CC BY-NC-ND
6
Discussion Articles / Дискуссионные статьи
Таким образом, система языка шина как бы отражает ситуацию в индоарийском до ведийского сдвига ударения на высокотональных платформах. Но в ведийском могло быть положение, при кото-
ром в порядке СуСуСу отсутствовала определенность в месте акцента, в отличие от порядка СуСу#, что могло отразиться в наличии акцентуационных дублетов.
Leonid Kulikov
Leiden University / Institute of Linguistics (Moscow)
Reply to replies
Replies of V. Dybo and A. Kassian offer a number of interesting historical observations, placing the issue of the history of the main accentual type of -ya-present in a new perspective. I will not enter here into a general discussion of the comparability of evidence provided by Balto-Slavic accent and Vedic verbal accentuation, which represents quite an intricate issue on its own, but goes far beyond the scope of the current discussion. Rather, I will confine myself to a few more specific remarks on the data and their interpretation provided by the discussants.
As rightly noticed by A.K., the explanation of several subgroupings within the system of the -ya-pres-ents in Vedic (largely) based on Kuryiowicz's analogical scenarios is not free from complications and several back and forth developments in the accentual history of the -ya-formations. Putting the accentual patterns in direct connections with the tonal schemes of the morphemic sequences in accordance with their accentual types (dominant/recessive) may, at first glance, spare some of such 'redundant' changes of my scenario (as outlined by A.K.).
Yet, this alternative explanation is not free from heavy problems either, while the lack of comparative evidence, quite unfortunately, makes this analysis less falsifiable than the (more traditional) explanation.
Let us take a closer look at the rule that forms the core of Dybo's tonal theory of the genesis of the Vedic accentuation as applied to the accent patterns of the -ya-presents (see p. 207): (i) the -ya-presents derived from the roots of the dominant tonal type should bear the accent on the suffix, whilst (ii) the -ya-presents derived from the roots of the recessive type should have the accent on the root. How could then this purely phonological distribution be dephonologized, so that, ultimately, the place of accentuation becomes conditioned
by the semantic types of -ya-presents? Developing the basic idea of V.D. and A.K., one might assume the following historical scenario: (I) a certain (semantically influential?) group of the -ya-presents of the former type (dominant roots = accent on the suffix) were mostly used as passives and therefore have formed the core group of the -ya-passives, whereas (II) a certain (se-mantically influential?) group of the -ya-presents of the latter type (recessive roots = accent on the root) mostly occurred in non-passive usages and therefore have given rise to the Old Indian 'class IV' presents, i.e. to the non-passive -ya-presents with the root accentuation. Subsequently, the first group attracted those -ya-passives which, by virtue of the tonal type of the root morpheme (recessive) had accent on the root (with the concomitant accent shift from the root to the suffix, in analogy with the core members of the class: *V'-ya- > V-ya-), while in another class we expect the opposite development: non-passive -ya-presents with the root accentuation attracted other non-passive -ya-presents that had accent on the suffix, with the concomitant accent shift from the suffix to the root, in analogy with the core members of the class: *V-ya- > V'-ya-).
Unfortunately, as noticed by A.K. (p. 199), the Balto-Slavic material furnishes as few as two reliable cognates of the Vedic -ya-presents that can be used for the reconstruction of the original tonal pattern of the Vedic stems. One of them, Slav. *topiti 'warm, make warm' (the exact cognate of the Vedic causative ta-payati id.) must testify to the dominant type of the root, which, in accordance with Dybo's rule, should result in Vedic suffix accentuation tapyate 'heats; suffers'. This accentuation is attested from the Athar-vaveda onwards, alongside with the root accentuation tapyate, which is met with, in particular, in the Yajur-vedic mantras; see p. 190 above. Another direct com-
Leonid KULIKOV. Reply to replies
parison is Slav. *dbr-ati (~ Ved. diryate 'cracks'), with a recessive root, which should point to the root accentuation (diryate, attested in the Taittiriya- and Mai-trayani Samhitas of the Yajurveda, alongside with the suffix accentuation diryate found in the Satapatha-Brahmana). Obviously, evidence is too scant for any decisive conclusion — however attractive the tonal hypothesis might appear for the explanation of the initial split of one single type into two accentual classes.
