Abstract
This study is concerned with finite subordinate clauses of reason in Circassian languages, putting them into a wider areal and typological context. These constructions, introduced by complex conjunctions comprising an interrogative pronoun and a conditional form of a verb of saying, are of special interest for the typology and diachrony of causal markers derived from speech act predicates. In addition to Circassian, they are found in East Caucasian languages, together with ‘say’-based causal markers of a more common type, thus adding North Caucasus to the list of areas for which such structures are typical. They reflect a relatively early stage of grammaticalization of a speech act verb and thus are able to shed light on general mechanisms of such diachronic developments. Being specifically dedicated to encoding of reason, they provide a counter-argument against a common claim that grammaticalized speech act verbs can only evolve into causal markers via the stage of complementizers. Finally, both the morphosyntactic composition of the markers in question and their distribution in the surveyed languages speak against treating them as syntactic borrowings from Russian, although language contact has most likely played a role in their propagation.
Key Words: Adyghe, Circassian, linguistic typology, causal clauses, subordination, grammaticalization
1. Introduction
This paper aims to contribute to the typology of causal/reason clauses (see (Diessel & Hetterle 2011; Martowicz 2011; Hetterle 2015; Kanetani 2019; Santana Covarrubias 2019; Zaika 2019) and references therein for recent advances) by focusing on a distinctive construction type found in Circassian languages, where a complex conjunction involving an interrogative pronoun and a conditional form of the verb of speech is used:
Adyghe
(1) dʷənaj-em-č̣ʼe a-nah bze ze-λ-a-ṣ̂e-re-m
world-obl-ins 1pl.abs-more tongue rfl.io-loc-3pl.a-know-dyn-obl
a-š’ə-š’, səd-a ṗ-ʔwe-me brjətanjəje-šxʷe-m-re
3pl.io-belong what-q 2sg.a-say-cond Britain-big-obl-coord
a.ša.ze-m-re j-a-bze-ŝhaʔe-w š’ə-t
USA-obl-coord 3pl.poss-tongue-main-adv loc-stand
‘It is one of the world’s best known languages, since it is the principal language of Great Britain and the USA.’
(https://ady.wikipedia.org/Инджылызыбзэ)
Kabardian
(2) je-wə-źereč̣ʼə-n xʷjej-š’, sət ŝha-č̣ʼe
loc-2sg.abs-hurry-pot aux-aff what head-ins
žə-ṗ-ʔwe-me zeman-ər q̇ə-p-pepλe-r-q̇əm
back-2sg.a-say-cond time-abs dir-2sg.abs-wait-prs-neg
‘You must hurry, for time does not wait.’ (Dzhaurdzhij & Syqun 1991: 107)
These constructions co-exist with other, presumably more prominent strategies for encoding causal relations, like subordinating suffixes (Rogava & Kerasheva 1966: 182, 424–426; Kumakhov 1989: 274; Zekokh 2002: 344–345; Höhlig 2007) or instrumental participles headed by postpositions (Zekokh 2002: 344–345; Gerasimov & Lander 2008: 293; Arkadiev & Gerasimov 2019: 12–13, 23–24). They appear somewhat atypical for Circassian languages, known for their preponderance of non-finite structures. Most authors have treated these constructions as recent innovations arisen under contact influence from Russian (cf. Kumakhov 1967: 163; 1989: 344).
In what follows, I am going to discuss causal constructions of the kind exemplified in (1)–(2) within a broader typological and areal context. Section 2 presents a general overview of causal markers derived from speech act verbs in the languages of the world and lists some issues they posit for modern linguistic typology and grammaticalization theory. Section 3 focuses on the specific features of Circassian constructions and their closest typological parallels. Section 4 wraps up the discussion.
2. Causal markers derived from speech act verbs
Causal constructions involving markers derived from speech act verbs are far from unique to Circassian. One well known example is, in fact, provided by Turkish (Göksel & Keslake 2005: 400):
Turkish
(3) Sadece Kıvanç var diye
only Kıvanç.nom exist.aor.3sg reas
izle-n-iyor dizi
watch-pass-prog.3sg tv.series.nom
‘This TV series is being watched only because Kıvanç (Tatlıtuğ) is in it.’ (Gündoğdu 2017: 50)
Heine and T. Kuteva (2002: 261) mention say > cause as one of recurrent grammaticalization scenarios for verbs of saying, citing examples from Baka (Ubangian) and Lezgian:
Lezgian
(4) Pul kwadar-na luhuz
money lose-aor saying
buba kwal-er-aj aqud-iz že-da-ni?
