CC BY
21
SPLIT AUXILIARY SYSTEM IN HITTITE?*
Резюме: В статье разбирается вопрос о статусе хеттских конструкций со вспомогательными глаголами hark- ‘держать, иметь’, сочетающимся с переходными и несколькими непереходными глаголами, и es-‘быть’, сочетающимся только с непереходными глаголами. Существуют две точки зрения: либо существуют две омонимичные конструкции, одна из которых выражает состояние, а другая - предшествование, либо у конструкции, выражавшей изначально состояние, впоследствии развилось результативное значение. Вторая точка зрения выглядит предпочтительней, так как она хорошо согласуется с типологическими данными. Можно предположить, что конструкции со вспомогательными глаголами hark- и es- оформились независимо друг от друга, а их источником могли послужить устоявшиеся выражения вроде lagan hark- ‘преклони (слух)’.
Ключевые слова: хеттский язык, глагол, вспомогательный глагол, перфект, результатив, статив, причастие.
There are several periphrastic constructions in Hittite. The one which is sometimes called ‘periphrastic perfect’ or ‘pluperfect’ is formed by hark- with a participle in N.-Acc. Neut. or by es- ‘to be’ and a participle in the Nominative, coordinated with the subject. The auxiliary hark- was used with transitive as well as some intransitive verbs, while es- was used with the rest of the intransitives.
It was already E. Benveniste (1960: 41-42, 62-63) who distinguished 2 different types of constructions with hark- and es-, a perfect one and a stative one. He also noted their resemblance to split auxiliary systems in European languages. This approach is accepted in Hoffner and Melchert 2008: 310, who, however, warn against a direct comparison between such systems in European languages and in Hittite due to the time gap between the development of these
* Статья написана в рамках проекта «Лексическое и морфологическое варьирование в традиционных индоевропейских текстах», выполняемого по Программе фундаментальных исследований Президиума РАН «Традиции и инновации в истории и культуре».
constructions in Hittite, and in Romance and Germanic languages.1 It was also stated that the distribution of auxiliary verbs for Hittite intransitive verbs is conditioned by the underlying type of intransitives in accordance with the Unaccusativity hypothesis, that is unaccusative verbs take es-, while unergative take hark- (Garrett 1996: 102ff.).2 The examples with hark- in preterite are rare and are first attested in copies of Old Hittite texts. The examples are as follows:
KUB 31.121 III 15-17
ammuk=ma UL kuitki sag[ahhi] memiyann=a=kan EGIR-anda arha UL istamassan harmi ‘I don’t know anything, and have not heard (anything) afterwards (about) the matter’ (Boley 1982: )
KBo 5.8 i 18-20
SA LU.MES URUTaggasta=ma kuies ERIN.MES NARARE anda warrissantes eser n=at arha parasesser ‘But the auxiliary troops of Taggasta who had come to help (my opponents) dispersed’ (Hoffner, Melchert 2008: 311)
Not all the instances, however, can be analyzed as pluperfects, or anteriors in terms of J. Bybee, and they are usually regarded as “stative constructions”. All the cases with imperative formds of hark-or es- belong here.
KBo 3.67 i 6
IBoT 1.29 Rs. 20-21 nu suppa sara danzi suppa karpan harkanzi ‘They take up the meat. They hold the meat’ (Boley 1982: )
1 The auxiliary systems in Germanic and Romance languages with the auxiliaries to be and to have must be an areal feature and may have originated after Latin and Greek models, cf. Drinka 2003.
2 There is a 3 Sg. pronominal clitic in Hittite; according to Watkins’ rule, transitive verbs take these clitics only in the Accusative, while some intransitive verbs require this clitic in the Nominative, that is the clitic denotes object with transitive verbs and subject with intransitive verbs. It is considered to be one of the proofs of split ergativity in Hittite. The intransitive verbs that take subject clitic, including e.g.pai- ‘to go’, arai- ‘to rise’, are considered to be unaccusative while the others are unergative. The list of Hittite verbs taking subject clitics is given by Hoffner and Melchert 2008:. It doesn’t completely match however with the distribution of hark-and es- auxiliary verbs, as karussiya- ‘to be silent’ takes subject clitic and takes hark- as an auxiliary verb (KUB 14.1 obv. 16-17 nu=wa karussiyan har(a)k ‘so keep quiet!’). Most verbs from the list are not attested in the constructions in question anyway. On split auxiliary systems see e.g. Aranovich 2005.
