Ronald I. Kim Greco-Armenian The persistence of a myth Abstract: It has been generally held since the beginning of the 20th century that Armenian is more closely related to Greek than to any other Indo-European branch. A more recent minority opinion posits an especially close relationship between Greek and Armenian, even going so far as to assume a period of Greco-Armenian unity. Following upon recent publications, above all Clackson 1994, this paper argues that the available evidence does not at all support this stronger hypothesis. In contrast to the lexical innovations common to Greek and Armenian, the phonological isoglosses shared by the two languages are extremely few and of an easily repeatable nature. The morphological features claimed as shared innovations may likewise represent independent developments and/or have parallels in other Indo-European branches, whereas other features of verbal morphology rather appear to connect Armenian with Indo-Iranian or Balto-Slavic. These considerations suggest that pre-Armenian belonged to a dialect continuum encompassing the ancestors of Greek, Phrygian, and Indo-Iranian for some time after the breakup of Proto-Indo-European, but made up a distinct speech community already by the late 3rd millennium BC. Keywords: Armenian, Greek, Indo-European, subgrouping, isogloss, relative chronology 1 Armenian and Greek: more than just neighbors? The hypothesis that Armenian is more closely related to Greek than to any other Indo-European (IE) branch goes back to the beginnings of the 20th century. In a lengthy article devoted to the relationship of Armenian with other IE languages, Pedersen (1906: 442; cf. 1924: 225) stated that “… eine betrachtung des schon früher gesicherten etymologischen materials zugleich mit den im vorliegenden aufsatz gewonnenen resultaten mich zu der ansicht geführt hat, dass das Armenische keiner anderen idg. sprache so nahe steht wie dem Griechischen.” Meillet (1925; 1927; 1936: 142–143) investigated lexical, phonological, and morphological isoglosses linking the two languages and similarly concluded that “ces deux Ronald I. Kim, Adam Mickiewicz University, Poznań/Charles University, Prague; ronald.kim@yahoo.com DOI 10.1515/if-2018-0009 Brought to you by | The University of British Columbia Library Authenticated Download Date | 9/24/18 2:51 PM 248 Ronald I. Kim groupes proviennent de parlers qui, dans le domaine indo-européen, devaient être voisins les uns des autres” (1925: 1). This view was shared by Bonfante (1937), who listed 24 phonological isoglosses linking Greek and Armenian (see Section 2 below), as well as Solta (1960: 462–466, 483; 1963: 125–126; 1990: 14–17), who confirmed Pedersen and Meillet’s hypothesis of a close relationship between Greek and Armenian based on both morphological and lexical isoglosses.1 Both Solta and Djahukian (1980) concluded that Armenian shares the most lexical isoglosses with Greek, with Indo-Aryan following in second place; Djahukian (1982: 201–205) arrived at the same finding on the basis of phonological and morphological isoglosses. Wyatt (1982: 40) likewise determined from his comparison of lexico-semantic fields that Greek and Armenian shared “at least pastoral items and perhaps more particularly, herding words having to do primarily with sheep and goats.” Schmidt (1980: 38; cf. 1985: 214–217; 1987: 39; 1988: 601) saw Armenian as an “eastern IE language” sharing isoglosses above all with Greek, Phrygian, and Indo-Iranian (see below), but also regarded “the particularly close connections between Armenian and Greek … as proven.” Today, it is generally accepted that Armenian shares areal affiliations with Greek, Phrygian, and Indo-Iranian, languages which agree in having the augment and, aside from Indo-Iranian, the so-called prothetic vowel.2 Such a dialectal position not only makes geographical sense, but also accords with the scant testimony of ancient Greek historians on the origins of the Armenians and Phrygians, most famously Herodotos’ description of the Armenians as Φρυγῶν ἄποικοι (VII, 73). Watkins (2001: 57) viewed these languages as belonging to a large subgroup “on the basis of shared grammatical features, like the ‘augment’-prefix, the prohibitive negation, and the whole structural organization of the verbal system,” adding that “[t]his group forms the basis on which the proto-language was first reconstructed, 1 Cf. also Greppin 1981: 124–125 (with reservations). 2 See Schwyzer 1939: 56–57; Godel 1975: 132; Schmidt 1980: 38ff.; Bonfante 1982: 163–166; de Lamberterie 1994: 148–149; 2011: 367–368; Morani 2001: 418–423; Martirosyan 2013: 86–87 and passim; cf. also Porzig 1954: 155–157, 161–164 for lexical isoglosses connecting Armenian with Greek and Indo-Iranian, respectively. A variant of this hypothesis holds that Greek and Armenian belong to a “Balkan Indo-European” subgroup also including Phrygian and Albanian: cf. Solta 1990: 17; Klingenschmitt 1994: 244–245; Ritter & Sowa 2004; Matzinger 2005a: 380–386; 2012 and in particular Holst 2009: 49–98 and de Lamberterie 2012; 2013. For possible lexical and phonological isoglosses shared by Armenian and Albanian, see Kortlandt 1980a; 1986. On the other hand, the old proposal of an especially close connection between Armenian and Phrygian cannot be maintained; see Matzinger 2005a: 377–380; de Lamberterie 2013: 25–28 for discussion and references. Recent suggestions of shared Armenian-Celtic isoglosses are weak at best and confined entirely to the lexical level: see Schmidt 2007; 2012 and especially Falileyev & Kocharov 2012. Brought to you by | The University of British Columbia Library Authenticated Download Date | 9/24/18 2:51 PM Greco-Armenian 249 and it is probably the most recent in time of the various ‘branches’ or subgroups of the family.”3 Over the past few decades, however, there has emerged a tendency to interpret Pedersen’s original hypothesis in terms of an especially close mutual relationship between Armenian and Greek: thus already Bonfante (1937: 33) wrote that “le grec et l’arménien forment parmi les langues indo-européennes un groupe extrêmement serré.” In its strongest form, this version posits a “Greco-Armenian” branch of the IE family, comparable to e.g. Indo-Iranian or Balto-Slavic. Hamp (1976: 91) went so far as to declare that the time is not far off “when we should speak of Helleno-Armenian,” and in the following years he repeatedly referred to “Helleno-Armenian” innovations (see e.g. Hamp 1979: 4–5; 1983a; 1983b; 1989; 1992: 58). Greppin (1982) explicitly stated that “proto-Greek and proto-Armenian were indeed one and the same language at a time somewhere before the second millennium” (347), i.e. “proto-Greek and proto-Armenian must be directly taken from the same immediate source” (352). On a more modest level, de Lamberterie (1992: 239) declared the possibility of reconstructing “de véritables mots grécoarméniens d’une certaine ampleur.” Several computational models based mainly on lexical isoglosses have also posited a Greco-Armenian node in the IE family tree, including Ringe, Warnow & Taylor 2002 (but see below); Gray & Atkinson 2003: 437, Figure 1; and Rexová, Frynta & Zrzavý 2003. As a result, scholars of IE historical linguistics often operate under the tacit assumption that Armenian must have undergone the same developments observable in Greek, and even posit Proto-Greco-Armenian preforms of apparently cognate formations. Occasionally this assumption is made explicit: thus Barton (1990–1991: 45) defended his derivation of denominal stative a-presents (e.g. mnam ‘remain, am left’) from *-h₁-ye- ← PIE *-eh₁- with the assertion that “[t]he extremely close dialectgeographical connection between Greek and Armenian makes it highly probable that innovations of early pre-Greek were shared by pre-Armenian.” Similarly, Klingenschmitt (1982: 136–137; cf. Peters 1997: 210) argued against a comparison of the Armenian weak aorist in -(e)acʿ with ā-preterites in other IE languages, specifically the Balto-Slavic preterite in *-ā- and the Italic imperfect *-ā- of Latin er-ā-, -bā- and Oscan pl. fufans, “wegen des Fehlens entsprechender Bildungen in archaischen Sprachen wie dem Indoiranischen und Griechischen.”4 Most recently, Jasanoff (forthc. §12), in discussing the origins of the reduplicated (thematic) aorist, speaks of “the typical Greco-Armenian introduction of ‘intensive’ reduplication … before 3 Interestingly, the pioneering lexicostatistical studies of the relationships among the IE branches by Czekanowski (1927: 50–57; 1928: 46–52; cf. Mańczak 1987: 19) found that Armenian is most closely related to Iranian and Indo-Aryan. 4 On the origin of the Armenian weak aorist, see Section 4 no. 4. Brought to you by | The University of British Columbia Library Authenticated Download Date | 9/24/18 2:51 PM 250 Ronald I. Kim the zero grade of roots beginning with a laryngeal,” even though the lone Armenian example is arar- ‘made’, corresponding to Gr. ἀραρε/ο- ‘joined, fit together’ < PIE *h₂e-h₂r-e/o-.5 But is the tacit assumption of a period of Greco-Armenian unity justified, and with it the methodological practice of looking to Greek in the first place to explain unsolved problems of Armenian historical grammar? It seems to me that much recent scholarship on Armenian historical-comparative linguistics has confused or conflated two quite distinct hypotheses: 1. The prehistoric dialects ancestral to Armenian and to Greek were in close enough geographical proximity and contact that they participated in a number of innovations. In many cases, these isoglosses are shared with one or more neighboring dialects, e.g. those ancestral to Phrygian or the even more poorly attested IE languages of the ancient Balkans, or others which died out without leaving any records of their existence. 