Phoenician and Luwian in Early Iron Age Cilicia

Anatolian Studies 65 (2015): 35–53 © British Institute at Ankara 2015 doi:10.1017/S0066154615000010 Phoenician and Luwian in Early Iron Age Cilicia Ilya Yakubovich Russian State University for the Humanities, Moscow, Russian Federation [email protected] Abstract The relationship between the Luwian and Phoenician versions of the bilingual texts emanating from Cilicia has never been systematically studied from the philological viewpoint. In this paper I endeavour to demonstrate that a converging set of formal arguments supports the primary character of the Phoenician versions of the ÇİNEKÖY and KARATEPE 1 bilinguals and the secondary character of their Luwian versions. I interpret this as a metaphor for the relationship between two ethnic constituents of the Neo-Hittite principality of Que, whose coexistence was earlier argued for on independent grounds. According to the proposed interpretation, the Phoenician language was emblematic of the rulers of Que, who claimed Greek descent and therefore attempted to distance themselves from the traditional elites of the neighbouring Neo-Hittite states. The use of the Luwian language was a concession to the indigenous population of Que. The adoption of Phoenician as a language of written expression by the Greek colonists in Cilicia happened at the point when the Linear B script had been forgotten and represented the first step toward the creation of the Greek alphabet. Özet Kilikya’da bulunan iki dilli metinlerin Luvice ve Fenikece versiyonları arasındaki ilişki hiçbir zaman filolojik bakış açısıyla sistematik olarak incelenmemiştir. Bu makalede, ÇİNEKÖY ve KARATEPE 1 iki dilli metinlerinin Fenikece versiyonlarının birincil özelliğini ve onların Luvice versiyonlarının ikincil özelliğini destekleyen bir dizi resmi argümanın kesiştiği gösterilmeye çalışılmıştır. Bu, daha önce başka sebeplerle tarafımızca savunulduğu gibi, Neo-Hitit dönemi Que prensliğinde birlikte yaşamış olan iki etnik yapının arasındaki ilişki için bir metafor olarak yorumlanmaktadır. Önerilen yoruma göre Fenike dili, Yunan soyundan geldiklerini iddia eden ve bu nedenle kendilerini komşu Neo-Hitit devletlerinin gelenekçi elitlerinden ayrı tutan Que yöneticilerinin bir simgesi olmuştur. Luvi dilinin kullanımı Que yerli halkı için istisnai bir durumdur. Kilikya’da Yunan kolonistlerce Fenike dilinin yazılı anlatım dili olarak kabul edilmesi, Linear B yazısının unutulduğu dönemde ortaya çıkmış ve Yunan alfabesinin oluşturulması yönünde ilk adımı temsil etmiştir. T were presumably known as ḥittīm, ‘Hittites’ in the Hebrew tradition, although the actual usage of the ethnonym ḥittīm in the Bible is somewhat blurred (Bryce 2012: 67–75). This designation is not ethnically based, but rather reflects the lingering geographic association of the Neo-Hittite polities with the empire of Hattusa. Judging by local inscriptions, the actual ethnic identity of the Neo-Hittite elites was rather heterogeneous. The ruling dynasties of Carchemish and Melid, for example, traced back their genealogies to the rulers of Hattusa and composed their hieroglyphic inscriptions in the Luwian language (Bryce 2012: 83–84, 98–99). These states can be regarded as the true cultural heirs to the empire of Hattusa, where the Luwian language came first to be associated he principality of Que, known under such a name from Neo-Assyrian sources and situated on the Cilician plain, was one of the so-called Neo-Hittite states. The term ‘Neo-Hittite’ is commonly applied to a number of polities which arose in the territory of southwestern Anatolia and northern Syria after the collapse of the empire of Hattusa in the early 12th century BC. Eventually most of them were absorbed into the Assyrian Empire by the end of the eighth century BC (with instances of sporadic re-emergences in the seventh century). This term is essentially an exonym, used in recollection of the fact that the lands stretching from Melid in the north to Hama in the south were collectively called Ḫatti by their eastern neighbours, the Assyrians (Bryce 2012: 52). Their inhabitants 35 Downloaded from https://www.cambridge.org/core. Universitaetsbibliothek Marburg, on 16 Sep 2017 at 11:25:10, subject to the Cambridge Core terms of use, available at https://www.cambridge.org/core/terms. https://doi.org/10.1017/S0066154615000010 Anatolian Studies 2015 with hieroglyphic inscriptions (Yakubovich 2010a: 297– 99). A very different sociolinguistic situation is attested in Sam’al, where the official inscriptions were composed in three Semitic languages: Phoenician, Ya’udic (Sam’alian) and standard Old Aramaic (Tropper 1993: 5). The names of the local rulers are in part Anatolian and in part Semitic, but the name of the country literally means ‘north’ in Aramaic, and therefore its formation must have been connected with the northward migrations of Aramaean tribes (Bryce 2012: 169–75). An intermediate situation can be observed in Hama, where rulers with Anatolian names issued Luwian inscriptions in the ninth