CC BY
24
DERIVATIONAL SEMANTICS
v __ 1
OF CLASSICAL ARMENIAN C-STEMS1
Статья посвящена исследованию древнеармянского глагольного класса, характеризующегося имперфективной основой на -с- и перфективной основой на -i-. Этимология обоих основообразующих суффиксов, а также история формирования данного парадигматического типа остаются спорными. В статье предпринята попытка привлечь к диахроническому анализу валентностные, акциональные и аспектуальных характеристики глаголов, засвидетельствованных для данного класса в ранних древнеармянских текстах, с целью установления возможных путей формирования особой глагольной парадигмы.
Ключевые слова: древнеармянский, диатеза, акциональные классы, аспект, имперфективные основы.
1. Morphology of the с- class
A number of Armenian verbal classes have the element -c -within the suffixes that marks the imperfective part of the paradigm (-c -, -ac -, -nc -, -anc -). None of these suffixes was productive by the time of the first written record of Armenian (5th c.). Three of the synchronically distinct morphological classes consist of few verbs: 1) pres. can-ac -em (aor. can-e-ay) tr. ‘to recognize’; 2) pres. erk-nc -im (aor. erk-e-ay) intr. ‘to be(come) afraid’, pres. mart-nc-im (aor. mart-e-ay) intr. ‘to fight’, pres. kor-nc-im (aor. kor-e-ay) intr. ‘to disappear’; 3) pres. mel-anc-em (aor. mel-ay) intr. ‘to sin’. The fourth class, with the suffix -c -, is much better attested since the earliest Classical Armenian texts : hangcim intr. ‘to take rest’ (y-angcim ‘to relax’), matcim intr. ‘to approach’, ostcim (dnd-ostc im)
1 The paper is based on the presentation delivered at the 9th International Conference on Armenian Linguistics (St. Petersburg, 2012). I am grateful to Prof. James Clackson and Rémi Viredaz for their valuable comments. I acknowledge the financial support from the Research Council of the President of the Russian Federation (grant nr. МК-3848.2013.6).
2 I rely on HAB, LALT, NBHL, and RADCA except for cases where the present stems mentioned in these editions do not find support in the early texts, e.g. karc im ‘to lay hold of; to be fixed on’ is first attested in the 7th c. text (Anania Sirakac'i) while caus. past part. karuc-eal (Pawstos Buzand) and caus. kar-ucanel (Bible) are better explained as by-forms of the imperfective nu-stem in view of the root final -r-.
intr. ‘to leap out’, pakcim intr. ‘to be(come) affected’, paxcim intr. ‘to flee’, f ak'cim intr. ‘to hide’; t'rcim intr. ‘to fly (away)’, xrtcim intr. ‘to be(come) anxious’, z-atcim intr. ‘to depart’ from *z-hatcim (beside z-atanem ‘to separate (smth.)’ and hatanem ‘to cut’), z-art'c im intr. ‘to awaken’ (beside z-art'num ‘to awaken’ and hart'num ‘to rebound’). The -c-, -ac-, and -nc- imperfective stems form the respective paradigmatic types with the perfective i-stem; mel-anc -em has the root perfective stem mei-ay. The -c- and -nc- suffixes are media tantum, while -ac- and -anc- (canacem, melancem) are activa tantum. All the four classes use the causative past participle in -uceal in intransitive meaning even when the corresponding causative is attested with the transitive meaning (hangc im intr. ‘to take rest’, hanguceal intr. ‘the one who had a rest’, hangucanem tr. ‘to give rest’); the aorist subjunctive has a mixed conjugation: 1sg. is mediopassive, all other forms are active.
In this paper, I will focus on one paradigmatic type characterized by the imperfective suffix -c -. The imperfective stem of this class is formed by adding the suffix to the root (which never ends in vowels in the c -class) and is followed by the mediopassive inflection in the present indicative, present subjunctive, and prohibitive; the imperfect has one inflection for both voices in all the paradigmatic types.
Formally, the suffix -c- can be analyzed as such or as a continuant of -dc-3 from PArm. *-ic- or *-uc- with the regular deletion of a schwa in a medial bi-consonant cluster: PArm. *-V.Ci.CV-/*-V.Cu.CV- > *-V.C^.CV- > Arm. -VC.CV- (see Vaux 1998: 21-34 for the Armenian syllabification and schwa
epenthesis/delition rules). Given that all roots of the c -class verbs result in -VC- codas, and the mediopassive inflections of the imperfective part of the paradigm (present indicative, imperfect, present subjunctive, prohibitive) always begin with a vowel, the schwa deletion rule would always apply to c -stems.
Within the perfective stem, the suffix -e- or -i- is added to the root and is followed by the mediopassive inflection. If the inflection begins with an -a- (aorist indicative, 1sg. aorist subjunctive in -ayk’, imperative in -aruk) the suffix is -e- according to the sandhi rule -i-a- > -e-a- that occurs at the morpheme boundary following a polysyllabic stem, cf. tari ‘year’, gen.pl. tare-ac but mi-ayn ‘only’). Accordingly, the rule did not apply to the remaining forms of the aorist subjunctive and imperative where the original -i- was retained or merged with the following -y- (3 sg. hangice < *hangi-yc -e, 2 sg.
3
According to the Classical Armenian orthographic rules, schwa was written only in certain initial clusters or at a line break.
hangir < *hangi-yr). Thus, the aorist subjunctive and imperative in -ir indicate that *-i- was the original Proto-Armenian perfective suffix of this paradigmatic type (henceforth i-perfective).
The account of the described phenomena led scholars (cf. Meillet 1936: 109 f.) to the conclusion that the imperfective stem was derived from the perfective one: PArm. perfect. *-i- ^ imperfect. *-i-c - > *-dc - (pre-tonic) > -c- (in the -VCdcV- environment). Note, however, that the imperfective stem shows the -i-l-d-alternation of the accented and non-accented vowel (PIE *kerd- > Arm. sirt, gen.sg. srti [ssr.tf]), while the perfective stem show the -e-l-i- alternation (PIE *woyd- > get ‘skilful’, gitem ‘to know’). In order to combine these two formants, one has to admit that the perfective stem originally had the -i-l-d- alternation and the accented penultimate syllable variant -i- had spread as a specific perfective marker to contexts where it would be lost. Otherwise, it is hard to explain, why the non-accented *-i- would be lost in the imperfective stem. The penultimate position of the perfective -i- is better explained if one starts with the secondary middle endings which should have characterized the aorist indicative of this verbal class already very early.
2. Etymologies of -c- and -i-
The i-perfective can go back to a choice of reconstructions: *-!-, *-e- (including long vowels of the laryngeal origin) multiplied by the combinations of these vowels with *-s- and *-y- in front of another vowel.
As for the c -imperfective, the temptation to derive it directly from PIE *-skelo- is moderated by the sound change PIE *-sk- > Arm. -c- (e.g. ctem ‘to graze’ < PIE skeyd- ‘to cut’ LIV 547f.). Taking into account this contradiction, Meillet (1936: 109 f.) proposed to derive c -imperfective from PIE *-i-sk-ye-. According to Meillet, the first two elements of this complex formant co-occur in Lat. -isco(r) (reminiscor ‘to remember’) and Gk. -ioKro (a^ioKopai ‘to be taken’), while PIE *-sk-ye- is attested in Gk. cyp^ooro (Hom., praes. tant.) ‘to stay awake’ and SsSiiTopai (Att.) ‘to frighten’. According to Meillet’s analysis, the *-i- of the present suffix should be kept apart from the *-i- of the optative *-i-ske/o- that presumable gave Arm. aor. subj. -ic- (1936: 122). However, the formal equation of Gk. pres. apapioKro ‘to fit together’ and Arm. aor. subj. araric ‘I willlwould make’ (pres. arnem ‘to make’) was taken by some scholars as an argument in favor of one “inchoative” stem behind both formations (cf. Godel 1975: 116f.; de Lamberterie 2005-2007: 34).
The reconstruction *-i-ske/o- looks attractive in view of certain semantic parallelism between such pairs as Arm. zartcim : Lat. expergiscdr ‘to awaken’, Arm. yangcim : Lat. paciscor ‘to relax’, Arm. matcim ‘to approach’ : Lat. apiscor ‘to attain’, Arm. ostcim ‘to leap out’ : Lat. proficiscor ‘to set out’. However, the history of the Greek and Latin *-i-ske/o- remains unclear (Duhoux 1992: 337; Sihler 1995: 520; Vine 1993; van de Laar 2000: 363-366)4.
However, the Proto-Indo-European antiquity of *-i-ske/o- has been convincingly disclaimed by Vine (1993). According to him, the Latin verbs in -Tsco(r) may be regarded as secondary derivatives from *ye/o-presents (cup-i-o ^ con-cup-i-sc-o ‘to long for’). Within this scenario, the direct comparison of Lat. -isco- to Arm. -c - leaves Arm. i-perfective without a coherent explanation. Note that unlike Armenian, *-i- does not appear in the aorist of the Greek verbs in -iokw. Instead, these verbs often have the thematic aorists in the Homeric Greek, cf. pres. a^ioKopai ‘to be captured’ : aor. ^Awv, pres. apap-iaKw : aor. apap-cTv, pres. svp-ioKro ‘to discover’ : aor.
e ~ 5 SUp-STV .
