Arseniy Vydrin, Saint Petersburg

Are there depictives in Adyghe?

1. Introduction1 The term “depictive secondary predicate” (or simply “depictive”) is used here in the sense specified by Eva Schultze-Berndt and Nikolaus P. Himmelmann (SCHULTZE-BERNDT & HIMMELMANN 2004; HIMMELMANN & SCHULTZEBERNDT 2005), who developed the following seven criteria for distinguishing depictive secondary predicates from simple predicates and adverbs (2004, 7778): A depictive secondary predicate construction is a clause-level construction which meets the following seven criteria: i. It contains two separate predicative elements, the main predicate and the depictive, where the state of affairs expressed by the depictive holds within the time frame of the eventuality expressed by the main predicate. ii. The depictive is obligatorily controlled, i.e., there exists a formal relation to one participant of the main predicate, the controller, which is usually interpreted as a predicative relationship (i.e., the depictive predicates an eventuality of the controller). The controller is not expressed separately as an argument of the depictive. iii. The depictive makes a predication about its controller which is at least in part independent of the predication conveyed by the main predicate, i.e., the depictive does not form a complex or periphrastic predicate with the main predicate. iv. The depictive is not an argument of the main predicate, i.e., it is not obligatory. v. The depictive does not form a low-level constituent with the controller, i.e., it does not function as a modifier of the controller. vi. The depictive is non-finite (it is not marked for tense or mood categories), or the dependency of the depictive on the main predicate is indicated in other formal ways.

1

Financial support for this study was provided by RGNF 06-04-00194а. I thank Yakov Testelets, Yuriy Lander, Christoph Schroeder and Winfried Boeder for their comments on the paper, as well as all participants of the second workshop on depictive secondary predicates (Oldenburg, December 7–10, 2005) for their discussion of the Adyghe material presented in the workshop. I also express my gratitude to all my consultants, Adyghe native speakers, especially to Rita Tlevtsezheva who provided most of the examples included in this paper. All errors are mine.

vii. The depictive is part of the same prosodic unit as the main predicate.

This concept of depictives is certainly useful for many European languages. Yet it is by no means evident that it should work for languages of a different typology. In this paper, I will discuss this issue using the data of Adyghe, a polysynthetic ergative language of the West Caucasian (Abkhaz-Adyghe) family. Adyghe is spoken by around 130 000 people, mainly in the Adyghe Republic, Russian Federation. Native Adyghe speakers are also found in Turkey, Syria and other countries (KOŽEMJAKINA 2000, 37-38; KORJAKOV 2006, 22-23). Adyghe is subdivided into the Abzakh, Shapsug, Bzhedugh and Temirgoi (Chamgui) dialects. This study is based on the Temirgoi dialect as spoken in the village of Haqurinohabl, where it is influenced by the Abzakh dialect spoken nowadays by some of the village residents.2 In the first section, I will examine the main functions of the suffix -ew, the only suffix which can express the depictive meaning. Then I will show that adjectives with depictive meaning obligatorily agree with their controllers in person and that forms in -ew with person agreement can be marked by different tense forms. The latter means that Adyghe forms in -ew violate the first criterion mentioned above (being in the time frame of the eventuality expressed by the main predicate). Thus according to Schultze-Berndt and Himmelmann’s criteria they cannot be considered as depictive secondary predicates. Person agreement is a property of predicates in Adyghe. As to participantoriented adjectives in -ew, only a small number of them can express depictive meaning without person agreement markers. This class of adjectives is unstable and varies with different speakers. Only these adjectives can be called true depictives in Adyghe. In the following section, I will examine the use of -ew with nouns (He came to the meeting a s a d ir ec tor ). Nouns with depictive meaning permit omission of person agreement with their controllers and cannot be marked by tense suffixes. They are more productive than adjectives, which permit the omission of person agreement while preserving their orientation to one of the participants. The only thing that might prevent us from classifying participant-oriented nouns in -ew as true depictives is the fact that they are not in the scope of negation of the main predicate. I conclude with the hypothesis, which needs to be tested on a wider range of material, that in a polysynthetic language with extensive agreement and weak

2

The material for this paper was collected during my fieldwork in Haqurinohabl, Shovgenovsky District, Adyghe Republic, Russian Federation in 2005 and 2006. The Abzakh variety represented there can be somewhat different from the variety described by Catherine PARIS (1989).

424

contrast between noun and verb, true depictives in the sense of Schultze-Berndt and Himmelmann may not be a productive phenomenon.