2. The ancestors of Armenian and Greek formed a single speech community (naturally with internal variation) for some time after the breakup of PIE. During that period, they underwent a set of exclusive innovations not shared with other IE languages. As a result, it is possible to reconstruct Proto-Greco-Armenian preforms for linguistic material inherited by both Greek and Armenian, just as one may reconstruct Proto-Indo-Iranian ancestors of cognate forms in IndoAryan and Iranian, or Proto-Balto-Slavic ancestors of cognate forms in Baltic and Slavic. Note that even if hypothesis 2 is valid, it does not follow that one can automatically assume the same prehistoric developments for Armenian as for Greek. By way of comparison, most scholars today consider Balto-Slavic a valid subgroup of IndoEuropean, yet the oldest attested stages of Slavic, Old Prussian, and the East Baltic languages have gone their own way in multiple points, from phonology to verbal morphology to the lexicon.6 These two hypotheses, which are naturally to be understood as lying on a spectrum of possible historical relations, require very different thresholds of evidence. For hypothesis 1, discovery of a sufficient number of lexical isoglosses, i.e. lexical items specific to Greek and Armenian, would by itself suggest a neigh5 This tendency can even subliminally influence those arguing against it, as e.g. when Viredaz (2015: §6.2.1), in a detailed elaboration of the parallel partial sigmatization of PIE root aorists in Armenian and Slavic, accidentally refers to “innovations gréco-arméniennes.” 6 The case of the Italic languages is especially instructive in this respect: despite forming a single branch, Latino-Faliscan and the various Sabellic languages have diverged markedly in e.g. their verbal system, most notoriously in the formation of the weak perfect. Brought to you by | The University of British Columbia Library Authenticated Download Date | 9/24/18 2:51 PM Greco-Armenian 251 boring relationship between the two groups of speakers in prehistoric times, a hypothesis which in turn could be strengthened by the identification of distinctive nonlexical features. In contrast, hypothesis 2 requires a list of exclusive shared innovations which are unlikely to have arisen independently.7 These innovations may be phonological, morphological, or lexical, but since many phonological changes are phonetically natural and therefore easily repeatable, while lexical items are easily diffused across already distinct speech communities, morphological isoglosses assume paramount importance. Lexical isoglosses should ideally entail not just root cognation, but word equations; it should therefore be possible to identify common innovations in word formation, i.e. in derivational as well as inflectional morphology.8 Does the available evidence justify this kind of prehistoric relationship for Armenian and Greek, i.e. the strong version of the Greco-Armenian hypothesis? In fact, two recent studies have argued that it does not. Clackson (1994), in by far the most detailed examination to date of the linguistic relationship between Greek and Armenian, asserts that “morphological developments … revealed little evidence in favour of the sub-group hypothesis,” and that “[c]lose scrutiny of the phonological and lexical agreements between Greek and Armenian … did not provide sufficient evidence to refute [this] conclusion” (199), though he does not exclude the possibility that the two languages descend from neighboring dialects of PIE (201). Given the number of lexical isoglosses shared by Armenian and Indo-Iranian, he concludes that “there is not sufficient evidence to suppose any closer link between Greek and Armenian than between either language and Indo-Iranian, and the reconstruction of a Greek-Armenian-Indo-Iranian dialect area is sufficient to account for these agreements” (202; cf. also Clackson 2004: 124). Following Clackson, Beekes (2003: 152–153) likewise was of the opinion that “a longer period of close contact can neither be demonstrated nor excluded at the present stage.” Morani (2001: 424–426) also emphasizes the primacy of isoglosses linking Armenian to both Indo-Iranian and Greek, and considers the mainly lexical innovations specific to Greek and Armenian to be from a more recent period. From a computational perspective, Ringe, Warnow & Taylor (2002) tentatively favor a Greco-Armenian subgroup on the basis of their cladistic models, but concede that “the evidence … is significantly poorer than for Italo-Celtic” (102), con7 So already Leskien 1876: vii. 8 On the pivotal role of morphological innovations in demonstrating subgrouping relationships, see among others Ringe & Eska 2013: 256–263; Ringe 2017b: 68 and passim. The limited value of lexical isoglosses for determining the linguistic position of Armenian was pointed out by Schmitt (1962: 151; 1972 [1974]: 34); cf. also the critical remarks of Makaev (1967: esp. 463–464) on the criteria for isoglosses adopted by Porzig (1954) and Solta (1960). Brought to you by | The University of British Columbia Library Authenticated Download Date | 9/24/18 2:51 PM 252 Ronald I. Kim sisting of the six lexical characters ‘day (= 24 hours)’, ‘husband’, ‘not’, ‘wind’, ‘grind’, and ‘young’. Furthermore, “the absence of any phonological or inflectional character makes it qualitatively poorer” than for Italo-Celtic (103), leading them to pronounce Greco-Armenian “by far the weakest of the larger subgroups our methodology has found” (106). Of the six lexical items, the authors note that parallel development can be excluded only for ‘day’ and ‘not’, and even the latter isogloss has been called into question, given the phonetic difficulties with the classic etymology of Gr. οὐκ(ί), Arm. očʿ < PIE *(ne) h₂óyu kʷíd (Cowgill 1960; cf. Clackson 1994: 158).9 That leaves only Gr. ἦμαρ ∼ Arm. awr ‘day’ as a (near-)word equation10 — hardly sufficient grounds for setting up a subgroup comparable in robustness to Italo-Celtic, let along Balto-Slavic or Indo-Iranian. The picture is not appreciably altered even if one includes word equations without secure IE etymologies or derivatives of PIE roots found exclusively in Greek and Armenian, e.g. the oft cited Gr. κίων ∼ Arm. siwn ‘pillar’ or Gr. ἄλευρον ∼ Arm. alewr ‘flour’.11 My intention in the following pages is not to challenge hypothesis 1, which I consider highly probable, but rather to argue that the phonological and morphological facts simply do not support hypothesis 2. Although Clackson has already argued as much in his exemplary 1994 study (see above), I will build on his arguments by considering additional forms of evidence. In particular, the relative chronology of sound changes suggests that pre-Armenian and pre-Greek were sufficiently divergent at an early stage that they could no longer be considered as belonging to a common speech community (Section 2). With respect to morphology, a reexamination of the data confirms Clackson’s view that those features apparently shared with Greek may reflect parallel developments (Section 3). On the basis of morphological features common to other neighboring IE branches, namely Indo-Iranian and Balto-Slavic (Section 4), I conclude that the IE dialect(s) ancestral to Armenian went their own way already from a very early stage (Section 5); 9 Clackson has since proposed that očʿ originally meant ‘no one’ and goes back to o- (cf. o-kʿ ‘anyone’, o-mn ‘someone’) + negator *čʿ (cf. čʿ-ikʿ ‘nothing’). See Clackson 2005: 155–156, followed by Martirosyan 2010: 531; Kim 2016: 45. 10 Arm. awr is traditionally derived from pre-Arm. *āmōr, standing in the same relationship to pre-Gr. *āmr̥ as τέκμωρ to τέκμαρ ‘mark, goal’ (Meillet 1903: 32; 1936: 55; Schmitt 1981: 166); note also the parallel of awr awowr ‘day by day’ and Myc. a-mo-ra-ma /amōr-amar/ (Dressler 1969). However, until the exact conditions for the lenition of *m > *w are determined, it cannot be excluded that awr rather continues pre-Arm. *āmr̥ (Hamp 1983b: 7–8). 