According to the alternative hypothesis, more acknowledged nowadays, the Armenian c-class resides on PIE imperfect. *-ehj-ske/o-6 : perfect. *-ehi-s- (Klingenschmitt 1982: 72-79). Indeed, PIE
4 The secondary replacement of -ok© ^ by -iok© may be conjectured for Ov'flGK© ‘to be dying’ (secondary to OvnaK©; both present stems are based on the perfect stem xe-Gvn-Ka, see DELG 422, van de Laar 2000: 165), 0p©GK© ‘to leap’ (secondary to OprooK©, along with Oopvupai, aor. eOopov, see DELG 444, van de Laar 2000: 167), ^i^v^ok® ‘to remind’ (secondary to pipvnoK©, along with non-reduplicated pvnoopai, aor. eVvnoa, see DELG 702; van de Laar 2000: 218).
5 Vine (1993: 53ff.) pointed out the correlation between the root-final liquids and the following -i- of the present suffix. He suggested to derive -p-(ok© from the *TRhr type roots which vocalized under the secondary stress as *TRh1-ske/o- > *ToRe-ske/o- or *ToRi-ske/o-, whence *TaRi-ske/o- by analogy to thematic aorist *TaR-e/o-. Besides the problematic accent and sonorant vocalization, in my opinion, such solution does not find enough support in the reconstructed proto-language — note that none of over fifty *-ske/o- stems reconstructed in LIV has the *TRh1- type root. Although a phonetically motivated distribution of present suffixes cannot be excluded for Greek (cf. the preference for the root final -y- and -p- in the verbal class characterized by the vu-present and sigmatic aorist stems noticed in van de Laar 2000: 349), the source of -p-i-oK© remains unclear.
6 According to Klingenschmitt 1982: 83f., the double reflex of PIE *-ske/o-depended on the left environment: c developed in the word internal position after *i (< *iH), *e (< *eH), *d (< *H), or *-n-; c - elsewhere including the initial position.
*-ehj- is a likely source for the Armenian paradigm in view of the lexical match between Arm. pres. t'ak'cim ‘I hide’, aor. t'ak'eay and Lat. pres. conticesco ‘I stop talking’, pres. taceo ‘I am silent’ .
The derivational semantics of PIE *-ehj- (*-e-) is itself an important problem. Two competing hypotheses should be mentioned. According to the hypothesis accepted in LIV, PIE *-ehj-l*-hj- was a fientive stem that served as a derivational basis for essives in *-h!-ye/o-. Alternatively, Jasanoff (2002) proposed to consider *-ehj-ye/o-and *-ehj-skelo- denominative stems derived from the instrumental case forms following the instrumental predicative constructions (cf. stative *hjes- + *X-ehj ‘be (with) X’, cf. Ved. guha as- ‘be hidden’, and inchoative *bhuhx- *X-ehj ‘become (with) X’, cf. Ved. guha bhu-‘become hidden’). The former hypothesis does not account for the imperfective uses of *-ehj- continuants in Latin, Slavic, and Armenian (cf. hangeay below), while the latter does not provide a convincing explanation of the Greek n-aorist. This issue has been recently reconsidered by Yakubovich (2013a), according to whom the easiest way to account for both present and preterit uses of *-e- is to start from deadjectival statives which had both fientive and essive values. Without entering the controversies of the competing analyses for the Proto-Indo-European *-ehj- and its traces in the daughter languages, I would like to present the Armenian evidence which is relevant in the following points: 1) i-stem is found exclusively in the preterit; 2) unlike the productive denominative inchoativelstative ana-class8, i-perfective verbs predominantly have verbal roots; 3) i-
In my opinion, the double reflex of PIE *-ske/o- can be explained taking into account the syllabification and schwa epenthesis rules that operated in Proto-Armenian. Whenever *sk occurred at the beginning of a syllable it resulted in -c-; when the syllabic boundary divided it in the word internal position, *k was subject to independent developments, cf. *prk-ske-[*psrkj.skje] > Arm. harc - ‘to ask’ but *gnh3-ske- [*gj3n.Has.kje] > canacem ‘to recognize’ and *ptke-ske- [*ptsk.e:s.kJe] > t'akcim ‘to hide’. The details on the relative chronology of *-s.k- > *-s.k- and *-k- + front vowel > c are a topic for another paper. See Viredaz 2012 for a different explanation of the imperfective suffixes -c - (canac em) and -ic -l-c -(take im).
The morphological match is consistent with either of the two available etymologies: 1) PIE *pt(h2)k-e-ske-, cf. Gk. ni^aa© ‘to shrink from fear’ (Pedersen 1906: 356; Solta 1960: 445f.; Klingenschmitt 1982: 70, 78-79; Clackson 1994: 169; LIV 495-496); 2) PIE *tak-e-ske-, cf. Goth. pahan ‘to keep secret’ (Sihler 1995: 497; de Vaan 2008: 604f.).
8 According to Yakubovich 2013a, the Vedic intransitive denominatives in -aya- continue *-eh1-ye/o-. Although I consider his argumentation very convincing, in my opinion, an alternative source for the intransitive zerograde aya-formations is possible. The mediopassive forms derived from the
perfectives denote both agentive (controllable) and non-agentive (uncontrollable) momentary events, processes, and states; 4) с -class verbs point to the paradigmatic or derivational ties between perfective *-é(s)- and imperfective *-ske/o- suffixes in Proto-Armenian. Therefore the Armenian perfective /-stem is best explained as an outcome of the primary verbal stative/resultative marker rather than denominative inchoative/stative.
Hittite provides an argument against the reconstruction of *-eh¡-ske/o- for common Proto-Indo-European. According to Hoffner & Melchert 2002: 378, some Hittite primary stative verbs,
denominative inchoatives in -ess-, and, importantly, denominative stative verbs in -e- are not attested with otherwise productive imperfective suffix -ske/a-. This does not mean, however, that such formation could not develop within the late Indo-European period while *-ehj- (or *-ё-) was gradually turning into a marker of the mediopassive perfect and thus switching from the domain of states to that of change-of-state events or telic inchoative processes in terms of actional classes.
The Old Armenian verbal system is traditionally described as having the binary aspectual opposition of imperfective (“present”) and perfective (“aorist”) stems; apart from the opposition between the imperfect and aorist in the preterit, the two stems served to distinguish between imperfective and perfective subjunctives (futures), prohibitive and imperative, cf. Meillet 1910-19119. Any attempt to further specify the functional contrast between various imperfective suffixes within the synchrony of Classical Armenian risks being superfluous. This notorious methodological issue is relevant for other ancient Indo-European languages and for the reconstructed proto-language (e.g. Melchert 1998 on Hitt. -ske/a-). However, a careful investigation of lexical and syntactic properties of the с -class verbs may prove to be relevant for the discussion of the unwritten history of that particular morphological class and provide evidence for further Indo-European comparison. Such investigation should take into account lexical meaning, argument
*-nh2-yé/ó- stems are probably at the origin of the Armenian inchoative/stative -ana- and Toch. -ññ- verbs (cf. Arm. tencana-, Toch. B tenkwáññá/e-; Kocharov 2011; see Jasanoff 2003: 125f. with an explanation of the double reflex in OInd. -anyá- and -aya-). Old Indic could inherit the intransitive middle type together with the transitive active grbhayáti (*grbh-nh2-yé-ti) type.
Cf. Weitenberg 1986 on the problem of alternating imperfective and perfective stems within past participles.
structure, actionality, and aspect uses of the imperfective c -stems, especially in verbs that attest variation of imperfective stems within
one lexeme10.
Following Meillet 1936: 110, the c-class verbs are often called “inchoative” or “stative inchoative” (cf. Klingenschmitt 1982: 72-79 who defines the term as subject’s entry into the state described, in the diachronic perspective, by the underlying i-stem of the aorist from *-eh1-s-). The principal goal of the present paper is to provide a systematic account of the internal Armenian evidence of the c -class verbs and to contribute to the understanding of how the lexical and derivational semantics fit together if one starts from PIE stativelresultative *-eh1- as a source of the i-perfective stem.
3. Synchronic evidence of the c -class verbs
For the sake of further analysis, the c -class verbs are arranged below into two groups according to whether or not they have valence on Agent.
3.1. Agentive c -class verbs
Hangc im denotes a state controlled by the Agent: ‘to take rest; to lodge’, cf. Job 39, 9: Ete kamici kez mieljerun carayel, kam hangcel ar msur k'o? (Pou^nosxai §8 ooi povoK8pro<; Sou^sfioai ^ Koi^n&^vai sni ^axvn^ cou) ‘Will the wild ox consent to serve you, or will he spend the night at your manger?’11.
