Philip Shushurin
Institute of Linguistic Studies, Moscow / Higher School of Economics, Moscow
phshushu@gmail.com
On verbs with lexicalized agreement markers in Chechen
Keywords: Nakh-Dagestanian, agreement, ergativity, alignment, Finno-Ugric, typology
Abstract. The squib discusses the behavior of two verbs in Chechen which contain lexically determined but syntactically active (object) agreement markers. I show that such verbs are true intransitives and do not contain a covert/elided direct object. The squib furthermore compares such verbs with a similar case in Moksha and raises the question of whether such phenomenon is typologically widespread.
Introduction
In the following Chechen sentence, the predicate is a one-place verb j-eša ‘read (intr.)’, ‘to be engaged in reading’.
|
|
|
|
|
|
|
|
|
|
‘Aslan is reading [is engaged with reading]’1
|
Non-typically for one-place verbs in the language, this verb (a) contains a lexically determined class prefix (j-) which does not agree with the subject, and (b) triggers Ergative, rather than Absolutive marking on the subject. The squib discusses the properties of this un(der)described construction and shows that these verbs are true intransitives with the class prefix containing lexicalized agreement features. The Chechen case is furthermore compared with a similar case in Moksha, a Finno-Ugric language.
Data
Chechen is an Ergative language of the Nakh-Dagestanian family. The language has predominant SOV basic word order and Ergative alignment. The verbs in Chechen exhibit quite rich verbal morphology, which includes prefixal agreement markers which cross-reference the gender (noun class) and number features of an Absolutive argument. Agreement with a DP in any other case (viz. Ergative) is impossible. As a rule, the predicative agreement in the language is universally with the Absolutive argument, irrespective of the syntactic position, grammatical function or semantic interpretation of the Absolutive argument. All Chechen verbs can be divided into two groups depending on whether they do or do not contain a morphological slot for a class/number prefix. There are no syntactic differences between the verbs of the two groups (Jakovlev 1960, Dotton & Doyle 2017).
The main focus of the paper are the following two Chechen verbs which contain a class prefix despite not containing a direct object: viz. j-eša ‘read (intr.)’ and d-eša ‘study (intr.)’. The lexically determined features in such verbs are nonetheless syntactically active and can show up on auxiliary verbs. Ergative marking on the subject is obligatory.
|
|
|
|
|
|
|
|
|
|
|
|
|
|
‘Aslan will read/will study’
|
The intransitive verb j-eša ‘read (intr.)’ can be contrasted with the transitive verb CL-eša, whose class/number prefix covaries with the features of the direct object.
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
‘Aslan read a book/ read a magazine’
|
The phenomenon of certain prototypical transitive verbs (such as read or eat) to have apparently object-less uses has been described in the literature, with the verbs partaking in this phenomenon being variously called “pseudointransitive”, “labile”, or “ambitransitive”. The intransitive uses of such verbs have been referred to as (context-independent) Indefinite Object Deletion (IOD) (see e.g. (Næss 2007), (Næss 2009) for a discussion).
There is evidence that from a syntactic point of view, the verbs in question are true intransitive verbs and do not project syntactically active direct objects. First, one can show that the examples such as (1) do not contain an elided NP kniga ‘a book’. The evidence comes from the difference in the semantic interpretation of (4) and (5). While (4), similarly to its English counterpart, implies that the book has been read to the end, no such implication is present in (5). An alternation hypothesis namely that the examples like (1) may contain an elided indefinite pronoun such as something can be dismissed given the fact that at least two different prefixes (viz. j- and d-) can be used in such construction – it is not clear why the elided pronoun would trigger different class/number agreement depending on the semantic context. (The possibility of the existence of two different, otherwise unattested null/elided indefinite pronouns should perhaps be dismissed as marginal and lacking independent ground).
|
|
|
|
|
|
|
|
|
|
|
|
‘Aslan has read a book/the book’
|
Both verbs historically derive from the transitive verb -eša, which is an ordinary transitive verb that takes object agreement markers depending on the gender/number of the Absolutive direct object. The lexicalized prefix j- presumably derives from the class/number prefix j- which is a result of agreement with the noun kniga, while the verb d-eša ‘study (intr.)’ presumably derives from the expression knigaš deša ‘read books’, where d- is a marker of plurality in the class to which the noun kniga belongs to (viz. class j/d).
It seems reasonable to assume that, from a diachronic perspective, the two Chechen verbs in question derive from their transitive analogues. This can be evidenced by the fact that a cognate transitive verb exists in the closely related Ingush language. More precisely, the verb j-eša in Ingush with the lexicalized agreement prefix j- can mean both ‘read’ and ‘study’ (Nichols 2011). These facts can be naturally explained if one assumes that Proto-Chechen-Ingush (Proto-Vainakh) possessed the transitive verb *eša and the two languages each individually have developed lexicalized agreement markers (LAMs) along unique lexicalization patterns.
Discussion
To sum up the data, we have seen several instances of intransitive verbs which contain a lexicalized agreement marker in the absence of the crossreferenced nominal argument.