11 See respectively the discussions in Clackson 1994: 140–143, 90–95 (also on ἀλέω ∼ ałam ‘grind’, ἀλετρίς ∼ aławri ‘woman who grinds corn’). Brought to you by | The University of British Columbia Library Authenticated Download Date | 9/24/18 2:51 PM Greco-Armenian 253 consequently, the language exhibits a unique combination of retained (late) PIE morphology and a distinctive set of defining innovations.12 2 Evaluating Greco-Armenian: phonological isoglosses The most famous phonological innovations shared by Greek and Armenian concern the treatment of PIE laryngeals, above all the vocalization of word-initial preconsonantal laryngeals (also in Phrygian, e.g. Gr. ἀνήρ, Phryg. αναρ, Arm. ayr ‘man’ < PIE *h₂nḗr ) and treatment of (at least word-final) PIE sequences *iH, *uH as *yV, *wV. However, there is continuing disagreement as to the exact reflexes of laryngeals in these positions in Armenian. Word-initially and -medially, most scholars today believe that Armenian does not share the “triple representation of schwa” with Greek (and probably Phrygian), and that the only lautgesetzlich development is to *a.13 Word-finally, it is generally assumed that PIE *-ih₂ > pre-Arm. *-ya, e.g. in sterǰ ‘sterile’ < PIE *ster-ih₂- (Ved. starı̄-́ , Gr. στεῖρα), but PIE *-ih₁ may also have given *-ya in duals such as PIE *h₃kʷ-ih₁ > *akʷya (*akʷye?) > *ačʰ → pl. ačʿkʿ ‘eyes’; here as well, then, there would be no counterpart of the Greek “triple representation” famously represented by Hom. ὄσσε ‘eyes’ < *okʷye < PIE *h₃kʷ-ih₁ vs. e.g. πότνια ‘mistress’ < PIE *pótn-ih₂.14 Determining other conditioned laryngeal developments, e.g. of *CRHC, is made even more difficult by the small number of relevant forms with clear morphological prehistories, which furthermore show 12 I should add that I am unconvinced by arguments for a closer relationship between Armenian and certain Greek dialects, as suggested by Ritter & Sowa (2004) and argued by Sowa (2006) on the basis of Arm. awjikʿ ‘collar’, Gr. αὐχήν ‘neck’ < post-PIE *angʷʰ-io-, *angʷʰ-en- beside Aeol. ἄμφην. The resemblance of the Armenian and dialectal Greek forms may be explained through borrowing, from pre-Armenian into pre-Greek or perhaps from a third language into both (Clackson 1994: 107–109; see Kim forthc. fn. 14 and the references listed there). 13 See Klingenschmitt 1970: 80, 86 no. 9; 1982: 105–106 fn. 27; Olsen 1985; 1989a; 1999: 762–764, 767–769; Clackson 1994: 33–36. For dissenting views, see Kortlandt 1987; Beekes 1988: 76–77; 2003: 185–187, 192–193 (*h₁/ ₂/ ₃C- > e/a/oC- but *CHC > CaC); Greppin 1989. 14 Other examples may reflect a typical “non-Greek” development of *iH, *uH > *ī, *ū, though the details are far from clear, e.g. hangist ‘peace, quiet’ < *sm̥-kʷih₁- (Lat. quiēs), jowkn ‘fish’ < *dʰĝʰuH- (Gr. ἰχθῦς). For additional examples and discussion, see Clackson 1994: 41–49; Olsen 1999: 770–773. Brought to you by | The University of British Columbia Library Authenticated Download Date | 9/24/18 2:51 PM 254 Ronald I. Kim conflicting outcomes: cf. e.g. gar̄n ‘lamb’ < *wr̥ H-ēn vs. čanačʿem ‘know’ ← *canačʿ< PIE *ĝn̥ h₃-ské̂ /ó-, with respective reflexes aR vs. aRa.15 2.1 Bonfante’s isoglosses Bonfante’s (1937) list of Greco-Armenian phonological isoglosses has also been subject to scrutiny by Clackson (1994: 49–56). I follow the latter’s judgments for the most part, but would go farther in entirely disregarding those isoglosses which involve retentions, rather than innovations. Bonfante’s shared features may be divided into four groups, of which the second and third overlap. Retentions: 1. maintenance of the distinction of PIE *ō̆ and *ā̆ ; 2. maintenance of PIE *e; 8. intervocalic *-wy- retained;16 16. maintenance of the distinction between voiced stops and voiced aspirates; 17. retention of voiceless aspirates; 22. retention of *-ēR, *-ōR in the nom.sg. of sonorant stems.17 Innovations also found in other IE branches: 3. *ə > *a, i.e. *CHC > *CaC (as in most other IE languages); 6. prothetic vowel before *r- (but absence of r- is a widespread areal feature, including Anatolian as well as the non-IE Hurro-Urartian); 7. intervocalic *-y- > *-∅- (also in Italic and many later IE languages); 10. diphthongization of *Vny, *Vry > *Vyn, *Vyr (also in Avestan and other Iranian languages, as well as Greek); 11. initial and medial *s > *h (also in Iranian and Insular Celtic); 23. *-m > *-n (also in Hittite, Insular Celtic, Germanic, Balto-Slavic, Tocharian); 24. *-t, *-d > *-∅ (also in Germanic, Balto-Slavic, Tocharian). Easily repeatable and/or phonologically trivial: 15 For additional examples and discussion, see Clackson 1994: 36–41; Olsen 1999: 775–778; Woodhouse 2015. 16 The example usually cited for Armenian is kogi ‘butter’ < PIE *gʷow-yo- ‘pertaining to a cow’, but it goes without saying that this noun could have been remodeled from the ancestor of kov ‘cow’ < *gʷow-. 17 This is technically a mixed retention/innovation, since PIE probably contrasted *-ḗn in hysterokinetic n-stems with *-ō < pre-PIE **-ōn in amphikinetic n-stems (Jasanoff 1989: 138 fn. 9; 2002: 34–35). The latter served as the model for the r-less nom.sg. of Indo-Iranian and Balto-Slavic, e.g. Ved. svásā, Lith. sesuõ ‘sister’. Brought to you by | The University of British Columbia Library Authenticated Download Date | 9/24/18 2:51 PM Greco-Armenian 9. 12. 13. 14. 15. 255 *-ln- > *-ll-; *sl- > l-, *-sl- > *-l(l)-; *sr- > r-; *sm- > m-, *-sm- > -m-; *sn- > n-, *-sn- > -n(n)-. Outdated (in square brackets) and/or doubtful: [4. *schwa secundum + sonorant > aR]; 5. *r̥ , *l̥ > ar, al; [18. greater resistance of Arm. to “satəmization” than other satəm languages];18 19. *-VKʷ- > *-VwKʷ- > *-VwK̂-; 20. *gʷ not palatalized before front vowels; Arm. *kʷi, but not *kʷe palatalized; [21. PIE *-TT- > “eastern IE” -st- vs. “western IE” -ss-]. Of this last category, no. 5 may be left aside, as the majority Greek dialects also have ρα, λα and the Aeolic dialects show a different vocalism in ορ/ρο, ολ/λο (cf. Myc. to-pe-za [torpeǰa] beside Att. τράπεζα < *kʷtr̥ -ped-ya), and there are Homeric scansions indicating survival of *r̥ into Proto-Greek, most famously Il. 16.857, 22.363 ἀ̆νδρο̆ τῆ̄τᾰ κᾰ ὶ ἥ̄βη̄ ν for *[anr̩ tɛːta].19 Similarly, alleged Greek examples of the first change in no. 19 are now understood as the result of Cowgill’s Law or other einzelsprachlich rounding, e.g. κύκλος ‘circle’, γυνή ‘woman’, νύξ ‘night’,20 and the second change of *-VwKʷ- > *-VwK̂- is in any case unique to Armenian (see below, Section 2.3). Finally, the patterning of palatalization of labiovelars before front vowels, by which Bonfante set so much stock, is hardly probative of a prehistoric connection; not only do the first-millennium Greek dialects famously disagree on the treatment of labiovelars, but Armenian does palatalize *kʷe in PIE *kʷetwóres → *kʷetores > *čʰeðor- > *čʰeor- > čʿorkʿ ‘four’; furthermore, there is no incontrovertible evidence against palatalization of *gʷ before front vowels, so that it may be that all labiovelars were palatalized before front vowels in pre-Armenian.21 The remaining isoglosses from Bonfante’s list belong to the category of “easily repeatable and/or phonologically trivial,” namely the assimilation of *-ln- > *-ll(no. 9) and reduction of clusters of *s + sonorant, for which secure examples are in any case scarce; note that similar changes affecting word-internal *sR clusters are found in a range of IE languages, including some which have otherwise preserved 18 19 20 21 I.e. the retention of affricate c, j as reflexes of PIE *ĝ, *ĝʰ, in contrast to Av. z, OCS z. On the treatment of the PIE syllabic liquids in Greek, see now in detail van Beek 2013. On Cowgill’s Law, see in particular Vine 1999. See the discussion and references in Kim 2016: 40–44. Brought to you by | The University of British Columbia Library Authenticated Download Date | 9/24/18 2:51 PM 256 Ronald I. Kim *s, e.g. Slavic, Germanic, or Italic.22 I conclude that none of Bonfante’s phonological isoglosses involves a significant, distinctive innovation shared exclusively by Greek and Armenian; and none of the scholarship since then has in my opinion discovered an incontrovertible example of such. 2.2 Relative chronology of pre-Armenian sound changes The relative chronology of sound changes from PIE to Classical Armenian also does not support the postulation of additional phonological innovations common to Armenian and Greek, much less of a Proto-Greco-Armenian stage. If the dialects ancestral to Greek and Armenian had enjoyed a period of relative unity for some time after the breakup of PIE, the sound changes assignable to that stage should be earlier than those specific to Armenian. However, the relative chronologies proposed to date, e.g. by Ravnæs (1991) or Kortlandt (1980b), do not support this hypothesis. Ravnæs (1991: 173–182) divides pre-Armenian sound changes into five categories, along with a few of indeterminate date: “Changes occurring in dialectal IE times” (1–2); “Changes occurring in the time of Armenian-Greek linguistic contact” (3–13); “Changes taking place in Proto-Armenian prior to the introduction of Iranian loan words” (14a–51); “Changes occurring after the introduction of Iranian loan words” (52–57); and “Changes occurring after the Sasanian period and prior to or during the Classical epoch” (58–63). Following the ruki-rule23 and word-internal *ə > *∅, which he assigns to stage I, he lists the changes below under stage II: 3. PIE *-m > *-n; 4. development of prothetic vowels before PIE *r-; 5. development of prothetic vowels from PIE *HC-; 6. *nKʷ > *wKʷ; 7. *uKʷ > *uK´; 8. palatals to sibilants and affricates, i.e. PIE *ḱ, *ǵ, *ǵʰ > *t́ś (> *ś), *d́ź, *d́ʰź; PIE *sḱ > *st́ś; 9. PIE *ḱw > *św > *š’w in ‘dog’; 22 For PIE *sm, cf. masc./neut.loc.sg. *tósmi > OCS tomĭ (Lith. tamè), → Goth. þamma (with generalized Verner’s Law treatment *zm > *mm; Ringe 2017a: 166); for PIE *sl, cf. *gʷʰiHslo- > Lat. fīlum ‘thread’, OCS žila ‘vein’ like Arm. ǰil, ǰił ‘sinew’ (vs. retention in Lith. gýsla ‘vein’). 