Cyril Aslanov, The Hebrew University of Jerusalem NORTHWEST SEMITIC STRUCTURAL INFLUENCES ON ARCHAIC GREEK: A REASSESSMENT In order to deal with the impact of non-Indo-European linguistic systems on the crystallization of the Greek linguistic system, it is necessary to distinguish between the pre-Hellenic Aegean substrate represented by linear A and the Northwest Semitic adstrate that is perhaps no less important than the substratic impact of the Aegean, especially as far as post-Mycenaean Greek is concerned. Since so little is known about the Aegean language once spoken in the region, it is preferable to focus on the Northwest Semitic structural infl uence on the development of Greek during the apogee of the Oriental infl uence on Archaic Greek civilization. Focusing on the Semitic infl uence on the development of Greek does not preclude the consideration that in some cases the convergence process between Archaic Greek and Phoenician within what we may call the Eastern Mediterranean Sprachbund relies on the same substrate constituted by the Prehellenic non-Indo-European and non-Semitic language(s) superseded by the expansion of Northwest Semitic westwards and Greek eastwards. To this Aegean language or language family belong the language represented by Linear A and probably also the Philistine language 1 and possibly even Etruscan 2. Th is paper examines some features of Archaic Greek that constitute an innova- tion with respect to Mycenaean Greek, attributing them to the impact provoked by intense contact with Northwest Semitic languages during the period of the Dark Ages, an epoch when the infl uence exerted by the Phoenician world was also felt outside the language system. My preference was to focus on structural features rather than lexical borrow- ing. Apart from the fact that the latter dimension has already been thoroughly 1 Giovanni Garbini, I filistei : gli antagonisti di Israele. Milan: Rusconi, 1997. P. 239, 241. 2 Giulio M. Facchetti, Appunti di morfologia etrusca con un’appendice sulla questione delle affinità genetiche dell’etrusco. Florence: Leo S. Olschki, 2002. P. 147–150. Judaica Petropolitana № 1’ 2013 17 investigated 3, it is not specifi c to Archaic Greek, as demonstrated by the presence of Semitic-borrowed lexical items in Mycenaean. Conversely, all the structural features that we examine here appear to be post-Mycenaean innovations. Th ese innovations can hardly be imagined as deriving from an internal development during the transition from Mycenaean to Archaic Greek, especially if we consider that some of them are restricted to Ionic or to the continuum represented by Ionic and Lesbian. As such, they appear to be conditioned by their location in a specifi c geographical area that constitutes the easternmost fringe of the Greek-speaking world, a region that was exposed more than any other to contact with Phoenician. Recent trends in the study of contact linguistics have stressed the importance of contact-induced diachronic changes 4. Th e reassessment of the importance of contact-induced change among dia- chronic changes in general could contribute to fi nding an external explanation for some features that have not received satisfactory internal explanation. 1. Phonological infl uences: from the impact of a common substrate to the constitution of an Eastern Mediterranean Sprachbund 1.1. Vowels So far, the attempt to reconstruct the vocalism of the pre-Hellenic substratic language has focused on qualitative alternations in Greek terms of non-Indo- European origin 5. We would like to transcend the narrow frame of isolated words, to analyze the vocalic system as a whole and to uncover what can be considered an innovation with respect to the Indo-European legacy. Th e main point pertains to an unusual correlation between vowel quantity and vowel quality, the Indo- European-inherited median vowels being reinterpreted as long open vowels (η and ω), whereas their short counterparts ε and ο seem to have been realized as mid-closed 6. In most dialects of ancient Greek, the Indo-European-inherited [ē] and [ō] were more open than their short counterparts [e] and [o]. Th is blatant dissym- metry between the Indo-European-inherited [ɛ:] and [ɔ:] on the one hand and 3 Heinrich Lewy, Die semitischen Fremdwörter im Griechischen. Berlin, 1895 (repr. Hil- desheim-New York, 1970); Émilia Masson, Recherches sur les plus anciens emprunts sémitiques en grec. Paris: Klincksieck, 1967; Cyril Aslanov, Emprunt enthousiaste et emprunt réticent: essai de typologie de différentes attitudes devant l’emprunt lexical à travers l’exemple du grec et du japonais / Meta. Translators’ Journal. 