Genitive Case and Possessive Adjective in Anatolian
Yüklə 161,56 Kb. Pdf görüntüsü
|
| Genitive Case and Possessive Adjective in Anatolian
H. Craig Melchert University of North Carolina at Chapel Hill
1. Introduction. One of the most famous features of the Anatolian Indo-European languages is the widespread use of an inflected adjective (that agrees with its head noun in gender, number and case) in place of the genitive case. In some languages use of this adjective leads to total or near total loss of the genitive. Less well-known is the variety of exponents of the genitive case itself in Anatolian. The following reexamination of this topic was directly stimulated by the article of Hajnal (2000). His analysis has led me to revise radically my own previous views, but in a very different direction from the one he proposes. I should make clear at the outset one fundamental difference in viewpoint. Hajnal argues (2000: 174ff.) that there was a consistent functional difference in PIE between the “individualizing genitive” and the “specifying” possessive adjective and (2000: 179ff.) that this difference is still discernible in the pattern of their use in Lycian and HLuvian. I cannot accept this claim for Lycian, where there is no functional difference between leθθi qlã “precinct of Leto” (adj.) and wazzije kbatra “daughter of Wazzije” (gen.), nor (pace Hajnal) between tebeija “of Tibe” (adj.) and Hlah “of Hla” (gen.), each of which respectively identifies the owner of the tomb. On the mixed syntax of possessive adjective and genitive
2 in HLuvian see note 12 below. This artificial distinction has led Hajnal to what are in my view very implausible historical derivations of some of the genitival endings. That the various denominative adjectives involved originally meant merely ‘pertaining to’ or ‘having the quality of’ the base noun is not in question. However, the crucial innovation of the western Anatolian languages (all those except Hittite and Palaic) was precisely to use these adjectives interchangeably with and in some cases in place of the genitive case (thus with Mittelberger 1966: 99, 101 and 103). 1
2. Genitive and Possessive Adjective in Anatolian—Synchrony. I must first briefly set forth the relevant data for the Anatolian languages. Old Hittite shows a genitive singular ending -aš and a genitive plural ending -an. 2
1 It is far from clear that the contrast Hajnal claims for the genitive and possessive adjective is valid even for PIE. See the very different characterizations of the genitive cited by Neumann (2001: 448). I thank Norbert Oettinger for this reference. 2 Despite false claims to the contrary (e.g. Gamkrelidze and Ivanov 1995: 233ff. with wholly unjustified further implications), there are no assured examples of -an in Hittite with a singular meaning! Most examples must be, and all can be, interpreted as plurals (see Laroche 1965: 33-40). The exclusively plural function is confirmed by the Anatolian cognates cited below. In later Hittite the genitive plural ending in -an is replaced by -aš, which may reflect syncretism either with the genitive singular or the dative-locative plural. The latter is far more likely, since there is other evidence for case syncretism in Hittite, but none for a loss of contrast between singular and plural. 3 Hittite does not make productive use of an adjective to express possession. 3 Palaic
attests a genitive singular in -aš cognate with the Hittite ending and a few possessive adjectives in -aša/i-, such as d
4
Cuneiform Luvian (CLuvian) has entirely replaced the genitive case with a possessive adjective in -ašša/i- (for the inflection with so-called ‘i-mutation’ see Starke 1990: 54ff.). Confirmation that the absence of the genitive is systematic and not due to chance comes from the fact that CLuvian alone developed a variant -anzašša/i- to mark plurality of the possessor (a distinction that otherwise cannot be expressed by the possessive adjective): see Melchert (2000: 173ff.). Hieroglyphic Luvian (HLuvian) has a genitive singular ending /-as/ spelled -Ca-