father house-pl-iness take.out-inf can-fut-q
‘Can we kick father out of the house because he has lost the money?’ (Haspelmath 1993: 390; Heine & Kuteva 2002: 261)
Further examples are discussed in (Ebert 1991; Lord 1993: 168–172; Güldemann 2008: 464–467) from languages of Africa and in (Saxena 1988, 1995; Chisarik & Wurrf 2003) from languages of Southeast Asia. In the latter area, the constructions in question are found in Dravidian, Indo-Aryan and Tibeto-Burman languages, thus constituting an areal feature. Anna Martowicz (2011: 131) reports 4 languages (out of her worldwide sample of 84) with ‘say’-based causal markers (Sango, Lezgian, Galo, and Thai). Outside of Africa and Eurasia, similar constructions have been reported in South America (Aguaruna, Matses) and the New Guinea (Telefol, Oksapmin).
(Ebert 1991: 87) proposes a universal cline of grammaticalization for quotative markers in clause-linkage:
(5) quote > complement > purpose > reason and/or condition
More detailed but essentially similar proposals are found in (Lord 1993; Saxena 1995; Chisarik & Wurrf 2003) and a number of other works (see references in (Martowicz 2011: 71)). They all assume that a grammaticalized unit evolved from a speech act verb or other quotation marker can not develop a causal meaning before it has acquired more basic functions of complementation and/or purpose, as testified, e.g., by Turkish diye or Besermyan šu(ǝ)sa (Serdobolskaya & Toldova 2011). Such hypotheses are, however, criticized in (Güldemann 2008: 444–446) as being too far-reaching and overgeneralizing certain area-specific patterns. A particularly strong counterexample can be found in Oksapmin (Trans-New-Guinean), where verbs li– ‘say’ and pl– ‘tell’ have developed a range of functional uses including reason, but excluding complementation (even with other speech act verbs, as all reported speech in Oksapmin is direct) (Loughnane 2009: 321, 426).
Another issue that arises in connection with causal markers descending from verbs of speech and that has been only cursorily tackled in the literature is the range of causal relations that they are able to express. At least in those languages for which we have enough data, such markers are not the primary means of encoding reason (just like in Circassian). Examples presented in reference grammars mostly involve not objective causation per se, but a motivation behind a decision made by a volitional agent. For Lezgian, M. Haspelmath writes explicitly:
“However, luhuz/lahana expresses an internal subjective motivation rather than an objective cause, reflecting the original meaning ‘saying/having said’” [Haspelmath 1993: 390].
Such restriction in the use of the structures in question seems only natural given their diachronic origins. They descend from constructions with speech act verbs, in which a proposition is stated by a participant (presumably as a motivation for a certain course of action or an argument for a certain claim) and are likely to first emerge in contexts involving a sentient participant whose actions call for justification in their own subjective terms. Further extension into the domain of non-sentient objects and objective causal relations is problematic, as grammaticalization in structures of this kind is believed to proceed unidirectionally along the subjectivity scale [Traugott 1995].
3. Constructions of the “Circassian” type
The constructions of the kind found in Circassian, exemplified in (1)–(2) above, stand out in a number of respects. They have a fairly transparent structure, with the morphological make-up of the speech act verb clearly discernible, and thus can be expected to reflect a relatively early stage of grammaticalization. In addition, they contain an interrogative pronoun and can easily traced back to a reanalysis of a conditional structure: ‘P, if you say why/what for, Q’. What’s more, unlike the majority of cases discussed in the literature, they can be related to constructions of reported discourse representing the speech not of a discourse participant, but of one of the locutors, either speaker or hearer (the grammaticalized verb is marked for 2nd singular, but this form is employed in impersonal meaning in Circassian). They thus represent not the subjective, but rather the intersubjective dimension.
Overall, with respect to their structural characteristics, diachronic development, and presumably semantics. the Circassian constructions form a distinct subtype of causal clauses introduced by markers derived from speech act verbs. To my knowledge, this distinction has never been explicitly made before.