1995. Vol. 40. No. 4. P. 540–547. 4 Frans van Coetsem, A General and Unified Theory of the Transmission Process in Language Contact. Heidelberg: Carl Winter Universitätsverlag, 2000. 5 Edzard J. Furnée, Die wichtigsten konsonantischen Erscheinungen des Vorgriechischen. The Hague; Paris: Mouton, 1972. P. 335–385. 6 Michel Lejeune, Phonétique historique du grec ancien et du mycénien. Paris: Klincksieck, 1972. P. 236–237. 18 Cyril Aslanov, The Hebrew University of Jerusalem their short counterparts on the other hand is confi rmed by the fact that the result of the compensatory lengthening of short [e] and [o] was the introduction of the two closed long median vowels [ē] and [ō] 7. Th e contrast between the inherited [ē] and [ō] and the “spurious diphthongs” ει and ου can be symbolized as follows: [ɛ:] = <η> vs. [e:] = <ει> < -εν(ς) or –εε- [ɔ:] = <ω> vs. [o:] = <ου> < -ον(ς ) or –εο-/ -οο- Now, the tendency to open the long vowel inherited from Proto-Indo-European contradicts a drift attested in many Indo-European languages. According to this general tendency median long vowels tend to be realized more closed than their short correlates. Th is departure from an Indo-European tendency could result from the interference of a Pre-Hellenic substrate where the correlation between openness and length was entirely diff erent from that of Indo-European. Th e fact that the Th essalian and Boeotian counterparts of [ɛ:] <η> and [ɔ:] <ω> were [e:] <ει> and [o:] <ου> shows that the correlation between openness and length shared by most Greek dialects was not kept in those Aeolic dialects 8. However, in most ancient Greek dialects that correlation was felt quite clearly, as shown by the merging of the Indo- European-inherited [ē], that is, [ɛ:], and [ae:] < [a:] and even more obviously by the opening of [ɛ:] <η> in ᾱ in Elean 9. If the result of the shift of ᾱ to η was phonemically identical with the etymological η and diff erent from the compensatory lengthening of ε, there is no doubt about the open nature of the refl exes of Indo-European [e:] in most ancient Greek dialects and consequently in common Greek. Th e correlation between openness and length as far as the median vowels [e] and [o] are concerned is probably the refl ection of a structural interference exerted by a language where the phonemes [ē] and [ō] tended to be pronounced [ɛ:] and [ɔ:] respectively. Th is special open nature of the long median vowels may be due to the action of the Aegean substrate where the openness of [ɛ:] might have been the result of the palatalization of [a:] in a way that anticipated the Ionic shift of ᾱ to η. As far as the Ionic shift is concerned, a strictly internal explanation has been proposed 10. According to this assumption, the palatalization of ᾱ to η was alleg- edly the consequence of the appearance of a new ᾱ as a result of the compensatory lengthening of α < -αν(ς) and of the contraction α + ε. However, this hypothesis is not very satisfactory since the appearance of a new ᾱ is by no means limited to the Ionic dialects. Th erefore, it is diffi cult to make a connection between the two 7 Philomen Probert, “Phonology” // A Companion to the Ancient Greek Language / Egbert J. Bakker (ed.). Malden, MA; Oxford: Wiley-Blackwell, 2010. P. 96–97. 8 Friedrich Bechtel, Die griechischen Dialekte, 2nd edition. Berlin: Weidmannsche Verlagsbuch- handlung, 1963. Vol. I. P. 136; 218–220; Carl D. Buck, The Greek Dialects. Chicago; London: University of Chicago Press, 1955. P. 25, 27. 9 Bechtel, Die griechischen Dialekte. Vol. III. P. 829; Buck, The Greek Dialects. P. 25. 