spelled -Ci-(i)-sa are also probably to be analyzed as /-is/ contracted from /-iyas/. HLuvian also makes wide use of the possessive adjective in /-assa/i-/ seen in CLuvian, as well as one in /-i(ya)-/. Finally, HLuvian also has examples of possessive forms spelled -Ca-si-(i) that do not show agreement with their head noun (e.g. pa-si-i-´ a-ta 5 -ma-za “his name” in ADIYAMAN 1, §8). While it is
3 The suffix -ašša/i- appears only in a few lexicalized substantives, such as genušša/i- “knee(- joint)” and d
4 Contra Hajnal (2000: 165) this stem is attested as an adjective with the meaning given, not as a noun meaning “cake of Z.”. The example cited from KUB 35.165 Vs 7 does not exist: here read d
a š=an=pa=ti takkuwāti “Zaparfa accepts it for himself”. 4 conceivable that these spellings represent the ordinary genitive singular in /-as/, it is more likely that we must assume a genuine alternate ending /-asi/, as first seen by Mittelberger (1966: 100). 5
among the Anatolian languages. First of all, for most appellatives it employs an adjective in -a/ehe/i- (appearing in Milyan/Lycian B as -a/ese/i-) cognate with Luvian /-assa/i-/. 6 Some personal names appear with a zero ending (e.g. Epñxuxa in TL 127,1), first identified by Neumann (1970: 61), who argues correctly that these examples should not be emended out of existence by adding an -h! Personal names and place names do attest a genitive ending -Vhe or -Vh. 7 Occasional accusative singulars in -Vhñ that precede(!) their modified noun are also merely secondarily inflected examples of the genitive ending -Vh(e), as per Hajnal (1995:
5 Unfortunately, HLuvian orthography does not permit us to tell whether any examples of genitives spelled -Ca-sa and -Ci-sa represent yet another ending /-sa/ matching Lycian -he (thus Bader 1991: 138ff.), since they may all be interpreted merely as /-as/ and /-is/. 6 There is also an isolated example of the possessive use of -i(je)-, matching HLuvian /-i(ya)-/: TL 100 ebe xupa me tibeija “This tomb (is that) of Tibe”. Pace Hajnal (2000: 180) there is no basis whatsoever for viewing the function of the adjective here as any different from that of the genitive in the same use (e.g. TL 129 Hlah): as always, the inscription names the owner of the tomb! 7 In appellatives this ending appears only in terihe “(the one) of three” = “third”. Milyan also shows a cognate ending -Vse (e.g. Kuprllese), with a few examples of secondarily inflected acc. sg. -zñ (Wizttasppazñ). I know of no examples of an apocopated variant matching Lycian -Vh. 5 197ff.), following Mittelberger (1966: 104) and Adiego (1994: 18), against Melchert (1994: 324 et aliter). Finally, Lycian also has a genitive plural in -ẽ cognate with Old Hittite -an. Lydian apparently attests a handful of cases of the ending -aν functioning synchronically as a genitive plural (see Gusmani 1964: 130 & 202). For the most part, however, this ending has taken on the function of a dative-locative plural. For possession Lydian otherwise employs an adjective in -Vl(i)- (on its inflection, once again with the ‘i-mutation’, see Starke 1990: 85). Possession is regularly marked in Carian by an ending -ś, where the exact nature of the sibilant remains open (see below). We also have a few cases of an inflected anim. acc. sg. such as pñmnnśñ (cf. Greek gen. sg. Πονµοοννου). In at least some cases the Carian ending -s (probably ordinary /s/, as per Melchert 2002: 307 with refs.) also appears to mark possession (see Adiego 1998: 18-19), although for the most part it indicates a recipient or beneficiary and thus functions synchronically as a dative (see Schürr 1996: 66 and apud Adiego 1998: 19). The southwestern Anatolian languages Pisidian and Sidetic, of very late and limited attestation, mark possession by means of a sibilant ending. For Pisidian -s, spelled with Greek sigma, see Brixhe (1988) and for Sidetic -z Nollé (2001: 632). 