Both perfective and imperfective stems (aor. hange-, pres. hangci-) are stative. The only noticeable difference between the imperfect and aorist indicative is that the latter denotes a state limited in time (limitative), cf. 1) Jn. 1, 32: Vkayeac Yovhannes ew ase t'e: Tesanei zhogin, zi ijaner ibrew zalawni yerknic, ew hangc ei [3 sg. impf.] i veray nora (Kai spapxtipncsv ’Iroavvn^ ^eyrov oxi xsOcapai
> n* n ^ > Dr ?!
xo nvsv^a Kaxapaivov rcspioxspav cq ovpavov, Kai c^sivcv sn
avxov) ‘John testified saying, I have seen the Spirit descending as a dove out of heaven, and He remained upon Him’; 2) Gen. 2, 2: Ew katareac Astuac yawurn vecerordi zgorcs iwr zor arar, ew hangeaw [3 sg. aor. ind.] yawurn ewt'erordi yamenayn gorcoc iwroc zor arar
10 See Бондарко 1987, Горбова 2010, Плунгян 2011, Croft 2012, Tatevo-sov 2002 with relevant bibliography on the actional classification of predicates.
1 Here and below the English Bible translation is cited according to the New American Standard Bible (the 1995 edition). The translation of passages from the Deuterocanonical Books is cited after the New Revised Standard Version Bible (the 1989 edition).
(каі 0uv8X8^80sv о Oso<; cv тп п^сра тп єктп та cpya awoft, a
Dr Л Г ~ t Г ^ {Л С» Л 9 Л Г <"w
cnoi^acv, каі Kaxcnauacv тп п^єРа тп єроо^п ano navTWV twv cpywv awoft, wv cnolnacv) ‘By the seventh day God completed His work which He had done, and He rested on the seventh day from all His work which He had done.’
Klingenschmitt (1982: 78f.) derives hangcim ‘to take rest’ from PIE *sm-kwihi-ske- without *-eh1- of the type *gnh3-ske/o- > canacem (see also LIV 393f.). However, as Olsen (l999: 480, n. 583) justly noted, *sm-kwih1-eh1-ske- is equally possible (Lat. con-quiesco). The sonorisation of *-kw- was regular after *-r- and *-n-, cf. PIE *penkwe > hing ‘five’.
Prepositional compound yangc im ‘to relax; become familiar’ is once attested in the Bible, cf. Ezek. 1, 24: ew i kaln iwreanc
Vf _ • r 1 Г • f / Л D r* t r D Л Г e
yangc ein tewk iwreanc (каі cv тф cornvai awa катспаиоу ai лтсриусд awwv) ‘whenever they stood still, they dropped their wings’. The imperfect denotes the habitual reading of the completive stage of telic process. The imperfective stem is further attested in Movses Xorenac'i 2003: 2084 2: ...ibr te aynpes hanapaz ar nma linelov naxararUn yangcicin 13 (3 pl. pres. subj.) xawsiwk', turewariwk\ xraxcanuteamb ew zbosanawk' orsoc... ‘...with the malicious intention that thus by being continually with him, the princes might become attached to him by social intercourse, festivities, and the pleasures of the chase...’ (trans. Thomson 1978: 323). A similar use is found in Lazar P'arpec'i 2003: 2217: E zi ew sovorut'iwn arnun yankc el [variant - ms. Mat. 2639: yakc el] yawrens mer... ‘And when this becomes the custom, they will conceive a liking for our faith...’ (trans. Bedrosian 1985). No i-per-fective is attested.
Matcim is well attested in the Bible in the meanings ‘to approach’ and ‘to be offered (of atonement, sacrifice)’. The latter meaning is secondary; it is a passive pair to the morphological causative derived from matcim in its primary meaning: mid. matcim ‘to approach’ ^ caus. mat-ucanem ‘to bring forth (to make approach), to offer’ ^ pass. mat-cim ‘to be offered’, cf. Ex. 32, 6: ...ew matoyc zohs p'rkut'ean... (каі npoonvsyKcv Ouolav оютпріои) ‘and offered burnt offerings’; Heb. 9, 9: yorum pataragkn ew zohk matcein, ork oc karein dst mits katarel zpastawneaysn (ка0’ nv 8юра ts каі 0иоіаі rcpo^spovTai ^n §uva^svai ката ouvsiSnciv
12 The present stem yankci-, cited in NBHL 2: 327 and RADCA: 141, is actually not attested.
13 The non-prefixed variant hangc icin (ms. Mat. 4351) should be recent.
xs^siwoai toy ^axps'uovTa) ‘Accordingly both gifts and sacrifices are offered which cannot make the worshiper perfect in conscience’.
In its primary meaning, matcim have a facultative semantic argument of Purpose regularly expressed by the prepositional constructions ily- + acc. or ar + acc. This argument determines the grammaticalized collocation mate el porjel lit. ‘to be going to try’ (Eznik Kolbac'i 2003: 452, 481). This argument also qualifies the volitional Agent that controls an intended action.
The imperfective stem mate i- describes the progressive stage of a telic process (e.g. 2 Mac. 11, 8) or the habitual aspect (Heb. 13, 11; Elise 2003: 636; Eznik Kolbac'i 2003: 448); the perfective stem denotes the completive stage of a telic process (e.g. Gen. 27, 22).
Arm. adj.ladv. mawt ‘near’ may go back to *mad-u- (or *mHd-u-), a *u-stem adjective to PIE *mad- (or *mHd-)u, cf. Goth. gamotjan ‘to meet’ (^ayKAH 1982: 182; Klingenschmitt 1982: 70f.; Lehmann 1986: 145; Olsen 1999: 49). The derivation from qualitative predicate *mad-u- ‘(is) close’ to stative *mad-eh1- ‘to be close’ may be further explained within the Caland-system (see Rau 2009 with further bibliography)15.
Alternatively, one may assume an archaic deadjectival *sk-formation *mad-sk- (cf. later Arm. mawt ‘near’ ^ mawt-em ‘to approach’). Note also PIE *gwm-ske- ‘to go’ (LIV 209f.), another agentive verb of motion, that can be safely reconstructed with the *ske/o-stem. It is possible that the i-perfective was introduced relatively late by analogy to other e -class verbs, cf. secondary aor. caneay ‘I got to know’ along with the inherited pres. eanae em ‘to recognize’ < PIE *gnh3-ske/o- ‘id.’ (Gk. yi-yvwoKro, Lat. (g)ndscd, OP xsnasa-, Alb. njoh).
Ostcim ‘to leap out’ is attested once in Movses Xorenaci 2003: 2057: ...ew i nocane i durs ostcein [v.l. ostneein] i meroc naxararacn k'aj zrahawork orpes paylatakanc earagayt'k' ‘out of the troops like rays of lightning rushed the bravest armed men among our princes’ (trans. Thomson 1978: 297). The wider context suggests
14 It is very probable, in my opinion, that *a in the root was an “allophone” of a schwa regularly developing in the (initial ?) sequence of a labial + plosive, cf. PIE *mag- ‘to be able’ (LIV 422), pIe *pdkw-ye/o- ‘to cook’ (LIV 468). If so, a-vocalism corresponds to the morphological zero-grade.
5 Note that the synchronic relation between Arm. i-perfective and -uc-eal causative past participle strongly reminds special paradigmatic relations between the PIE morphological perfect (functionally equivalent of stativelresultative *eh1-stems) and *u-stem adjectives (resultative
participles), see de Lamberterie 1990.
that the verb describes distributed punctual events that took place at the same reference time rather than a series of consequent events.
No aorist indicative forms are attested.
According to NBHL 2: 523, the stem ostani- is attested in the 5th c. Armenian translation of Gregory of Nazianzus’ homily (Girk' ew car hogesah, Polis 1722): Dipesci kacnoyn i же1вхёп i bac ostanel ‘and the iron head slips off the handle’, cf. Deut. 19, 5 where the verb ankanel ‘to fall down’ is used to translate Gk. склжтю ‘to fall off’ in the same context. The infinitive marks the punctual event. The meager evidence does not allow deciding what was the grammatical contrast between -c - and -an- imperfective morphs in this verb.
The verb has no established etymology, see HAB, 3: 569.
The prepositional compound dnd-ostcim ‘to leap out’ is attested in the 6th c. (dating apud LALT) Armenian translation of Philo’s Questions and Answers on Genesis, Book IV, § 243 (Philo 1826: 439): Zi yapales ew oc sarzis ew sndostcis? ‘Why dost thou delay and dost thou not move and leap up?’ (trans. Marcus, Philo 1953: 545). The punctive event co-referential with the moment of speaking receives the progressive reading.
Paxcim ‘to flee’ is well attested in the corpus of the Classical Armenian texts (RADCA: 141). In the Bible, the indicative forms based on the c-stem denote the habitual meaning (e.g. Tob. 13, 2; Jn. 10, 13; Elise 2003: 636). An exception is the distributive use in 1Mac. 10, 64: Ew elew ibrew tesanein zpars nora, ew Isiin zjayn barbaroy karozin, ew tesaniin zi zgeceal er ciranis, paxcein, t'ak'ciin, korncein amenayn caraxosk'n (каі cycvcxo w<; ciSov oi cviuyxavovxe^ x^v 8o^av айхоїї, каОю^ ¿K^pu^cv, каі
nepiPeP^n^Cvov arnov лорф'брау, каі сфиуоу navxc^) ‘When his accusers saw the honor that was paid him, in accord with the proclamation, and saw him clothed in purple, they all fled [, concealed themselves, and disappeared].’ Another particular application of the present form, found in 1Sam. 22, 17, is a calque from Greek: Matik ew spanek zkahanaysn Tearn, zi jern noca 3nd Dawt'i ё, zi gitacin ete paxci na, ew oc yaytnecin yunkn im (каі 0аvаxovx8 хой^ icpcT^ xoft кирши, оті ц xcip айт^ ^єха Даиі8, каі оті єууюоау оті фє'буєі ашо^, каі ойк dл;гкaAuyav то wxiov ^ou) ‘Turn around and put the priests of the Lord to death, because their hand also is with David and because they knew that he was fleeing and did not reveal it to me.’