10 Stephen Colvin, A Historical Greek Reader: Mycenaean to the Koiné. Oxford; New York: Oxford University Press, 2007. P. 10. Judaica Petropolitana № 1’ 2013 19 phenomena, that is, the almost Pan-Hellenic creation of a secondary ᾱ on the one hand and the strictly Ionic-Attic shift from α to η on the other hand. Interestingly enough the closing of [a:] is attested in Canaanite though along the range of the back vowels rather than the front vowels. In this Northwest Semitic language, almost every long [a:] was rounded in [o:]. If both shift s — the non- conditioned Ionic closing of ᾱ to η and the conditioned Canaanite shift from [a:] to [o:] — are to be ascribed to the same substratic infl uence, it could be that the non-Indo-European and non-Semitic language that interfered with both Ionic and Canaanite had a vowel system where the feature of vocalic length was compatible only with the median vowels but not with [a]. Th is language seems to have exerted a deep substratic infl uence on Ionic Greek where all the occurrences of inherited ᾱ were closed to η and a weaker substratic infl uence on Canaanite where the shift from [a:] to [o:] was restricted to certain conditions. In spite of these diff erences, it is tempting to assume that a common substrate was responsible for three shift s: 1) the reinterpretation of the Indo-European-inherited [e:] and [o:] as open long vowels as a result of the adoption of common Greek by people speaking a language where the feature of length was compatible only with open median vowels (probably as a result of a previous shift from [a:] to [ae:] and [ɛ:] and of a velar [a:] to [ɔ:]). Th is feature seems to have been felt less in Th essalian and Boeotian, probably because of their belonging to the Aeolic dialect considered by some scholars a branch of Achaean in Mycenaean times 11. To be sure, Lesbian seems to have behaved diff erently, since it was infl uenced by its contact with Ionic, a dialect that had manifold impact on the crystallization of Lesbian as a special dialect within the Aeolic group. Th e Th essalian and Boeotian treatment of [e:] and [o:] is suffi cient reason for believing that the opening of the Indo-European-inherited [e:] and [o:] was not a Panhellenic feature already extant in Common Greek, but rather a feature confi ned to a certain area in spite of its extension to an overwhelming majority of dialects. It is likely that at the stage of Mycenaean Greek, a diff erence was already felt in the realization of the long median vowels. Whereas part of the Mycenaean world did share the trend toward the opening of the Indo-European-inherited [e:] 11 On the relationship between Mycenaean Greek and (Proto-) Aeolic, see the manifold views expressed by Ernst Risch. La position du dialecte mycénien // Études mycéniennes (Actes du Colloque International sur les textes mycéniens, Gif-sur-Yvette, 3–7 avril 1956) / Michel Lejeune (ed.). Paris: CNRS, 1956. P. 167–172; Vladimir Georgiev. Mycenaean among the Other Greek Dialects // Mycenaean Studies (Proceedings of Third International Colloquium for Mycenaean Studies Held at ‘Wingspread’, 4–8 September 1961) / Emmett L. Bennett, Jr. (ed.). Madison: The University of Wisconsin Press, 1964. P. 125–139; Mycénien et homérique: le problème du digamma // Proceedings of the Cambridge Colloquium on Mycenaean Studies / Leonard R. Palmer and John Chadwick (eds.). Cambridge: The University Press, 1966. P. 104–124; C. J. Ruijgh. Études du grec mycénien. Amsterdam: Adolf M. Hakkert, 1967. P. 35–41; Antonín Bartoněk, Greek Dialectology after the Decipherment of Linear B // Studia Mycenaea: Proceedings of the Mycenaean Symposium / Antonín Bartoněk (ed.). Brno: Universita J. E. Purkyně,1968. P. 37–51. 20 Cyril Aslanov, The Hebrew University of Jerusalem and [o:], other parts therein — later continued by Aeolic, and more specifi cally by Th essalian and Boeotian — did not display this feature and kept realizing the [e:] and [o:] as closed median vowels. 2) Th e Ionic shift from ᾱ to η can be considered a prolongation of the hypo- thetic Prehellenic shift from [a:] to [ae:] and [ɛ:]. Th e fact that Ionic partially reverberated this substratic trend indicates that the epicenter of the substratic infl uence must have been related to the shores of the Aegean where the various Ionic dialects and the Attic developed in post-Mycenaean times. 3) Th e possibility of ascribing the Canaanite shift from [a:] to [o:] to the same substratic or areal infl uence that was at work in the Ionic shift from [a:] to [ɛ:] is given additional credence by the fact that this process seems to have started already in the fourteenth century BCE, as evidenced by the Tel al-Amarna tablets 12. 