23 N. B. that the ruki-rule is not shared with Greek, and is in any case questionable for Armenian itself: as I hope to show elsewhere, apparent examples of -r < *[-ž] < *[-š] < *-is, *-us (including *-is, *-us < *-ēs, *-ōs; see Winter 1975: 115–116; Olsen 1989b: 5–15 and already Pedersen 1905: 228–231) may be explained otherwise. Brought to you by | The University of British Columbia Library Authenticated Download Date | 9/24/18 2:51 PM Greco-Armenian 257 10. *š’w > *š’y in ‘dog’ (note the contrast with “new *św” in no. 11!); 11. assimilation of *sw > *św in PIE *sweḱurā > *sweśurā > *śweśurā ‘stepmother’ (> Arm. skesowr); 12. (a) *ms, *ns > *ss; (b) *rs > *rr > r̄ ; 13. *s > *h in certain positions. It is immediately apparent that most of these changes are not shared by Greek, namely nos. 6, 7, 8, and 9, as well as 12a (since *ns was retained word-internally as well as in Auslaut into Proto-Greek) and 12b (ditto; cf. Gr. ὄρρος ∼ ὄρσος* ‘rump’). Nos. 10 and 11 do not apply to Greek, since affrication of PIE palatal stops did not take place there; and nos. 4 and 5 have already been discussed above. That leaves nos. 3 and 13, respectively Bonfante’s isoglosses 23 and 11, which are phonologically trivial and easily repeatable, therefore hardly probative of a close dialectal relation between Greek and Armenian at an early post-PIE date. Kortlandt’s relative chronology of approximately 21 sound changes (Kortlandt 1980b; see Beekes 2003: 207–211) agrees in general with that of Ravnæs, but wisely avoids all periodicizing with respect to Greek. After the loss of aspiration in PIE voiced aspirated stops (no. 1), the following distinctive combinatory changes involving PIE stop + *w clusters and labiovelars are assigned to his change no. 2: – the “rise of new labialized stops,” i.e. PIE *ḱw, *tw, *dw > *cʷ, *tʷ, *dʷ (> Arm. š, kʿ, k, respectively in šown ‘dog’, kʿo ‘your’, mełk ‘soft’);24 – PIE *uK > *uK̂, i.e. merger of velar with palatal stops after *u, e.g. *lewkos (or *lowkos) > *lewk-̂ > Arm. loys ‘light’; – and finally, the “awcanem-rule” of *VnKʷ > *VwK̂ in PIE *h₂n̥ gʷ- > awc- ‘anoint’, PIE *(a)ngʷʰi- > awj ‘snake’.25 These are succeeded in turn by assibilation of PIE *ḱ > *ś and of PIE *sk > *ć (no. 3), the lexically restricted assimilation in *sweśurā > *śweśurā (> skesowr ‘stepmother’; no. 4), lenition of PIE *s > *h (no. 5), depalatalization of *ś, *ć > *s, *c (> Arm. s, cʿ; no. 6), and finally “[r]edistribution of labialization” (no. 7), by which *cʷ > *sʷ (> š in šown ‘dog’) and *hw > *hʷ (> kʿ in kʿoyr ‘sister’). 24 The last of these is contested, most scholars today preferring PIE *dw- > Arm. erk- as in PIE dwō > erkow ‘two’, PIE *dwi- > *erki- in erknčʿim, aor. erkeay ‘fear’. The literature on this “Great Armenian Puzzle” is vast; among recent treatments, see especially Viredaz 2003 and DeLisi 2013. 25 Kortlandt also proposes *VuK > *VKʷ in PIE *h₂ewg- > ačem ‘grow’ and *h₂óyu-kʷe > *oukʷe > *okʷe > očʿ ‘not’, but these may be explained otherwise, by a conditioned loss of *w in ‘grow’ (Martirosyan 2010: 43) or a different preform for ‘not’ (see above, Section 1 with fn. 9). Brought to you by | The University of British Columbia Library Authenticated Download Date | 9/24/18 2:51 PM 258 Ronald I. Kim Thus of the first seven changes reconstructed by Kortlandt for the prehistory of Armenian, the only one shared by Greek is *s > *h, and this had already been preceded by the loss of aspiration in PIE voiced aspirates and the specifically Armenian consonant developments grouped under no. 2. Even if one insists on treating the crosslinguistically unremarkable lenition of *s as an innovation common to Greek and Armenian, Kortlandt’s relative chronology implies that it would have to have spread over a prehistoric dialect continuum that was already diversified enough to in effect encompass distinct languages. 2.3 The awcanem-rule: uniquely Armenian? The distinctive Armenian changes of *uK > *uK̂ and *VnKʷ > *VwK̂ have also been discussed at length by Kümmel (2007: 319–327). On the basis of a broad crosslinguistic database of consonant changes, he argues that a shift *uK > *uK̂ makes phonetic sense only if the PIE “labiovelars” were still the labial counterpart of the “palatals” at the time, i.e. the latter were actually pure velars. The awcanemrule can then be understood as a leftward spreading or assimilation of rounding from the labiovelar to the nasal and then to the vowel, producing a diphthong, followed by a simple dissimilation: *[VŋKʷ] > *[VŋʷKʷ] > *[VwŋʷKʷ] > *[VwŋʷK] > *[VwŋK] > *[VwK]. If valid, Kümmel’s argument constitutes further evidence that the speech variety ancestral to Armenian was distinct not only from pre-Greek, but from all other dialects already in late PIE times. The changes described above must date from a very early period, before the PIE “palatals” [= velars] and “velars” [= postvelars] were fronted in place of articulation, so that the rounding dissimilation of labiovelars in the environment of *u caused them to fall together with the former rather than the latter. The restriction of such an idiosyncratic sequence of changes to Armenian naturally does not preclude the possibility that other changes followed which were shared with the dialects ancestral to Greek, but in that case they must have been diffusing across a dialect continuum that was already differentiated to a significant extent. Brought to you by | The University of British Columbia Library Authenticated Download Date | 9/24/18 2:51 PM Greco-Armenian 259 3 Evaluating Greco-Armenian: morphological isoglosses A number of morphological innovations have also traditionally been considered as shared by Greek and Armenian, including: 1. locative (singular) Arm. -oǰ ∼ Gr. -οθι; 2. instrumental Arm. -b, -w (pl. -bkʿ, -wkʿ) ∼ Gr. -φι; 3. deverbal noun suffix Arm. -oł ∼ Gr. -ολης; 4. augmented preterite forms, e.g. Gr. impf. ἔφερε, Arm. aor. eber (also Phrygian and Indo-Iranian); 5. nasal presents in Gr. -ν-ανω, -αινω, -νῡμι and Arm. -anem, -nowm; and 6. Ionic Gr. iterative imperfects in -σκ- beside Arm. weak aorists in -cʿ-. Here too, the value of these innovations for the Greco-Armenian hypothesis has been challenged by Clackson (1994: 60–85), to which I would add the following comments and updates: 1. The distribution of -oǰ in Armenian strongly suggests that the ending was taken over from gen./dat./loc. mioǰ, knoǰ (to mi ‘one’, kin ‘woman’; cf. also gełǰ to gewł ‘village’) as a useful unambiguous locative marker of wo- and ea/wostems, e.g. ordwoǰ (beside ordi), tełwoǰ to ordi ‘son’, tełi ‘place’ (Clackson 1994: 60–64). There is thus no necessary connection with Gr. -οθι, and no need to take -oǰ from pre-Arm. *-o-dʰy-V(-) or a prevocalic sandhi variant *-o-dʰy; cf. the alternative derivation from an amphikinetic i-stem in gełǰ < PIE gen. *-l-y-e/os, dat. *-l-y-ey (Klingenschmitt 1982: 154; Rasmussen 1985 [1987]: 31–34 [1999: 105–109]; Olsen 1999: 172), knoǰ < *kinoǰ ← *kinǰ < PIE gen. *gʷen-y-e/os, dat. *gʷen-y-ey (Matzinger 2005b: 106–109). 2. Despite continuing debate over the original value of -φι in Greek (singular and plural in Homeric vs. plural in Mycenaean), the majority view today seems to be that Homeric forms such as ἶφι ‘by force’, ἠνορέηφι ‘with manhood’ reflect the older situation. Recent rethinking of the evolution of the IE case system (Jasanoff 2009; Melchert & Oettinger 2009) suggests that *-bʰi was an adverbial marker that only became grammaticalized as a case ending after the departure of Anatolian, in the dialects ancestral to Celtic, Italic, Greek, Armenian, and Indo-Iranian. Within this group, Greek and Armenian are distinguished only in employing the suffix in the singular, thereby ousting the inherited *-(e)h₁ (but cf. Balto-Slavic, where originally adverbial *-m(i) became the marker of the instr.sg. outside the o-stems). 3. As Clackson points out (1994: 74–75), neither Gr. -ολης nor Arm. -oł is well attested in the earliest texts (see for Armenian de Lamberterie 1982: 37–45), Brought to you by | The University of British Columbia Library Authenticated Download Date | 9/24/18 2:51 PM 260 Ronald I. Kim and reflexes of deverbal adjectives in *-lo- with the meaning ‘habitually X-ing, prone to X’ are known from other branches, e.g. Lat. bibulus, credulus. The only possible common innovation of Greek and Armenian is the inflectional class (ā-stem and a-stem, respectively), and this too can be an independent development. 4. The status of the augment as a post-PIE innovation is not beyond debate, and some scholars continue to allow the possibility that it existed in PIE and was lost in the branches outside the “augment zone” (see e.g. Meier-Brügger 2010: 315–316).26 In any case, it is also present in Indo-Iranian and so would not be exclusive to Greco-(Phrygio-)Armenian. 5. The productivity of nasal presents, mostly beside strong aorists (< PIE root aorists), is certainly noteworthy, but the exact prehistory of Arm. presents in ‑anem (-anim), -am, -(a)nam remains unclear in details; see Greppin 1973; Hamp 1975; Klingenschmitt 1982: 84–127; and especially Kocharov 2008.27 Furthermore, Greek and Armenian are not the only IE branches to make widespread use of nasal presents beside root aorists; they also enjoyed considerable productivity in Indo-Iranian as well as Tocharian. 