The aorist indicative forms are not attested. Instead, analytic perfect paxeal em is found in the Bible, cf. 1Mac. 6, 10: paxeal em i
sirt im hogoc imoc (оицпсптшка тп кардіа and x^ ^epi^vn^) “I am downhearted with worry”.
In the Bible, the verb is also attested in the present subjunctive forms of the stem paxnu-16, cf. Amos 5, 19: Zor orinak t'e paxice mard yeresac ariwcu, ew patahesce nma arj, ew paxnucu i tun, ew
t f_ • — • -J -f * t— r tr I r
yecusce zjers iwr yormn, ew harkanice zna oj (ov xponov oxav ф^уп avOprono^ ск npoownou xoft ^covxo^ каі cancan arnw ц арко^, каі 8шгсп$ПаП Tdv oikov айхоїї каі апсрсіоптаї xaq X8Tpaq айхоїї спі
xdv xoT%ov каі 8акп avxdv о офід) ‘As when a man flees from a lion and a bear meets him, or goes home, leans his hand against the wall and a snake bites him’. Note the semantic difference between ablative paxice ‘to ran away (from a lion)’, consistent with the use of paxcim, and allative paxnucu ‘to run into (a house)’; Gk. (oxav) фйуп (ск) and (siq) are not co-referent in the original
passage. The nu-stem denotes the progressive stage of a telic process. Mt. 23, 33: Awjk\ cnundk izic, ziardpaxnucuk i datastane gehenin (офск; ycw^axa c%i§vwv, nwq ф'буптс and x'nq Kpiocwq x'nq yccvvnq) ‘You serpents, you brood of vipers, how will you escape the sentence of hell?’ The negative phrase underlying the rhetoric question favors the habitual reading.
The etymology is missing (HAB 4: 470).
Takcim ‘to hide’. The i-perfective expresses the punctive aspect of the autocausative change-of-state event.
The indicative forms derived from the imperfective stem have secondary distributive (1Mac. 10, 64; Lazar P'arpec'i 2003: 2308, 2373) and habitual (Job 29, 8; Mandakuni 2003: 1188) readings of the underlying completive events. Other uses, functionally equivalent to habitual, are assertion in unreal conditional (Ps. 54, 13)17, negation or interrogation with a determiner amenayn ban ‘anything’ as the subject (Jer. 32, 27 and 2Kings 18, 13), impersonal negated debita-tive (Agat'angelos 2003: 1420). Within the examined contexts, c-imperfective never describes the actual present progressive meaning.
The imperfective stem t'ak'nu- is first attested in the Armenian translation of John Chrysostom’s Commentary on Isaiah (6 c.) with a habitual meaning, cf. Chrysostom 1880: 280: ...vasn aynorik snd
16 According to NBHL 2: 923, the nu-stem is present in the post classical texts (Grigor Magistros, Grigor Narekaci, Nerses Shnorhali, and John Climacus). HAB 4: 470 lists paxci- and paxnu- with reference to John Chrysostom’s Homilies on the Epistles to the Philippians (Chrysostom 1862, I: 342-521); both are missing.
17 The use of imperfect is determined by the regular syntactic pattern of the unreal past conditional, cf. Lyonnet 1933: 118-143.
anuambkn nocin mert' t ak nu, mert' yaytni. ‘therefore sometimes he (Christ) hides, sometimes reveals himself by these names.’ 18.
Trcim ‘to fly, fly away’. The i-stem aorist indicative expresses the inchoative stage of atelic process ‘flew away’ (and never the completive stage of telic process), cf. 2Sam. 22, 11: El ew nstaw i veray КвгоуЬёж ew treaw, ew erewecaw i veray t'ewoc holmoc (каі cncKdOiocv ¿пі xepouPiv каі cncxdoOn каі юф&п ¿пі птєр'бушу avc^ou) ‘And He rode on a cherub and flew; and He appeared on the wings of the wind.’ In the Bible, the three indicative imperfective forms express habitual aspectual meaning (Job. 5:7, Ps. 54: 7, Ps. 90: 5). Note, that the imperfect form used in Ps. 54, 7 (zi t'rc ёi ‘I would fly away’) is a standard way to express the consequent of the unreal conditional. The Target argument is expressed in the two other contexts which increases the telicity of the underlying predicate, cf. Job. 5, 7: jawk arcueac i barjuns t'rc in ‘as sparks fly upward’. It is possible that the c -stem denoted the progressive stage of a telic process like in matcim ‘to approach’. Note, that the two forms derived from the competing imperfective ani-stem (Prov. 26, 2 and Prov. 27, 8), again with the habitual meaning, have no expressed target participant.
Arm. t'rcim ‘to fly’ and Gk. nxepov ‘wing’ are etymologically connected, but little can be said with confidence about the exact derivational process behind the Armenian verb (HAB 2: 184f.; Clackson 1994: 169; Klingenschmitt 1982: 70; see also Martirosyan 2010: 287-294 on t’ir- ‘to fly’ and t'it'ern ‘butterfly’). Formally, t'r-may go back to *р1ё^- or *р1ёт especially if one assumes a nasal suffix *р1ё^ш- or *р1ё^Н- (intr. *t'ranam > mp. t'ranim). However, the root long vowel is aberrant for both nasal formations, and the details of the development *-VrnH- > -r(a)n(a)- remain unclear. In my opinion, aor. *р1ё^- > t'ir—> pres. t'r-anim ‘to fly’ is a more promising solution in view of aor. *sed-s- > hec—> pres. hec-anim ‘to ride’; both are verbs of processual motion. It is hard to say whether the *-s- stem should be interpreted here as perfective (sigmatic “aorist”) or imperfective (iterative/habitual “present”, cf. PIE *peh2-s- ‘to pasture; to protect’: Hitt.pahs ‘to protect’, Lat. pasco ‘pasture’, OCS ПАСТИ ‘id.’)19.
18 Later evidence from Apophthegmata Patrum, Venice 1855 (unknown date) and Yovhannes Sarkawag, Venice 1853 (12 c.), cited in NBHL 1: 803, is not taken into account here. Ireneus’ Against Heresies (1910) does not contain the t'ak'nu- stem (contra RADCA 144).
19 Cross-linguistically the inchoative and iterative markers are often combined (Hega^KOB 1984: 52). This opens a new perspective for the analysis
The c -stem is best explained here as a relatively late innovation introduced to mark the progressive stage of telic process by analogy to other c -class motion verbs.
Zatcim ‘to depart’. In the Bible, the verb belongs to a different paradigmatic type - imperf. -an-e/i- (zatanem/zatanim): perfect. -0-(zati/zatay), itself a prefixed form of hatane- tr. ‘to separate, to divide’ (Ex. 5, 8; Ex. 39, 3) / hatani- intr. ‘to stand out, to depart’ (Gen. 14, 17).
The preposition z- seems to have the ablative meaning ‘from, off’ within the compound, cf. Num. 16, 25: Zataruk' surj zzolovrdeambd Korxay ew Datanay ew Abironi! (’Ava/wp^aaxs кбк^ф and ouvayroy^q Kops [sic.]) ‘Get back from around the dwellings of Korah, Dathan and Abiram!’; Gen. 30, 40: ew zataner iwr hawts aranjinn (каі 8ic%wpiG8V ¿аихф noi^via каО' ¿auxdv) ‘and he put his own herds apart’. The intransitive uses of hatani- and zatani-(both are indirectly attested in forms of the imperfect in the Bible) are quasi-synonymous, cf. the above-cited Num. 16, 25 and Gen. 31, 40: hateal er kun yacac imoc (каі афштато o vnvoq and xwv офОаА^ету ^ou) ‘and my sleep fled from my eyes’. The only noticeable difference is that, unlike zatani-, the Subject of hatani- is never a volitional Agent. This seems to indicate that z- somehow combined spatial meaning with the reflexive connotation, cf. zgenum ‘to dress oneself’ (from PArm. *genumai) and the direct object article z-.
The agentive value of zat- was inherited by zatc i-/zate- attested two times in Movses Xorenac'i in the form of infinitive with punctive aspectual value (Movses Xorenac'i 2003: 1918, 2073). Thus, it may well be a morphological innovation developed under the influence of other c -class motion verbs. The i-perfective stem is not attested, the causative past participle is used instead, cf. Movses Xorenac'i 2003: 1918: Bayc Samueli Mamikoneni yinkn dmbrneal zt'ult'n Xosrovu ew zpatcen t'lt'oyn naxararacn zatuceal i nocane gnac ar kaysrn Arkad ‘But Samuel Mamikonean acquired Khosrov’s letter and a copy of the princes’ letter, and abandoning them went to the Emperor Arcadius.’ (trans. Thomson 1978: 312).