1.2. Consonants Another feature shared by both Archaic Greek and Canaanite is the dropping of digamma in Ionic-Attic and Lesbian 13 and in some Doric dialects of the Aegean (Th era, Rhodes, Cos). Th e fact that Cyprian did not share in this trend may be ascribed to its belonging to the Arcado-Cyprian dialectal group that is far more conservative than the highly innovative Ionic group. Some data seem to contradict the assumption of a correlation between the substratic impact of the Prehellenic language and the disappearing of digamma. If we suppose that from a graphemic vantage-point there was a good deal of con- tinuity between linear A and linear B, then the language represented by linear A certainly had a phoneme [w], as shown by the existence of an almost complete series <wa>, <we>, <wi>, <wo> in linear B, not to speak of the syllabograms with an initial consonantal cluster: <nwa>; <dwe>; <dwo>; <twa>/ <two>; <swa>; <swe>; <swi>; <twe> 14. Furthermore, the syllabograms <wa> (54) and <wi> (40) of linear B are found with an identical value in linear A 15. Th e dropping of [w] might not be due to the substratic infl uence exerted by the Prehellenic Aegean language. Nevertheless, it is beyond doubt that in this evolution the aforementioned Eastern dialects of Archaic Greek and Canaanite are united by an isogloss that consists in the dropping of the bilabial semi-consonant 12 Anson F. Rainey, Canaanite in the Amarna Tablets: A Linguistic Analysis of the Mixed Dialect Used by the Scribes from Canaan. Leiden: E. J. Brill, 1996. Vol. I. P. 48. 13 Buck, The Greek Dialects. P. 46; Albert Thumb, Handbuch der griechischen Dialekte, 2nd edn. Heidelberg: Carl Winter, 1959. Vol. II. P. 93, 261–262. 14 On the representation of [w] in linear B, see Carlo Gallavotti, Le grafie del wau nella scrittura micenea // Mycenaean Studies (Proceedings of Third International Colloquium for Mycenaean Studies Held at ‘Wingspread’, 4–8 September 1961) / Emmett L. Bennett, Jr. (ed.), Madison: The University of Wisconsin Press, 1964. P. 57–65. 15 David W. Packard, Minoan Linear A. Berkeley; Los Angeles; London: University of California Press, 1974. P. 96. Judaica Petropolitana № 1’ 2013 21 [w] in certain positions within the word. Moreover, it seems that in Mycenaean itself, there were already some tendencies to drop [w] in certain positions 16. Th ose isolated phenomena may have been the prodromes of a tendency that asserted itself in post-Mycenaean times in the dialects spoken in an area that happened to be in contact with some representatives of Northwest Semitic. It should be noted that in Canaanite in general and in Phoenician in particular 17, [w] was a fragile phoneme, as shown by its disappearing in intervocalic position (e.g. Hebrew *qāwam > qām “he rose”) and its shift to [j] at the beginning of the word (e.g. Hebrew *wayn > yayn “wine”; *wāted > yāted “peg”). Both phenomena are the result of a restriction in the distribution of [w] in intervocalic position with the diff erence that within the word, [w] was dropped altogether whereas at the begin- ning of the word, the need for word boundary probably precluded its complete dropping. However, it did not prevent [w] from shift ing into [j]. As far as Greek is concerned, the dropping of digamma was felt more strongly and earlier in Ionic, that is, precisely the area where the shift ᾱ to η took place in post-Mycenaean times. Curiously enough, in Canaanite itself, the disappearing of intervocalic [w] coexists with the shift that aff ected [a:]. We already mentioned the similarity between the Ionic-Attic shift from ᾱ to η and the Canaanite shift from [a:] to [o:] and we ascribed it to the same substrate. Here, however, we can only state that two similar processes occurred in the area of contact between the easternmost edge of Archaic Greek and the westernmost fringe of Semitic. Even though the dropping of [w] is probably not connected with the Aegean substrate, it could nevertheless be considered the eff ect of a convergence process of two languages in contact in the same Eastern Mediterranean Sprachbund. It is diffi cult to determine whether the epicenter of the dropping of intervocalic [w] was located in Greek or in Canaanite. We already overturned the assumption that a substratic infl uence could have been infl uential in this shift . However, in another instance at least, there is little doubt that the innovation came from the Greek side. I refer to the Phoenician shift from [∫] to [s] that seems to have started in the fi ft h century BCE or even earlier 18. It is likely to have started in Cyprus, a well-known point of contact between the Greeks and the Phoenician 19. Th e unidirectionality of this infl uence is further confi rmed by the fact that Greek never possessed a postalveolar fricative whereas this phoneme is well attested in all Semitic languages. Th e result of the contact between Archaic Greek and Phoeni- cian, the northwest outpost of Canaanite, was a singular lisping that consisted in pronouncing every shin as sin. It is important to note that this Phoenician shift from 16 Vladimir Georgiev, Mycénien et homérique: le problème du digamma // Proceedings of the Cambridge Colloquium on Mycenaean Studies / Leonard R. Palmer and John Chadwick (eds.). Cambridge: The University Press, 1966. P. 104–124. 