6. Despite repeated attempts to connect the weak aorist suffix with PIE s-aorists (e.g. Klingenschmitt 1982: 286–287; Kortlandt 2003: 81, 108–109, 114–116), Arm. -cʿ- must go back to PIE *-skê /o-. But if the original function of the suffix was to form derived imperfectives, as in Hittite, the only actual innovation shared by Armenian and Greek would be its restriction to the past. Note also that the productive shape of the iterative suffix in Ionic is -εσκ-, with athematic formations making up only a small minority of all tokens, e.g. ἵστασκε, στάσκε ‘was standing’. In contrast, Armenian has *-acʰ- in all but a few weak aorists, and the exceptions (e.g. elicʿ ‘filled’, zgecʿaw ‘put on (clothing)’) are probably late creations, rather than inherited (pace Karstien 1956: 227; Godel 1965: 37; 1975: 127; Clackson 1994: 82; see Kim forthc. and Section 4 no. 4 below). Thus none of the morphological features adduced as evidence for a special relationship between Greek and Armenian can be demonstrated to be exclusive innovations. One may nevertheless agree with the conclusion of Martirosyan (2013: 92) that “the cumulative strength of these morphological (and a few phonological) features and a great number of such lexical agreements gives additional weight 26 The problem is naturally bound up with the source of the augment, for which numerous hypotheses exist. I continue to find attractive the proposal of Delfs (2006; cf. Ringe 2017a: 30) that it originated as an evidential particle, which would explain its restriction to past indicative forms. 27 On factitives in -anam, see also Section 4 no. 2 below. Brought to you by | The University of British Columbia Library Authenticated Download Date | 9/24/18 2:51 PM Greco-Armenian 261 to the evidence” that pre-Greek and pre-Armenian were in contact in prehistoric times (see below, Section 5). 4 Thinking outside the (Greco-Armenian) box The widespread assumption of a special relationship between Armenian and Greek has had the unfortunate consequence that potential isoglosses shared with other IE languages have not always received proper consideration. Some such isoglosses have of course long been recognized, e.g. the grammaticalization of PIE verbal adjectives in *-lo- as participles (sir-eal/sirecʿ-eal ‘loved’, gt-eal ‘found’), otherwise found only in Slavic and Tocharian; but these have for the most part been treated as chance similarities or parallel innovations, of limited relevance for the position of Armenian within the Indo-European family. However, an unprejudiced examination of the historical morphology of Armenian reveals that the number of innovations clearly shared with Greek is in fact quite limited. In contrast, several significant features of Armenian verbal morphology appear to be shared with other geographically neighboring branches, namely Indo-Iranian and Balto-Slavic: 1. The present passive conjugation in -i-, corresponding to active presents in -e(e.g. sirem ‘love’, sirim ‘am loved’), has been connected with PIE stative *-eh₁and specifically with Greek aorists in -η- (see the discussion in Klingenschmitt 1982: 9–11). The functional and distributional problems with this equation were pointed out by Barton (1990–1991: 50 fn. 69), who argued instead for -i- < *-iye/o-, the Sievers’s Law allomorph of the *-yé/ó- reflected in Indo-Iranian passives such as Ved. kr-iyá-te ‘is made’, chid-yá-te ‘is cut off’. The comparison with Indo-Iranian goes back to Meillet (1903: 79; 1936: 107–108), where however the i-presents of Baltic and Slavic are to be left aside; it is likewise upheld by Watkins (1969: 75) and Stempel (2000: 517–518). 2. The productive Arm. fientives in -ana- (e.g. lowsana- ‘become light’, mecana‘become big’ ← loys ‘light’, mec ‘big’) have been examined by several scholars (see e.g. Klingenschmitt 1982: 119–127; Kocharov 2011), but their origin and exact relationship to other nasal formations remain unclear. One possibility that has not yet to my knowledge been seriously pursued is a connection with “Northern IE” intransitives of the type of Goth. fullnōn ‘fill (up), become full’, Lith. liñka ‘is left, remains’, OCS lęgǫ ‘lie down’, niknǫ ‘appear’. The prehistory of the latter formations remains uncertain, but even if they go back to a PIE Brought to you by | The University of British Columbia Library Authenticated Download Date | 9/24/18 2:51 PM 262 Ronald I. Kim prototype,28 the specialization of this suffix in fientive function could be a noteworthy innovation shared by Armenian with Balto-Slavic and Germanic. 3. Bonfante (1942) first compared Arm. aor.1sg. edi ‘put’, etow ‘gave’ with OCS děxŭ, daxŭ and derived both from a sigmatized root aorist *(e-)dʰeh₁-s-om, *(e-)deh₃-s-om ← PIE *(e-)dʰeh₁-m, *(e-)deh₃-m; this account has since been accepted by Schmitt (1981: 54) and Barton (1989: 146–147). Viredaz (2015) has now argued that all PIE root aorists underwent what he calls “semi-sigmatization” in Armenian, e.g. 1sg. *mer-m̥ → *mer-om → *mers-om > OCS mrěxŭ, → Arm. mer̄ay ‘died’; he considers this to be a significant innovation shared with Slavic. 4. The weak aorist suffix *-acʰ- goes back to an earlier stem in *-a- to which the originally imperfective *-cʰ- < PIE *-skê /o- was added (see above, Section 3 no. 6). By far the simplest hypothesis is to identify this *-a- with Balto-Slavic preterites in *-ā́- and Latin impf. er-ā-, -bā-, all reflecting a (post-)PIE morpheme *-eh₂-, perhaps originally an optative marker.29 The weak aorist suffix in -ecʰ- ∼ -eacʰ found with the overwhelming majority of e-presents therefore goes back to a sequence *-eyā-, originally proper to denominatives in *-e-yé/óto thematic nominal bases, e.g. gorcem ‘make, do’, gnem ‘buy’ ← gorc (o) ‘work’, gin (o) ‘price’ (Kim forthc. §5). The agreement between Armenian and Balto-Slavic on the last three points (specifically with Slavic on no. 3) raises the possibility that these two branches were in contact in early post-PIE times. This hypothesis is far from new — cf. the lists of possible Armenian-Slavic lexical isoglosses adduced by Karstien (1956: 212–220) and Saradževa (1980; 1986), and recall the conclusion of Hübschmann (1877: 39) that “[d]as armenische steht im kreise der arisch-slavo-lett. sprachen zwischen iranisch und slavolettisch”30 — but has long been downplayed in comparison with the connections with Greek and Indo-Iranian, which as seen above are of a largely lexical nature. To be sure, the likely prehistoric route of migration of the pre-Armenians from the Balkans across Asia Minor to the Armenian highlands and southern Caucasus rather disfavors any recent direct interaction with the speakers 28 See e.g. Gorbachov (2007), who reconstructs a PIE formation with zero-grade root and stressed h₂e-conjugation endings, e.g. *li-n-kʷ-é ‘is left’. 29 See also Stempel (2000: 519–520), although he confuses these formations in *-ā- with āintensives of the type of Lat. cēlāre ‘hide, conceal’, Gr. νωμάω ‘deal out, distribute’, which were rather one of the sources of Armenian present stems in -a-. 30 Similarly Meillet 1896: 149 (“la situation dialectale de l’arménien est intermédiaire entre l’indoiranien, le letto-slave et le grec”); Pedersen 1924: 225 (“… daß das Arm. unter den lebendigen idg. Sprachzweigen etwa nach drei Seiten hin nähere verwandtschaftliche Beziehungen hat: w[estlich] zum Griech., ö[stlich] zum Indisch-Iran., n[ördlich] zum Slavisch-Balt.”). Brought to you by | The University of British Columbia Library Authenticated Download Date | 9/24/18 2:51 PM Greco-Armenian 263 of Proto-Balto-Slavic, which would have to have taken place while the speakers of pre-Armenian were still located to the north and west of the Black Sea. Nevertheless, given the generally accepted importance of morphological innovations in determining subgrouping, the existence of these isoglosses could well point to a closer affinity between Armenian and Balto-Slavic than emerges from cladistic studies based exclusively or primarily on lexicostatistics. 5 Summary and conclusions The largely negative result of the preceding sections confirms the position laid out in the introduction, namely that the list of linguistic innovations exclusively shared by Greek and Armenian is overwhelmingly composed of lexical items. Furthermore, most of these involve general root cognations, not full word equations allowing for reconstruction of an intermediate preform, which raises the possibility that they are either (partial) independent creations or even borrowings from a third language. In this respect, the relationship between Greek and Armenian differs greatly from that of Indo-Aryan and Iranian, or Baltic and Slavic, where it is possible to reconstruct dozens of distinct lexical preforms for Proto-Indo-Iranian and Proto-Balto-Slavic, respectively.31 Among the lexical isoglosses identified by Ringe, Warnow & Taylor (2002), for instance, only Gr. ἦμαρ ∼ Arm. awr ‘day’ qualifies as a (near-)equation; and despite the long lists of alleged Greco-Armenian lexical innovations, the actual number of assured word equations is not large (Section 1). This picture is consistent with the hypothesis that the dialects ancestral to Greek and Armenian were in geographical proximity in prehistoric times, say up to the late 3rd millennium BC, but in no way justify a unitary Proto-Greco-Armenian intermediate in time between Proto-Indo-European and the emergence of the Greeks (and later, the Armenians) in the historical record. I thus concur with the findings