More substantial evidence for hypothesizing about the original accentuation of the (Old) Indo-Aryan -ya-presents is, allegedly, provided by the Dardic language Shina, which, according to V.D., preserves the original accentuation (on the root) in the cognates of all -ya-presents, irrespectively of the accentuation in Vedic. The question on whether we can reduce all -ya-presents to a single accentual system on the basis of evidence from Dardic remains open, however (see A.K.'s objections, p. 199).
Furthermore, the presentation of the Vedic material on p. 208-209 is not free from inaccuracies or unlikely assumptions. Thus, there are no good reasons to trace two homonymous (albeit perhaps genetically related) roots IT, 'cling, adhere' and 'dissolve', and the root ri (which, as Praust (2000) has demonstrated, has the meaning 'whirl, swirl', not merely 'flow') to the same historical source. Notice that IT 'disappear, dissolve' (as of salt in water) normally refers to a solid substance that becomes liquid; by contrast, ri can only be constructed with the subject of a liquid.
Problematic is the comparison of Sh. Gil. paruzhei 'hears, listens' ['слышит, слушает'] with budhyate, which originally could only mean 'awakens'; the mean-
ing 'perceives, notices' must represent inner Vedic development.
Sh. Gil. razei 'is cooked' etc. is compared to the nonexistent Sanskrit form *radhyate 'is softened'; its reconstruction on the basis of act. radhyatu 'let (him) subdue, be subject' AV etc. is implausible: these two meanings can hardly be reconciled with each other.
Positing such a monstrous form as the alleged passive *misryate1 ('is mixed'?) is based on mere misunderstanding: misrayati 'mixes' is a late Vedic (from the Sutras onwards) denominative derived from the adjective misra- 'mixed, mingled'. Passives based on denominatives are very late and by no means could exist in early Vedic, let alone in Proto-Indo-Iranian.
Last but not least, the prevalence of non-passives in the list of Shina verbs is remarkable and considerably weakens V.D.'s hypothesis.
Notice also the amazing parallelism between the accent shift from the first short syllable to the second (suffixal) syllable and the accent shift2 from the root to suffix in -ya-presents made from the roots of the type Cr, as in mriyate 'dies' < *mr-ia-te,3 which I discussed elsewhere (Kulikov 1997).
To conclude, accentuation in Shina, however archaic it might be, can hardly corroborate our hypotheses on the original accentual patterns in of the Vedic -ya-presents.
1 Such form is hardly possible by virtue of the rules of Sanskrit phonetics; a more probable structure would be perhaps *misriyate.
2 Discarded by V.D. with no good reasons; see p. 209, fn. 6.
3 Notice that this is the only type of -ya-stems, where the root syllable is short, i.e. has a short vowel and is open.
Abbr
AB Aitareya-Brähmana
AV(S) Atharvaveda (Saunakiya recension)
BÄU(K) Brhad-Äranyaka-Upanisad (Känva
recension)
BÄUM Brhad-Äranyaka-Upanisad,
Mädhyandina recension
Br. Brähmana(s)
ChUp Chändogya-Upanisad
DhP Dhätupätha
KathÄ Katha-Äranyaka
KätySS Kätyäyana-Srauta-Sütra
KS Käthaka(-Samhitä)
[X]m mantra part of the text [X]
MänGS Mänava-Grhya-Sütra
ManuSmr. Manu-Smrti (= Mänava-Dharma-Sästra)
MBh. Mahä-Bhärata
MS Maiträyani Samhitä
iations
[X]p prose part of the text [X]
Pan. Panini (Astadhyayi)
PB Pancavimsa-Brahmana (= Tandyamaha-
Brahmana) RV Rgveda
RVKh. Rgveda-Khilani Sh. Gil. Shina, Gilgit dialect SBK Satapatha-Brahmana, Kanva recension
SB(M) Satapatha-Brahmana (Madhyandina
recension) TA Taittiriya-Aranyaka
TB Taittiriya-Brahmana
TS Taittiriya-Samhita
VS(M) Vajasaneyi-Samhita (Madhyandina recension)
YV Yajurveda(-Samhita) (= VS(K), MS, KS,
KpS, TS)