2. The suffix -ew The only candidate for a depictive marker is the suffix -ew, which has at least four functions.3 2.1 Main functions of the suffix -ew 1) The suffix -ew converts adjectives into adverbs (1)-(3): (1) psEnB’e — psEnB’-ew ‘fast (adj.); light, easy (adj.)’ ‘quickly; easily’ (2) fabe — fab-ew ‘warm’ ‘warmly’ (3) daxe — dax-ew ‘beautiful’ ‘beautifully’ 2) -ew optionally marks the second argument ((5) vs. (4)) or the third argument ((7) vs. (6), (8) vs. (7)) of some predicates such as ‘become’, ‘nominate’, ‘consider’. Grammars call -ew in this function an “adverbial case” (JAKOVLEV & AŠXAMAF 1941; ROGAVA & KERAŠEVA 1966; ZEKOX 2002). (4) maSjEnjEst sE-HwE-S’t engine.driver 1SG-become-IRR ‘I am going to be an engine driver (when I grow up).’ (5) maSjEnjEst-ew sE-HwE-S’t engine.driver-ADV 1SG-become-IRR ‘I will become an engine driver (when I grow up).’ (6) djErektor s-a-S&E-R-ep director 1SG-3PL-make-PST-NEG ‘I was not nominated director.’ (7) djErector-ew s-a-S&E-R-ep director-ADV 1SG-3PL-make-PST-NEG ‘I was not nominated director.’

3

In this paper I will not discuss the functions of this suffix in the closely related Kabardian language, where it may have somewhat different functions. For -ew in Kabardian, see JAKOVLEV (1948); ABITOV et al. (1957); KUIPERS (1962); BAGOV (1970); COLARUSSO (1992).

425

ze-se-LEte-Z’E se sE-{wS-ew I 1SG-clever-ADV REFL-1SG-consider-RE ‘I consider myself clever.’ 3) -ew can be used to form “adverbial participles” (converbs) from verbal stems. Giorgi Rogava and Zainab Keraševa (ROGAVA & KERAŠEVA 1966, 170-236) state that adverbial participles cannot be marked for mood categories, though they can be used with all tense suffixes of dynamic verbs: (9) Kwen ‘to go’ w K -ew ‘while he is going’ (go-ADV) Kwa-R-ew ‘while he was going’ (go-PST-ADV) w K e-Ra-R-ew ‘while he had been going’ (go-PST-PST-ADV) KwE-St-ew ‘while he will go’ (go-IRR-ADV) Tense in Adyghe adverbial participles does not obligatorily coincide with the tense of the main predicate. Thus they do not satisfy Schultze-Berndt & Himmelmann’s first and sixth criteria for a depictive secondary predicate construction. As Adyghe adverbial participles cannot be candidates for depictives, I do not discuss them in this paper. Adyghe also possesses converbs marked by -ze (dialectal -re, -zere), -me (conditional suffix), -B’e (labelled “conditional second” in some grammars) and their various derivates. All of them can be marked by various tense forms and will not be discussed here. 4) -ew can mark relative construction heads (10) se s-S&e-r-ep njepe gWEpSEs-ew E-gWE jE-LE-r I 1SG-know-DYN-NEG today idea-ADV 3SG-heart LOC-lie-ABS ‘I don’t know the ideas he has today’. (8)

2.2 -ew constructions with depictive meaning The following example shows that -ew can form constructions with depictive meaning. It can have two readings: ‘Ramazan is walking around Moscow joyfully’, where C’ef-ew is an event-oriented adverb; and ‘Ramazan is joyful as he walks around Moscow’ where C’ef-ew is participant-oriented. (11) ramazan C’ef-ew moskwa qE-S’-je-KwEhe Ramazan joyful-ADV Moscow DIR-LOC-LOC-walk ‘Ramazan is walking around Moscow being joyful / joyfully.’ This brings up the issue of how one can formally distinguish between participant-oriented and event-oriented forms in -ew. 426