century BC, but a dynasty of Aramaean origin held sway a century later (Bryce 2012: 134–38). In comparison with the above cases, the sociolinguistic situation in Que is considerably harder to interpret. Apart from a handful of seals, whose authenticity and provenance cannot always be reliably established, it can be studied on the basis of the monumental inscriptions (or their groups), namely İNCİRLİ, KARATEPE, ÇİNEKÖY, HASSAN-BEYLİ and CEBELİREİS DAĞI. The first four of these monuments are commonly dated to the eighth century BC, while CEBELİREİS DAĞI probably originated a century later, at a point when one is no longer sure about the existence of the principality of Que. All five of them have Phoenician versions, İNCİRLİ, KARATEPE and ÇİNEKÖY have Luwian versions and the İNCİRLİ inscription also features an Akkadian version. On the other hand, no Semitic personal names are attested in these inscriptions in connection with local individuals, while the majority of names in the monumental deed of transfer CEBELİREİS DAĞI appear to be Luwian in origin (Payne 2006: 130). The name of Azatiwada, a de-facto ruler of Que and commissioner of the great KARATEPE inscription (KARATEPE 1), is also clearly Luwian, and so are the names of scribes mentioned in the related Luwian inscription KARATEPE 4. By contrast, no Luwian etymologies impose themselves for the royal names Awarku (KARATEPE, HASSAN-BEYLİ) and Waraika (İNCİRLİ, ÇİNEKÖY, CEBELİREİS DAĞI). A large group of scholars prefers to regard these two names as variant spellings of the same personal name *Awarika (cf. Gander 2012: 292–94), but even such a solution does not yield a convincing Luwian etymology of the hybrid name. It is a matter of general agreement that the indigenous population of Cilicia spoke a form of Luwian before the collapse of the empire of Hattusa (Bryce 2012: 153–54; cf. Yakubovich 2010a: 272–85). As mentioned above, the Luwian language continued to be in use in those NeoHittite principalities that maintained their local elites and harkened back to the old imperial traditions. If one assumes that Que represented one such state, it is reasonable to expect that Luwian was its principal official language. The written use of Phoenician could then be explained as an attempt to accommodate the international lingua franca of the time, the simpler and more accessible form of writing. This can be called the indigenist hypothesis. By contrast, if the Luwian-speaking groups lost their grip of Cilicia in the Neo-Hittite period, then one possible interpretation of the written use of Phoenician would be the assertion of a separate cultural identity by the new elites, in contrast to the rulers of the neighbouring states. The written use of Luwian alongside Phoenician could then be taken as a concession to the native population groups of Que. The assumption that outsiders took control of the Cilician plain after the collapse of the empire of Hattusa can be called the migrationist hypothesis. In order to discriminate between these two hypotheses, it is useful to take a closer look at the multilingual inscriptions emanating from the principality of Que. Since the Luwian version of the İNCİRLİ inscription remains unpublished and is apparently badly preserved, my analysis will focus on the KARATEPE 1 and ÇİNEKÖY bilinguals. I shall endeavour to demonstrate that a convergent set of arguments pleads for the primary character of the Phoenician version in both cases under consideration. I interpret this result as being better compatible with the migrationist hypothesis, i.e. supporting the existence of substantial non-Luwian components among the elites of Que. But before moving to a philological analysis it is appropriate to undertake a brief review of the migrationist hypothesis. According to the current state of the debate, the ethnic group commonly seen as supplying the new elites to the principality of Que is not that of the Phoenicians but of the Greeks. Although no local inscriptions in the Greek language have been found for the relevant place and period, a number of new discoveries and methodological advances made since the year 2000 have strengthened the case for Greek presence in Early Iron Age Cilicia. By contrast, there seem to be no arguments for significant Phoenician presence in Cilicia in the same period, apart from trade connections (cf. Lipiński 2004: 138–43). The strongest philological argument for the Greek migration hypothesis comes from the local attestations of the name of Mopsus. The seer Mopsus (Μόψος), who allegedly lived at the time of the Trojan War, was mentioned already apud Hesiod according to Strabo 14.1.27, but the later Greek tradition also portrays him as a founding figure. Although the mythological activities of Mopsus stretch all along the western and southern coasts of Asia Minor, his specific connection with Cilicia is reflected in the local Hellenistic toponyms Μοψουκρήνη and Μοψουεστία (Vanschoonwinkel 1990). The discovery of the KARATEPE inscriptions provided an earlier testimony for this