3. Genitive and Possessive Adjective in Anatolian—Diachrony The prehistoric source of the genitive singular ending /-as/ of Hittite, Palaic and HLuvian and of the zero ending of Lycian personal names is straightforward: 6 PIE *-os (for the Lycian thus also Adiego 1998: 13). 8 Likewise, the genitive plural seen in Old Hittite -an, Lycian -ẽ and Lydian -aν (the last mostly shifted to dat.-loc. plural) reflects PIE gen. pl. *-ōÃm. As per Adiego (1994: 14ff.), the usually uninflected genitive ending of Lycian in -Vh(e), which is notably restricted to personal names, is best derived from the PIE thematic genitive ending *-(o)so (my earlier objections were ill- founded). The recessive pattern of the ending’s use is prima facie evidence for an archaism (contra Hajnal 2000: 177). The argument against this derivation by Hajnal (2000: 178, note 48) is entirely circular. Having declared (without argument) that the ending -Vh(e) must be an innovation, he then says that there is no evidence in Anatolian for the *-(o)so ending! 9 We may dispense with the 8 The PIE gen. sg. ending *-s is preserved only in the relic nekuz < *nek w t-s “of evening” in the set phrase nekuz me æur. 9 His further argument that *-(o)so is attested elsewhere in Indo-European only as a pronominal ending is, of course, falsified by Germanic languages (e.g. Runic -as, ON -s, OE -æs in o-stem nouns). The analogical spread of the ending from the o-stems to other stem classes is trivial. The shorter variant in -Vh does present a genuine problem, since there is no regular apocope of unaccented short *-o in Lycian (see Melchert 1994 323f.). Nevertheless, apocope (as per Mittelberger 1966: 104) remains the likely explanation. I explicitly withdraw my account (1994: 325), and that of Hajnal (1995: 200ff.) is suspiciously complex. Given other evidence for “phrasal stress” in Anatolian (see Melchert 1998, esp. 484 with reference to Durnford), the apocope here
7 scenario of Hajnal (2000: 177f.), whereby Lycian took a single inherited ending *-os, which it kept while at the same time creating from it two new endings, for some reason restricting all of them only to personal and place names, at a time when it was adopting the inflected adjective in *-eh 2 so/i- as its productive marker of possession. As per Melchert (2002: 309), Carian -s may be equated with Lycian -h(e) and also derived from *-oso. As indicated in note 5 above, we cannot affirm whether HLuvian has endings /-asa/ and /-isa/ also reflecting *-(o)so. 10
The clear evidence for limited secondary inflection of reflexes of the *-oso ending in Lycian also supports the account of Bader (1991: 99) of the Palaic inflected adjectives in -aša/i-: they merely reflect the same process taken to its conclusion of a fully developed inflection. For reasons to reject the alternate account of Hajnal (2000: 166) see the discussion below on the adjective suffix /-assa/i-/. The presence of PIE genitive ending *-oso in Anatolian raises the possibility that the HLuvian ending /-asi/ is a reflex of the PIE thematic genitive ending