of Martirosyan (2013: 85), who on the basis of a detailed review of the lexical evidence concludes … that Armenian, Greek, (Phrygian) and Indo-Iranian were dialectally close to each other. Within this hypothetical dialect group, Proto-Armenian was situated between Proto-Greek (to the west) and Proto-Indo-Iranian (to the east). The Indo-Iranians then moved eastwards, while the Proto-Armenians and Proto-Greeks remained in a common geographical region for a long period and developed numerous shared innovations. At a later stage, together or independently, they borrowed a large number of words from the Mediterranean/Pontic 31 Cf. for Proto-Balto-Slavic *gálwā́ ‘head’, *gwai(g)zdā́ ‘star’, *rankā́ ‘hand’, *wárnā́ ‘crow’, *wainikas ‘wreath’, *źeimā́ ‘winter’, etc. Brought to you by | The University of British Columbia Library Authenticated Download Date | 9/24/18 2:51 PM 264 Ronald I. Kim substrate language(s), mostly cultural and agricultural words, as well as animal and plant designations. This conclusion is supported by recent studies which have identified isoglosses connecting Armenian with Indo-Iranian rather than with its western neighbors, e.g. de Lamberterie 1986: 49–57 on the preverb *ni- (Arm. nstim ‘sit (down)’ ← *niste< PIE *ni-si-sd-e/o-), or Stempel 2000 on the present passives in -i- (see above, Section 4 no. 1) as well as Armenian lexical items with cognates in Indo-Iranian poetic language, e.g. Arm. arew ‘sun’ ∼ Skt. ravi- ‘sun(-god)’ or Arm. erg ‘song’ ∼ Ved. arká- ‘song (of praise)’.32 The continuing popularity of the Greco-Armenian hypothesis is no doubt due in large part to the inertia of established scholarly opinion, as well as the deeply rooted temptation to interpret the often difficult Armenian facts through the prism of the far better understood historical grammar of Greek. This practice has on more than one occasion influenced scholars in favor of or against a particular hypothesis on the grounds that the sound change, morphological formation, distribution of stems, etc. in question is or is not attested in Greek, rather than because it explains the Armenian data better or worse than alternative models. One must therefore resist the temptation to give preference to “Hellenoid” explanations and dismiss too quickly comparisons with other (neighboring) IE branches such as Balto-Slavic, which may in fact be of some importance and even point to shared innovations, as seen above in the case of verbal morphology (Section 4). At any rate, one can hardly disagree with the opinion of de Lamberterie (2011: 367) that “les isoglosses entre l’arménien et le grec mises en évidence, dans les années 1920 et 1930, par Meillet et Pedersen doivent être aujourd’hui restituées dans un ensemble plus vaste.” Examination of the relative chronology of sound changes (Section 3) lends further support to the view that the linguistic ancestors of the Armenians made up a distinct speech community already at an early prehistoric stage, at any rate by the end of the 3rd millennium BC. Any phonological and morphological changes that may have been shared with neighboring IE dialects after that time were either of an easily repeatable type (e.g. partial merger of PIE velar and labiovelar stops, loss of the perfect) or were subsequently obscured by later sound changes, loss of morphological categories, or lexical replacement of all relevant forms. Possible exceptions are the features of verbal morphology discussed in Section 4, which may point to very early contacts with Balto-Slavic and in any case should be accorded at least as much weight as lexical isoglosses in determining the position of Armenian in the IE family. However, while these and the other innovations adduced above 32 Some of these could however represent retentions rather than common innovations; see Schmitt 1967: 259–260 with refs. Brought to you by | The University of British Columbia Library Authenticated Download Date | 9/24/18 2:51 PM Greco-Armenian 265 could have been shared with any of the neighboring dialects with which Armenian was in contact, from Balto-Slavic and (Indo-)Iranian to Greek and Phrygian, there is insufficient evidence at present to demonstrate any such shared innovations. The great majority of features that together make Armenian so distinct within the Indo-European family are best understood as developments internal to the language itself, the outcome of independent and often idiosyncratic processes of language change operating over millennia. Acknowledgment: I thank Götz Keydana and Paul Widmer for organizing the successful and thought-provoking workshop on Indo-European subgrouping in Göttingen, and my colleagues for their stimulating comments and discussions on this topic, including Petr Kocharov, Daniel Kölligan, Martin Kümmel, Reiner Lipp, Hrach Martirosyan, Birgit Anette Olsen, Dariusz Piwowarczyk, and Wojciech Sowa. The research for this paper has been supported by grant no. GA17-19686S from the Czech Science Foundation, whose generosity is hereby gratefully acknowledged. Bibliography Barton, Charles R. (1989). “PIE. *mer-, Arm. mer̄anim ‘die’”. In: Indogermanische Forschungen 94, 135–157. Barton, Charles R. (1990–1991). “On the denominal a-statives of Armenian”. In: Revue des Études Arméniennes N. S. 22, 29–52. van Beek, Lucien (2013). The Development of the Proto-Indo-European Syllabic Liquids in Greek. PhD thesis. Leiden University. Beekes, Robert S. P. (1988). “Laryngeal developments. A survey”. In: Die Laryngaltheorie und die Rekonstruktion des indogermanischen Laut- und Formensystems. Ed. by Alfred Bammesberger. Heidelberg: Winter, 59–106. Beekes, Robert S. P. (2003). “Historical phonology of Classical Armenian. A survey”. In: Kortlandt, Frederik H. H. Armeniaca. Comparative Notes. With an appendix on the historical phonology of Classical Armenian by Robert S. P. Beekes. Ann Arbor: Caravan, 133–211. Bonfante, Giuliano (1937). “Les isoglosses gréco-arméniennes”. In: Mélanges linguistiques offerts à M. Holger Pedersen à l’occasion de son soixante-dixième anniversaire, 7 avril 1937. København: Levin & Munksgaard, 15–33. Bonfante, Giuliano (1942). “The Armenian aorist”. In: Journal of the American Oriental Society 62.2, 102–105. Bonfante, Giuliano (1982). “The place of Armenian among the Indo-European languages”. In: Annali dell’Istituto Orientale di Napoli 4, 151–169. Clackson, James (1994). The Linguistic Relationship between Armenian and Greek. Oxford & Cambridge, MA: Blackwell. Clackson, James (2004). “Classical Armenian”. In: The Cambridge Encyclopedia of the World’s Ancient Languages. Ed. by Roger D. Woodard. Cambridge: Cambridge University Press, 922–942. Brought to you by | The University of British Columbia Library Authenticated Download Date | 9/24/18 2:51 PM 266 Ronald I. Kim Clackson, James (2005). Rev. of Kortlandt 2003. In: Annual of Armenian Linguistics 24–25, 153– 158. Cowgill, Warren (1960). “Greek ou and Armenian očʿ”. In: Language 36.3, 347–350. Czekanowski, Jan (1927). Wstęp do historii Słowian. Perspektywy antropologiczne, etnograficzne, prehistoryczne i językoznawcze. [Introduction to the History of the Slavs. Anthropological, Ethnographic, Prehistoric, and Linguistic Perspectives]. Lwów: Jakubowski. Repr. Poznań: Instytut Zachodni, 1957. Czekanowski, Jan (1928). “Na marginesie recenzji p. K. Moszyńskiego o książce Wstęp do historii Słowian”. [“Marginalia to K. Moszyński’s review of the book Introduction to the History of the Slavs”]. In: Lud 27, 41–56. Delfs, Lauren C. (2006). “Evidence for the origin of the augment”. Paper read at the 25th East Coast Indo-European Conference, The Ohio State University, 19–22 June 2006. DeLisi, Jessica (2013). “Feature metathesis and the change of PIE *du̯ to Classical Armenian ‑rk‑”. In: Diachronica 30.4, 469–491. Djahukian, Geworg B. (1980). “The position of Armenian in the Indo-European languages”. In: First International Conference on Armenian Linguistics. Proceedings, the University of Pennsylvania, Philadelphia, 11–14 July 1979. Ed. by John A. C. Greppin. Delmar, NY: Caravan, 3–16. Djahukian, Geworg B. (1982). Сравнительная грамматика армянского языка. [Comparative Grammar of Armenian]. Ереван: Издательство AH Армянской ССР. Dressler, Wolfgang (1969). “Altarmenisch awr awur ‘Tag für Tag’”. In: Revue des Études Arméniennes N. S. 6, 19–21. Falileyev, Alexander & Petr Kocharov (2012). “Celtic, Armenian, and eastern Indo-European languages. Comments on a recent hypothesis”. In: Ireland and Armenia. Studies in Language, History and Narrative. Ed. by Maxim Fomin, Alvard Jivanyan & Séamus Mac Mathúna. Washington, DC: Institute for the Study of Man, 65–84. Godel, Robert (1965). “Les origines de la conjugaison arménienne”. In: Revue des Études Arméniennes N. S. 2, 22–41. Godel, Robert (1975). An Introduction to the Study of Classical Armenian. Wiesbaden: Reichert. Gorbachov, Yaroslav (2007). Indo-European Origins of the Nasal Inchoative Class in Germanic, Baltic and Slavic. PhD thesis. Harvard University. Gray, Russell D. & Quentin D. Atkinson (2003). “Language-tree divergence times support the Anatolian theory of Indo-European origin”. In: Nature 426 (27 Nov 2003), 435–439. Greppin, John A. C. (1973). “The origin of Armenian nasal suffix verbs”. In: Zeitschrift für vergleichende Sprachforschung 87.2, 190–198. Greppin, John A. C. (1981). “Concerning