17 Stanislas Segert, A Grammar of Phoenician and Punic. Munich: C. H. Beck, 1976. P. 71–72. 18 Charles R. Krahmalkov, A Phoenician-Punic Grammar. Leiden; Boston; Cologne: Brill, 2001. P. 25. 19 Segert, A Grammar of Phoenician and Punic. P. 64. 22 Cyril Aslanov, The Hebrew University of Jerusalem [∫] to [s] has nothing to do with the Shibboleth episode related in Judges 12:6. What was at stake there was most probably a contrast between the interdental [θ] and [∫] 20. Another possible infl uence of Greek on Phoenician is the realization of the aff ricate [dz] as [zd]. Whatever the original value of the letter zain in Proto- Canaanite — a spirant [z] or an aff ricate [dz] — , it seems to have had the value of [zd] in Phoenician 21, probably as a result of the metathesis of the two elements of the aff ricate. A similar reversal occurred in Archaic Greek where the original aff ricate [dz] represented by <ζ> was realized [zd], as shown by the occasional spelling <σδ> that was even retained in the textual tradition of Lesbian lyric and as confi rmed by the use of <ζ> in order to represent the sequence <σδ> in Ἀθήναζε<Ἀθήνας-δε “to Athens”. Since the metathesis [dz] > [zd] is attested only in Phoenician, it is tempting to ascribe it to contact with Archaic Greek where the same metathesis seems to characterize all the dialects since the post-Mycenaean time at least (in Mycenaean Greek, the original value [dz] is likely to have been retained) 22. 2. Two possible manifestations of Semitic infl uences in the fi eld of Greek morphology 2.1. Nominal morphology: syncretism in the declension system A possible manifestation of the impact resulting from the contact of Archaic Greek with Semitic languages is the drastic reduction of the eight cases of Proto-Indo-European or of the six cases of Mycenaean Greek 23 to only fi ve in Archaic and Classical Greek as a result of a syncretistic process that is not necessarily attributable to strictly internal factors. In our opinion, the contact of Indo-European languages with non-Indo-European languages was responsible for the preservation of a rich case system or conversely to its reduction or even disappearance. A valuable touchstone that may help evaluate the amplitude of the commotion provoked by the encounter between Indo-European and Semitic languages is 20 Ephraim A. Speiser, The Shibboleth Incident (Judges 12:6) // Bulletin of the American Schools of Oriental Research. 1942. Vol. 85. P. 10–13 (= Oriental and Biblical Studies: Collected Writings of E. A. Speiser / J. J. Finkelstein and Moshe Greenberg (ed.). Philadelphia: University of Pennsylvania Press, 1967. P. 143–150); Ian Young, Diversity in Pre-Exilic Hebrew. Tübingen: Mohr-Siebeck, 1993. P. 188. 21 Segert, A Grammar of Phoenician and Punic. P. 113–114; Krahmalkov, A Phoenician-Punic Grammar. P. 21–22. 22 Michel Lejeune, Les sifflantes fortes du mycénien // Lejeune. Mémoires de philologie mycénienne, deuxième série (1958–1963). Rome: Edizioni dell’Ateneo, 1971. P. 132–133; Phonétique historique du mycénien et du grec ancien, 1972. P. 113. 23 For an interpretation of the Mycenaean case system, see Oswald Szemerényi, The development of the –o-/ -ā- stems in the Light of the Mycenaean Evidence // Proceedings of the Cambridge Colloquium on Mycenaean Studies / Leonard R. Palmer and John Chadwick (eds.). Cambridge: The University Press, 1966. P. 222–225; Ivo Hajnal, Studien zum mykenischen Kasussystem. Berlin; New York: Walter de Gruyter, 1995. Judaica Petropolitana № 1’ 2013 23 the process of erosion that aff ected nominal declension in other Indo-European languages in contact with allogeneous systems. Th e most striking parallel is provided by the Indo-Iranian branch. Whereas Middle Indo-Aryan languages still preserve a wide range of cases in spite of the syncretism between genitive and dative as well as between instrumental and ablative, Middle Persian already reduced its