46th Poznań Linguistic Meeting Poznan, September 16, 2016 ANTIPASSIVES IN NAKH-DAGHESTANIAN LANGUAGES Bernard Comrie, Diana Forker, & Zaira Khalilova comrie@linguistics.ucsb.edu, diana.forker@uni-bamberg.de, zaira.khalilova@googlemail.com 1. Introduction - Nakh-Dagestanian languages are overall transitivizing (Nichols et al. 2004), but a number of detransitivizing processes are attested, incl. antipassivization:     all Tsezic languages except Khwarshi Avar (Bokarev 1949, Mallaeva 2007) at least some Andic languages (e.g. Godoberi) Dargwa languages (van den Berg 2001, 2003; Sumbatova & Mutalov 2003) - absent (on the basis of explicit published statements or personal communications) or probably absent (on the basis of lack of mention in extensive grammar) from     some Andic languages (e.g. Bagvalal) Lezgian, Agul, Tabasaran, Tsakhur, Archi, Udi (all from the Lezgic branch) Khinalugh Nakh languages - in Lak, the former antipassive has become a progressive (Schulze 2011) - properties in Dargwa rather different from those in Avar-Andic-Tsezic - more generally, different languages show interesting micro-variation 4. Tsezic languages: Bezhta and Hinuq (Comrie, Khalilov & Khalilova 2015; Forker 2013) 4.1 Overview (i) marked alternation - derivation via suffixes added directly to the verbal stem - Bezhta: -la/ā, -da/ā, -ya/ā; Hinuq: -liː, -doː - choice of allomorphs seems to be largely lexical (ii) can be formed from intransitive, unergative (Bezhta), transitive, affective, ditransitive verbs, but not fully productive (e.g. in Bezhta 77 verbs have been tested, of which 45 form the antipassive) (iii) valency change possible with unergative and transitive verbs, but not with intransitive or affective verbs 1 (iv) general meaning durative/iterative (perhaps basically durative, with iterative interpretation from interaction with semelfactive lexical meaning), but with idiosyncratic exceptions, e.g. Hinuq -oge‘open widely, stretch’ > -oge-liː- ‘squeeze through, edge through’ (v) restricted combinatory potential with other valency changing derivations: - in Hinuq it can co-occur with the potential suffix (detransitivizing), and it can follow the causative suffixes, but not precede them - in Bezhta it cannot be formed from derived potential verbs, and (as in Hinuq) it cannot serve as input to causative suffixes, though it can form analytic causatives (vi) antipassive/reflexive syncretism in Bezhta and Hinuq only with ‘wash’ - in Bezhta the antipassive form can have the regular antipassive interpretation, but also a purely reflexive interpretation ‘wash (oneself)’ without durative/iterative semantics and no separate overt expression of P of source verb - in Hinuq only reflexive interpretation 4.2 Syntax - antipassives of intransitive: no change in valency (1) Bezhta ‘jump’ (a) öžd̈ b-ogi<ba>c’-iyo boy.PL HPL-jump<PL>-PST ‘the boys jumped once’ (b) öžd̈ b-ogi<ya-ba>c-ca boy.PL HPL-jump<ANTIP-PL>-PRS ‘the boys jump many times’ (2) Hinuq ‘be’ ɬe r-exƛ’es r-iči-doː-ho water(V) V-warm V-be-ANTIP-PRS ‘the water is usually warm’ - antipassive of unergative verb in Bezhta: SA in Ergative shows up as S in Absolutive (Hinuq has no unergative verbs) (3) Bezhta ‘cough’ (a) öždi öhƛö-yö boy.OBL(ERG) cough-PST ‘the boy coughed (once)’ (b) öžö öhd̄-yö boy cough.ANTIP-PST ‘the boy was coughing’ 2 - antipassive of transitive verb: - ergative A of source verb marked with absolutive - Bezhta: omission of absolutive P of source verb or P marked with instrumental or, less frequently, interessive - Hinuq: obligatory omission of P of source verb (4) Bezhta ‘eat’ (a) öždi y-̈ⁿq-č̈. xo boy.ERG meat(IV) IV-eat-PRS ‘the boy eats the meat’ (b) öžö xo-lo-d Ø-̈ⁿq-d̈-̌. boy(I) meat-OBL-INS I-eat-ANTIP-PRS ‘the boy is busy eating the meat’ (5) Hinuq ‘plough’ (a) di-žo moči obu-y b-eƛ-o I.OBL-GEN2 father-ERG field(III) III-plough-PRS ‘my father ploughs the field’ (b) di obu Ø-eƛe-liː-ho I.GEN1 father(I) I-plough-ANTIP-PRS ‘my father ploughs (repeatedly)’ - antipassive of ditransitive verb: - Bezhta: the same as with transitive verbs (P omitted or oblique) - Hinuq: either the same as with transitive verbs (P omitted) or no change of valency pattern (P in absolutive) (6) Bezhta ‘give’ (a) öždi t’ek kib-ba-l niƛ-iyo boy.OBL(ERG) book girl-OBL-LAT give-PST ‘the boy gave the book to the girl’ (b) öžö kib-ba-l t’ek-lā-d niƛ-da-s boy girl-OBL-LAT book-PL.OBL-INS give-ANTIP-PRS ‘the boy is giving books to the girl’ (7) Hinuq ‘give’ hayɬoy elu-z ʕǎi neƛ-doː-ho // neƛ-o sayɣat-be he.ERG we.OBL-DAT often give-ANTIP-PRS // give-PRS present-PL ‘he gives us often / many presents // he gives us presents’ 3 - antipassive of affective verbs:  Bezhta: the construction is at best highly marginal  Hinuq: either change of valency pattern from two-place to one-place (standard intransitive verb) or no change of the valency pattern (but note the idiosyncratic meaning) (8) Hinuq ‘hear’ (a) hayɬo-z kečʼ toq-o he.OBL-DAT song hear-PRS ‘he hears a song’ (b) hago tʼek-mo-za-ɬ he toq-liː-ho book-OBL-OBL.PL-CONT hear-ANTIP-PRS ‘he often rummages in the books’ (9) Hinuq ‘forget’ diž debe roži ̌uƛʼ-o / ̌uƛʼe-doː-ho I.DAT you.SG.GEN1 word forget-PRS / forget-ANTIP-PRS ‘I (often) forget your words’ 5. Godoberi 5.1 Overview (i) marked alternation - suffix -a - with some verbs additional marking: reduplication and addition of (other?) suffixes - formation is morphologically restricted it can be derived only from the verbs with thematic vowel -i for verbs with thematic vowel -a the resulting form would be identical to the perfect converb and the relevant form can only be interpreted as a perfect converb - the formation is irregular; it is impossible to predict the form of the antipassive from the form of the original stem (ii) can be formed from intransitive, transitive as well as some A-labile verbs - not fully productive: attested for around 60 verbs - semantically, the majority of these verbs can be characterized as follows: they tend to be activities (implying no natural limit) to have animate As as their primary arguments to have inanimate Ps (if transitive) (iii) valency change from transitive to intransitive (iv) general meaning durative (‘be busy, be engaged’) - close to the meaning of biabsolutive constructions: the A is highlighted but with the antipassive the P is suppressed 4 - speakers consistently comment that antipassive clauses are readily interpretable as answers to questions of the type “What is Ali doing? Where is Ali?” - antipassives sometimes acquire a semi-lexicalized additional meaning; for instance, girgisada (antipassive of ‘shake’) frequently means ‘is feverish’ 5.2 Syntax - antipassives of intransitive: no change in valency (10) a. mat'u girgis-at-a-da glass b. mat'u girgis-a-da shake-PRS-CVB-COP ‘the glass shakes’ glass shake-ANTIP.CVB-COP ‘the glass shakes’ - antipassive of transitive verb: ergative A of source verb marked with absolutive; omission of P of source verb (11) a. ˁali(-di) q'iru b. ˁali w-ol-a-da b-el-at-a-da Ali(-ERG) wheat N-thresh-PRS-CVB-COP ‘Ali is threshing wheat’ Ali M-thresh-ANTIP.CVB-COP ‘Ali is threshing’ 6. Dargwa languages: Akusha (Standard) Dargwa