the reply of Kerns and Schwartz to Austin”. In: Bono Homini Donum. Essays in Historical Linguistics in Memory of J. Alexander Kerns. Ed. by Yoël L. Arbeitman & Allan R. Bomhard. Vol. 1. Amsterdam: Benjamins, 119–126. Greppin, John A. C. (1982). “Arm. tʿ – Gk. πτ-”. In: Journal of Indo-European Studies 10.3–4, 347–354. Greppin, John A. C. (1989). “A rejoinder to Birgit Olsen on the sole color of the laryngeal in Armenian being a”. In: Revue des Études Arméniennes N. S. 21, 477–480. Hamp, Eric P. (1975). “On the nasal presents of Armenian”. In: Zeitschrift für vergleichende Sprachforschung 89.1, 100–109. Hamp, Eric P. (1976). “*gʷeiHo - ‘live’”. In: Studies in Greek, Italic, and Indo-European Linguistics Offered to Leonard R. Palmer on the Occasion of his Seventieth Birthday, June 5, 1976. Ed. by Anna Morpurgo Davies & Wolfgang Meid. Innsbruck: Institut für Sprachwissenschaft, 87–91. Brought to you by | The University of British Columbia Library Authenticated Download Date | 9/24/18 2:51 PM Greco-Armenian 267 Hamp, Eric P. (1979). “Indo-European *gʷen-Ha ”. In: Zeitschrift für vergleichende Sprachforschung 93.1, 1–7. Hamp, Eric P. (1983a). “On the Helleno-Armenian shared lexicon”. In: Annual of Armenian Linguistics 4, 63–64. Hamp, Eric P. (1983b). “Philologica varia”. In: Revue des Études Arméniennes N. S. 17, 5–12. Hamp, Eric P. (1989). “Two Armenian etymologies”. In: Revue des Études Arméniennes N. S. 21, 25–27. Hamp, Eric P. (1992). “On subgrouping and Armenian”. In: Annual of Armenian Linguistics 13, 55–61. Holst, Jan Henrik (2009). Armenische Studien. Wiesbaden: Harrassowitz. Hübschmann, Heinrich (1877). “Ueber die stellung des armenischen im kreise der indogermanischen sprachen”. In: Zeitschrift für vergleichende Sprachforschung 23.1, 5–49. Repr. in Hübschmann 1976, 1–45. Hübschmann, Heinrich (1976). Kleine Schriften zum Armenischen. Ed. by Rüdiger Schmitt. Hildesheim & New York: Olms. Jasanoff, Jay H. (1989). “Old Irish bé ‘woman’”. In: Ériu 40, 135–141. Jasanoff, Jay H. (2002). “The nom.sg. of Germanic n-stems”. In: Verba et Litterae. Explorations in Germanic Languages and German Literature (Festschrift for Albert L. Lloyd). Ed. by Alfred R. Wedel & Hans-Jörg Busch. Newark, DE: Linguatext, 31–46. Jasanoff, Jay H. (2009). “*-bhi, *-bhis, *-ōis. Following the trail of the PIE instrumental plural”. In: Internal Reconstruction in Indo-European. Methods, Results, and Problems. Section papers from the XVI International Conference on Historical Linguistics, University of Copenhagen, 11th–15th August, 2003. Ed. by Jens E. Rasmussen & Thomas Olander. København: Museum Tusculanum, 137–149. Jasanoff, Jay H. (forthc.). “The prehistory of the IE perfect. A contribution to the theory of the h₂e-conjugation”. Karstien, Hans (1956). “Das slavische Imperfekt und der armenische -acʿe- Aorist”. In: Festschrift für Max Vasmer zum 70. Geburtstag am 28. Februar 1956. Ed. by Margarete Woltner & Herbert Bräuer. Wiesbaden: Harrassowitz, 211–229. Kim, Ronald I. (2016). “Studies in Armenian historical phonology (II). Early raising of mid vowels in Auslaut”. In: Indogermanische Forschungen 121, 39–51. Kim, Ronald I. (forthc.). “The prehistory of the Classical Armenian weak aorist”. In: The Armenian and Indo-European Preterite. Forms and Functions. Ed. by N. Kazansky et al. SanktPeterburg: Nauka. Klingenschmitt, Gert (1970). “Griechisch ῑ̔λάσκεσθαι”. In: Münchener Studien zur Sprachwissenschaft 28, 75–88. Klingenschmitt, Gert (1982). Das altarmenische Verbum. Wiesbaden: Reichert. Klingenschmitt, Gert (1994). “Die Verwandtschaftsverhältnisse der indogermanischen Sprachen”. In: In honorem Holger Pedersen. Kolloquium der Indogermanischen Gesellschaft vom 25. bis 28. März 1993 in Kopenhagen. Ed. by Jens E. Rasmussen. Wiesbaden: Reichert, 235–251. Kocharov, Petr (2008). Отражение индоевропейских назальных презенсов в древнеармянском языке. [The Reflexes of Indo-European Nasal Presents in Old Armenian]. PhD thesis. Saint Petersburg University. Kocharov, Petr (2011). “On ana-presents of Armenian”. In: Indogermanistik und Linguistik im Dialog. Akten der XIII. Fachtagung der Indogermanischen Gesellschaft vom 21. bis 27. September 2008 in Salzburg. Ed. by Thomas Krisch & Thomas Lindner. Wiesbaden: Reichert, 271–278. Brought to you by | The University of British Columbia Library Authenticated Download Date | 9/24/18 2:51 PM 268 Ronald I. Kim Kortlandt, Frederik H. H. (1980a). “Albanian and Armenian”. In: Zeitschrift für vergleichende Sprachforschung 94.1–2, 243–251. Repr. in Kortlandt 2003, 13–19. Kortlandt, Frederik H. H. (1980b). “On the relative chronology of Armenian sound changes”. In: First International Conference on Armenian Linguistics. Proceedings, the University of Pennsylvania, Philadelphia, 11–14 July 1979. Ed. by John A. C. Greppin. Delmar, NY: Caravan, 97–106. Repr. in Kortlandt 2003, 26–33. Kortlandt, Frederik H. H. (1986). “Armenian and Albanian”. In: La place de l’arménien dans les langues indo-européennes. Ed. by M. Leroy & F. Mawet. Louvain: Peeters, 38–47. Repr. in Kortlandt 2003, 68–74. Kortlandt, Frederik H. H. (1987). “Notes on Armenian historical phonology V”. In: Studia Caucasica 7, 61–65. Repr. in Kortlandt 2003, 75–78. Kortlandt, Frederik H. H. (2003). Armeniaca. Comparative Notes. With an appendix on the historical phonology of Classical Armenian by Robert S. P. Beekes. Ann Arbor: Caravan. Kümmel, Martin J. (2007). Konsonantenwandel. Bausteine zu einer Typologie des Lautwandels und ihre Konsequenzen für die vergleichende Rekonstruktion. Wiesbaden: Reichert. de Lamberterie, Charles (1982). “Poids et force. Reconstruction d’une racine indo-européenne”. In: Revue des Études Arméniennes N. S. 16, 21–55. de Lamberterie, Charles (1986). “Deux isoglosses entre l’arménien et l’indo-iranien”. In: La place de l’arménien dans les langues indo-européennes. Ed. by M. Leroy & F. Mawet. Louvain: Peeters, 48–61. de Lamberterie, Charles (1992). “Introduction à l’arménien classique”. In: LALIES 10. Actes des sessions de linguistique et de littérature (Aussois, 29 août – 3 septembre 1988 / 28 août – 2 septembre 1989). Ed. by Jean Lallot. Paris: Sorbonne Nouvelle, 234–289. de Lamberterie, Charles (1994). “L’arménien”. In: Langues indo-européennes. Ed. by Françoise Bader. Paris: CNRS, 139–165. de Lamberterie, Charles (2011). “Grammaire comparée des langues indo-européennes”. In: Annuaire de l’École Pratique des Hautes Études, Section des Sciences Historiques et Philologiques 141, 364–368. de Lamberterie, Charles (2012). “L’arménien et l’indo-européen balkanique”. In: IXth International Conference on Armenian Linguistics. Abstracts. Sankt-Peterburg: Nestor-Istoriya, 109–111. de Lamberterie, Charles (2013). “Grec, phrygien, arménien. Des anciens aux modernes”. In: Journal des Savants janvier-juin 2013, 3–69. Leskien, August (1876). Die Declination im Slavisch-Litauischen und Germanischen. Leipzig: Hirzel. Makaev, È. (1967). “Армяно-индоевропейские лексические изоглоссы и ареальная лингвистика”. [“Armenian-Indo-European lexical isoglosses and areal linguistics”]. In: Zeitschrift für Phonetik, Sprachwissenschaft und Kommunikationsforschung 20.5–6, 449–464. Mańczak, Witold (1987). “The relationship of Armenian to Greek and Persian”. In: Annual of Armenian Linguistics 8, 19–28. Martirosyan, Hrach K. (2010). Etymological Dictionary of the Armenian Inherited Lexicon. Leiden & Boston: Brill. Martirosyan, Hrach K. (2013). “The place of Armenian in the Indo-European language family. The relationship with Greek and Indo-Iranian”. In: Journal of Language Relationship / Вопросы языкового родства 10, 85–138. Brought to you by | The University of British Columbia Library Authenticated Download Date | 9/24/18 2:51 PM Greco-Armenian 269 Matzinger, Joachim (2005a). “Phrygisch und Armenisch”. In: Sprachkontakt und Sprachwandel. Akten der XI. Fachtagung der Indogermanischen Gesellschaft, 17.–23. September 2000, Halle an der Saale. Ed. by Gerhard Meiser & Olav Hackstein. Wiesbaden: Reichert, 375–394. Matzinger, Joachim (2005b). Untersuchungen zum altarmenischen Nomen. Die Flexion des Substantivs. Dettelbach: Röll. Matzinger, Joachim (2012). “‘Zwischensprachen’. Areallinguistische Bemerkungen aus dem Bereich des Balkanindogermanischen”. In: Iranistische und indogermanistische Beiträge in memoriam Jochem Schindler (1944–1994). Ed. by Velizar Sadovski & David Stifter. Wien: Akademie der Wissenschaften, 137–159. Meier-Brügger, Michael (2010). Indogermanische Sprachwissenschaft. 