It must be noted that similar constructions (albeit, understandably, without the person marking) are found in neighbouring East Caucasian languages:
Ingush
(6) Earzii eanna c’i tyllaai
(name) say.cv.ant name bestow.nw
hana ealcha ciga=chy earzii daaxandea.
why say.cv.temp there=in eagle live.cv.reas
‘It’s was named “Earzii” (‘eagle’) because an eagle lived there.’ (Nichols 2011: 530)
Agul (Burshag-Ooshan dialect)
(7) čin fikir q’.u-ndaj s:ul q:ˁät:e-j wu-ji-r,
we.excl thought make.pf-aor:pst:neg fox lame-sub cop-pc-sub
fas p.i-t’en, iča-s gič’ x.i-naj.
why say.pf-cond we.excl-dat fear become.pf-aor:pst
‘We hadn’t thought that the fox was lame, because we were afraid.’ (Maisak 2014: 236)
However, in some Daghestanian languages we find ‘say’-based causal markers of the more common type, without the interrogative pronoun and any identifiable conditional marker, cf. (4) above and (8) below:
Tabasaran
(8) čwe ula-z ʁaf-un-dar k’uri,
brother.abs home-dat come.pf-aor-neg say.ipf.cv
či jarχi jis̊di nivk’u-z ʁuš-un-dar.
sister long all_night dream-dat go.pf-aor-neg
‘Because the brother didn’t come home, the sister could not fall asleep all night.’ (Xanmagomedov 1970: 205; glosses per Arkadiev & Maisak 2018: 138)
In sum, causal markers derived from speech act verbs are typical for North Caucasian languages, which is not surprising, given their propensity to grammaticalize such verbs (cf. (Ershova 2012) specifically for Kabardian, (Arkadiev & Maisak 2018: 137–140) for North Caucasian in general, (with special reference to Lezgic). Yet we clearly see that two distinct types of ‘say’-based constructions are present; their precise distribution across the languages of the area is worth investigating.
Outside of the Caucasus, to our knowledge, the “Circassian”-like strategy has only been attested in Buryat as a recent innovation:
Buryat
(9) Ganzar jexel jaara-na juundeb gexe-de
Ganzar much hurry-prs[3sg] why say-cond
radiator-aj-nj uhan xüre-še-že bolo-xo baj-gaa.
radiator-gen-poss.3 water freeze-intns-cv aux-pc.fut aux-pc.prs[3]
‘Ganzar is in a great hurry, because water in the radiator may freeze’. (Darzhaeva 2017: 88)
Thus, at the very least, this type of construction is not a uniquely North Caucasian, let alone Circassian feature. Notably, such constructions are specialized for encoding of reason, presenting another empirical argument against the claim for universality of grammaticalization clines in (Ebert 1991) and others.
4. Conclusion
While causal markers derived from speech act verbs are attested in languages from diverse parts of the world, the data discussed above show that North Caucasus should be added to the list of areas for which such connectors are especially typical (Southeast Asia, Volga-Kama area, probably some regions of Africa). The clauses introduced by the complex conjunctions səd-a ṗ-ʔwe-me in Adyghe, sət ŝha-č̣ʼe žə-ṗ-ʔwe-me in Kabardian form a distinct subtype, also attested in East Caucasian and Buryat.
Such constructions are of considerable interest in many respects. They are dedicated to expression of reason, “bypassing” more basic functions like complementation and purpose that have been claimed to be a necessary prerequisite for development of causal meaning. In this they testify to a special affinity between speech act verbs and causal clauses, the latter typically being used in spoken discourse to support a potentially problematic statement (Ford & Mori 1994; Diessel & Hetterle 2015), thus calling a specific attention to the act of communication itself. While any proposals that tie grammaticalization scenarios to specific socio-cultural factors by necessity stumble into a highly speculative territory, I would not refrain from noting that a particular Circassian (and, more generally, Caucasian) etiquette, with its emphasis on positive politeness, attention towards one’s interlocutors, and frequent explicit references to the on-going act of communication might have provided a fertile ground for precisely this kind of morphosyntactic development.
The constructions in question also represent a relatively early stage of grammaticalization, as testified by their transparent and non-reduced structural form. They may shed further light on the general mechanisms of development of speech act verbs into causal markers, but it must be kept in mind that they constitute a special subtype that has most probably emerged via its own distinctive scenario. It appears tempting to investigate whether difference in origins is reflected in the range of uses. According to my preliminary observations, at least in Adyghe səd-a ṗ-ʔwe-me can easily be used in contexts of objective causal relations, but the full picture can only be discovered through further study, for which the Adyghe Corpus (http://adyghe.web-corpora.net) offers a suitable tool.