and Sanzhi Dargwa 6.1 Overview - no derivational morphology or other formal marking on the verb - can only be formed from transitive verbs (we have no information on ditransitive verbs) - not fully productive (e.g. in Standard Dargwa 234 verbs have been tested, out of which 93 form the antipassive and 17 are ambiguous between genuine antipassive and ‘antipassive look-alike’) - largely (if not fully) restricted to imperfective verb stems and consequently to those tenses that are available for verbs with imperfective stems (e.g. progressives, habituals, futures); unavailable in the aorist or resultative; exceptions: the verbs ‘eat’ and ‘drink’, which are generally much more flexible in their morphosyntactic behavior than other verbs and allow for antipassive constructions with more tenses than other verbs - valency change from transitive to intransitive - restriction on possible P arguments (person, animacy) - general meaning habitual/iterative - antipassives may describe an event as incomplete, or non-punctual, but this is due to the fact that they occur only with imperfective verb stems 6.2 Syntax - valency change from transitive to intransitive: - ergative A of source verb marked with absolutive - absolutive P of source verb marked with ergative (also used for instruments) and in Standard Dargwa occasionally with a spatial case 5 - in Standard Dargwa the Future tense has different suffixes for intransitive and transitive verbs, and with the antipassive construction the suffix for intransitive verbs is used (12) Standard Dargwa džuz (a) dudě.li b-uč’-ul sa.y father.ERG book.ABS N-read.IPFV-CVB COP.M ‘Father is reading a book’ [transitive] džuz.li // džuz.li-zi-w // dzuz-a // Ø (b) dudě uč’-ul sa.y father.ABS book.ERG// book-ILL-M // book-PL.ERG // Ø M.read.IPFV-CVB COP.M ‘Father is engaged in reading (a book // in a book // books)’ [antipassive] - in Sanzhi Dargwa it seems that only the A argument, but never the P argument can be omitted - imperative and prohibitive suffixes show that antipassivization is a detransitivizing operation: -ut is used with intransitive verbs, whereas -it is used for transitive verbs (13) Sanzhi Dargwa (a) dig ma-b-uk-it! meat PROH-N-eat.IPFV-PROH.SG ‘Do not eat the meat!’ [transitive] (b) dig-li ma-w-k-ut! meat-ERG PROH-M-eat.IPFV-PROH.SG ‘Do not eat (always) meat!’ [antipassive] - verbs for which the antipassive is not available are those verbs for which it is unclear what the result of the action that they denote would be - in Sanzhi Dargwa the antipassive is not available with first or second person P - there are animacy restrictions: it is impossible for both A and P to be animate or inanimate - the last two constraints are not really syntactic in nature since the resulting clauses are normally grammatical, but with an awkward meaning (van den Berg calls them ‘antipassive look-alikesʼ) (14) Sanzhi Dargwa (a) it-i-l t'ult' b-uk-unne ca-b this-OBL-ERG bread N-eat.IPFV-CVB be-N ‘he is eating bread’ [transitive] (b) it t'ult'-li Ø-uk-unne ca-w this bread-ERG M-eat.IPFV-CVB be-M ‘he is eating bread’ [antipassive] 6 (15) Sanzhi Dargwa (a) aždaha-l du Ø-ukː-unne=da monster-ERG 1SG M-eat.IPFV-ICVB=1 ‘the monster is eating me (M) [transitive] aždaha (b) du-l b-ukː-unne=da 1SG-ERG monster N-eat.IPFV-ICVB=1 ‘I am eating the monster’ NOT: ‘the monster is eating me’ [transitive] 6.3 Semantics & pragmatics - the oblique (demoted) object has a low degree of identifiability in the proposition, as it is indefinite and/or non-referential (usually plural or omitted) - reduction of the