case system to an opposition between casus rectus and obliquus, like in Modern Indo-Aryan languages where the reduction of the range of cases may be the result of internal factors, as in the evolution of ancient to modern Greek. However, one should also consider attributing those relatively recent phenomena to the contact of Modern Indo-Aryan languages with Farsi, especially as far as Hindustani is concerned. To be sure, other Middle Iranian languages like Sogdian and Khotanese pre- served a more articulated nominal declension. However, this conservatism may be ascribed to the relative isolation of those languages relegated to a peripheral zone where the dynamics of linguistic change and grammatical erosion may have been somewhat hampered. More specifi cally, those Iranian languages of Central Asia were not in contact with Semitic languages excluding the occasional encounter on the Silk Road of Sogdian and Aramaic during the fi rst centuries of the Christian era. In order to properly evaluate the impact of language contact on the preser- vation vs. disappearance of nominal declension, one has to think of an Iranian language like Ossetian that has no less than eight cases. Th is exceptional resistance to the structural drift that consists in the gradual decay of the case system is prob- ably due to the contact of Ossetian with Caucasian languages that are character- ized by a strong declension system (though agglutinative rather than fl exional). Aft er this excursus through a cognate branch of the Indo-European language family, let us apply the causality of language contact in order to understand the erosion that aff ected the Greek declension system since the earliest contacts between Greek and Northwest Semitic. Th e fi rst step in the reduction occurred at the stage of Mycenaean Greek, probably the result of contact with the Proto- Aegean substrate. However, the next step in the reduction occurred when Archaic Greek was already exposed to contact with Phoenician, a language that had almost dropped the last remnants of the Proto-Semitic-inherited declension 24. During the transition from Mycenaean to Archaic Greek two concrete cases (instrumental and locative) merged with the dative, a more abstract case. Th e system thus obtained was obviously not identical with the Proto-Semitic three- fold system consisting only in nominative, accusative and genitive. However, we may perceive in the simplifi cation of the declension system toward a more abstract set of morphological oppositions the result of a convergence process that reached its peak with the loss of the dative during the transition from the Koiné to medieval Demotic Greek 25. 24 Krahmalkov, A Phoenician-Punic Grammar. P. 123. 25 David Holton and Io Manolessou, Medieval and Early Modern Greek // A Companion to the Ancient Greek Language / Bakker (ed.). P. 546–547. 24 Cyril Aslanov, The Hebrew University of Jerusalem Th e gradual reduction of the gamut of cases during the various stages of the history of Greek led to a situation where the fi nal result (a three-fold system) is identical with the Proto-Semitic system that was no longer actual at the time of contact between Archaic Greek and Phoenician. 2.2. Verbal morphology: A Semitic-like combination of Indo-European-inherited morphemes? One of the most vexing quandaries of the historical morphology of Greek is the question of the origin of the passive aoriste morpheme -θ- that was combined with the ending -η- in order to create the very productive infi x –θη-. Chantraine suggested two possible Indo-European etymologies for this strange morpheme, but he stressed that neither was really demonstrable 26. I would like to propose a hypothesis that attributes the use of this mysterious morpheme to the accusional abolition of the boundaries between the nominal and the verbal systems. By itself, the suffix –η- was sufficient to express the value of the aorist intransitive, the passive aorist being one of the specifications of this basic intransitive value. Thus the adding of the infix -θ- may be interpreted as a way of expliciting the special passive value of the aorist. According to Émile Benveniste 27, the origin of this redundant agglutinative morpheme is identi- cal with the nominal suffix -θ- < Proto-Indo-European *–dh-. In his turn, Chantraine interpreted this morpheme as expressing an instrumental value or at least referring to a tool, especially when combined