3.2. Non-agentive c -class verbs
Pakcim ‘to be(come) affected’. The imperfective stem is attested only once in the Bible and can be interpreted either as a distributive punctive event or as a state (both aspectual meanings are typical for the imperfect forms), cf. Wis. 17, 14: Er zi nsanawk' cnoricn
of punctive change-of-state inchoatives (Hitt. -ess-), sigmatic aorists, and iterative *s-presents.
pake ein (3 pl. impf.)... ‘and now were driven by monstrous specters’. The non-stative reading is typical for the perfective forms (Josh. 7, 5: ew pakeaw sirt zolovrdeann ew elen ibrew zjur / каі спто^Оп Ц карбіа хоїї Аяоїї каі cycvcxo юапер ftSwp ‘so the hearts of the people melted and became as water’) and imperfect subjunctive forms attested in the early classical texts, cf. the Armenian translations of Ephrem (Ep'rem 1836: 267, with reference to Deut. 4, 19): Tesanicek zaregakn ew zlusin, pakeicik (2 pl. pres. subj.), jimaricik “You will see the sun and the moon, will be drawn away and go mad”.
No secure etymology (HAB 4: 7f.).
Xrteim ‘to be(come) anxious’. The imperfective stem is attested with the habitual meaning in Movses Xorenac'i (2003: 2085): ...kanzi yarajeln jis im xrtei. ‘Because if I go first my horse will shy [lit. because my horse gets nervous from being first].’ (trans. Thomson 1978: 324) It is further attested in John Chrysostom’s Commentary on Isaiah (Yovhann Oskeberan 1880: 375): ...et'e oc eki zi datecayc zasxarh, ayl zi kecucic: zi paxcik, zi xrteik? ‘for I came not to judge the world, but to save the world: why you run away, why you are/become anxious?’; and in The Commentary on Mathew (Yovhann Oskeberan 1826, II: 547): Zi xrteein imn i Kristose asakertkn Yovhannu... ‘Because John’s disciples were/got anxious about Jesus...’20. Both contexts allow for distributive punctive and stative readings. The stative and punctive inchoative aspects could be alternatively expressed by analytic noun predicates xirtxirt linel ‘to be anxious’ and xirt i xirts mate el ‘to become anxious’ both attested e.g. in John Chrysostom’s Homilies on the Epistles to the Philippians (Yovhann Oskeberan 1862, I: 387f., 529).
The competing imperfective stem xrtnu- is found once in the Armenian translation of Severian of Gabala’s writings (Seberianos 1827: 58): Jilk erkatT zanhnazand paranocs ase, or taltkayce ew xusapice mtanel 3nd lcov astuacpastut’ean, or yawrinacn xrtnucu... ‘He says that iron sinews do not obey the neck so that it would become weary, would escape from coming under the yoke of piety, and would get frustrated by law.’
No secure etymology (HAB 2: 371).
Zart e im ‘to awaken’ is a prepositional compound (from *z-hart'c im)21. The imperfective stem occurs only twice in the non-
20
LALT confuses the two citations from NBHL 1: 995, leaves out one of the two examples and, as a result, wrongly spells and attributes the source text of another one.
21 Pres. art'ci- is first attested in the undated Armenian translation of
biblical texts (Movses Xorenac'i, Yacaxapatum), while the perfective stem is well represented already in the Bible. Besides the Experiencer and Stimulus arguments, this verb has the facultative Source argument expressed by the ablative prepositional construction i/y- + abl. (e.g. i k'noy ‘out of sleep’).
The imperfective stem is attested in the infinitive form in Movses Xorenac'i 2003: 1834: Ayl ew i canaparhi nnjeln Hrudenay, ew Biwraspeay k'arseln i blurn, ew zartceln [note variants: (i) zart'(n)c el(n)] Hrudenay, ew tanel zna yayrs inc lerinn ew kapel... ‘and [how] on the journey Hruden fell asleep and Biurasp dragged him to the hill; and Hruden woke up and led him to a cave in the mountain and bound him...’ (Thomson 1978: 126). A series of nominalized infinitives describing past events suggests the punctual reading of zart'c eln.
The verb is further attested in Yacaxapatum, a collection of homilies attributed to Grigor Lusaworic or Mastoc' and dated to 7th c. in LALT (Абегян 1975: 81-83; Thomson 1995: 126f.). The passage from Yacaxapatum is a quote of Ps. 2, 12 that deviates from the standard Bible edition in the use of 3 sg. pres. subj. zart'c ici instead of 3 sg. aor. subj. borbokesci to translate Gk. 3 sg. aor. subj. сккаиб^ ‘shall be kindled’, cf. Lusaworic' 2003: 80: dnkalaruk' zxrat nora zi mi barkasci Tër, ew korncijik' i canaparhacn ardarut'ean i zamanaki yorum zart'c ici barkut'iwn nora (8ра^ао0е naiôciaç, оплоте ôpyia0^ KÔpioç каі алоА,еїо0е sÇ ôôoïï ôiKaiaç. ornv еккаив^ [sv Ta%ei] ô 0u^ôç айтоїї) ‘Do homage to the Son, that He not become angry, and you perish in the way, for His wrath may soon be kindled.’
The stem zart nu- is attested four times in the Bible (not in Agat'angelos, P'awstos Buzandac'i, Elise, and Movses Xorenac'i contra RADCA 143). The infinitive zartnul (Num. 23, 19; Jer. 47, 7; Rom. 13, 11) always translates the Greek aorist infinitive and denotes a punctual event. The only personal form is found in Zech. 4, 1 with the habitual meaning: zor awrinak yorzam zartnucu ayr i k'noy iwrmë (ov Tponov ornv ê^syspO'n avôpronoç sÇ ïïnvou ашоїї) ‘as a man who is awakened from his sleep’. The non-prefixed verb
Aphrahat’s homilies (“Sermo unicus” of the cited edition that is labeled as “19” in NBHL 1: 554): Tesanen zna amenayn imastumk', ew art’ein i nmanë, ew oc' arnen zkams nora / Vident eum omnes sapientes, et cavent ab ipso, et non implant voluntatem ejus ‘All wizards see him and restrain from him, and do not fulfill his will.’ (Aphrahat 1756: 441 = 1824: 378, wrongly attributed to Jacob of Nisibis; cf. Thomson 1995: 30). The imperfective stem has habitual meaning (distributive-punctive or distributive-durative).
hartnum is attested twice in the sense ‘to move away’ in Severian 1827: 126: Bayc é ew ayl inc orinak tesanel; yorzam urkan arkanici, juknn hartnu, ew i mijin cocn dimé. ‘But another example may be observed: when a net is cast, fish rushes away towards the bosom’. Ew tés zskancelisn, zi zarajinn satanay arkanér urkan asxarhi, ew dewkn jgéin, ew ardarkn hartnum, ew artak's kan zurkann elanéin. ‘And see the wonder: at first Satan casts a net to the world, and the devils stretch [it], and the righteous people rush away and escape the net.’ In both cases, the nu-stem is used with the habitual meaning and refers to hasty ablative motion.
The aspectual properties of the examined c-class verbs are summarized in the table below.
1 Aorist indicative Present/Imperfect indicative
Agentive
dndostc im ‘to leap out’ hangc im ‘to take rest’ mate im ‘to approach’ ostc im ‘to leap out’ paxc im ‘to flee’ t'ak'c im ‘to hide’ t'rc im ‘to fly’ zat'c im ‘to part’ yangc im ‘to relax’ n/a limitative completive n/a n/a punctive inchoative n/a n/a progressive (punctive) stative progressive, habitual (completive) distributive (punctive) habitual / distributive (completive) habitual / distributive (punctive) habitual (progressive) distributive (punctive) habitual (completive)
Non-agentive
pakc im ‘to be(come) affected’ xrtc im ‘be(come) anxious’ zart'c im ‘to awaken’ inchoative n/a inchoative distributive (inchoative/stative) habitual (inchoative /stative) distributive (inchoative /stative)
4. Derivational semantics of PArm. *-e-ske/o-
The three verbs of the c -class with established Indo-European etymologies (hangc im, mate im, take im) are agentive; two of them have *-ehj- (*-e-) and *-ehj-ske/o- (*-e-ske/o-) cognates in Latin. This is an argument in favor of the assumption that the volitional Agent was typical for at least a part of the intransitive verbs that belonged to the morphological type in question from the earliest stages of Proto-Armenian.
The present stem of Lat. taceo most probably continue the PIE *-ehj-ye/o- formation; the Armenian continuant of PIE *-ehj-ye/o- is the imperfective mediopassive marker -i-, cf. Arm. nstim ‘to sit’
from *ni-sed-e-ye/o- ‘to remain seated’ (cf. OCS. C%fl%TH, see Meillet 1936: 108; Godel 1975: 120; Klingenschmitt 1982: 9-11). The telic eventive verb PIE *sed- ‘to sit down’ is an autocausative reflexive verb the Subject of which corresponds to the co-referent Actor and Undergoes. The internal limit of this telic predicate is expressed in the prepositional compound stem *ni-sed-. The stative *-ehj- suffix would then denote the resultant state of the Agent22. I assume, that the fact that the volitional Agent keeps control over the resultative predicate is responsible for the eventive (dynamic) interpretation of the state, whence the secondary intransitive marker *-ye/o-23. Note that the aorist of nstim is nst-ay not xnst-e-ay; it goes back to the root aorist of the original eventive verb.