KUB 24.1 i 16-17
nu=mu DINGIR-LUM istamanan lagan hark ‘O god, keep your ear inclined to me’. (Hoffner, Melchert 2008: 311)
In KBo 3.67 karpan hark- cannot have a pluperfect meanings, as the original action is expressed by sara danzi ‘to take up, lift’. Karpan hark- may only denote here an ongoing process r state: ‘to keep holding’.
Despite Hoffner and Melchert’s warning, it is tempting to consider those constructions in Hittite and in European languages to be related or having a common ancestor in PIE, since the auxiliary ‘to have’ is quite rare in the languages of the world. However the analysis of hark-forms by Boley, only briefly mentioned and in fact discarded in the grammar by H. Hoffner H. Craig Melchert, makes it rather implausible.
In her monograph of 1984, J. Boley thoroughly studied all the contexts of hark-constructions, and she summarized the findings in her article of 1992. Her general conclusion was quite different from that of Benveniste and Hoffner and Melchert. First she argues that hark- is never to be regarded as an independent verb in such clauses, as the negation is always attested before both finite form of hark- and the participle (Boley 1992: 39ff.). For a few Old Hittite attestations she rather suggests a stative meaning or at least a reference to the current situation, similar to early Greek and Sanskrit perfects, cf.:
KBo 17.15 Vs. 14-16 [(UGULA , lümeSMUHALDIM)] (15) hassas katta ket arta 6 harnaiSAR harzi L[Ühesta] (16) hassas katta edi parsanan harzi 6 harnaiSAR harzi ‘the head cook stands near the hearth on this side (and) holds 6 H., the LÜhesta crouches by the hearth on that side (and) holds 6 H.’ (Boley 1992: 40).
Note that parsanan harzi ‘crouches’ is parallel to arta ‘stands’.
However she also adduces one context with a resultative meaning already in Old Hittite original, KBo 22.1 Vs 23 natta=smas LÜMESDUGUD-as tuppi hazzian harzi ‘(As my father lets you to your house and as he keeps writing to you), has he not written/engraved the tablet of the LÜME DUGUD for you? (Boley 1992: 43).
In Middle and New Hittite we still find ‘stative’ usage of these forms, as in KUB 17.21 Vs. I 14-23 namma sumenzan DINGIRmeS-as kue ALAMfflA=KUNU SA KU.BABBAR GUSKIN (15) nu=ssan kuedani DINGIRLIM-ni kuit ^ tuekkissi (16) anda wizzapan DINGIRMES-sa kue ÜNUTEmeS wizzapanta (17) n=at anzel iwar EGIR-pa ÜL kuiski (18) newahha[n hart]a (19) namma=smas=san
SISKUR.SISKURHIA-as parkuyannas uddani (20) nahsarattan kissan UL kuiski tiyan harta (21) nu=smas UD-as ITU-as MU-ti meyaniyas SISKUR.SISKURfflA (22) EZENHIA kissan sara UL kuiski (23) tittanuwan harta
‘Further the images of you gods of silver and gold, and whatever has become old to whatever god in his body, and what vessels of the gods have become old, no-one like us kept them renewed. Furthermore, no-one had fear set for the matter of your rituals’ purity in this manner; no-one had/kept the daily, monthly and New Year rituals and festivals set up in this manner for you.’ (Boley 1984: 43)3.
The hark-constructions are better translated as ‘used to do’ or ‘kept doing’. Later in the text it is stated that the situation changed after the war with Gaska people and these actions are no longer carried out.