One of the criteria that could be used for distinguishing between event-oriented adverbs and participant-oriented adverbs in Adyghe seems to be the capacity of the latter to take the plural agreement suffix -x. Usually, this suffix is not found with event-oriented adverbs (12), while it is obligatory with participant-oriented adverbs when the controller is in the plural (13). adEga-bze-r (12) SE-re SEpHwE-re C’ef-ew brother-& sister-& joyful-ADV Adyghe-language-ABS zer-a-Ra-I&e RECP-3PL-CAUS-know ‘The brother and the sister learn the Adyghe language with enjoyment [“joyfully”].’ (13) SE-re SEpHwE-re C’efE-x-ew adEga-bze-r brother-& sister-& joyful-PL-ADV Adyghe-language-ABS zer-a-Ra-I&e RECP-3PL-CAUS-know ‘The brotheri and the sisterj learn the Adyghe language with enjoyment [“joyfuli+j”].’ Nonetheless (14) and (15) show that this contrast is not strong, since here we find the plural marker with e v e n t - o r i e n t e d forms in -ew. However, note that in these examples the adverbials are separated from their possible controllers. This suggests that p a r t i c i p a n t - o r i e n t e d forms in -ew with the plural marker cannot be separated from their controllers. By contrast, e v e n t o r i e n t e d forms in -ew can be marked with the plural marker only if they are placed near the predicate. (14) SE-re SEpHwE-re adEga-bze-r C’efE-x-ew brother-& sister-& Adyghe-language-ABS joyful-PL-ADV zer-a-Ra-I&e RECP-3PL-CAUS-know ‘The brotheri and sisterj learn the Adyghe language joyfully / *joyfuli+j.’ (15) SE-re SEpHwE-re jeGaPe-m C’efE-x-ew brother-& sister-& school-ERG joyful-PL-ADV

427

adEga-bze-r zer-a-Ra-I&e Adyghe-language-ABS RECP-3PL-CAUS-know ‘The brother and sister joyfully learn the Adyghe language at school.’ *‘The brotheri and sisterj learn the Adyghe language at school joyfuli+j.’ The question we have to answer is what C’ef-x-ew ‘joyful’ in (13) is. Is it a specialized participant-oriented adverb (i. e. a depictive) or a general subordinated converb similar to what we saw in 2.1? In the next section I will show that in Adyghe, the number agreement of participant-oriented forms in -ew like C’ef-x-ew ‘joyful’ in (13) and their adjacency to the controller are not sufficient to classify them as depictives.

3. Person agreement of participant-oriented forms in -ew 3.1 Person agreement in Adyghe Adyghe verbs are divided into dynamic and static classes. Table 1 shows the way the static verb S’EtEn ‘to stand’ agrees with its absolutive argument. SG 1. se sE-S’Et ‘I am standing’ 2. we wE-S’Et ‘You are standing’ 3. ar Ø-S’Et ‘S/he is standing’

PL te tE-S’Et ‘We are standing’ Swe SwE-S’Et ‘You are standing’ axer Ø-S’EtE-x ‘They are standing’

Table 1: Verbal person agreement

Notice that the 3rd person singular remains unmarked. The plural is conveyed by the suffix -x. The same person markers occur in predicative nouns: SG 1. se sE-pIaI ‘I am a girl’ 2. we wE-pIaI ‘You are a girl’ 3. ar Ø-pIaI ‘She is a girl’

PL te tE-pIaI ‘We are girls’ Swe SwE-pIaI ‘You are girls’ axer Ø-pIaIe-x ‘They are girls’

Table 2: Nominal person agreement

428

Participant-oriented forms in -ew obligatorily agree with their controller in person. With 1st and 2nd person plural controllers, they can have an optional plural marker -x. (16) sE-VEKw-ew bjEbljEje-m s-jE-Ga-R 1SG-little-ADV Bible-ERG 1SG-3SG-read-PST ‘I read the Bible being a child’ (‘When I was a child I read the Bible’). (17) te wolejbol LeS-ew tE-rE-raze-[x]-ew we volleyball strong-ADV 1PL-INSTR-pleased-[PL]-ADV t-je-I&e 1PL-3SG-play ‘We play volleyball with great pleasure.’ (18) we wE-LapV-ew wE-q-jEB’e-R se you 2SG-barefooted-ADV 2SG-DIR-go.out-PST I sE-b-Ra-Kwete-Z’E-n-ew 1SG-2SG-CAUS-go-RE-POT-ADV ‘You went out barefooted to see me off.’ (19) Iwe IwE-VEne-[x]-ew konfet qE-Iw-a-tE-S’t-ep you.PL 2PL-wet-[PL]-ADV sweets DIR-2PL-3PL-give-IRR-NEG ‘Being wet you will not get the sweets.’ (‘While/since you are wet they will not give you the sweets.’) We may conclude that in examples like (13) (repeated here for convenience as (20)) forms in -ew are zero-marked for the 3rd person: (20) SE-re SEpHwE-re Ø-C’efE-x-ew jeGaPe-m brother-& sister-& 3-joyful-PL-ADV school-ERG adEga-bze-r zer-a-Ra-I&e Adyghe-language-ABS RECP-3PL-CAUS-know ‘The brotheri and the sisterj learn the Adyghe language with enjoyment [“joyfuli+j.”].’ 3.2 Controllers Most Adyghe predicates can be categorized as monovalent intransitive (Kwen ‘to go’), bivalent intransitive (jeZen ‘to wait’), trivalent intransitive (tE-qE-dE-we-ZaR 1PL-DIR-COM-2SG-wait-PST ‘we were waiting for you with him’), bivalent transitive (edEn ‘to sew’), or trivalent transitive (qE-z-de-p-S’a-R DIR-1SG-COM-2SGbring-PST ‘you brought him with me’).