The *-ehj-ske/o- (or *-e-ske/o-) formant could develop along the same lines in the case of stative hangc im ‘to take rest’. The question then arises what was the grammatical or chronological difference between PArm. *-ehj-ye/o- (Arm. xhangim) and *-ehj-ske/o- (Arm. hangc im). A similar derivational pattern is provided by the Greek verbs noBsro ‘to regret’ ^ Hom. rcoBssoKov, and ^i^sro ‘to love’ ^ Hom. 9i^880K8 (Sihler 1995: 521). Significantly, the eoK-derivatives are quasi-synonymous to the underlying stative e-stem.
Besides a stative verb, the agentive verbs of the c -class comprise verbs denoting controlled motions, both processual (mat'cim ‘to come close’, paxcim ‘to flee’, t'rcim ‘to fly (away)’, yangcim ‘to relax’) and punctual (ostcim, dndostcim ‘to leap out’, t'ak'cim ‘to hide’, zatcim ‘to depart’).
The verbs that belong to the agentive type of c -verbs semantically overlap with the an-e/i-class, cf. t'rcim / t'ranim ‘to fly’, ostc im / ostanim ‘to leap out’, and zatcim ‘to part’ / zatanem ‘to separate’. Note that all the intransitive ane-class verbs denote motions: ancanem ‘to come up, pass’, elanem ‘to go out; to mount’, ijanem ‘to go down’, hasanem ‘to reach’, mtanem ‘to come in’. Prepositional compound pres. act. y-anc -anem, aor. mp. y-anc-e-ay ‘to transgress’ (derived from pres. act. anc-anem, aor. act. anc-i ‘to
22 The resultative value of PIE *-ehj- fits well into Cowgill’s claim (1973) that the primary function of the suffix was to derive “stative” stems out of telic dynamic verbs (“aoristic roots”). It may be illustrated by such reconstructions as PIE *ghehxbh- ‘take’ ^ *ghdxbh-eh1-(ye-)ti > Lat. habeo ‘to have’ (LIV 195), and PIE *swelH-/*swlH- ‘to swell’ ^ *swlH-eh1- ‘to be(come) swollen’, Hitt. sulle-/sulla- ‘id.’ (Melchert 2005).
23 See Yakubovich 2013b: 64f. with an up-to-date overview of the PIE secondary intransitive suffix *-ye- and its traces in Anatolian and Indo-Iranian.
pass’) is a morphologically intermediate type . The volitional Agent of these verbs is certainly responsible for the active inflection of the markedly transitive ane-class which comprises morphological causatives in -uc-ane-. The autocausative (reflexive) interpretation of the motion verbs may be the reason for their partial transfer to the c-class motion; the autocausative semantics would explain such morphological deviations of the c-class as the anomalous use of the causative past participle -uc-eal to express basic intransitive meaning, and the active inflection of the aorist subjunctive conjugation (except mp. 1sg.) with the reflexive middle meaning.
Klingenschmitt (1982: 78) evokes the morphological parallelism between Gk. àÀôn^Kœ ‘to grow’ : okbaivro ‘to make to grow’ and Arm. zatcim ‘to depart’ : zatanem ‘to separate’. However this comparison does not seem probative, since zatanem is attested with the intransitive meaning of a motion verb (see above) and the motion verbs in -ane- do not add valence to the base verb in Armenian. Here, the reflexive value of z- (cf. hatanem ‘to depart’ with non-volitional Subjects) goes hand in hand with the suggested autocausative interpretation of the c -class verbs of motion.
This scenario finds further support in anticausative zart'c im ‘to awaken’ ^ zart'num from an agentive motion verb hart'num ‘to rebound’. The change is best explained through the intermediation of a reflexive derivational model: z- + -c -.
In order to account for the non-stative agentive verbs of the c -class, I consider the following possibility. The gradual grammaticalisation of the resultative markers to the markers of perfect, aorist, and preterit is a commonplace of the linguistic typology25. If such process happened in a subclass of PIE *-ehj- stems they could replace the older root “aorists” under specific conditions and would subsequently serve as basis for derived *ske/o-stems expressing a
24 The prefix y- accompanies actional derivation of a telic process contingent with the state expressed by the non-prefixed stative predicate, cf. hange im ‘to take rest’ ^ y-angcim ‘to relax’, y-awelum ‘to add to’ from aweli ‘more’. The terminative force of the prefix y- is consistent with the directional value of the prepositional construction i/y- + accusative. When added to an eventive predicate, which is the case of ancanem ‘to pass by’ ^ yancanem ‘to transgress, to commit a crime’, y- marks the telicity.
25 The grammaticalization from resultative to perfect (or from stative to inchoative) markers is widely discussed in the Indo-European studies with regards to the reduplicated perfect, e.g. Маслов 1983, Nedjalkov 1988, Kümmel 2000. Just like in the reduplicated perfect forms in Ancient Greek, the gradual process of such grammaticalization could affects different lexical classes and even individual verbs at varying speed.
wide range of secondary imperfective meanings proper for the actional classes of punctual events or telic processes. Optionally, one can assume that the perfectivised resultative received an additional perfective marker *-s- (or that the resultative was perfectivized by *-s-)26 yielding *-i-s- consistent with Klingenschmitt’s etymology of /-perfective in Armenian. Thus, *tak-i- ‘to be silent’ > *tak-i(-s)- ‘to get silent/hide’ ^ tak-iske/o- ‘to be getting silent/hide’ > t'akcim 27, *mad-i- ‘to be close’ > *mad-i(s)- ‘to get close’ ^ *mad-iske/o- ‘to be getting close’ > mate im.
The non-agentive verbs of the e'-class are anticausative predicates denoting “psychological” events or states (psych verbs); their Subject is an Experiencer and the facultative Oblique Object is a Stimulus. The case marking of a Stimulus coincides with the marking of an Agent in the passive construction (instrumental case or the prepositional construction i/y- + abl.), on the one hand, and with the Source argument of motion verbs (the prepositional construction i/y- + abl.). This may be one of the morphological ties between the agentive verbs of motions and non-agentive anticausative verbs of the с -class.
26 Note that *-5- marked the change-of-state event preceding the resultant state in the inchoative *-ehj-s- stems (cf. Hitt. -ё-ss-, Watkins 1971; see also Kloekhorst 2008: 255 who reconstructs *-ehj-shi- to account for -ss-). In this way, *-5- acquired the prospective value with regards to the resultative stem to which it was added. This prospective value of *-5- may be related to the derivational semantics of the sigmatic future formations in *-ehr5-attested in a number of Indo-European branches (Gk. -по-, Balt. -e5iu, see Jasanoff 2003); cf. the derivational semantics of the preposition по- in Russ. imperfect. pret. (“imperfect”) нёс ‘I was bearing’ : perfect. pret. (“aorist”) по-нёс ‘I started bearing’ along with pres. несу ‘I bear’ : fut. по-несу ‘I will bear’. If the sigmatic *-ehj-5- (*-ё-5-) marked the inchoative telic events, the question arises whether its progressive aspect was marked by a derived (*-ehj-5-ke/o-) or substituted (*-ehj-ske/o-) morpheme.
27According to Haverling 1996: 177, the Late Latin tacёsc5 is an unprefixed derived form the corresponding Classical Latin cont^sc5. It is noteworthy that Lat. cont^5co is characterized by the prefix con-. “Con- indicates that the action or process described by the 5co-verb is terminative without emphasizing its beginning or end...” (Haverling 1996: 171; see also Rosen 1992). Along the same lines, P. Chantraine assumed that -(i)oko- marked accomplishment of a telic process and was often accompanied by a prefix as in Hom. ало-, каха-Оу'пакю (DELG 62, 422, 702). PIt. *-ё5c- may have comparable derivational background; the *-ё- suffix would then be a perfective maker just like *-i- (whatever be its origin) in Gk. -г-ако-.
The c -class anticausatives semantically overlap with the «-class psych verbs, cf. ateam tr. ‘to hate’ (aor. atec'ay) from *ati-am)28 from PArm. *H(o)d-e- (cf. PIE stative *h3e-h3(o)d-, Lat. pf. odi, LIV 296). The grammatical contrast consists in the following properties: the a-class psych verbs are transitive (cf. cankam tr. ‘I wish’), do not have valence on a Stimulus, and their imperfective stem denotes a state.