However, the resultative usages are attested much more often and became a primary meaning of this construction. J. Boley maintains that these constructions never have clear anterior or pluperfect meaning (e.g. Boley 1992: 39, 55-56), as it is never attested in temporal clauses and can always be explained as a resultative. She also assumes a similar development from stative to resultative for es-+ participle constructions (Boley 1992: 46ff.).
The constructions made of a participle of an intransitive verb and es- have never been studied in much detail. In the Old Hittite originals there are a few participles used predicatively, but the verbs involved are mostly transitive so they are better described as passive. As to the intransitive verbs, only huwai- ‘to run’ is attested as a predicate, regretfully in broken contexts (cf. also Boley 1992: 46). The examples are: KBo 8.74 + Vs. II 12 ]x heyawes=a n=e man ser huyantes (13) [ m]a?nikuwantes ma[n]=e kattann=a huyantes, KBo 25.23 8 [... p]eran huyan, KBo 20.19 I 13 [ ]x huyanza[ n]=e 3-SU wahanzi.
In Middle Hittite texts, where we have huyant- in preserved contexts, there seem to be no difference between huyant- + es- and simple present^ and preterite forms, cf. e.g. IBoT 1.36 II 22-23 GAL DUMU.MES.E.GAL-ma=at LUGAL-i pai nu GIShuluganniya piran GAL LUMESsalashas huyanza giSPA harzi ‘and the chief-of-palace-
3
This is preferable to Singer’s translation “<...> no one had ever renewed them as we have, <.> .no one had established’ (Singer 2002: 41).
attendants gives it to the king. In front of the cart walks the chief-of-grooms and he holds the staff’ (Guterbock, van den Hout 1991: 17).
It is not always possible to say whether we deal with a perfect with the function of a pluperfect, or a resultative4. The examples like HKM 25:15-19 Nu=ssan man halkies arantes n=as=kan arha warsten ‘When the crops have ripened (lit., ‘arrived’), harvest them! (Bring them to the granary!)’ (Translated as a perfect by Hoffner, Melchert 2008: 311). This ambiguity may be due to the fact that Hittite participles express a state, a result of the action, as was established by E. Neu (1968: 120ff.), cf. also Hoffner, Melchert 2008: 339. This can be supported by the fact that many Hittite periphrastic passives also seem to have a perfect meaning as well.
In general, Boley’s arguments seem preferable. If there was originally a hark- construction with a ‘stative’ meaning, that later developed into a resultative and further into a true perfect, it matches well the available typological data where possessive constructions and constructions with a copula may develop into a resultative which in turn may become a perfect (Bybee, Dahl 1989: 67-78, CHHHHaBa 2005: 87ff.). The idea of two synchronically homonymous and unrelated constructions with stative and pluperfect meanings, on the other hand, seems rather improbable. However, it is quite possible that in the Late Hittite we may deal with true perfects (cf. the passages from the Bronze Tablet in Boley 1992: 54ff.).
In this case an independent development of the periphrastic constructions with hark- and es- is to be assumed, which later merged into one split auxiliary system. Due to the high frequency and persistence of several expressions like karpan hark-, lagan hark-and huyant- + es- it may be assumed that it was these expressions that set the stage for the development of resultative constructions in Hittite.
While the construction with the auxiliary hark- is well attested since Old Hittite originals, the es- + participle of an intransitive verb is more or less rare even in the New Hittite texts. Beside huwai- ‘tu run’ and ar- ‘to come, arrive’, there are predicative participles ofpai-‘to go’, watku- ‘to jump, flee’, wariss- ‘to come to help’ and a few more that are usually attested since Suppiluliuma I. That may mean
4 For the criteria for distinguishing perfects and resultatives see Hega^KOB 1983.
that es + participle became a distinct resultative construction later than the one with hark- .