Although the phenomenon in question may be rare, I am aware of at least one language outside of the Nakh-Dagestanian family which possesses verbs with LAMs. In Moksha, a Finno-Ugric language, the verb pozdandams ‘be late’, despite being intransitive, contains an object agreement marker. While the construction is intransitive, the verb pozdandams contains the 3SG object marker.
|
|
|
pozdanda-z'/*pozdanda-s'-t'
|
|
|
|
|
|
|
be.late-PST.3.O.3PL.S/be.late-PST.3-PL
|
|
|
|
|
|
‘They were five minutes late for the bus’
|
[Toldova et al. 2018]
Lexicalized agreement markers should be sharply contrasted with agreement markers attached to bound stems. In such case, the lexical entry of such predicate may contain a (perhaps, default) value of the agreement marker. However, this value is not fixed with respect to the lexical entry and is subject to varying along with the phi-features of the cross-referenced argument. For instance, the class markers in Chechen verbs attach to morphologically bound stems, with the verbs appearing in dictionaries together with a class marker (d-, by convention). However, this marker is not lexicalized as the value of this marker would covary along with the agreed-with argument. In LAM predicates, in contrast, the choice of a particular agreement marker is associated with particular lexical meaning.
Yet another case of morphologically bound but crucially not lexicalized agreement markers can be found in the nominal domain with possessive agreement. In Navajo, for instance (see (7) below), the nominal stem -be' ‘milk’ is morphologically bound as it cannot appear without the possessive agreement marker (7c). However, no value of this agreement marker is exclusively associated with the lexical entry ‘milk’. To conclude, such cases are not instances of LAM either.
|
|
Navajo (Young & Morgan 1987:3)
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
Although the squib specifically discusses verbs which have lexicalized (direct) object agreement markers, other syntactic configurations are of course theoretically possible. For instance, one may think of a one-place predicate with lexicalized subject agreement marker. Whether such verbs exist is an intriguing open question.
Although the fact of the existence of lexicalized agreement markers is by no means new in the literature (see for instance Chumakina & Corbett (2008) on fossilized gender markers on Archi nouns), I am not aware of any works dedicated to typology of lexicalized agreement markers on verbs, in particular of those that result in instances of ambitransitivity. Meanwhile, the described facts raise several potentially far-reaching questions, such as the following.
- What are some languages which possess lexicalized agreement markers (LAMs), besides Nakh and Moksha languages?
- Do all one-place LAM verbs pattern with unergatives rather with unaccusatives?
- Do lexicalized agreement markers universally derive from Object, rather than Subject Agreement? If not, are Subject and Object LAMs in an implicational hierarchy, that is to say, if there are any subject LAM, are there any verbs with subject LAMs and syntactically active object Agreement markers?
- Are there any two-place verbs with LAM?
- Are there any verbs with Object LAM in Ergative languages which trigger Absolutive rather than Ergative marking on the subject?
Abbreviations
3 — 3 person; ABL — ablative; AUX — auxiliary; DAT — dative; DEF — definite; ERG — ergative; IRR — irrealis; O — object; PL — plural; PRF — perfect; PST — past; S — subject; SG — singular
Bibliography
Chumakina & Corbett 2017 — Chumakina, Marina & Greville G. Corbett. 2008. Archi: The challenge of an extreme agreement system. In Aleksandr V. Arkhipov, Leonid M. Zakharov, Andrej A. Kibrik, Aleksandr E. Kibrik, Irina M. Kobozeva, Olga F. Krivnova, Ekaterina A. Ljutikova, Olga V. Fedorova (eds.). Fonetika i nefonetika. K 70-letiju Sandro V. Kodzasova (Фонетика и нефонетика. К 70-летию Сандро В. Кодзасова.) [Phonetics and nonphonetics. For the 70th anniversary of Sandro V. Kodzasov], 184-194. Moscow: Jazyki slavjanskih kul’tur.
Dotton & Wagner 2017 — Dotton, Zura & John Doyle Wagner. 2017. A Grammar of Chechen. Durham, NC: Duke University Slavic and East European Language Resource Center.
Jakovlev 1960 — Jakovlev, Nikolaj F. (Яковлев, Николай Ф.) 1960. Morfologiya chechenskogo yazyka (Морфология чеченского языка) [Morphology of the Chechen language]. Groznyj: Checheno-ingushskoe knizhnoe izdatel’stvo.
Næss 2007 — Næss, Åshild. 2007. Prototypical Transitivity (Typological Studies in Language 72). Amsterdam: John Benjamins Publishing Company.
Næss 2009 — Næss, Åshild. 2009. How transitive are EAT and DRINK verbs. In John Newman (ed.). The linguistics of eating and drinking, 27-43. Amsterdam: John Benjamins Publishing Company.
Nichols 2011 — Nichols, Johanna. 2011. Ingush grammar. Berkeley — Los Angeles — London: University of California Press.
Toldova et al. 2018 — Toldova, Svetlana Y., Maria A. Kholodilova, Sergej G. Tatevosov, Egor V. Kashkin, Aleksej A. Kozlov, Lev S. Kozlov, Anton V. Kukhto, Maria Y. Privizentseva & Ivan A. Stenin (eds.). 2018. Elementy mokšanskogo jazyka v tipologičeskom osveščenii (Элементы мокшанского языка в типологическом освещении) [Elements of the Moksha language from the typological perspective]. Moscow: Buki Vedi.
Young & Morgan 1987 — Young, Robert W. & William Morgan. 1987. The Navajo Language: A Grammar and Colloquial Dictionary. Albuquerque: University of New Mexico Press. (revised edition).
1The results are obtained in the course of fieldwork with three native speakers from Shatoy, Chechnya.