns? \ftx \dt 27/Apr/1998 \key Qord_cl_case \shd Constituent order in main clauses--general case \fln For Help jump to: \cf Hord_cl_case \ftx \q What is the pragmatically neutral order of constituents (A/S, P and V) in basic clauses of the language? \ftx \dt 27/Apr/1998 \key Qord_cl_caut \shd Constituent order in main clauses--Cautionary note \fln For Help jump to: \cf Hord_cl_caut \ftx \txt For a cautionary note about jumping to conclusions relating to constituent ordering, jump to the help screen. \ftx \key Qord_cl_flex \shd Constituent order in main clauses--`Flexible' constituent order languages \fln For Help jump to: \cf Hord_cl_flex \ftx \txt For examples of languages with so-called flexible constituent order, jump to the help screen. \ftx \dt 27/Apr/1998 \key Qord_cl_prag \shd Constituent order in main clauses--Pragmatic factors \fln For Help jump to: \cf Hord_cl_prag1 \ftx \q Is the basic constituent order governed by grammatical relations? \ftx \q Does your language show similarities to other languages where pragmatic factors seem to influence constituent order? \ftx \dt 27/Apr/1998 \key Qord_cl_rigd \shd Constituent order in main clauses--'Rigid' constituent order languages \fln For Help jump to: \cf Hord_cl_rigd \ftx \txt For examples from languages which have a rigid consitituent order, jump to the help screen. \ftx \dt 27/Apr/1998 \key Qord_cmpar \shd Comparatives \fln For Help jump to: \cf Hord_cmpar \ftx \q What is the order of the Standard, the Marker, and the Quality by which an item is compared to the Standard? \ftx \dt 27/Apr/1998 \key Qord_hist_cl \shd Historical and theoretical background--Clauses \fln For Help jump to: \cf Hord_hist_cl \ftx \q How are propositions expressed naturally in clauses? \ftx \shd Constituent order \fln For Help jump to: \cf Hord_hist_ord \ftx \q What is the normal constituent order in clauses? \ftx \shd Syntactic correlates of constituent order \fln For Help jump to: \cf Hord_hist_syn \ftx \txt For a discussion of Greenberg's observations of the syntactic correlates which tend to go with certain constituent orders, jump to the help screen. \ftx \shd Problems with Greenberg's work \fln For Help jump to: \cf Hord_hist_prob \ftx \txt For a discussion of the problems which emerged from Greenberg's treatment of the syntactic correlates of constituent order, jump to the help screen. \ftx \shd Concluding remarks \fln For Help jump to: \cf Hord_hist_cncsl \ftx \txt For some summary remarks on the usefulness of the Greenberg typology, jump to the help screen. \ftx \fln ------------------------------------------------ \ftx \shd Distribution of constituent order types around the world--common orders \fln For Help jump to: \cf Hord_dist_com \ftx \q What is the normal constituent order in the language in terms of the three way distinction between A S and P? \ftx \shd Verb initial clauses \fln For Help jump to: \cf Hord_dist_v-init \ftx \txt For a discussion of the difficulties that arise with verb initial languages, jump to the help screen. \ftx \dt 27/Apr/1998 \key Qord_np \shd Noun phrase (Det, Num, Gen, Mod, RC, N) \fln For Help jump to: \cf Hord_np \ftx \q Describe the order(s) of elements in the noun phrase. \ftx \key Qord_q_part \shd Question particles \fln For Help jump to: \cf Hord_q_part \ftx \q In yes/no questions, if there is a question particle, where does it occur? \ftx \key Qord_q_wd \shd Question words \fln For Help jump to: \cf Hord_q_wd \ftx \q In information questions, where does the question word occur? \ftx \key Qord_sum \shd Summary \fln For Help jump to: \cf Hord_sum \ftx \q How does this language compare in its constituent orders to universal expectations, as represented by Greenberg (1966), Hawkins (1983), or some other well-known typology? \ftx \dt 27/Apr/1998 \key Qord_sum_chart \shd Greenberg Universals Chart \fln For Help jump to: \cf Hord_sum_chart \ftx \txt For a chart which summarizes Greenberg's expectations for adpositions, jump to the help screen. \ftx \dt 27/Apr/1998 \key Qord_vp_v-aux \shd Verb phrase (Verb, Aux) \fln For Help jump to: \cf Hord_vp_v-aux \ftx \q Where do auxiliaries occur in relation to the semantically 'main' verb? \ftx \key Qord_vp_v-aux-adv \shd Verb phrase (Verb, Aux, adverbs) \ftx \q Where do verb phrase adverbs occur with