may have arisen in the cases of attributive use of the genitive, where the noun in the genitive was weakly accented or unaccented in the syntagm of noun (gen.) + noun. 10 Likewise it is impossible to determine whether the sibilant endings of Pisidian and Sidetic, which appear to be functioning as genitive endings synchronically, continue the *-oso ending or the /-assa/i-/ adjective, reanalyzed due to loss of inflection. Melchert (1994: 44) and Hajnal (2000: 182) assume the latter, but Brixhe (1988: 52, note 29) and Nollé (2001: 632) prudently entertain both possibilities. 8 *-osyo, with apocope. 11 Lack of parallels makes it impossible to confirm or refute this derivation. Evidence that the -ś of the Carian possessive marker has a palatal quality (see Melchert 2002: 310 with refs.) allows us to entertain the same source for that ending (the limited inflection seen in pñmnnśñ may be secondary), although a reflex of the originally inflected adjective in /-assa/i-/ cannot be excluded (cf. Melchert 2002: 311f.). We may derive the HLuvian possessive adjectives in /-i(ya)-/ without difficulty from PIE *-iyo-. For the remarkable HLuvian syntax that combines these adjectives with nominal genitives see Melchert (1990: 202ff.). 12 The Lydian possessive adjective in -l(i)- reflects PIE denominative *-(o)lo- (well attested in
11 The assumption of a ‘reinforcing’ -i to a genitive in *-Vs (Bader 1991: 139 and Hajnal 2000: 178) is entirely unmotivated. 12 The attempt of Hajnal (2000: 179ff.) to explain this usage in terms of “conjunction reduction” is ingenious, but cannot account for the overall pattern of use in HLuvian, where there is no prohibition against having a possessor and its apposition both appear in the genitive (see e.g. the openings of BOHÇA, BULGARMADEN, BOYBEYPINARI 1&2, etc.). Nor does it account for the usage in the long genealogy of MARAŞ 1, where personal names appear in the form of /-assa/i-/ adjectives, while their appositions are in the genitive. Nor is the HLuvian usage of /-iya-/ adjectives as attested inherited, as claimed by Hajnal. What is remarkable about the HLuvian construction is not that a possessive adjective may occur conjoined with or in apposition to a nominal genitive, but rather that a genitive may depend on a possessive adjective in a ‘nested’ construction, something not shown in the parallels he cites.
9 Hittite as a derivational suffix). The prehistory of the suffix /-assa/i-/ is more problematic. 13 Two facts, however, are paramount. First, in the two languages where the orthography may show it (Hittite and CLuvian) this suffix has consistently a geminate /-ss-/. Second, in Lycian it is clear that the first vowel of the suffix is inherently /-a-/, and that all cases of /-e-/ may be due to the Lycian umlaut rule whereby /a/ becomes /e/ before a front vowel in a following syllable (see Meriggi 1980: 217). The only viable preform for the /-assa/i-/ suffix is *-eh
2 so- (so also Hajnal 2000: 164; cf. Bader 1991: 137). There is simply no positive evidence for the existence of a variant *-os(s)o- in Luvian or Lycian. Neither of the arguments presented by Hajnal (2000: 167ff.) is remotely compelling. First, the occasional “plene” spelling of the first vowel of the suffix in CLuvian (e.g. tar-ma-a-aš-ša/i-) may reflect merely that accented short *ă was regularly lengthened in CLuvian in closed syllables. There is no counterevidence to this assumption, and it is supported by the fact that a < accented *ĕ is lengthened in this position (Melchert 1994: 263f.). Second, his claim that consistent -ehe/i- with e-vocalism in Lycian always reflects *-osso- to o- or e- stems is refuted by the cited examples in -ẽñnehe/i- to stems in -ẽñn(i)- (sic!), since the latter reflect athematic stems in *-wen-. In any case the alleged ‘resistance’ of -ahe/i- to umlaut is a very weak foundation on which to posit an
13 For an excellent treatment of the syntax of this suffix in cases where it is used for a series of ‘nested’ genitives see Neumann (1982). 10 otherwise unsupported preform *-osso-. 14 As per Hajnal (2000: 174, note 41), it is hard to resist the oft-cited comparison of Anatolian /-assa/i-/ with Latin -ārius, although we must now make explicit that the Anatolian forms with i-vocalism reflect the so-called ‘i-mutation’ and have nothing directly to do with the -i- of Latin -ārius. I forgo here any further analysis of the preform *-eh 2 so-.