association. It had been known since 36 Downloaded from https://www.cambridge.org/core. Universitaetsbibliothek Marburg, on 16 Sep 2017 at 11:25:10, subject to the Cambridge Core terms of use, available at https://www.cambridge.org/core/terms. https://doi.org/10.1017/S0066154615000010 Yakubovich Fig. 1. The northeastern Mediterranean in the Early Iron Age. 1948 that the ruling house of Que was called bt mpš, ‘the house of Mopsus’, in the Phoenician version of the KARATEPE 1 inscription, while later the name mpš was also identified in the ÇİNEKÖY and İNCİRLİ texts. The Luwian phrase, muksassan parni, ‘to the house of Mopsus’, yields the stem muksa- as an equivalent of Phoenician mpš (KARATEPE 1 § 21). As long as one resorts to purely philological arguments, the ‘Cilician Mopsus’ could be regarded as an instance of interpretatio graeca based upon an accidental similarity of the Greek and local Cilician names (thus, for example,Vanschoonwinkel 1990; Gander 2012). But an application of historical linguistic methodology yields different results. Since the counterparts of Hebrew ‫( שׂ‬sin) and ‫( שׁ‬shin) have been neutralised in the Phoenician sound system, Phoenician mpš can be regarded as a direct borrowing of Greek Μόψος (cf. Krebernik 2007: 129). By contrast, Luwian muksa- can only represent the adaptation of an earlier form of the same name, which is attested as mo-qo-so in Mycenaean transmission (KN De 1381.B and PY Sa 774: Ventris, Chadwick 1973: 562). The inherited labiovelars are preserved in Luwian up to the eighth century BC, and therefore the character of correspondence between the Phoenician and Luwian forms suggests that they reflect diverging paths of phonological adaptation of a Greek labiovelar (Oettinger 2008). One has to assume that the Greek name *mokwso- was taken over by the Luwian speakers before the disappearance of the Greek labiovelars (*mokwso- > muksa-), while the borrowing of this name into Phoenician postdates this process (*mopso> mpš). The former inference is all the more likely, given that the personal name Muksus is attested in an admittedly unclear context (KUB 14.1 rev. 75) in the Hittite 14thcentury composition known as ‘The Indictment of Madduwatta’, which makes frequent references to the activities of the land of Aḫḫiya- (Mycenaean Greece or its part) in southwestern Anatolia and Cyprus. In this case, too, one has to assume the adaptation of a Greek labiovelar before its disappearance (*mokwso- > mokso-, spelled muuk-šu), since the inherited Anatolian labiovelar would have been preserved in some form in Hittite and yielded the spelling **mu-uk-ku-šu (vel sim.). 37 Downloaded from https://www.cambridge.org/core. Universitaetsbibliothek Marburg, on 16 Sep 2017 at 11:25:10, subject to the Cambridge Core terms of use, available at https://www.cambridge.org/core/terms. https://doi.org/10.1017/S0066154615000010 Anatolian Studies 2015 Thus although the earliest attestation of the name Mopsus may come from Anatolian sources, linguistic considerations plead for its Greek origin. Furthermore, they find a good correlation in the archaeological record. The Late Helladic IIIc pottery, commonly dated to the 12th to 11th century BC, made its way in large quantities to the Cilician plain, whereas only small numbers of the earlier Late Helladic IIIa–b potsherds have been found in this region. Gunnar Lehmann (2007: 512) summarises this evidence with the claim that the ‘impact of Late Helladic ceramic styles on the material culture of Cilicia is comparable only to the Land of the Philistines and the ‘Amuq plain’. While the intrusive origin of the Philistines in southern Levant is reflected in the Biblical record, scholars lacked for a long time comparable internal textual confirmation for the ‘Amuq plain, and this complicated the direct link between philological and archaeological data in the instance of Cilicia. The situation has changed with linguistic advances leading to the identification of the land of Walistina in the Luwian texts (Rieken 2010; Rieken, Yakubovich 2010: 215–16) and the discovery of a variant Palistina for the same toponym (Hawkins 2011: 41a; but cf. already Harrison 2007). Part of the kingdom of Palistina/Walistina, probably to be interpreted as Falistina, was situated in the ‘Amuq plain, since this area yielded the Luwian inscription TELL TAYINAT 1 containing a mention of a Walistinean king (frs 3–5, l.1; Hawkins 2000: 2.366). Thus the philological record supports the presence of intrusive population groups in all the areas with the highest concentrations of Late Helladic IIIc pottery in the eastern Mediterranean. The migration of new population groups to the Levantine coast in the early 12th century BC is described in Egyptian sources as the ‘Sea People’ invasion. According to the Medinet Habu historical texts, the pharaoh Ramesses III fought a motley coalition of diverse ethnic groups, who had arrived from unspecified islands, crushed a number of states on their way and eventually set their camp in the land of Amurru. Many details