respect to the verb and auxiliaries? \ftx \dt 27/Apr/1998 \key Qpnom \shd ***Predicate nominals and related constructions \fln The following is a list of section headings in this chapter: \ftx \ftx H6.0.A Introduction--Definitions \ftx H6.0.B Introduction--Other types \ftx H6.0.C Introduction--Chart of types \ftx H6.0.D Introduction--Practical considerations \ftx H6.0.E Introduction--Examples of various types \ftx H6.0.F Introduction--Cross linguistic comparison \ftx H6.1.A Predicate nominals--inclusion vs equation \ftx H6.1.B Predicate nominals--Copulas \ftx H6.1.C Predicate nominals--Copula as verb I \ftx H6.1.D Predicate nominals--Copula as verb II \ftx H6.1.E Predicate nominals--Copula as verb III \ftx H6.1.F Predicate nominals--Copula as verb IV \ftx H6.1.G Predicate nominals--Copula as verb V \ftx H6.1.H Predicate nominals--Characteristics of copula verbs \ftx H6.1.I Predicate nominals--Copula as pronoun \ftx H6.1.J Predicate nominals--Copula as particle \ftx H6.1.K Predicate nominals--Copula as operation \ftx H6.1.L Predicate nominals--Copula in non-present tenses \ftx H6.1.M Predicate nominals--Summary of typology \ftx H6.2.A Predicate adjectives--attributive clauses \ftx H6.2.B Predicate adjectives--contrast with nouns \ftx H6.2.C Predicate adjectives--other possibilities \ftx H6.3 Predicate locatives \ftx H6.4.A Existentials \ftx H6.4.B Existentials vs Copulas \ftx H6.5 Possessive constructions \ftx H6.6.A Summary of Pred Nom and EPL relationships--Charts \ftx H6.6.B Summary of Pred Nom and EPL relationships \ftx \txt By highlighting any of the above topics and using the jump feature you should be able to jump directly to the help notes on any of these items. Be sure your search path is set correctly, then if you highlight the section number or heading including the initial H you should go directly to the Help file. If you highlight it excluding the initial H, it should take you to that specific question. It doesn't matter how much of the rest of the section title you highlight, provided you start with either the initial H or the first digit of the section number. \ftx \dt 27/Apr/1998 \key Qpnom_exst \shd Existentials \fln For Help jump to: \cf Hpnom_exst \ftx \q How are existential clauses formed? (Give examples in different tense/aspects if there is significant variation). Be especially aware of how existentials compare to locational and possessive clauses. \ftx \shd Existentials vs Copulas \fln For Help jump to: \cf Hpnom_exst_vscop \ftx \q Is there a distinction between copulas and existentials in the language? \ftx \dt 27/Apr/1998 \key Qpnom_intro \shd Introduction \fln For Help jump to: \cf Hpnom_intro \ftx \q Does the language have predicate nominal constructions? \ftx \shd2 Other types \fln For Help jump to: \cf Hpnom_intro_other \ftx \q What other predicate types occur in the language? \ftx \shd2 Chart of types \fln For Help jump to: \cf Hpnom_intro_chart \ftx \q What is the relationship between predicate types and semantically rich verbs in the language? \ftx \shd2 Practical considerations \fln For Help jump to: \cf Hpnom_intro_consd \ftx \q List the various predicate nominal constructions found in the language. \ftx \shd2 Cross-linguistic comparison \fln For Help jump to: \cf Hpnom_intro_cmp \ftx \q Do the predicate nominal constructions in your language match the most frequently occurring structures found in other languges, or are they different? \ftx \dt 27/Apr/1998 \key Qpnom_p-adj \shd Predicate adjectives \ftx \fln For Help jump to: \cf Hpnom_p-adj \ftx \q How are predicate adjectives formed? (Include a separate section on predicate adjectives only if they are structurally distinct from predicate nominals). \ftx \shd2 Contrast with nouns \ftx \fln For Help jump to: \cf Hpnom_p-adj_ncontr Predicate adjectives--contrast with nouns \ftx \q In what ways, if any, do nouns and adjectives contrast in predicate constructions in your language? \ftx \shd2 Other possibilities \ftx \fln For Help jump to: \cf Hpnom_p-adj_other Predicate adjectives--other possibilities \ftx \q What other kinds of phrases act like predicate adjectives in your language? \ftx \dt 27/Apr/1998 \key Qpnom_p-loc \shd Predicate locatives \fln For Help jump to: \cf Hpnom_p-loc \ftx \q How are locational clauses formed? \ftx \dt 19/Aug/1997 \key Qpnom_p-nom_cop \shd Predicate nominals--Copulas \fln For Help jump to: \cf Hpnom_p-nom_cop \ftx \q What restrictions are there, if any, on the TAM marking of predicate nominal constructions? \ftx \dt 27/Apr/1998 \key Qpnom_p-nom_cop_op \shd Predicate nominals--Copula as operation \fln For Help jump to: \cf Hpnom_p-nom_cop_op \ftx \q Does the language use a derivational operation to form a verb from a noun rather than using a regular predicate nominal construction with a copula? \ftx \dt 27/Apr/1998 \key Qpnom_p-nom_cop_part \shd Predicate nominals--Copula as particle \fln For Help jump to: \cf Hpnom_p-nom_cop_part \ftx \q Does the language use an invariant particle as a copula to join nominals in a predicate nominal construction? \ftx \dt 27/Apr/1998 \key Qpnom_p-nom_cop_pron \shd Predicate nominals--Copula as pronoun \fln For Help jump to: \cf Hpnom_p-nom_cop_pron \ftx \q Does the language employ pronouns as copulas to link nominals in predicate nominal constructions? \ftx \dt 27/Apr/1998 \key Qpnom_p-nom_cop_tns \shd Predicate nominals--Copula in non-present tenses \fln For Help jump to: \cf Hpnom_p-nom_cop_tns \ftx \q Does the language use a copular morpheme or verb in non-present tenses? \ftx \dt 19/Aug/1997 \key Qpnom_p-nom_incl/eq \shd Predicate nominals--inclusion vs equation \fln For Help jump to: \cf Hpnom_p-nom_incl/eq \ftx \q How are proper inclusion and equative predicates formed? \ftx \dt 19/Aug/1997 \key Qpnom_p-nom_sum \shd Predicate nominals--Summary of typology \fln For Help jump to: \cf Hpnom_p-nom_sum \ftx \txt For a summary of the typology of predicate nominal constructions, jump to the help screen. \ftx \dt 27/Apr/1998 \key Qpnom_p-nom_vcop \shd Predicate nominals--Copula as verb \fln For Help jump to: \cf Hpnom_p-nom_vcop \ftx \q Do copulas in your language show the morphosyntactic properties of normal verbs? \ftx \fln For examples, jump to the help screen. \ftx \dt 27/Apr/1998 \key Qpnom_poss \shd Possessive constructions \fln For Help jump to: \cf Hpnom_poss \ftx \q How are possessive clauses formed? Remember that possessive clauses ('I have a dollar') are distinct from possessive noun phrases ('my dollar). Describe possessive clauses here. \ftx \dt 27/Apr/1998 \key Qpnom_sum \shd Summary of Pred Nom and EPL relationships \fln For Help jump to: \cf Hpnom_sum \ftx \txt For a summary of the various existential, locative, and possessive constructions, jump to the help screen. \ftx \shd Charts \fln For Help jump to: \cf Hpnom_sum_chart \ftx \txt For charts summarising the characteristics of the different constructions, jump to the help screen. \ftx \dt 19/Aug/1997 \key Qprag \shd ***Pragmatically marked structures \fln The following is a list of section headings in this chapter: \ftx \ftx H10.0.1.A Pragmatic statuses--Introduction \ftx H10.0.1.B Pragmatic status and grammatical relations \ftx H10.0.1.C Specially marked Pragmatic statuses \ftx H10.0.2.A Identifiability--articles \ftx H10.0.2.B Identifiability--proper names \ftx H10.0.2.C Identifiability--implicit in situation \ftx H10.0.2.D Objective Referentiality \ftx H10.0.2.E Discourse Referentiality \ftx H10.0.3.A Focus--Overview \ftx H10.0.3.B Focus--First approach \ftx H10.0.3.C Focus--Second approach \ftx H10.0.3.D Focus--Third approach \ftx H10.0.3.E Scope of Focus \ftx H10.0.4 Topic \ftx H10.0.5 Contrast \ftx H10.1 The morphosyntax of pragmatic prominence \ftx H10.1.1.A Constituent order--General \ftx H10.1.1.A.1 Constituent order--Aghem example \ftx H10.1.1.B Constituent order--Dislocation \ftx H10.1.1.B.1 Constituent order--Other kinds of movement \ftx H10.1.1.B.2 Constituent order--Rules relating to dislocation \ftx H10.1.1.B.3 Constituent order--Examples \ftx H10.1.1.B.4 Constituent order--Use of particles \ftx H10.1.2 Formatives \ftx H10.1.2.A Formatives--Examples \ftx H10.1.2.B Formatives--Overlay systems \ftx H10.1.3 Cleft constructions \ftx H10.1.3.A Cleft constructions--English examples \ftx H10.1.3.B Cleft constructions--Examples from other languages \ftx H10.2.A Negation--Introduction \ftx H10.2.B Negation--Lexical \ftx H10.2.C Negation--Morphological \ftx H10.2.D.1 Negation--Analytic by particle \ftx H10.2.D.1.1 Negation--Analytic by particle--Examples I \ftx H10.2.D.1.2 Negation--Analytic by particle--Examples