effect of the verbal activity onto the object, which does not necessarily undergo a change of state - oblique P is usually plural and can with some verbs also be in the singular; it predominantly has an inanimate referent; preference hierarchy for referents occurring as oblique objects: (16) plural (inanimate > animate) > inanimate singular - activity understood as occurring habitually / frequently, e.g. antipassive ‘drink’ may be interpreted as ‘being drunk (of alcohol)’, ‘read’ as ‘study (of pupils/students)’, ‘steal’ as ‘thieve, live dishonestly’ (17) Sanzhi Dargwa heχ-tːi piwa-l b-učː-ul ca-b this.down-PL beer-ERG HPL-drink.IPFV-ICVB COP-HPL ‘they are (regularly) engaged in drinking beer’ [antipassive] (18) Sanzhi Dargwa hež-i-l χabar b-urs-ul this-OBL-ERG story N-tell-ICVB ‘He is telling the story.’ [transitive] ca-b COP-N (19) Sanzhi Dargwa hež χabur-t-a-l ux-ul ca-w this story-PL-OBL-ERG tell.M.IPFV-ICVB COP-M ‘He is narrating.’ [antipassive] 7 7. Summary - Nakh-Daghestanian antipassives differ considerably from canonical antipassive constructions - only Dargwa and Bezhta allow the P to be expressed in an oblique case - in Standard Dargwa, expression of the P is optional; we do not have information on whether it is optional or obligatory in Sanzhi Dargwa. - in all languages other than Dargwa, the derivation is also available for intransitive verbs - the antipassive is not very productive and generally only available with a limited number of verbs; this is at least partially due to its semantics - the antipassive commonly has a durative, iterative or habitual meaning (Comrie & Khalilov 2007) - its use is (almost) entirely determined by its semantic load - accessibility to major syntactic processes in Daghestanian languages is not generally constrained by grammatical relations, so the antipassive does not serve to feed syntactic processes as in, for instance, Dyirbal (Dixon 1972). References Bokarev, Evgenij A. 1949. Sintaksis avarskogo jazyka. Moskva: Izd-vo AN SSSR. Comrie, Bernard & Madzhid Khalilov. 2007. Durativnye i iterativnye glagoly v dagestankix jazykax. In G. G. Gamzatov (ed.): Nasledie kak sistema cennostej, 531-541. Makhachkala: IJaLi DNC RAN. Comrie, Bernard, Madzhid Khalilov & Zaira Khalilova. 2015. Valency and valency classes in Bezhta. In Andrej Malchukov & Bernard Comrie (eds.), Valency classes: a comparative handbook. Berlin: De Gruyter Mouton, 541-570. Dixon, R. M. W. 1972. The Dyirbal language of north Queensland. Cambridge: Cambridge University Press. Forker, Diana. 2013. A grammar of Hinuq. Berlin: De Gruyter Mouton. Kibrik, Andrej A. 1996. Transitivity in lexicon and grammar. In Aleksandr E. Kibrik (ed.): Godoberi, 108-147. München: Lincom. Mallaeva, Zulajxat M. 2007. Glagol avarskogo jazyka: struktura, semantika, funkcii. Makhachkala: Institut jazyka, literatury i iskusstva im. G. Cadasy. Nichols, Johanna, David A. Peterson & Jonathan Barnes. 2004. Transitivizing and detransitivizing languages. Linguistic Typology 8: 149-211. Schulze, Wolfgang. 2011. The Lak language. Nová filologická revue 3: 11-36. Sumbatova, Nina R. & Rasul O. Mutalov. 2003. A grammar of Icari Dargwa. München: Lincom Europa. Tatevosov, Sergei. 2011. Detelicization and argument suppression: evidence from Godoberi. Linguistics 49: 135-174. van den Berg, Helma. 2001. Dargi folktales: oral stories from the Caucasus with an introduction to Dargi grammar. Leiden: Research School of Asian, African and Amerindian Studies. van den Berg, Helma. 2003. Antipassive constructions in Nakh-Daghestanian languages. Leipzig: MS. 8