of this account are apparently fanciful, but the Amurru connection is probably not, since the Late Bronze Age land of Amurru roughly overlaps with the Early Iron Age kingdom of Palistina/Walistina (cf. Kahn 2011). One of the ‘Sea People’ groups mentioned in the Medinet Habu texts is dnỉn, which is traditionally rendered as Denyen in secondary literature, but frequently identified with the Akkadian term danuna of the Amarna Letter 151, line 52. This mid 14th-century letter of Abi-Milku, king of Tyre, informs the Egyptian pharaoh, among other things, of the death of the king of Danuna and a fire in the royal palace of Ugarit (Moran 1992: 238). Since the reports of the vassal Levantine princes collected in the Amarna archive primarily deal with regional affairs, Danuna must have been located in the eastern Mediterranean, perhaps not far from Ugarit. This, in turn, corroborates its comparison with the dnnym, the designation of the population of Que in the Phoenician versions of the KARATEPE 1 and ÇİNEKÖY inscriptions (cf. Hawkins 2000: 2.40, n.21). As shown by Emmanuel Laroche, this Phoenician term goes back to the Luwian possessive adjective addanawan(ni)-, ‘Adanean’ (cf. Vanschoonwinkel 1990: 295–96 with references). This adjective is derived from the name of the Anatolian town which is attested as Adaniya in Hittite sources and still known as Adana today. If the triple comparison Denyen – Danuna – dnnym can be maintained, this implies that the inhabitants of Cilicia participated in the ‘Sea People’ coalition according to the opinion of Ramesses III (Bryce 2012: 154; cf. Adams, Cohen 2013: 659, n.12). This need not be seen as a contradiction with the etymologies proposed for other members of the ‘Sea People’ coalition, which support their connection with the Aegean (for example Ekwesh = ‘Achaeans’?) or even the western Mediterranean (for example Shekelesh = ‘Sicilians’?). The migrations of the ‘Sea Peoples’ are normally not perceived in modern scholarship as a coordinated long-range offensive campaign, but rather as a sequence of local ethnic movements caused by the collapse of respective complex societies (Killebrew, Lehmann 2013: 5–6; cf. Oettinger 2010). It is thus perfectly possible than the plain of Adana merely represented the last intermediate stop of the displaced population groups, before some of them continued their way east and pitched their camps in the ‘Amuq plain, while others consolidated in Cilicia under the rule of the ‘house of Mopsus’. At any rate, the proposed identification of Denyen is linguistically superior to its frequently cited alternative, which implies comparison with Greek Δαναοί, ‘Danaeans’, since the latter ethnonym is probably rendered as tnjw in the earlier Egyptian transmission (Adams, Cohen 2013: 658, n.11; cf. Latacz 2004: 130–31). A new possible argument for the Aegean origin of the Que elites emerged with the publication of the ÇİNEKÖY inscription. As we shall see below, the familiar ethnonym dnnym in the Phoenician version of the ÇİNEKÖY inscription turned out to correspond to the toponym Hiyawa (or Hiyawi) in its Luwian version. The similarity of this toponym with Aḫḫiyawa-, which was the designation of Mycenaean Greece or a part of it in the Hittite sources of the Empire period, did not elude the first editors of the inscription (Tekoğlu, Lemaire 2000: 981–84). They also took notice of the fact that the Cilicians were formerly called Ὑπαχαίοι, literally ‘Sub-Achaeans’, according to Herodotus 7.91. The mention of the Hiyawi-people in connection with Lukka/Lycia in an Akkadian letter sent to Ugarit by the last king of Hattusa, Suppiluliyama II, was 38 Downloaded from https://www.cambridge.org/core. Universitaetsbibliothek Marburg, on 16 Sep 2017 at 11:25:10, subject to the Cambridge Core terms of use, available at https://www.cambridge.org/core/terms. https://doi.org/10.1017/S0066154615000010 Yakubovich ‘Adana’, while the toponym á-*429-wa/i- is never written without <wa/i> in the same inscription. It is also worth mentioning that the toponym Adanawa is not attested with such a suffix in any other sources, while Luwian á-ta-nacorresponds to Phoenician ’dn. The attempts to find additional Greek elements in the Que inscriptions are to be judged each on its own value and probably would not have been advanced if their authors were not convinced on independent grounds that Greeks were present in Cilicia in the Early Iron Age. The least problematic is perhaps the suggestion that the name of the king Awarku (Phoen ’wrk, Luwian á-wa/i+ra/i-kusa) mentioned in the KARATEPE 1 and HASSAN-BEYLİ inscriptions represents an adaptation of the Greek personal name Ἔυαρχος, literally ‘fit for ruling’ (Krahmalkov 2000: 38–39). The main drawback of this hypothesis is the lack of independent evidence for such a name in a Cilician context, but it was frequent enough in various parts of the Greek-speaking world, including Cyprus (Lipiński 2004: 120–21). Also noteworthy is the attempt to find a form of the Greek name Ῥοῖκος, syllabic Cypriot wo-ro-i-ko, literally ‘crooked, lame’, behind the royal name Waraika (Phoenician wryk(s), Luwian wa/i+ra/i-i-ka-sá) in İNCİRLİ, ÇİNEKÖY and CEBELİREİS DAĞI, which is also attested as Urikki in Neo-Assyrian transmission (Lipiński 2004: 122). One may object that this would be a rather peculiar name to give to a royal offspring, but Edward Lipiński mentions that it was common among the Cypriote kings of Amathus in the Classical period. Philip Schmitz (2009) offers the interpretation of Phoenician krntryš, an epithet of Baal in KARATEPE 1, as Greek *κορυνητήριος, ‘mace-bearing’, with a reference to the attested Phoenician divinity b‘l ṣmd, ‘Baal of the Mace’. This etymology is somewhat weakened by the lack of actual attestations of the Greek form *κορυνητήριος, but the latter would represent a regular adjectival derivative of the attested κορυνήτης, ‘mace-bearer’. Less grounded, in my opinion, are the attempts of Schmitz (2008) to attribute Greek origin to the personal names nww˹l˺[x?]nmš, ṣṣš and ˹kl˺[x]pyš in the inscription Pho./S.I. belonging to the KARATEPE complex (Çambel 1999: 69). As the author himself acknowledges, his interpretation of these names as Ϝουλαμώνυμος, Τιτυός and the common noun κλινοποιός, ‘maker of beds or bedsteads’, competes with Anatolian etymologies in two of the three cases, and in each of these three instances his reasoning involves restitution of fragmentary names or linguistic speculations. On the other hand, one can point to a number of facts that cannot be regarded as direct arguments in favour of the migrationist hypothesis, but receive an elegant explanation if it is accepted. It is not to be forgotten in this connection that the territory of Que corresponded to a part initially perceived as an additional piece of supporting evidence for the etymological connection between the place names Hiyawa and Ahhiyawa (Singer 2006; cf Singer 2013). The Luwian endonym Hiyawa almost certainly underlies the exonym Que known to us from Assyrian sources (Simon 2011: 263). Recently, however, the proposed etymology of Hiyawa came under attack by Max Gander (2011: 48–56). Gander justly points out that the aphaeresis */axxijawa/ > /xijawa/ would have been irregular in the Luwian language of the 13th century BC, when the toponym or ethnonym Hiyawi is first attested. Even more importantly, he draws historical conclusions from the observation of Ivo Hajnal (2003: 41) that the fragmentary toponym Hiya[...] is attested in a Cilician context in the annals of Arnuwanda I, king of Hattusa (early 14th century BC). A restoration of this toponym as Hiyawa militates against its Greek origin, since there is no independent evidence for Greek settlements in southeastern Anatolia at such an early date. The arguments of Hajnal and Gander (further elaborated in Simon 2011) undermine to some extent the original etymological identity of Hiyawa and Ahhiyawa. Much, however, will depend on the reception of the recent proposal of Rostislav Oreshko (2013), who argues that the name of Ahhiyawa is attested without aphaeresis in KARATEPE 1. The sequence á-*429-wa/i- was traditionally read as *Adanawa- because of its Phoenician correspondences dnnym, ‘Danuneans’, and ‘mq ’dn, ‘plain of Adana’. At the same time, the use of different names – Adanawa (KARATEPE 1) and Hiyawa (ÇİNEKÖY) – for the same principality of Que, in the same language and within the same tradition, would be unique in the NeoHittite world and is acknowledged as a puzzle (Bryce 2012: 154). Oreshko argues that the sign *429, which does not occur in the first millennium corpus outside KARATEPE 1, need not be read as <DANA>, but can be assigned a value <HIYA> in this inscription. In support of his suggestion he adduces the graphic similarity between *429 and the second-millennium sign *306, which has the syllabic value <hí>. Oreshko’s arguments cannot be fully evaluated before the publication of the recently found ARSUZ inscriptions where the logogram *429 and the syllabic spelling Hiyawa occur side by side (see now Dinçol et al. 2015). Philological discussion will determine whether these can be taken as variant spellings of the same toponym or must refer to two different entities, which would then falsify Oreshko’s claim. On the other hand, an argument in favour of Oreshko’s theory, which its author does not use, is the nonexistence of a phonetic spelling á-ta-na-wa/i-. The Luwian phrase á-ta-na-wa/i-za(URBS) TERRA+LA+LA-za in KARATEPE 1 § 37 is to be interpreted as attana-wan-za wal(i)l-an-za, ‘Adanean plain’, a derivative of á-ta-na-, 39 Downloaded from https://www.cambridge.org/core. Universitaetsbibliothek Marburg, on 16 Sep 2017 at 11:25:10, subject to the Cambridge Core terms of use, available at https://www.cambridge.org/core/terms. https://doi.org/10.1017/S0066154615000010 Anatolian Studies 2015 §1 ’nk w[ryk(s) bn ...] [EGO-mu] wa/i+ra/i-i-[ka-sá x-x-x-x(-x) (“INFANS”)ni-]mu-wa/i-za-sa ’špḥ mpš [mlk dnnym] [mu-ka]-sa-sa |INFANS.NEPOS-si-sà |hi-ia-wa/i[-ni]-sá[(URBS)] |REX-ti-sa hbrk b‘l |(DEUS)TONIT[RUS]-hu-t[a-sa SERVUS-la/i-sá] §2 ’š [yrḥbt] [á-mu-wa/i] wa/i+ra/i-i-ka-sá “[TER]RA”?