II \ftx H10.2.D.1.3 Negation--Analytic by particle--Examples III \ftx H10.2.D.2.1 Negation--Analytic by initial finite verb \ftx H10.2.D.2.2 Negation--Analytic by final finite verb \ftx H10.2.E Negation--Secondary modifications \ftx H10.2.E.1 Negation--Secondary modifications--word order \ftx H10.2.E.2 Negation--Secondary modifications--tone \ftx H10.2.E.3 Negation--Secondary modifications--tense/aspect \ftx H10.2.E.4 Negation--Secondary modifications--inflection \ftx H10.2.E.5 Negation--Secondary modifications--case \ftx H10.2.F Non-clausal Negation \ftx H10.2.F.1 Non-clausal Negation--derivational \ftx H10.2.F.2 Non-clausal Negation--quantifiers \ftx H10.2.F.3 Non-clausal Negation--scope \ftx H10.3 Non-declarative Speech acts \ftx H10.3.A Non-declarative Speech acts--theory \ftx H10.3.1. Interrogatives \ftx H10.3.1.1. Yes/No Questions \ftx H10.3.1.1.A Yes/No Questions--Intonation \ftx H10.3.1.1.B Yes/No Questions--other phonological markers \ftx H10.3.1.1.C Yes/No Questions--word order \ftx H10.3.1.1.D Yes/No Questions--special particles \ftx H10.3.1.1.E Yes/No Questions--Tag questions \ftx H10.3.1.1.F Functions of Yes/No Questions \ftx H10.3.1.2. Question word (information, content) questions \ftx H10.3.1.2.A Functions of question words \ftx H10.3.1.2.B Question words in VO languages \ftx H10.3.1.2.C Question words in OV languages \ftx H10.3.1.2.D Question words and case markers \ftx H10.3.2. Imperatives \ftx H10.3.2.A Imperatives and verb forms \ftx H10.3.2.B Imperatives and negation \ftx H10.3.2.C Imperatives and mode \ftx H10.3.2.D Imperatives and case \ftx \txt By highlighting any of the above topics and using the jump feature you should be able to jump directly to the help notes on any of these items. Be sure your search path is set correctly, then if you highlight the section number or heading including the initial H you should go directly to the Help file. If you highlight it excluding the initial H, it should take you to that specific question. It doesn't matter how much of the rest of the section title you highlight, provided you start with either the initial H or the first digit of the section number. \ftx \dt 27/Apr/1998 \key Qprag_cntr \shd Contrast \fln For Help jump to: \cf Hprag_cntr \ftx \q How is the concept of contrast encoded in the language? \ftx \dt 19/Aug/1997 \key Qprag_foc \shd Focus \fln For Help jump to: \cf Hprag_foc \ftx \q Are you using 'focus' to describe some characteristic of your language? In what sense are you using the term? \ftx \q Are you using the term 'focus' to describe some poorly understood morphosyntactic feature of the language? \ftx \q Are you using 'focus' to refer to the new information in any given sentence? \ftx \q Are you using 'focus' to refer to some special pragmatically marked feature of certain sentences? \ftx \q If you are using 'focus' in this way, how is the scope of the focus indicated in the language? \ftx \dt 27/Apr/1998 \key Qprag_iden \shd Identifiability \fln For Help jump to: \cf Hprag_iden \ftx \q Does your language use articles in determining identifiability? \ftx \q In what ways does your language use proper names in determining identifiability? \ftx \q Can you describe the ways in which implicit information in the communication situation is used by the language to establish identity? \ftx \q Does your language make a recognizable contrast between referentiality and identifiability? \ftx \q How does your language deal with referentiality in discourse? \ftx \dt 27/Apr/1998 \key Qprag_ndecl \shd Non-declarative Speech acts \fln For Help jump to: \cf Hprag_ndecl \ftx \q How does the language mark constructions as being indicative, interrogative, imperative, etc? \ftx \dt 27/Apr/1998 \key Qprag_ndecl_imp \shd Imperatives \fln For Help jump to: \cf Hprag_ndecl_imp \ftx \q How are imperatives formed? \ftx \q Are there 'polite' imperatives that stand in contrast to more direct imperatives? \ftx \q Are there '1st person' imperatives (e.g. 