In my opinion, the non-agentive c -class verbs are best explained as outshoots of the object-oriented resultatives in *-ehj- (*-e-) which grammaticalized to the perfective marker within in Proto-Armenian. Object-oriented resultatives (like *swelH-/*swH- ‘to swell’ ^ *swlH-ehj- ‘to be(come) swollen’, Hitt. sulle-l sulla- ‘id.’, see n. 20) would be a natural source for Greek passive aorists in -n-, cf. c^iyn ‘(is) mixed’ (mainly primary verbs). The hypothesis that *-ehj- alone (not *-ehj-ye/o-) could shape a stativelresultative stem has been recently criticized in Jasanoff 2003. However, the evidence of the Armenian c -class seems to support Cowgill’s hypothesis of the stative value of *-ehj-. The derivational path from object-oriented resultative *-ehj- to punctive inchoative event finds further support in Middle Iranian decausative inchoative intransitives in -äs-, recently studied in Yakubovich 2013, cf. Chor. yr’s l yr’d ‘to awaken’. According to Yakubovich, the inchoative *-ä-sa- could replace the earlier *-ä-iia-. This replacement should probably be analyzed in terms of actional derivation rather than a formal analogical change.
There is a wide-spread opinion that quite a number of Armenian verbs have both c- and nu-stems (e.g. Meillet 1936: 110, Jahukean 1987: 378). In fact, only paxcim and paxnum prove to be by-forms in the Bible (unlikepaxcim, the latter stem is not found elsewhere). None other verb shows variation of -c - and -nu- within the securely attributed 5th c. texts (Eznik, Koriwn, Agat'angelos, Lazar P'arpec'i, P'awstos Buzandac'i; apud RADCA: 5). Besides paxci-lpaxnu-, three verbs show variation in the later texts dated in RADCA and LALT up to, but not including, Grigor Narekac'i (t'ak'ci-lt'ak'nu-, xrtci-lxrtnu-, zart'ci-lzart'nu-); even then, the two stems are distributed among different authors and are unequally represented. The stems ostnu- (NBHL 2: 523; HAB 3: 569 and RADCA: 144) and paknu- (NBHL 2: 585; HAB 4: 7; RADCA: 143) are not attested within the scope of the early classical literature considered in RADCA. No t'ak'nu- is attested in the 7th-8th c. Armenian translation
28
Barton hypothesis that the Armenian a-stems continue the PIE *-hj-ye/o-(Barton 1990-1991) is dubious; note in particular that Arm. mnam ‘to remain’ from PArm. *men-a- is not compatible with the ablaut of *mn-hj-ye/o-.
of Ireneus’s Adversus Haereses (1910) referred to in RADCA: 144; instead, one finds forms derived from the c-imperfective and i-perfective stems. The nu-stem characterizes three morphological classes that have: 1) root perfecive, 2) c--perfective, 3) i-perfective. The latter type has the causative past participle in -uc-eal and is undoubtedly the source of confusion with the respective c -class.
The verbal class characterized by the nu-imperfective and i-perfective stems comprise agentive (e.g. zbalnum ‘to be occupied’, psnum ‘to stare’, paxnum ‘to run up to’, hartnum ‘to retreat’) and non-agentive (e.g. casnum ‘to be angry’, k'alc'num ‘to starve, get hungry’) verbs. The autocausative semantics is prominent in zbalnum and is consistent with the motion verbs; the non-agentive verbs are clear anticausatives. Again, the autocausative semantics might explain the i-perfective, causative participle and the active subjunctive inflection in intransitive contexts for this sub-set of the nu-class.
It is difficult to pinpoint the grammatical contrast between the c -and nu-class verbs with the i-perfective stem. It should be noted however that while c-verbs of motion are predominantly used in ablative (source-oriented) contexts, the nu-class verbs of motion are directive (goal-oriented), cf. Amos 5, 19 above. The Stimulus of the anticausative c-verbs is external and is marked like the Agent of passives, while the Stimulus of the anticausative nu-verbs is either internal and is regularly omitted (k'alc'num ‘to starve’) or is marked differently than that of the c -verbs, e.g. the Stimulus of casnum ‘to be angry’ is expressed by the dative (Tob. 5, 19), snd + accusative (Jn. 7, 23), i veray + genitive (lEsdras 1, 52), ar + accusative (lEsdras 4, 31), vasn + genitive (Movses Xorenaci 3: 17), but never as i/y- + ablative. These phenomena may be ultimately related to the aspectual contrast between inchoative and completive reading of the i-perfective stem resulting in the source-oriented c -imperfectives and goal-oriented nu-imperfectives respectively. This tentative suggestion requires further study.
To sum up, the imperfective c -stem expressed a variety of aspectual meanings of intransitive verbs depending on the lexical meaning. The progressive aspectual meaning is very rare for c -imperfective indicative forms. On the contrary, the stative and secondary imper-fective aspects of the underlying punctive/completive events are well attested. The tendency of c -stems to describe secondary aspectual meanings derived from punctive and completive ones of the underlying i-perfective stems may be considered as an argument in favor of the hypothesis that Proto-Armenian or late Indo-European
*ske/o-fformations had a period of productivity while *-ehj- (*-ë-) was either gradually turning into a marker of the mediopassive perfect or were “perfectivised” by means of *-s-.
Abbreviations
DELG - Chantraine P. Dictionnaire étymologique de la langue grecque. Histoire des mots / Taillardat J., Masson O., Perpillou J.-L. (eds.). Paris, 2009.
HAB - Acaryan Hr. Hayeren armatakan bararan. I-IV. 2nd ed. Yerevan, 1971-1979.
LALT - The Leiden Armenian Lexical Textbase / Weitenberg J. J. S. (ed.), 2003. Scholarly Digital Editions: http://www.sd-editions.com/LALT. LIV - Lexicon der indogermanischen Verben / Rix., H. et al. (eds.). 2 ed.
Wiesbaden, 2001.
NBHL - Nor bargirk' haykazean lezui. I-II. Yerevan, 1979-1981.
RADCA - Jungmann P., Weitenberg J. J. S. A Reverse Analytical Dictionary of Classical Armenian. NY; Berlin, 1993.
Primary sources and translations
Agat'angelos 2003 - Iwzbasean K., Muradean P. (eds.). Agat'angelos. Patmut'iwn hayoc' / Ekawean Z. (ed.). Armenian Classical Authors. Vol. II. Antelias, 2003. P. 1289-1736.
Aphrahat 1756 - Antonelli N. (ed.). Opera S. Jacobi Nisibi. Rome, 1756 (= Jakob Mcbnay. Girk' or koc'i Zgawn. Kosdandnupolis, 1824). Bedrosian 1985 - Bedrosian R. Ghazar P'arpets'i’s History of the Armenians. New York, 1985. http://rbedrosian.com/gpintro.htm.
Bible - Hovhannes Zöhrapean (ed.). Astowacasownc' matean hin ew nor ktakaranac', Venetik 1805.
Elise 2003 - Iwzbasean K., Muradean P. (eds.). Elise. Vasn Vardanay ew Hayoc' paterazmin // Ekawean Z. (ed.). Armenian Classical Authors. Vol. I. Antelias, 2003. P. 515-738.
Ep'rem 1836 - Srboyn Ep'remi Matenagrut'iwnk'. Vol. I. Commentary on Deuteronomy. Venice. 263-295.
Eznik Kolbac'i 2003 - K'eoseean Y. (ed.). Eznik Kolbac'i. Elc alandoc' // Ekawean Z. (ed.). Armenian Classical Authors. Vol. I. Antelias, 2003. P. 431-513.
Ireneus 1910 - Ter-Mekerttschian K., Ter-Minassianz E. (eds.). 1910. Irenaeus gegen die Häretiker. Buch IV und V in Armenischer Version. Leipzig, 1910.
Lazar P'arpec'i 2003 - Iwzbasean K., Muradean P. (eds.). Lazar P'arpec'i. Patmut'iwn hayoc' // Ekawean Z. (ed.). Armenian Classical Authors. Vol. II. Antelias, 2003. P. 2201-2294.
Lusaworic' 2003 - K'eoseean Y. (ed.). Grigor Lusaworic'. Yacaxapatum Cark' // Ekawean Z. (ed.). Armenian Classical Authors. Vol. I. Antelias, 2003. P. 1-153.
Mandakuni 2003 - Step'anean L. (ed.). Yovhan Mandakuni. Cark' // Ekawean Z. (ed.). Armenian Classical Authors. Vol. I. Antelias, 2003. P. 1153-1288.
Philo 1953 - Marcus R. (trans.). Philo: Questions and Answers on Genesis. 1953 (Loeb Classical Library No. 380).
Movses Xorenac'i 2003 - Muradean P., Iwzbasean K. (eds.). Movses Xorenac'i. Patmut'iwn hayoc' // Ekawean Z. (ed.). Armenian Classical Authors. Vol. II. Antelias, 2003. P. 1737-2196.
Philo 1826 - Awgerean M. P'iloni Ebrayec'woy Mnac'ordk' i Hays // Philonis Judaei Paralipomena Armena. Venice, 1826.
Seberianos 1927 - Awgerean M. (ed.). Seberianosi kam Sewerianosi Emesac'woy Gabalac'woy episkoposi Cark'. Venice, 1927.
Thomson 1978 - Thomson R.W. Moses Khorenats'i. History of the Armenians. Harvard, 1978.
Yovhann Oskeberan 1826 - Yovhannu Oskeberani yAwetaranagirn Matt'eos girk' krkin. I-III. Venice, 1826.