(-)la-tara/i-ha bt ’rṣ ‘mq [’dn] [DOMUS-na-za TERRA-sa-za hi-]ia-wa/i-za(URBS) TERRA+ || [b‘br] b‘l wb‘br ’[l ’bt] |(DEUS)TONITRUS-hu-ta-ti |á-mi-ia-ti-ha |tá-ti-ia-ti |DEUS-na<-ti> §3 [wp’]l ’nk ’p ss [‘l ss] |wa/i-ta (EQUUS.ANIMAL)zú-na (EQUUS)zú-wa/i |SUPER+ra/i-ta |i-zi-ia-ha §4 [(w)m]ḥnt ‘l mḥnt EXER[CITUS-la/i/u-za-pa-wa/i-ta] EXERCITUS[-la/i/u-ni] |SUPER+ra/i-ta |i-z[i]-ia-h[a] §5 wmlk [’šr] |kwa/i-p[a]-wa/i-mu-u |su+ra/i-wa/i-ni-sa(URBS) |REX-ti-sa [w]kl bt ’šr |su+ra/i-wa/i-za-ha(URBS) |DOMUS-na-za |ta-ni-ma-za kn ly l’b [wl]’m |tá-[ti-sa MATER-ni-sa-ha] i-zi-ia-si §6 wdnnym w’šrym |hi-ia-wa/i-sa-ha-wa/i(URBS) |su+ra/i-ia-sa-ha(URBS) kn lbt ’ḥd |“UNUS”-za |DOMUS-na-za |i-zi-ia-si §7 wbn ’nk ḥmy[t] kwa/i-pa-wa/i *274-li-ha (CASTRUM)ha+ra/i-na-sà bmṣ’ šmš šmnt III III II wbmb’ šmš šb‘t III III I wkn X III II |ORIENS-mi-ia-ti |x-i?-ni? 8 OCCIDENS-mi-ti-ha 7 CASTRUM-za Table 1. Text of the initial parts of the Luwian and Phoenician versions of the ÇİNEKÖY bilingual. divine epithet pihassass(i)-, ‘of lightning’. The Storm-god pihassass(i)- was the patron deity of the Hittite king Muwattalli II, who took residence in Tarhuntassa, situated somewhere in Cilicia (cf. Hutter 2003: 269–70 with references). In addition, the late Luwian aphaeresis in the word for ‘Assyria’, seen in Suriya- (ÇİNEKÖY § 6), may have paved the way for the transmission of the shorter form Συρία, ‘Syria’, to the Cilician Greeks and later on to Europe (Rollinger 2006). Summing up, substantial archaeological evidence for the influence of Aegean material culture upon the plain of Cilicia in the Early Iron Age (Lehmann 2007 with the sources cited) correlates with a number of philological of the Bronze Age kingdom of Kizzuwatna, which was characterised by a close cultural symbiosis between the Luwians and Hurrians (Hutter 2003: 251–52 with references). Thus the Greek residents of Cilicia would represent the most suitable intermediaries for transmitting the fabula of the Hurrian Kumarbi epic, which is now commonly recognised to underlie the plot of Hesiod’s Theogony (West 1997: 279–80). The colonised Cilicia would also provide an ideal milieu for the transmission of certain loan-words into Greek. This holds, for example, for the winged horse Πήγασος, ‘Pegasus’, who is first mentioned in the Theogony, where he carries the lightning and thunderbolt of Zeus, and therefore can be connected with the Luwian 40 Downloaded from https://www.cambridge.org/core. Universitaetsbibliothek Marburg, on 16 Sep 2017 at 11:25:10, subject to the Cambridge Core terms of use, available at https://www.cambridge.org/core/terms. https://doi.org/10.1017/S0066154615000010 Yakubovich §1 Luwian [I am] Waraika [s]on [of X], grandson of [Muk]sa, [king of Hiyawa, man loved by] Tarhunt. Phoenician I am W[araika son of X], descendant of Mopsos, [king of the Adaneans,] the blessed one of Baal, §2 Luwian I, Waraika, extended [the house of the land] of the plain of [Hi]yawa, by Tarhunt and my paternal gods. Phoenician (I) who [extended] the house of the land of the plain [of Adana, by the grace of] Baal and by the grace of the g[ods of my father(s)]. §3 Luwian I made horse on top of horse, Phoenician And I [ma]de horse [on top of horse] §4 Luwian [and] I made ar[my] on top of ar[my]. Phoenician [and ar]my on top of army. §5 Luwian Furthermore, the Assyrian king and all the Assyrian house became father and mother to me, Phoenician And the king [of Assyria and] all the house of Assyria became father [and mo]ther to me, §6 Luwian and Hiyawa and Assyria became one house. Phoenician and Adaneans and Assyrians became one house. §7 Luwian Furthermore, I destroyed fortresses: on the east … 8 and on the west 7 fortresses. Phoenician And I built walled fortress[es]: on the sunrise eight (8), on the sunset seven (7), and there were (altogether) 15. Table 2. Translation of the initial parts of the Luwian and Phoenician versions of the ÇİNEKÖY bilingual. observations that have been advanced to yield support to the suggestion of Greek presence in Que with varying degrees of confidence and plausibility (for example Lipiński 2004; Oettinger 2008; Oreshko 2013; Singer 2013). At the same time, there are scholars who are inclined to dismiss some or all of these observations as a chain of coincidences (for example Hajnal 2003; Gander 2012). It remains to be seen whether the sociolinguistic analysis of the bilingual inscriptions from Que can make a meaningful contribution to this debate. applaud the pioneering efforts of the editors of this inscription, which saw it published a mere three years after its discovery; but the down side of this quick publication is the lack of sufficient synthesis in the presentation of the Luwian and Phoenician versions. In essence, we are dealing with two separate editions, which make the Luwian and Phoenician texts appear more different from each other than they really are. In most cases, this is due to difficulties with the interpretation of the Luwian version. A