In sum, the western Anatolian languages do largely or wholly replace the inherited genitive with a possessive adjective, but the Anatolian languages as a whole preserve not only the PIE athematic singular and plural endings *-os and *-ōÃm, but also the thematic ending *-oso and probably also *-osyo. References Adiego, Ignacio-Javier. 1994. “Genitiu singular en lici i protoluvi”. Anuari di Filologia 17:11-23. Adiego, Ignacio-Javier. 1998. “La nueva bilinguë greco-caria de Cauno y el desciframiento del cario”. Kadmos 34:18-34. Bader, Françoise. 1991. “Problématique du genitif thématique sigmatique”. Bulletin de la Société Linguistique de Paris 86/1:89-157. Brixhe, Claude. 1988. “La langue des inscriptions épichoriques de Pisidie”. A Linguistic Happening in Memory of Ben Schwarz, ed. Yoël Arbeitman, 131- 155. Peeters: Louvain-la-Neuve.
14 Almost all of the cited cases of ‘consistent’ -ehe/i- are in fact attested only once or twice each. their ‘consistency’ may therefore be a mirage. 11 Gamkrelidze, Thomas & Vyačeslav Ivanov. 1995. Indo-European and the Indo- Europeans (= Trends in Linguistics. Studies and Monographs, 80). Berlin & New York: Mouton de Gruyter. (English version by Johanna Nichols) Gusmani, Roberto. 1964. Lydisches Wörterbuch. Heidelberg: Winter. Hajnal, Ivo. 1995. Der lykische Vokalismus. Graz: Leykam. Hajnal, Ivo. 2000. “Der adjektivische Genitivausdruck der luwischen Sprachen (im Lichte neuerer Erkenntnis)”. 125 Jahre Indogermanistik Graz, ed. by Michaela Ofitsch & Christian Zinko, 159-184. Graz: Leykam. Laroche, Emmanuel. 1965. “Études de linguistique anatolienne”. Revue hittite et asianique 23:33-54. Melchert, H. Craig. 1990. “Adjectives in *-iyo- in Anatolian”. Historische Sprachforschung 103:198-207. Melchert, H. Craig. 1994. Anatolian Historical Phonology. Amsterdam & Atlanta: Rodopi. Melchert, H. Craig. 1998. “Poetic Meter and Phrasal Stress in Hittite”. Mír Curad. Studies in Honor of Calvert Watkins, ed. by Jay Jasanoff et al., 483-493. Innsbruck: Institut für Sprachwissenschaft der Universität Innsbruck. Melchert, H. Craig. 2000. “Aspects of Cuneiform Luvian Nominal Inflection”.
12 Melchert, H. Craig. 2002. “Sibilants in Carian”. Novalis Indogermanica. Festschrift für Günter Neumann zum 80. Geburtstag, ed. by Matthias Fritz & Susanne Zeilfelder, 305-313. Graz: Leykam. Meriggi, Piero. 1980. “La declinazione dei nomi propri e dei pronomi in licio”.
Mittelberger, Hermann. 1966. “Genitiv und Adjektiv in den altanatolischen Sprachen”. Kratylos 11:99-106. Neumann, Günter. 1970. “Beiträge zum Lykischen IV”. Sprache 16:54-62. Neumann, Günter. 1982. “Die Konstruktionen mit Adiectiva genetivalia in den luwischen Sprachen”. Investigationes Indogermanicae. Gedenkschrift Heinz Kronasser, ed. by Erich Neu, 149-161. Wiesbaden: Harrassowitz. Neumann, Günter. 2001. “Der adverbale Genetiv im Althethitischen”. Akten des IV. Internationalen Kongresses für Hethitologie Würzburg, 4.-8. Oktober 1999, ed. by Gernot Wilhelm, 446-455. Wiesbaden: Harrassowitz. Nollé, Johannes. 2001. Side im Altertum: Geschichte und Zeugnisse. Band II. Bonn: Habelt. Schürr, Diether. 1996. “Bastet-Namen in karischen Inschriften Ägyptens”. Kadmos 35:55-71. Starke, Frank. 1990. Untersuchungen zur Stammbildung des keilschrift-luwischen Nomens (= Studien zu den Boğazköy-Texten, 31). Wiesbaden: Harrassowitz. Yüklə 161,56 Kb. Dostları ilə paylaş:
|