Markedly different from more typical Circassian non-finite subordinate clauses, the structures of the kind discussed in the present paper have often been claimed a case of syntactic borrowing from Russian. (Matasović 2010: 105) goes as far as to suspect sət ŝha-č̣ʼe žə-ṗ-ʔwe-me and other such complex conjunctions in Kabardian to be calques. Given their distribution in North Caucasian languages, their morphosyntactic composition, drastically different from that of their Russian counterparts, and the typologically common pattern of grammaticalization, I seriously doubt this view. Following the three-stage model of language change [Croft 2006], it appears likely that contact with Russian has facilitated propagation of such constructions throughout the speech community and probably also their diffusion within the language system. They are likely to further gain in frequency in the future, with the considerable number of heritage speakers, for whom complex converbial and participial structures are notoriously difficult to master. But their initial emergence, the stage of innovation itself, must be rooted in the inner potential of the Circassian language system.
Abbreviations
a – agent; abs – absolutive; adv – adverbial; aff – affirmative; ant – anterior; aor – aorist; aux – auxiliary; cond – conditional; coord – coordination; cop – copula; cv – converb; dat – dative; dir – directive; dyn – dynamic; excl – exclusive; fut – future; gen – genitive; iness – inessive; inf – infinitive; ins – instrumental; intns – intensive; io – indirect object; ipf – imperfective; loc – locative preverb; neg – negation; nom – nominative; nw – nonwitnessed; obl – oblique case; pass – passive; pc – participle; pf – perfective; pl – plural; poss – possessive; pot – potential; prog – progressive; prs – present; pst – past; q – question; reas – reason; rfl – reflexive; sg – singular; sub – substantivizer; temp –temporal.
REFERENCES
Arkadiev, P. M. & Gerasimov, D. V. (2019). Ot relativizacii k aspektu: konstrukcii s prefiksom zere– v adygskix jazykax [From relativization to aspect: constructions marked with the prefix zere– in Circassian]. In D. V. Gerasimov, S. Yu. Dmitrenko & N. M. Zaika (Eds.). Sbornik statej k 85-letiju V. S. Khrakovskogo [A collection of papers presented to Dr. Viktor Khrakovsky on his 85th birthday] (pp. 11–42). Moscow: YaSK.
Arkadiev, P. M. & Maisak, T. A. (2018). Grammaticalization in the North Caucasian languages. In H. Narrog & B. Heine (Eds.). Grammaticalization from a typological perspective (pp. 116–145). Oxford: Oxford University Press.
Chisarik, E. & van der Wurff, W. (2003). From ‘say’ to ‘because’: Grammaticalisation and reanalysis. Paper presented at the Conference on Comparative Diachronic Syntax, UCLC, Leiden, 29-30 August 2003.
Croft, William. 2006. Evolutionary models and functional-typological theories of language change. In A. van Kemenade & B. Los (Eds.), The handbook of the history of English (pp. 68–91). Oxford: Blackwell.
Darzhaeva, N. B. (2017). Buryatskie konstrukcii prichiny: ot sinkretichnyx form k specializirovannym [Buryat causative constructions: From syncretic forms to special ones]. Filologicheskie nauki. Voprosy teorii i praktiki [Philological Sciences: Issues of Theory and Practice], 11(77)/3, 86–88.
Diessel, H. & Hetterle, K. (2011). Causal clauses: A cross-linguistic investigation of their structure and use. In P. Siemund (Hrsg.). Linguistic universals and language variation (pp. 23–54). Berlin: Mouton de Gruyter.
Dzhaurdzhij, H. Z. & Syqun, H. H. (1991). Urys-Adyge shkol psalale [Russian-Adyghe school dictionary]. Nalchik: Nart.
Ebert, K. H. (1991). Vom Verbum dicendi zur Konjunktion – Ein Kapitel universaler Grammatikalisierung. In W. Bisang & P. Rinderknecht (Hrsg.), Von Europa bis Ozeanien – von der Antonymie zum Relativsatz, Gedenkschrift für Meinrad Scheller (pp. 77–95). Zurich: Universität Zürich.
Ershova, K. (2012). Reported speech and reportative grammaticalization in Besleney Kabardian. In B. Surányi (Ed). Proceedings of the Second Central European conference in linguistics for postgraduate students (pp. 71–87). Budapest: Pázmány Péter Catholic University.