number of Luwian restorations made in the editio princeps have been silently rectified in the German translation of the bilingual provided by David Hawkins (2005). This prompts me to provide a new synoptic edition of the initial parts of the Luwian and Phoenician versions of the ÇİNEKÖY bilingual (tables 1, 2). I am limiting myself to the first seven clauses of the inscription, because after that the Phoenician text becomes too fragmentary for a meaningful contrastive analysis, while the rest of the Luwian version will be treated shortly by Hawkins (forthcoming). The clause division above follows the syntax of the Luwian version. An attempt has been made to match the corresponding Phoenician and Luwian phrases within each clause. Following the conventions of the editio princeps, the uncertain readings and restorations in the Phoenician version are italicised, while uncertain restorations in the Luwian version are underlined. KARATEPE 1 and ÇİNEKÖY Before one can approach the contrastive study of the Luwian and Phoenician inscriptions, it is necessary to ascertain the reliability of their editions. In the case of the KARATEPE 1 bilingual, we are lucky to have the synoptic edition of its versions (Hawkins 2000: 1.45–67), which summarises the results of the 50-year-long investigation of this monument by a variety of scholars. My contrastive study of different versions of this bilingual is essentially based on this publication, as well as the close treatment of the Phoenician versions in Çambel 1999. By contrast, the ÇİNEKÖY bilingual was found too late to make its way into Hawkins 2000, and the discussion of this fragmentary text is still commonly based on its editio princeps (Tekoğlu, Lemaire 2000). One must 41 Downloaded from https://www.cambridge.org/core. Universitaetsbibliothek Marburg, on 16 Sep 2017 at 11:25:10, subject to the Cambridge Core terms of use, available at https://www.cambridge.org/core/terms. https://doi.org/10.1017/S0066154615000010 Anatolian Studies 2015 The overall goal of the commentary is to demonstrate that the Phoenician and Luwian versions of the ÇİNEKÖY inscription closely resemble one another, and therefore this bilingual reflects the process of translation and not free adaptation. The direction of translation will be addressed later on in this paper. (DEUS)SOL-mi-sá CAPUT-ti-i-sá (Hawkins 2000: 1.48). According to Petra Goedegebuure (2009), the Luwian title (DEUS)SOL-mi-sá = tiwadamis can be literally translated as ‘the steward of the Sun’ = ‘His Majesty’s steward’; cf. the Hittite royal title dUTU-ŠI, ‘My Sun’ = ‘My Majesty’. If one assumes close correspondence between the Phoenician and Luwian titles, the KARATEPE 1 inscription pleads for the understanding of hbrk b‘l as ‘the steward of (his) lord’. By contrast, in our case the counterpart of Phoenician hbrk b‘l in ÇİNEKÖY begins with the name of the Luwian Storm-god, the functional equivalent of the Phoenician Baal, which implies a different understanding of the Phoenician phrase hbrk b‘l in this case. Hawkins (2005) proposes the restoration TONIT[RUS]-hu-t[a-sa SERVUS-la/i-sá], ‘servant of Tarhunt’, on the assumption that it could represent the approximate equivalent of hbrk b‘l, ‘steward of Baal’. Note, however, that the Luwian phrase ‘servant of Tarhunt’ corresponds to Phoenician ‘bd b‘l, ‘servant of Baal’, in KARATEPE 1 § 1. As a likely alternative, one can consider (DEUS)TONIT[RUS]-hut[a-ti (LITUUS)á-za-mi-sa], ‘loved by Tarhunt’; cf., for example, KARKAMIŠ A23 § 1 DEUS-ni-ti (LITUUS)áza-mi-sa, ‘loved by gods’, at the end of the chain of titles. This restoration would imply the understanding of hbrk b‘l as ‘blessed by Baal’. The sequence of two epithets |(DEUS)TONIT[RUS]-hu-t[a-sa SERVUS-ta4-sa (DEUS) SOL-mi-sa CAPUT-ti-i-sa], ‘servant of Tarhunt, Sunblessed man’, which is restored in the editio princeps, is precluded by reasons of space. Consequently, one can give up the matching Phoenician reconstruction hbrk b‘l ’š [’lm], ‘steward of Baal, man [of the gods]’ (thus Lipiński 2004: 127). § 2. The analysis of the editio princeps dividing the Luwian version of this segment into two separate clauses is unfounded, as already implied in Hawkins 2005. Its only attested predicate is Luwian “[TER]RA”?(-)la-tara/i-ha, whose Phoenician equivalent can be supplied in the lacuna based on the parallel KARATEPE 1 § 5. On the assumption that the Luwian and Phoenician clauses are similar in structure, the lacuna in the Luwian clause must contain a part of its direct object, corresponding to Phoenician bt ’rṣ ‘mq [’dn], ‘house of the land of the Adana plain’ (Tekoğlu, Lemaire 2000: 997). The suggestion formulated in Hawkins 2005, according to which the Luwian equivalent of this phrase was simply ‘the Hiyawa plain’, does not leave sufficient material for filling the lacuna in the Luwian texts (cf. Tekoğlu, Lemaire 2000: 971, fig. 11 and Tekoğlu, Lemaire 2000: 985, the autograph). In my restoration I assume that the Luwian phrase was very close in structure to its Phoenician counterpart and contained a