with -ρο-/ -ρᾱ — or with -λο- / -λᾱ - 28, as in βέρεθρον/ βάραθρον “abyss” derived from the root βερ- appearing in βιβρώσκω “to swallow”; κολυμβήθρα “swimming-place” derived from κολυμβάω “to swim”; γένεθλον/ γενέθλη “birth, descent” from the root γεν- of γίγνομαι “to come into being; to become”. Now, the adoption of -θ- in order to reinforce the stative-passive value of -η- is also an interesting way of crossing the boundaries between the nominal and the verbal systems. Th is blurring of the diff erences between parts of speech is something that is well known in Semitic languages where the preformants m- and t- are widely used in order to produce substantives (mainly with a resul- tative meaning). Th ose very same preformants are also very productive in the formation of the passive, precisely: e.g. the Arabic noun maf ‘alun formed with the same m- as the past participles maf ‘ūlun, mufa‘‘alun, mufā‘alun. To be sure, the participles are still part of the nominal system. However, the recursive use of the same morpheme in the nominal and verbal systems is beyond doubt in the case of the nouns preformed with the morpheme t- like taf ‘īlun. Indeed, 26 Pierre Chantraine, Morphologie historique du grec. Paris: Klincksieck, 1980. P. 167–170. 27 Émile Benveniste, Origines de la formation des noms en indo-européen. Paris: Maisonneuve, 1984. P. 196–197. 28 Pierre Chantraine, La Formation des noms en grec ancien. Paris: Champion, 1933. P. 372–375. Judaica Petropolitana № 1’ 2013 25 the morpheme t- reappears in the fi ft h, sixth and tenth form: tafā‘alа; ift a‘ala; istaf ‘ala (Akkadian iptaras; Hebrew hitpa‘‘el; Aramaic etpə‘el; etpa‘‘al; ettaf ‘al; eštaf ‘al) not only in the participle, but even in the conjugated forms of the verb. In Akkadian, in Hebrew and in the sixth and tenth form of Arabic the morpheme –t- is used as an infi x. Now, how should we consider the insertion of –θ- between the semantic base of the verb and the stative morpheme –η-? Isn’t it a clear case of infi xation, the result of which was the explicitation of the passive value of the verbal form? To be sure, the morphological implications of the infi xation are diff erent in Greek and in the various Semitic languages where –t- is used as a mark of passive or refl exive diathesis. However, the very principle of infi xing a morpheme in order to confer a non-active value to the verb may be considered a creative imitation of Semitic verbal morphology. To sum up, the infi xation of a nominal morpheme in order to explicit or to mark a diathetic value may be considered a manifestation of the convergence process by dint of which some morphological practices that were usual in Semitic were sometimes adopted by Archaic Greek. Th e double dimension involved in this imitation of mechanisms characteristic of Semitic grammar — namely the cross- ing of the boundaries between noun and verb and the infi xation with a diathetic value — is a clear example of the transtypological move that occurred in Greek during the transition from Mycenaean to Archaic Greek. 3. Syntactic calques 3.1. In the nominal microsyntagm Th e contact with Northwest Semitic languages may be responsible for a major innovation of Greek with respect to its Indo-European background, or more precisely, of Archaic Greek with respect to Mycenaean. It is the developing of the article on the basis of the demonstrative element ὁ, ἡ, τό (< *so, *sā / to) as a defi - nite article 29. Th is microsyntactic element was not really required by the internal structure of prehistoric Greek. I assume that this innovation was not motivated by internal factors and that it is rather the result of the protracted contact between Archaic Greek and Phoenician. In Canaanite in general and in Phoenician in particular, the article ha- con- tinuing a former *han is well attested since 1000 BCE 30. It is possible that this marking innovation of Canaanite was motivated by the need to compensate for the aforementioned loss of the declension ending. However, such motiva- tion was irrelevant in Greek where the nominal declension has been preserved 29 On the move from the deictic value of ὁ, ἡ, τό to its use as an article, see Antoine Meillet, Aperçu d’une histoire de la langue grecque. 6th edn. Paris: Hachette, 1948. P. 179–185. 30 Segert, A Grammar of Phoenician and Punic. P. 107–108; Krahmalkov, A Phoenician-Punic Grammar. P. 86–87. 26 Cyril Aslanov, The Hebrew University of Jerusalem