'let's eat')? If so how are they used? \ftx \q What special verb forms are used to form imperatives in your language? \ftx \q How are imperatives negated in your languge? \ftx \q Are there special relationships between mode and imperative forms in your language? \ftx \q How does case interact with imperatives in your language? \ftx \dt 27/Apr/1998 \key Qprag_ndecl_q \shd Interrogatives \fln For Help jump to: \cf Hprag_ndecl_q \ftx \q How are interrogatives formed in the language? \ftx \dt 19/Aug/1997 \key Qprag_ndecl_q_qwd \shd Question word (information, content) questions \fln For Help jump to: \cf Hprag_ndecl_q_qwd \ftx \q How are information questions formed in your language? \ftx \dt 19/Aug/1997 \key Qprag_ndecl_q_qwd_case \shd Question words and case markers \fln For Help jump to: \cf Hprag_ndecl_q_qwd_case \ftx \q In what ways do question words, case markers and adpositions interact in your language? \ftx \dt 27/Apr/1998 \key Qprag_ndecl_q_qwd_func \shd Functions of question words \fln For Help jump to: \cf Hprag_ndecl_q_qwd_func \ftx \q In what ways does your language use question words? \ftx \dt 19/Aug/1997 \key Qprag_ndecl_q_qwd_ov \shd Question words in OV languages \fln For Help jump to: \cf Hprag_ndecl_q_qwd_ov \ftx \txt For consideration of the way question words operate in OV languages, jump to the help screen. \ftx \dt 27/Apr/1998 \key Qprag_ndecl_q_qwd_vo \shd Question words in VO languages \fln For Help jump to: \cf Hprag_ndecl_q_qwd_vo \ftx \txt For consideration of the way question words operate in VO languages, jump to the help screen. \ftx \dt 27/Apr/1998 \key Qprag_ndecl_q_ynq \shd Yes/No Questions \fln For Help jump to: \cf Hprag_ndecl_q_ynq \ftx \q How are yes/no questions formed? \ftx \dt 19/Aug/1997 \key Qprag_ndecl_q_ynq_func \shd Functions of Yes/No Questions \fln For Help jump to: \cf Hprag_ndecl_q_ynq_func \ftx \q How do yes/no questions function in your language? \ftx \dt 19/Aug/1997 \key Qprag_ndecl_q_ynq_inton \shd Yes/No Questions--Intonation \fln For Help jump to: \cf Hprag_ndecl_q_ynq_inton \ftx \q Can yes/no questions be marked simply by change of intonation? \ftx \dt 19/Aug/1997 \key Qprag_ndecl_q_ynq_ord \shd Yes/No Questions--word order \fln For Help jump to: \cf Hprag_ndecl_q_ynq_ord \ftx \q Are yes/no questions marked by change of word order? \ftx \dt 19/Aug/1997 \key Qprag_ndecl_q_ynq_other \shd Yes/No Questions--other phonological markers \fln For Help jump to: \cf Hprag_ndecl_q_ynq_other \ftx \q Are there any other phonological markers which mark yes/no questions in your language? \ftx \dt 27/Apr/1998 \key Qprag_ndecl_q_ynq_part \shd Yes/No Questions--special particles \fln For Help jump to: \cf Hprag_ndecl_q_ynq_part \ftx \q Do yes/no questions utilize special particles? \ftx \dt 19/Aug/1997 \key Qprag_ndecl_q_ynq_tag \shd Yes/No Questions--Tag questions \fln For Help jump to: \cf Hprag_ndecl_q_ynq_tag \ftx \q Does your language employ tag questions? \ftx \dt 19/Aug/1997 \key Qprag_ndecl_theory \shd Non-declarative Speech acts--theory \fln For Help jump to: \cf Hprag_ndecl_theory \ftx \q How do you identify declarative type clauses in your language? \ftx \dt 19/Aug/1997 \key Qprag_neg_anal_part \shd Negation--Analytic by particle \fln For Help jump to: \cf Hprag_neg_anal_part \ftx \q Does your language implement analytic type negation by the use of special particles? \ftx \dt 27/Apr/1998 \key Qprag_neg_anal_part_ex1 \shd Negation--Analytic by particle--Examples I \fln For Help jump to: \cf Hprag_neg_anal_part_ex1 \ftx \txt For an example of an analytic particle in Tagalog, jump to the help screen. \ftx \dt 19/Aug/1997 \key Qprag_neg_anal_part_ex2 \shd Negation--Analytic by particle--Examples II \fln For Help jump to: \cf Hprag_neg_anal_part_ex2 \ftx \txt For an example of an analytic particle in Mandarin, jump to the help screen. \ftx \dt 19/Aug/1997 \key Qprag_neg_anal_part_ex3 \shd Negation--Analytic by particle--Examples III \fln For Help jump to: \cf Hprag_neg_anal_part_ex3 \ftx \txt For an example of an analytic particle in Iraqi Arabic, jump to the help screen. \ftx \dt 27/Apr/1998 \key Qprag_neg_anal_v-fin \shd Negation--Analytic by final finite verb \fln For Help jump to: \cf Hprag_neg_anal_v-fin \ftx \q Does your language use a finite verb and a complement clause to encode analytic negation? (Examples from verb final languages) \ftx \dt 27/Apr/1998 \key Qprag_neg_anal_v-init \shd Negation--Analytic by initial finite verb \fln For Help jump to: \cf Hprag_neg_anal_v-init \ftx \q Does your language use a finite verb and a complement clause to encode analytic