429

Participant-oriented forms in -ew with person agreement can be used with different types of main predicates (intransitive, bivalent transitive, trivalent transitive) and can be controlled by subject of intransitive predicate (21), indirect object of intransitive predicate (22), subject of bivalent transitive predicate (23), direct object of bivalent transitive verb (24), subject of trivalent transitive predicate (25), direct object of trivalent transitive predicate (26) and indirect object of trivalent transitive verb (27). (21) se sE-{wES-ew we sE-b-de-Kwe I 1SG-clever-ADV you 1SG-2SG-COM-go ‘Being clever I go with you.’ (‘Because I’m clever I’m going to marry you’ / ‘I’m clever and I’m going to marry you.’ (ironic)) (22) se we wE-{wES-ew sE-b-de-Kwe I you 2SG-clever-ADV 1SG-2SG-COM-go ‘Ii go with youj being cleverj.’ (‘I’m going to marry you because you are so clever.’ (ironic)) (23) we wE-dax-ew qepLanE-r w-e-wEB’E you 2SG-beautiful-ADV tiger-ABS 2SG-3SG-kill ‘Being beautiful you are killing the tiger.’ (‘Though you are beautiful, you are killing the tiger.’) (24) se we wE-nEbZ’EB’-ew wE-sE-wEB’E-St I you 2SG-young-ADV 2SG-1SG-kill-IRR ‘Ii will kill youj being youngj.’ (‘I will kill you while you are young.’) (25) we wE-dax-ew a-r se qE-z-de-p-S’a-R you 2SG-beautiful-ADV that-ABS I DIR-1SG-COM-2SG-bring-PST ‘You are beautiful and you brought him with me.’ (lit.: ‘Being beautiful you brought her/him with me.’) (26) tE-m a-S’ se sE-nEbZEB’-ew sE-rjE-tE-R father-ERG that-ERG I 1SG-young-ADV 1SG-3SG-give-PST ‘My father gave mei to somebody when I was youngi.’ (lit.: ‘The father gave mei being youngi to somebody.’) (27) a-S’ a-r se sE-dax-ew that-ERG that-ABS I 1SG-beautiful-ADV qE-s-f-jE-S’a-R DIR-1SG-BEN-3SG-bring-PST ‘S/hei brought her/himi to mek when I was beautifulk.’ (lit.: ‘S/hei brought herj/himj to mek being beautifulk.’)

430

3.3 Semantics of forms with -ew Depictive-like forms marked with -ew can refer to manner (13), lifetime (16) and color (29). (28) [S’aj I&wEVe]NP4 wE-jeIwe tea black 2SG-drink ‘You are drinking black tea.’ (29) S’aj I&wEV-ew wE-jeIwe tea black-ADV 2SG-drink ‘You drink tea black.’ (‘You drink tea if it is black.’) -ew can also be used with expressions of similarity (30) and time (31). (30) bl-ew zE-{wanTe-S’tE.Re snake-ADV REFL-curve-IPFV ‘S/he was coiling like a snake.’ (31) VEf-me C’eS’E-ew I&wemEB’ q-a-wERwejE-R man-ERG.PL night-ADV coal DIR-3PL-collect-PST ‘The people were collecting coal at night.’ However these forms are event-oriented. Witness the following examples, where the similitive phrase bl-ew does not agree with its controller in person (32) and number (33). (32) [we] bl-ew / *wE-bl-ew zE-p-{wanTe-S’tE.Re you snake-ADV 2SG-snake-ADV REFL-2SG-curve-IPFV ‘You were coiling like a snake’. (33) axer bl-ew / *ble-x-ew z-a-{wanTe-S’tE.Re they snake-ADV snake-PL-ADV REFL-3PL-сurve-IPFV ‘They were coiling like snakes’. -ew is ungrammatical with group numerals (34), though it is allowed with quantified nominals (35). (34) jure-re se-re t-jE-Tw / *t-jE-Tw-ew twEC’anE-m 5 Yura-& I-& 1PL-LNK -two 1PL-LNK-two-ADV shop-ERG tE-Kwa-R 1PL-go-PST ‘Yura and I both went to the shop.’ 4 5

The structure of NP will be discussed in 3.4. A linker is a special morpheme which is obligatory with numerals. In its form it is identical to the possessive marker. However we consider them different morphemes, because possessive jE is used only with some nouns, while the linkerjE is obligatory with all numerals.