Ford, C. E. & Mori, J. (1994). Causal markers in Japanese and English conversations: A cross-linguistic study of interactional grammar. Pragmatics, 4, 31–61. doi: 10.1075/prag.4.1.03for
Gerasimov, D. V. & Lander, Yu. A. (2008). Reljativizacija pod maskoj nominalizacii i faktivnyj argument v adygejskom jazyke [Relativization under the guise of nominalization and the factive argument in Adyghe]. In V. A. Plungian, S. G. Tatevosov (Eds.). Issledovanija po glagol’noj derivacii [Studies of Verbal Derivation] (pp. 290–313). Moscow: YaSK.
Göksel, A. & Kerslake, C. (2005). Turkish: A comprehensive grammar. London; NY: Routledge.
Güldemann, T. (2008). Quatative Indexes in African Languages: A synchronic and diachronic survey. Berlin: Mouton de Gruyter.
Gündoğdu, H. Y. (2017). The structure of diye clauses in Turkish. MA Thesis, Boğaziçi University.
Heine, B. & Kuteva, T. (2002). World Lexicon of Grammaticalization. Cambridge: Cambridge University Press.
Hetterle, K. (2015). Adverbial Clauses in Cross-Linguistic Perspective. Berlin: Mouton de Gruyter.
Höhlig, M. (2007). Non-finite verbal constructions in Adyghe text: verbal suffix ‑ŝ. Paper presented at the Conference on the Languages of the Caucasus, MPI EVA, Leipzig, 7–9 December.
Kanetani, M. (2019). Causation and reasoning constructions. Amsterdam; Philadelphia: John Benjamins.
Kumakhov, M. A. (1967). Adygejskij jazyk [Adyghe]. In Jazyki narodov SSSR [Languages of the USSR]. Vol. IV (pp. 145–164). Moscow: Nauka.
Kumakhov, M. A. (1989). Sravnitel’no-istoricheskaya grammatika adygskix (cherkesskix) jazykov [A comparative grammar of Circassian languages]. Moscow: Nauka.
Lord, C. D. (1993). Historical change in serial verb constructions. Amsterdam; Philadelphia: John Benjamins.
Loughnane, R. (2009). A grammar of Oksapmin. Ph.D. Thesis, University of Melbourne.
Maisak, T. A. (2014). Agul’skie teksty 1900–1960-x godov [Agul texts of 1900–1960-ies]. Moscow: Academia.
Martowicz, A. (2011). The Origin and Functioning of Circumstantial Clause Linkers: A Cross-linguistic Study. Ph.D. Thesis, University of Edinburgh.
Matasović, R. (2010). A short grammar of East Circassian (Kabardian). Ms., Zagreb University.
Nichols, J. (2011). Ingush Grammar. Berkeley, LA & London: University of California Press.
Rogava, G. V. & Kerasheva, Z. I. (1966). Grammatika adygejskogo jazyka [A grammar of Adyghe. Krasnodar; Majkop.
Santana Covarrubias, A. C. (2019). Is porque more like because or like omdat? An exploration of causality and subjectivity in Spanish backward causal connectives. Ph.D. Thesis, Universiteit Utrecht.
Saxena, A. (1988). The case of the verb ‘say’ in Tibeto-Burman. Berkeley Linguistic Society, 14, 375–388.
Saxena, A. (1995). Unidirectional grammaticalization: diachronic and cross-linguistic evidence. Sprachtypologie und Universalienforschung, 48 (4), 350–372.
Serdobolskaya, N. V. & Toldova, S. Yu. (2011). Grammaticalization of the verb of speech in Finno-Ugric languages. In S. Csúcs et al. (Eds.). Congressus XI. Internationalis Fenno-Ugristarum (pp. 285–293). Budapest: Reguly Társaság.
Traugott, E. C. (1995). Subjectification in grammaticalization. In D. Stein & S. Wight (Eds.). Subjectivity and subjectivisation: Linguistic perspective (pp. 31–54). Cambridge: Cambridge University Press.
Zaika, N. M. (2019). Polipredikativnye prichinnye konstrukcii v jazykax mira: prostranstvo tipologicheskix vozmozhnostej [Causal clauses in the languages of the world: The space of typological possibilities]. Voprosy jazykoznanija [Issues in Linguistics], 2019, 4, 7–32.
Xanmagomedov, B. G-K. (1970). Ocherki po sintaksisu tabasaranskogo yazyka [Essays on the syntax of Tabasaran]. Makhachkala: Daguchpedgiz.
Zekokh, U. S. (2002). Adygejskaja grammatika [Adyghe grammar]. Majkop: Adygeja.