The earlier emergence of the resultative construction with hark-may explain its usage with several intransitive verbs as well, as hark-may have started to spread on certain types of intransitives. Besides parsnai- ‘to crouch’ J. Boley (1992: 40ff.) lists lalukisnu- ‘to give light’, wahnu- ‘to turn’, wastanu- ‘to sin’, es- ‘to sit’5, kururiyahh-‘to make war (on)’, karussie- ‘to be silent’. It is remarkable that 4 of these intransitive verbs have causative/factitive suffixes -nu- and -ahh- that could justify the involvement of hark- as an auxiliary. Boley argues that these verbs with the exception of karussie- and es-have an internal object (1992: 41f.), which doesn’t seem plausible. We may note, however, that the verbs that take es- as an auxiliary are mostly verbs of movement, while the verbs that take hark- denote controlled actions. This distribution is in accord with the Auxiliary Selection Hierarchy, established for Western European languages by A. Sorace (2000), cf. also Aranovich 2007: 15ff, where non-motional controlled verbs universally take have while motional verbs may take both have and be, depending on the language.
References
Недялков, Яхонтов 1983 - Недялков В. П., Яхонтов С. Е. Типология результативных конструкций // Типология результативных конструкций. Ред. В. П. Недялков. Л.: Наука. C. 5-41.
Сичинава 2005 - Сичинава Д. В. Типология глагольных систем с синонимией базовых элементов парадигмы. Дисс. ... канд. филол. наук. М., МГУ им. М. В. Ломоносова.
Aranovich 2007 - Aranovich R. (ed.). Split-auxiliary systems: a cross-linguistic perspective. Amsterdam.
Benveniste 1962 - Benveniste E.. Hittite et indo-européen: études comparatives. Paris.
Boley 1984 - Boley J. The Hittite hark-Construction. IBS 44. Innsbruck. Boley 1992 - Boley J. The Hittite periphrastic constructions. Per Una Grammatica Ittita. Ed.: O. Carruba. Pavia. P. 33-59.
Bybee, Dahl 1989 - Bybee J., Dahl E. The creation of tense and aspect systems in the languages of the world. Studies in Language 13-1. P. 51-103.
5 J. Boley considers appa asan hark- to be obscure; J. Puhvel, however, translates KUB 32.121 II 22 ]-kan appa tuk-pat asan harweni as ‘we are seated behind you’ (Puhvel 1991: 149).
Drinka 2003 - Drinka B. The formation of periphrastic perfects and passives in Europe. Historical linguistics 2001. Amsterdam: Benjamins. P. 105-128.
Garrett 1996 - Garrett A. Wackernagel’s Law and Unaccusativity in Hittite. Approaching second: Second position clitics and Related Phenomena. Eds.: A.L. Halpern, A.M. Zwicky. Stanford. P. 85-133.
Güterbock, van den Hout 1991 - Güterbock H., van den Hout Theo. The Hittite instruction for the royal bodyguard. Chicago.
Hoffner, Melchert 2008 - Hoffner H., Melchert H. C. A Grammar of the Hittite Language. Winona Lake: Eisenbrauns.
Neu 1968 - Neu E. Das hethitische Mediopassiv und seine indogermanischen Grundlagen. StBoT 6. Wiesbaden.
Puhvel 1991 - Puhvel J. Hittite Etymological Dictionary. Volume 3; Words beginning with H. Berlin; New-York.
Singer 2002 - Singer I. Hittite prayers. Leiden: Brill.
Sorace 2000 - Sorace A. Gradients in auxiliary selection with intransitive verbs. Language 76. P. 859-890.
A. V. Shatskov. Split auxiliary system in Hittite?
According to the Hittite grammar by H. Hoffner and H. Craig Melchert (2008: 310ff.) there are 2 homonymous types of periphrastic constructions with the verbs hark- ‘to have’ and es- ‘to be’ with participles in Hittite. J. Boley however (1984, 1992) who thoroughly studied Hittite constructions with the auxiliary hark- argues that originally hark- + participle had a stative meaning and later it developed a resultative meaning. This latter approach is preferable since it is supported by cross-linguistic parallels. The constructions with the auxiliaries hark- and es- are likely to have developed independently from possessive constructions and copula + participle.