431

(35)

a-xe-r S’EsE-Re-x nebgEr-jE-S’-ew that-PL-ABS sit-PST-PL man-LNK-three-ADV ‘They were sitting together the three of them.’ (lit.: ‘They were sitting as three man.’) In SCHULTZE-BERNDT & HIMMELMANN’S (2004) list of possible depictives one can find expressions with comitative and locative meanings. In Adyghe, in the case of comitatives and locatives, either the -ew construction requires the presence of a subordinated predicate specifying the relation (cf. (36)-(37)) or it is not used at all (examples (38)-(39)). (36) milicionerE-r ha-xe-r jERwEs-ew a-S’ policeman-ABS dog-PL-ABS accompany-ADV that-ERG LEHwE-S’tE.Re seek-IPFV ‘The policeman was seeking her/him with dogs.’ (lit. ‘The policemani was seeking her/himj with dogs accompanying himi.’) (37) *milicionerE-r ha-x-ew a-S’ LEHwE-S’tE.Re policeman-ABS dog-PL-ADV that-ERG seek-IPFV (38) mafe qes qale-m meze-m-B’e we-Kwe day every town-ERG forest-ERG-INS 2SG-go ‘Every day you go to the town through the forest.’ (39) *mafe qes qale-m mez-ew we-Kwe day every town-ERG forest-ADV 2SG-go 3.4 Attributive and -ew constructions When asked to translate Russian sentences with depictive expressions into Adyghe, native speakers often provide attributive constructions. This suggests the idea that attributive and -ew constructions are in some way related to each other. Here I will try to establish what type of attributive construction can be transformed into constructions with -ew. Adyghe has two attributive constructions.6 The first is structured as “semantic head + attributive”, where the attributive is usually an adjective (wEne fEZ’E-r house white-ABS ‘the white house’). The second is structured as “attributive + semantic head”, where the attributive is usually a noun denoting for example 6

It should be noted that the constituents of both constructions phonologically, morphologically and syntactically form a single whole. Arguably, it is more correct to label these constructions “compounding” or “nominal incorporation”. However, a discussion of this alternative analysis would go far beyond the scope of this paper.

432

material (pHe pHenTeKwE-r wood chair-ABS ‘the wooden chair’).7 The former attributive construction can be transformed into a construction with -ew, as in (40)-(41), but the latter cannot (42)-(43). (40) Iwe [bwEtElke neB’E]NP-r IwE-qwEta-Re you.PL bottle empty-ABS 2PL-break-PST ‘You broke the empty bottle.’ (41) Iwe bwEtElke-r neB’-ew IwE-qwEta-Re you.PL bottle-ABS empty-ADV 2PL-break-PST ‘You broke the bottle being empty.’ (‘You broke the bottle while/when it was empty.’) (42) kol’e [pHe pHenTeKwE]NP-r E-Sta-R Kolja wood chair-ABS 3SG-take-PST ‘Kolja took the wooden chair.’ (43) *kol’e pH-ew pHenTeKw-r E-Sta-R Kolja wood-ADV chair-ABS 3SG-take-PST *kol’e pHenTeKw-r pH-ew E-Sta-R Kolja chair-ABS wood-ADV 3SG-take-PST We thus see that the possibility of a presumed depictive formation correlates to some extent with the corresponding type of attributive construction: mainly those depictives are appropriate and meet the criteria which correspond to the “semantic head + attributive” pattern. 3.5 -ew-predicates formed from pronouns Forms in -ew with person markers can even be formed from personal pronouns and the reflexive pronoun jež’ ‘-self’. It is more typical, however, for such forms to serve as arguments of predicates such as HwEn ‘become’ and I&weI&En ‘seem’. Therefore they do not behave as depictives proper. sE-HwE-S’t jEnE sE-HwE-me (44) se sE-we-r-ew 1SG-become-COND I 1SG-you-COP-ADV 1SG-become-IRR big ‘I will be [like] you when I grow up.’

7

The distinction between the first and the second attributive constructions is not clear-cut, nor is the distinction between adjectives and nouns, both being based on the tradition of Adyghe language descriptions. There is a class of “adjectives” which can be postposed and preposed to the semantic head (e. g. names of colors). There is also a class of “nouns” which can be preposed and postposed to the semantic head. Arguably, it is more correct not to distinguish between adjectives and nouns and to use the term “nominal stem” for both.

433

(45)

a-S’ sE-je