9., durchgesehene und ergänzte Auflage. Berlin & New York: de Gruyter. Meillet, Antoine (1896). “Varia. V. Position dialectale de l’arménien”. In: Mémoires de la Société de Linguistique 9, 149–155. Meillet, Antoine (1903). Esquisse d’une grammaire comparée de l’arménien classique. Vienne: Imprimerie des PP. Mékhitharistes. Meillet, Antoine (1925). “Remarques sur l’étymologie de quelques mots grecs. À propos de ἐρύκω, τμήγω, νήχω”. In: Bulletin de la Société de Linguistique 26.1–2, 1–22. Repr. in Meillet 1977, 215–220. Meillet, Antoine (1927). “De la prothèse vocalique en grec et en arménien”. In: Bulletin de la Société de Linguistique 27, 129–135. Repr. in Meillet 1977, 233–239. Meillet, Antoine (1936). Esquisse d’une grammaire comparée de l’arménien classique. Seconde édition entièrement remaniée. Vienne: Imprimerie des PP. Mékhitharistes. Meillet, Antoine (1977). Études de linguistique et de philologie arméniennes. Vol. 2: Études linguistiques. Louvain: Imprimerie Orientaliste. Melchert, H. Craig & Norbert Oettinger (2009). “Ablativ und Instrumental im Hethitischen und Indogermanischen. Ein Beitrag zur relativen Chronologie”. In: Incontri Linguistici 32, 53–73. Morani, Moreno (2001). “Armeno e altre lingue indeuropee. Alcune riflessioni”. In: Cinquant’anni di ricerche linguistiche. Problemi, risultati e prospettive per il terzo millennio. Atti del IX Convegno internazionale di linguisti. Ed. by R. B. Finazzi & P. Tornaghi. Alessandria: Edizioni Dell’Orso, 417–430. Olsen, Birgit A. (1985). “On the development of Indo-European prothetic vowels in Classical Armenian”. In: Revue des Études Arméniennes N. S. 19, 5–17. Olsen, Birgit A. (1989a). “A surrejoinder to J. A. C. Greppin’s remarks on prothetic vowels in Armenian”. In: Revue des Études Arméniennes N. S. 21, 481–483. Olsen, Birgit A. (1989b). “Three notes on Armenian historical phonology”. In: Annual of Armenian Linguistics 10, 5–25. Olsen, Birgit A. (1999). The Noun in Biblical Armenian. Origin and Word-Formation, with Special Emphasis on the Indo-European Heritage. Berlin & New York: de Gruyter. Pedersen, Holger (1905). “Zur armenischen sprachgeschichte”. In: Zeitschrift für vergleichende Sprachforschung 38.2, 194–240. Repr. in Pedersen 1982, 56–102. Pedersen, Holger (1906). “Armenisch und die nachbarsprachen”. In: Zeitschrift für vergleichende Sprachforschung 39.3, 334–485. Repr. in Pedersen 1982, 112–263. Pedersen, Holger (1924). “Armenier. B. Sprache”. In: Reallexikon der Vorgeschichte. Ed. by Max Ebert. Vol. 1. Berlin: de Gruyter, 219–226. Repr. in Pedersen 1982, 302–309. Pedersen, Holger (1982). Kleine Schriften zum Armenischen. Ed. by Rüdiger Schmitt. Hildesheim & New York: Olms. Brought to you by | The University of British Columbia Library Authenticated Download Date | 9/24/18 2:51 PM 270 Ronald I. Kim Peters, Martin (1997). “Der armenische Flexionstyp gitem, gitacʿi und das ion.-att. Plusquamperfekt”. In: Sound Law and Analogy. Papers in Honor of Robert S. P. Beekes on the Occasion of his 60th Birthday. Ed. by Alexander Lubotsky. Amsterdam & Atlanta: Rodopi, 209–217. Porzig, Walter (1954). Die Gliederung des indogermanischen Sprachgebiets. Heidelberg: Winter. Rasmussen, Jens E. (1985 [1987]). “Miscellaneous morphological problems in Indo-European languages I–II”. In: Lingua Posnaniensis 28, 27–62. Repr. in Rasmussen 1999, 100–145. Rasmussen, Jens E., ed. (1999). Selected Papers on Indo-European Linguistics. With a Section on Comparative Eskimo Linguistics. København: Museum Tusculanum. Ravnæs, Erling (1991). The Chronology of the Sound Changes from Proto-Indo-European to Classical Armenian. PhD thesis. University of Oslo. Rexová, Kateřina, Daniel Frynta & Jan Zrzavý (2003). “Cladistic analysis of languages. IndoEuropean classification based on lexicostatistical data”. In: Cladistics 19.2, 120–127. Ringe, Don (2017a). From Proto-Indo-European to Proto-Germanic. 2nd ed. New York & Oxford: Oxford University Press. Ringe, Don (2017b). “Indo-European dialectology”. In: Handbook of Comparative and Historical Indo-European Linguistics. Ed. by Jared Klein et al. Vol. 1. Berlin: de Gruyter, 62–75. Ringe, Don & Joseph F. Eska (2013). Historical Linguistics. Toward a Twenty-First Century Reintegration. Cambridge: Cambridge University Press. Ringe, Don, Tandy Warnow & Ann Taylor (2002). “Indo-European and computational cladistics”. In: Transactions of the Philological Society 100.1, 59–129. Ritter, Ralf-Peter & Wojciech Sowa (2004). “Greek-Armenian isoglosses and the Ancient Greek dialects”. In: Proceedings of the 24th Annual Meeting of the Department of Linguistics, Aristotle University of Thessaloniki, 9–11 May 2003. Ed. by Marianna Margariti-Rogka, Georgios Papanastasiou & Nikos Liosis. Thessaloniki: Aristotle University of Thessaloniki, 645–656. Saradževa, L. A. (1980). Армянo-славянские лексико-грамматические параллели. [ArmenianSlavic Lexical-Grammatical Parallels]. Ереван: Издательство AH Армянской ССР. Saradževa, L. A. (1986). Армянo-славянские лексико-семантические параллели (сравнительноисторическое исследование генетически общей лексики). [Armenian-Slavic LexicalGrammatical Parallels (Comparative-Historical Investigation of Genetically Common Lexicon)]. Ереван: Издательство AH Армянской ССР. Schmidt, Karl Horst (1980). “Armenian and Indo-European”. In: First International Conference on Armenian Linguistics. Proceedings, the University of Pennsylvania, Philadelphia, 11–14 July 1979. Ed. by John A. C. Greppin. Delmar, NY: Caravan, 35–58. Schmidt, Karl Horst (1985). “Die indogermanischen Grundlagen des altarmenischen Verbums”. In: Zeitschrift für vergleichende Sprachforschung 98.2, 214–237. Schmidt, Karl Horst (1987). “The Indo-European background of the Classical Armenian noun declension”. In: Annual of Armenian Linguistics 8, 35–47. Schmidt, Karl Horst (1988). “On Armenian phonology and Indo-European reconstruction”. In: Languages and Cultures. Studies in Honor of Edgar C. Polomé. Ed. by Mohammad Ali Jazayery & Werner Winter. Berlin, New York & Amsterdam: de Gruyter, 601–610. Schmidt, Karl Horst (2007). “Armenian and Celtic. Towards a new classification of early IndoEuropean dialects”. In: Bulletin of the Georgian Academy of Sciences 175.1, 199–203. Schmidt, Karl Horst (2012). “Armenian and Celtic. Towards a new classification of early IndoEuropean dialects”. In: Ireland and Armenia. Studies in Language, History and Narrative. Ed. by Maxim Fomin, Alvard Jivanyan & Séamus Mac Mathúna. Washington, DC: Institute for the Study of Man, 51–63. Revised version of Schmidt 2007. Brought to you by | The University of British Columbia Library Authenticated Download Date | 9/24/18 2:51 PM Greco-Armenian 271 Schmitt, Rüdiger (1962). Rev. of Solta 1960. In: Kratylos 7, 149–155. Schmitt, Rüdiger (1967). Dichtung und Dichtersprache in indogermanischer Zeit. Wiesbaden: Harrassowitz. Schmitt, Rüdiger (1972 [1974]). “Die Erforschung des Klassisch-Armenischen seit Meillet (1936)”. In: Kratylos 17, 1–68. Schmitt, Rüdiger (1981). Grammatik des Klassisch-Armenischen mit sprachvergleichenden Erläuterungen. Innsbruck: Institut für Sprachwissenschaft der Universität Innsbruck. Rpt. as 2., durchgesehene Auflage. Innsbruck: Institut für Sprachen und Kulturen der Universität Innsbruck, 2007. Schwyzer, Eduard (1939). Griechische Grammatik. Auf der Grundlage von Karl Brugmanns Griechischer Grammatik. Vol. 1: Allgemeiner Teil. Lautlehre – Wortbildung – Flexion. München: Beck. Solta, Georg R. (1960). Die Stellung des Armenischen im Kreise der indogermanischen Sprachen. Eine Untersuchung der indogermansichen Bestandteile des armenischen Wortschatzes. Wien: Mechitharisten. Solta, Georg R. (1963). “Die armenische Sprache”. In: Armenisch und kaukasische Sprachen. Ed. by B. Spuler. Leiden & Köln: Brill, 80–128. Solta, Georg R. (1990). “Die Stellung des Armenischen im Kreise der indogermanischen Sprachen. Ein Überblick”. In: Armenien. Frühzeit bis 1. Jahrtausend. Sprache, Kunst und Geschichte. Ed. by Erwin M. Ruprechtsberger. Linz: Stadtmuseum Linz, 7–18. Sowa, Wojciech (2006). “Greckie αὐχήν, ἀμφήν i ormiańskie awjikʿ”. [“Greek αὐχήν, ἀμφήν and Armenian awjikʿ”]. In: In the Orient Where the Gracious Light … Satura Orientalis in honorem Andrzej Pisowicz. Ed. by Anna Krasnowolska, Kinga Maciuszak & Barbara Mękarska. Kraków: Księgarnia Akademicka, 173–177. Stempel, Reinhard (2000). “Armenisch und Indoiranisch”. In: Indoarisch, Iranisch und die Indogermanistik. Arbeitstagung der Indogermanischen Gesellschaft vom 2. bis 5. Oktober 1997 in Erlangen. Ed. by Bernhard Forssman & Robert Plath. Wiesbaden: Reichert, 515–522. Vine, Brent (1999). “On ‘Cowgill’s Law’ in Greek”. In: Compositiones Indogermanicae in honorem Jochem Schindler. Ed. by H. Eichner & H. C. Luschützky. Praha: Enigma, 555–600. Viredaz, Rémy (2003). “The Great Armenian Puzzle is not thorny”. In: Annual of Armenian Linguistics 22–23, 59–81. Viredaz, Rémy (2015). “Les désinences arméniennes de prétérit singulier actif”. Paper presented at the Armenian Aorist Workshop, Saint Petersburg, 14 May 2015. Watkins, Calvert (1969). Indogermanische Grammatik. Vol. 3.1: Formenlehre. Geschichte der indogermanischen Verbalflexion. Heidelberg: Winter. Watkins, Calvert (2001). “An Indo-European linguistic area and its characteristics. Ancient Anatolia. Areal diffusion as a challenge to the comparative method?” In: Areal Diffusion and Genetic Inheritance. Problems in Comparative Linguistics. Ed. by Alexandra Y. Aikhenvald & R. M. W. Dixon. New York & Oxford: Oxford University Press, 44–63. Winter, Werner (1975). “Die Personalendungen des Imperfekts und Aorists im Armenischen”. In: Zeitschrift für vergleichende Sprachforschung 89.1, 110–122. Woodhouse, Robert (2015). “Some observations on the putative dual reflexes of PIE *CRHC in Greek and Armenian, Francis’ Law and Greek αὐχήν ‘neck, etc.’” In: Studia Etymologica Cracoviensia 20, 257–272. Wyatt, William F. (1982). “Lexical correspondences between Armenian and Greek”. In: Annual of Armenian Linguistics 3, 19–42. Brought to you by | The University of British Columbia Library Authenticated Download Date | 9/24/18 2:51 PM Brought to you by | The University of British Columbia Library Authenticated Download Date | 9/24/18 2:51 PM