negation? (Examples from verb initial languages). \ftx \dt 27/Apr/1998 \key Qprag_neg_intro \shd Negation--Introduction \fln For Help jump to: \cf Hprag_neg_intro \ftx \q What is the standard means of forming a negative clause in ypur language? \ftx \q What secondary strategies are there? When are they used? Is there constituent negation? Derivational negation? \ftx \q How is morphology normally associated with negation employed in creative ways in discourse? \ftx \dt 27/Apr/1998 \key Qprag_neg_lex \shd Negation--Lexical \fln For Help jump to: \cf Hprag_neg_lex \q Does your language use lexical negation? \dt 19/Aug/1997 \key Qprag_neg_mor \shd Negation--Morphological \fln For Help jump to: \cf Hprag_neg_mor \q Does your language employ morphological negation? \ftx \dt 19/Aug/1997 \key Qprag_neg_n-cl \shd Non-clausal Negation \fln For Help jump to: \cf Hprag_neg_n-cl \ftx \q How is non-clausal negation handled by your language? \ftx \dt 19/Aug/1997 \key Qprag_neg_n-cl_drv \shd Non-clausal Negation--derivational \fln For Help jump to: \cf Hprag_neg_n-cl_drv \ftx \q Does your language show evidence for derivational negation? \ftx \dt 19/Aug/1997 \key Qprag_neg_n-cl_quan \shd Non-clausal Negation--quantifiers \fln For Help jump to: \cf Hprag_neg_n-cl_quan \ftx \q How does the language handle quantifiers with respect to negation? \ftx \dt 19/Aug/1997 \key Qprag_neg_n-cl_sco \shd Non-clausal Negation--scope \fln For Help jump to: \cf Hprag_neg_n-cl_sco \ftx \q How does your language indicate the scope of negation? \ftx \dt 19/Aug/1997 \key Qprag_neg_s-mod \shd Negation--Secondary modifications \fln For Help jump to: \cf Hprag_neg_s-mod \ftx \q Does your language show evidence for secondary modifications which accompany negative constructions? \ftx \dt 27/Apr/1998 \key Qprag_neg_s-mod_case \shd Negation--Secondary modifications--case \fln For Help jump to: \cf Hprag_neg_s-mod_case \ftx \q Does your languages show changes in case relating to negative constructions? \ftx \dt 27/Apr/1998 \key Qprag_neg_s-mod_infl \shd Negation--Secondary modifications--inflection \fln For Help jump to: \cf Hprag_neg_s-mod_infl \ftx \q Does your language show changes in inflection correlating with negative constructions? \ftx \dt 19/Aug/1997 \key Qprag_neg_s-mod_ord \shd Negation--Secondary modifications--word order \fln For Help jump to: \cf Hprag_neg_s-mod_ord \ftx \q Does your language employ changes in word order relating to negative constructions? \ftx \dt 27/Apr/1998 \key Qprag_neg_s-mod_tone \shd Negation--Secondary modifications--tone \fln For Help jump to: \cf Hprag_neg_s-mod_tone \ftx \q Does your language show changes of tone in conjunction with negative constructions? \ftx \dt 27/Apr/1998 \key Qprag_neg_s-mod_t/a \shd Negation--Secondary modifications--tense/aspect \fln For Help jump to: \cf Hprag_neg_s-mod_t/a \ftx \q Are there changes in tense/aspect which correlate with negative constructions? \ftx \dt 27/Apr/1998 \key Qprag_prom \shd The morphosyntax of pragmatic prominence \ftx (e.g. focus, contrast and 'topicalization') \fln For Help jump to: \cf Hprag_prom \ftx \q Are there special devices for indicating pragmatic statuses in basic clauses? e.g. marked constituent orders, left- and/or right-dislocation, affixes or particles indicating referentiality, specificity, topic, focus, contrast, etc.? \q Are there different types or strengths of pragmatic prominence? \ftx \dt 27/Apr/1998 \key Qprag_prom_clft \shd Cleft constructions \fln For Help jump to: \cf Hprag_prom_clft \ftx \q Describe cleft constructions. If possible, give a characterization of their discourse functions. \ftx \dt 27/Apr/1998 \key Qprag_prom_clft_ex_eng \shd Cleft constructions--English examples \fln For Help jump to: \cf Hprag_prom_clft_ex_eng \ftx \txt For some examples of the way cleft constructions are used in English, jump to the help screen. \ftx \dt 19/Aug/1997 \key Qprag_prom_clft_ex_other \shd Cleft constructions--Examples from other languages \fln For Help jump to: \cf Hprag_prom_clft_ex_other \ftx \txt For some examples of the way cleft constructions are used in other languages, jump to the help screen. \ftx \dt 19/Aug/1997 \key Qprag_prom_form \shd Formatives \fln For Help jump to: \cf Hprag_prom_form \ftx \q Are there special formatives that ascribe