Yovhann Oskeberan 1862 - Yovhannu Oskeberani meknut'iwn t'lt'oc'n Pawlosi. I-II. Venice, 1862.
Yovhann Oskeberan 1880 - Tiroyean, A. (ed.). Yovhannu Oskeberani meknut'iwn Esayeay margarei. Venice, 1880.
Works cited
Абегян 1975 - Абегян М. История древнеармянской литературы. Ереван, 1975.
Бондарко 1987 - Бондарко А. В. (отв. ред.). Теория функциональной грамматики: Введение. Аспектуальность. Временная локализован-ность. Таксис. Л., 1987.
Горбова 2010 - Горбова Е. В. Акциональность глагольной лексики и аспектуальные граммемы. Вопросы взаимодействия. СПб., 2010.
Джаукян 1982 - Джаукян Г. Б. Сравнительная грамматика армянского языка. Ереван, 1982.
Маслов 1983 - Маслов Ю. С. Результатив, перфект и глагольный вид // Недялков В. П. (ред.), Типология результативных конструкцией (результатив, статив, пассив, перфект). Л., 1983.
Недялков 1984 - Недялков В. П. Заметки по типологии начинательных конструкций // Сусов И. П. (ред.). Прагматика и семантика синтаксических единиц: сборник научных трудов. Калинин, 1984.
Плунгян 2011 - Плунгян А. В. Введение в грамматическую семантику. Грамматические значения и грамматические системы языков мира. М., 2011.
Barton 1990-1991 - Barton Ch. R. On the denominative a-statives of Armenian // Revue des études arméniennes. 1990-1991. Vol. 22. P. 2952.
Clackson 1994 - Clackson J. The linguistic relationship between Armenian and Greek. Oxford, 1994.
Cow gill 1973 - Cow gill W. The source of Latin stare, with notes on comparable forms elsewhere in Indo-European // The Journal of Indo-European studies. 1973. Vol. 1. P. 271-303.
Croft 2012 - Croft W. Verbs: Aspect and Causal Structure. Oxford, 2012.
Duhoux 1992 - Duhoux Y. Le verbe grec ancien. Éléments de morphologie et de syntaxe historiques. Louvain-la-Neuve, 1992.
Godel 1975 - Godel R. An Introduction to the Study of Classical Armenian. Wiesbaden, 1975.
Haverling 1996 - Haverling G. On sco-verbs, on prefixes and on semantic functions // Hannah Rosén (ed.). Aspects of Latin, Papers from the 7th International Colloquium on Latin Linguistics, Jerusalem, April 1993. Innsbruck, 1996. P. 169-180.
Hoffner, Melchert 2002 - Hoffner H. A., Melchert H. C. A practical approach to verbal aspects in Hittite // Stefano de Martino and Franca Pecchioli Daddi (eds.). Anatolia Antica. Studi in memoria di Fiorella Imparati. Firenze, 2002. P. 377-390.
Jahukyan 1987 - Jahukyan G. B. Hayoc lezvi patmut'yun: naxagrayin zhamanakasryan. Yerevan, 1987.
Jasanoff 2003 - Jasanoff J. H. “Stative” *-e- revisited // Die Sprache. 2003. Bd. 43/2. P. 127-170.
Jasanoff 2003 - Jasanoff J. H. Hittite and the Indo-European Verb. Oxford University Press, 2003.
Klingenschmitt 1982 - Klingenschmitt G. Das altarmensiche Verbum. Wiesbaden, 1982.
Kloekhorst 2008 - Kloekhorst A. Etymological Dictionary of the Hittite Inherited Lexicon. Leiden, 2008.
Kocharov 2011 - Kocharov P. On ana-presents of Armenian // Indogermanistik und Linguistik im Dialog. Akten der XIII. Fachtagung der Indogermanischen Gesellschaft vom 21. bis 27. September 2008 in Salzburg. Wiesbaden, 2011. P. 271-278.
Kümmel 2000 - Kümmel M. Das Perfekt im Indoiranischen: eine Untersuchung der Form und Funktion einer ererbten Kategorie des Verbums. Wiesbaden, 2000.
Laar 2000 - Laar H.van de. Description of the Greek individual verbal systems. Amsterdam, 2000.
Lamberterie 1990 - Lamberterie Ch. de. Les adjectifs grecs en -uç: sémantique et comparaison. Vol. I-II. Louvain-la-Neuve, 1990.
Lamberterie 2005-2007 - Lamberterie Ch. de. L’augment dans le texte arménien de l’Évangile // Revue des études arméniennes. 2005-2007. Vol. 30. P. 31-57.
Lehmann 1986 - Lehmann W. P. A Gothic Etymological Dictionary. Leiden, 1986.
Lyonnet 1933 - Lyonnet S. Le parfait en arménien classique. Paris, 1933.
Martirosyan 2010 - Martirosyan H. Etymological Dictionary of the Armenian Inherited Lexicon. Leiden, 2010.
Meillet 1910-1911 - Meillet A. La syntaxe comparée de l’arménien. IV. Emploi des forms personnelles des verbes // Mémoires de la société de linguistique de Paris. 1910-1911. Vol. 16. P. 92-131.
Meillet 1936 - Meillet A. Esquisse d’une grammaire comparée de l’arménian classique. 2 éd. Vienne, 1936.
Melchert 1998 - Melchert H. C. Aspects of verbal aspect in Hittite // S. Alps and A. Süel (eds.). Uluslararasi Hititoloji Kongresi Bildirleri. Acts of the 3rd International Congress of Hittitology. Çorum, September 1622, 1996. Ankara, 1998. P. 413-418.
Melchert 2005 - Melchert H. C. Latin însolêscô, Hittite sulle(ss)- and PIE Statives in -e- // Казанский Н. Н. (ред.). Hrda Manasa. Сборник статей к 70-летию со дня рождения проф. Леонарда Георгиевича Герценберга. СПб., 2005. P. 90-98.
Nedjalkov 1988 - Nedjalkov V. P. (ed.). Typology of resultative constructions. Amsterdam, 1988.
Olsen 1999 - Olsen B. A. The Noun in Biblical Armenian. Origin and Word-Formation. NY; Berlin, 1999.
Pedersen 1906 - Pedersen H. Armenisch und Nachbarsprachen // Zeitschrift für vergleichende Sprachforschung. 1906. Bd. 39. S. 334-485.
Rau 2009 - Rau J. Indo-European Nominal Morphology: The Decads and the Caland System. Innsbruck, 2009.
Rosen 1992 - Rosen H. B. Die Kompositamit co(n)- in funktioneller und vergleichender Sicht // O. Panagl, Th. Krisch (eds.). Latein und Indogermanisch. Akten des Kolloquiums der Indogermanischen Gesellschaft. Salzburg; Innsbruck, 1992. S. 357-367.
Sihler 1995 - Sihler A. New comparative grammar of Greek and Latin. Oxford, 1995.
Solta 1960 - Solta G. Die Stellung des Armenischen im Kreise der indogermanischen Sprachen. Eine Untersuchung der indogermanischen Bestandteile des armenischen Wortschatzes. Wien, 1960.
Tatevosov 2002 - Tatevosov S. The parameter of actionality // Linguistic typology. 2002. Vol. 6. P. 317-401.
Thomson 1995 - Thomson R. A bibliography of classical Armenian literature to 1500 AD. Turnhout, 1995.
Vaan 2008 - Vaan M. de. Etymological dictionary of Latin and the other Italic languages. Leiden, 2008.
Vaux 1998 - Vaux B. The phonology of Armenian. Oxford, 1998.
Vine 1993 - Vine B. Greek -юкю and Indo-European “*-iske/o-” // Historische Sprachforschung. 1993. Bd. 106. S. 49-60.
Viredaz 2012 - Viredaz R. Origine du type canawt', canacem et questions connexes (Materials for the presentation at the 9th International Conference on Armenian Linguistics). St. Petersburg 7-9 November. Watkins 1971 - Watkins C. Hittite and Indo-European studies: the denominative statives in *-e- // Transactions of the Philological Society. 1971. P. 51-93.
Weitenberg 1986 - Weitenberg J. J. S. Infinitive and participle in Armenian // Annual of Armenian Linguistics. 1986. Vol. 7. P. 1-26.
Yakubovich 2013a - Yakubovich I. Reflexes of Indo-European ‘e-statives’ in Old Indic // Transactions of the Philological Society. 2013. P. 1-23. Yakubovich 2013b - Yakubovich I. Middle Iranian Intransitives in -as- // Лурье П. Б., Тохтасьев С. Р. (ред.). Commentationes Iranicae:
сборник статей к 90-летию В. А. Лившица. СПб., 2013. P. 64-73.
P. A. Kocharov. Derivational semantics of Classical Armenian с-stems
The paper presents a study on the Classical Armenian verbal class characterized by the imperfective с -stem and perfective i-stem. The etymology remains controversial for both stems. The paper aims to show how the account of the argument structure and aspect can contribute to the diachronic analysis of the imperfective -с - and perfective -i- suffixes and their paradigmatic relations.
Keywords: Classical Armenian, voice, actionality, aspect, imperfective stems.