pragmatic prominence to various parts of sentences? \ftx \dt 27/Apr/1998 \key Qprag_prom_form_ex \shd Formatives--Examples \fln For Help jump to: \cf Hprag_prom_form_ex \ftx \txt For some examples of the use verb morphology and a focus particle to convey pragmatic nuances, jump to the help screen. \ftx \key Qprag_prom_form_ovl \shd Formatives--Overlay systems \fln For Help jump to: \cf Hprag_prom_form_ovl \ftx \q Does your language use a system of overlaying case markers for marking pragmatic status? \ftx \dt 27/Apr/1998 \key Qprag_prom_ord \shd Constituent order--General \fln For Help jump to: \cf Hprag_prom_ord \ftx \q What are the functions of unusual constituent orders? \ftx \key Qprag_prom_ord_dis \shd Constituent order--Dislocation \fln For Help jump to: \cf Hprag_prom_ord_dis \ftx \q Does the language use the feature of dislocation to encode topicalization or similar phenomena? \ftx \dt 27/Apr/1998 \key Qprag_prom_ord_dis_ex \shd Constituent order--Examples \fln For Help jump to: \cf Hprag_prom_ord_B_3 \ftx \txt For some examples of the various kinds of movement, jump to the help screen. \ftx \dt 19/Aug/1997 \key Qprag_prom_ord_dis_other \shd Constituent order--Other kinds of movement \fln For Help jump to: \cf Hprag_prom_ord_dis_other \ftx \q Does dislocation of constituents, i.e. placing them outside the predication, contrast with other kinds of movement such as apposition and fronting? \ftx \dt 27/Apr/1998 \key Qprag_prom_ord_dis_part \shd Constituent order--Use of particles \fln For Help jump to: \cf Hprag_prom_ord_dis_part \ftx \q Does your language use special particles to mark dislocation of constituent structure? \ftx \dt 27/Apr/1998 \key Qprag_prom_ord_dis_rule \shd Constituent order--Rules relating to dislocation \fln For Help jump to: \cf Hprag_prom_ord_dis_rule \ftx \txt For some rules of thumb to help in identifying the various kinds of movement, jump to the help screen. \ftx \key Qprag_prom_ord_ex \shd Constituent order--Aghem example \fln For Help jump to: \cf Hprag_prom_ord_ex \ftx \txt For an example of the way the West African language Aghem handles constituent order, jump to the help screen. \ftx \key Qprag_stat \shd Pragmatic statuses--Introduction \fln For Help jump to: \cf Hprag_stat \q What are the pragmatic characteristics of the language you are studying? \q Does your language employ grammatical relations to convey pragmatic information? \q Does your language employ special morphosyntactic devices to mark particular pragmatic statuses? \dt 27/Apr/1998 \key Qprag_top \shd Topic \fln For Help jump to: \cf Hprag_top \ftx \q Are you using the term 'topic' to describe some characteristic of the language? \ftx \dt 27/Apr/1998 \key Qv \shd Verb and verb-phrase operations \fln For a complete listing of the Help topics covered, jump to: \cf Hv Verb and verb-phrase operations \ftx \fln There are no Questions written yet for this section. \dt 25/Aug/1997 \key Qval \shd ***Valence changing operations \fln The following is a list of section headings in this chapter: \ftx \ftx H8.0. Valence and predicate calculus \ftx H8.0.A Valence and predicate calculus--types of zero \ftx H8.0.B Valence and predicate calculus--Pro-drop languages \ftx H8.0.C Valence and predicate calculus--Transitivity \ftx H8.0.D Valence and predicate calculus--Participant roles \ftx H8.0.E Valence and predicate calculus--Examples \ftx H8.0.F Valence and predicate calculus--Sources \ftx H8.0.G Valence as predicate calculus \ftx H8.0.H Valence as predicate calculus--Examples \ftx H8.1 Valence increasing operations \ftx H8.1.1 Causatives \ftx H8.1.1.A Causatives--Definitions \ftx H8.1.1.B Causatives--Transitivity \ftx H8.1.1.C Lexical Causatives \ftx H8.1.1.D Morphological Causatives \ftx H8.1.1.D.1 Morphological Causatives--Eskimo \ftx H8.1.1.D.2 Morphological Causatives--Quechua \ftx H8.1.1.D.3 Morphological Causatives--Punjabi \ftx H8.1.1.E Analytic Causatives \ftx H8.1.1.F Cause & Effect--I \ftx H8.1.1.G Cause & Effect--II \ftx H8.1.1.G.1.A Coding principles--Structural distance--I \ftx H8.1.1.G.1.B Coding principles--Structural distance--II \ftx H8.1.1.G.1.C Coding principles--Structural distance--III \ftx H8.1.1.G.1.D Coding principles--Structural distance--IV \ftx H8.1.1.G.1.E Coding principles--Structural distance--V \ftx H8.1.1.G.2 Coding principles--Finiteness \ftx H8.1.1.G.3 Coding principles--Case \ftx H8.1.2 Applicatives \ftx H8.1.2.A Applicatives--Examples I \ftx H8.1.2.B Applicatives--Examples II \ftx H8.1.2.C Applicatives--Examples III \ftx H8.1.2.D Applicatives versus Causatives \ftx H8.1.3.A Dative shift and other valence increasing operations \ftx H8.1.3.B Dative of interest \ftx H8.2.A Valence decreasing operations \ftx H8.2.B Valence decreasing operations--Panare \ftx H8.2.1.A Reflexives/reciprocals--prototypical reflexive \ftx H8.2.1.B Reflexives/reciprocals--lexical reflexive \ftx H8.2.1.C Reflexives/reciprocals--morphological reflexive \ftx H8.2.1.C.1 Morphological reflexive--Examples I \ftx H8.2.1.C.2 Morphological reflexive--Examples II \ftx H8.2.1.D Analytic reflexives \ftx H8.2.1.E Reflexives/reciprocals--prototypical reciprocal \ftx H8.2.1.E.1 Reflexives/reciprocals--lexical reciprocal \ftx H8.2.1.E.2 Morphological Reciprocals \ftx H8.2.1.E.3 Morphological Reciprocals--Examples I \ftx H8.2.1.E.4 Morphological Reciprocals--Examples II \ftx H8.2.1.F Other Coreference \ftx H8.2.2.A Passives--prototypical \ftx H8.2.2.B Passives--personal \ftx H8.2.2.C Passives--lexical \ftx H8.2.2.D Passives--morphological \ftx H8.2.2.E Passives--analytic \ftx H8.2.2.F Passives--impersonal \ftx H8.2.3 Middle constructions \ftx H8.2.3.A Middle constructions contrasted with passives \ftx H8.2.3.B Middle constructions contrasted with reflexives \ftx H8.2.4 Antipassives \ftx H8.2.5 Object demotion \ftx H8.2.6.A Object incorporation--noun incorporation defined \ftx H8.2.6.B Object incorporation \ftx \txt By highlighting any of the above topics and using the jump feature you should be able to jump directly to the help notes on any of these items. Be sure your search path is set correctly, then if you highlight the section number or heading including the initial H you should go directly to the Help file. If you highlight it excluding the initial H, it should take you to that specific question. It doesn't matter how much of the rest of the section title you highlight, provided you start with either the initial H or the first digit of the section number. \ftx \dt 27/Apr/1998 \key Qval_decr \shd Valence decreasing operations \fln For Help jump to: \cf Hval_decr \ftx \q Describe the operations which detransitivize a verb or clause. The major possibilities are covered in subsequent subheadings. \ftx \dt 27/Apr/1998 \key Qval_decr_anti-p \shd Antipassives \fln For Help jump to: \cf Hval_decr_anti-p \ftx \q Does the language show evidence for a distinct antipassive construction? \ftx \key Qval_decr_demot \shd Object demotion \fln For Help jump to: \cf Hval_decr_demot \ftx \q Does the language exhibit a distinct process of object demotion? \ftx \key Qval_decr_mid \shd Middle constructions \fln For Help jump to: \cf Hval_decr_mid \ftx \q Does the language manifest a distinct middle construction? \ftx \q Do middle and passive constructions contrast in your language? \ftx \q Do middle and reflexive constructions contrast in your language? \ftx \key Qval_decr_ncorp \shd Object incorporation \fln For Help jump to: \cf Hval_decr_ncorp \ftx \q Does your language show evidence for noun incorporation? \ftx \q Does the language show evidence of object incorporation? \ftx \key Qval_decr_panare \shd Valence decreasing operations--Panare \fln For Help jump to: \cf Hval_decr_panare \ftx \txt For a table listing some typical Panare transitive verbs and their detransitivized counterparts grouped according to the detransitivizing prefix employed, jump to the help screen. \ftx \dt 27/Apr/1998 \key Qval_decr_pass \shd Passives \fln For Help jump to: \cf Hval_decr_pass \ftx \q Which type(s) of passive construction does the language have? (Exemplify each type, and describe its function or functions.) \ftx a. Lexical \ftx b. Morphological \ftx c. Analytic \q Are there 'impersonal' passives? i.e. passives of intransitive verbs, or passives where there is not necessarily an AGENT implied? \q Is a passive construction obligatory in any particular environment? e.g. when a PATIENT outranks an AGENT on some pragmatically defined hierarchy? \q Does your language have personal passive constructions? \q Does your language show evidence for lexical passives? \dt 27/Apr/1998