chain of possessive adjectives: parnanza taskwarassanza hiyawanza wal(i)li(ya)nza, ‘house of the land of the Hiyawean plain’. In principle, one can think of possible Commentary on ÇİNEKÖY § 1. The Luwian version names the author of the inscription as wa/i+ra/i-i-ka- (restored from § 2), which corresponds to w[ryk(s)] in its Phoenician version. The same name is spelled w˹ryks˺ and wryk in İNCİRLİ and CEBELİREİS DAĞI respectively, although in the second instance it is likely to refer to a later king (Hallo, Younger 2003: 3.138, n.25). One can identify King Waraika of our inscription with Urikki, king of Que, attested several times in Assyrian sources as Ú-ri-ik, Ú-ri-ik-ki and Ú-ri-ia-ik-ki (Kaufman 2007: 22), for whose prosopography one can consult Lanfranchi 2005. The second syllable of the original name must have contained a diphthong, because the Phoenician inscriptions of Cilicia do not feature matres lectionis except in word-final position. Following Lipiński 2004 (119–23) and Simon 2014, I see no plausible way of equating Waraika with the King Awarku mentioned in the KARATEPE 1 and HASSAN-BEYLİ inscriptions, whose name is transmitted as á-wa/i+ra/i-ku- in Luwian and ’wrk in Phoenician (cf. Yakubovich 2010a: 152, n.94). Accordingly, Awarku cannot be identified with King Urikki under the present state of our knowledge. For Greek etymologies offered for the personal names Awarku and Waraika, see above. The comparison of both names with Greek ‘Ράκιος, father of Mopsos, in a Greek tradition, is preferred by Hajnal (2011) but cannot be reconciled with the known sound laws. Lipiński (2004: 127) stresses the difficulties of filling in the lacuna at the end of the first line of the Phoenician version. The large size of the letters belonging to this line does not appear to leave enough space for the phrase ‘son of X’ after the name of Waraika, which can be reconstructed for the Luwian version. Yet, it seems unnecessary to posit an intentional difference in content between the Luwian and Phoenician texts based on this ground alone. Perhaps the scribe found an irregular way of squeezing the name of Waraika’s father at the end of the line, for example modifying the spaces between letters. Alternatively, the missing part of the Phoenician line could feature a scribal error of omission. The Phoenician version contains the formula hbrk b‘l, which is variously interpreted as ‘blessed by Baal’ and ‘steward of Baal’ or ‘steward of (his) lord’ (Tekoğlu, Lemaire 2000: 996 with references; cf. Goedegebuure 2009: 2). The formula reoccurs in the KARATEPE 1 inscription, where it corresponds to the Luwian expression 42 Downloaded from https://www.cambridge.org/core. Universitaetsbibliothek Marburg, on 16 Sep 2017 at 11:25:10, subject to the Cambridge Core terms of use, available at https://www.cambridge.org/core/terms. https://doi.org/10.1017/S0066154615000010 Yakubovich alternatives featuring less precise translation – for example amanza parnanza hiyawanza wal(i)lanza, ‘my house, the Hiyawean plain’ – but such a solution does not impose itself. Likewise, there seems to be no sure way of knowing whether the Phoenician version contains the short formula ‘by Baal and the gods’, as in KARATEPE 1 §§ 10, 58, or a closer equivalent of the longer Luwian formula ‘by Baal and my paternal gods’. I am inclined to embrace the latter solution, since the Luwian versions of KARATEPE 1 §§ 10, 58 also have ‘by Baal and the gods’, and there are no obvious reasons to believe that the correspondence in our case should be less precise. § 3. This clause displays a perfect match between its Luwian and Phoenician versions already in the editio princeps. For the transliteration <zú> instead of the earlier <sù>, see already Melchert 1987: 201–02 and the elaboration of his views in Yakubovich 2010a: 66–67, n.58. § 4. The Phoenician counterpart of this clause is assumed to feature the ellipsis of the repeated predicate, and therefore can be considered an appendix to the previous clause. By contrast, the Luwian version clearly contains the fully-fledged clause. The discrepancy between the Luwian and Phoenician texts is assured here not only for reasons of space but in view of a similar mismatch in KARATEPE 1 § 9. § 5. For various ways of translating the Luwian sentence-initial particle kwippa=, see the partially divergent views of Goedegebuure 1998 and Melchert 2002. In this inscription I consistently translate it as ‘furthermore’ on the assumption that it divides the Luwian texts into thematic blocks, although I must concede that such a translation would not always fit the KARATEPE 1 inscription. Gianni Lanfranchi (2007: 186) avoids attempts to draw any distinction between sentence-initial wa= and kwippa=, observing that both tend to correspond to w- in the Phoenician version. The unisex parental metaphor ‘the king became father and mother to me’ appears to represent a hybrid of Mesopotamian and Anatolian traditions. The most ancient Hittite text, the so-called ‘Deeds to Anitta’, informs us that Anitta’s father Pithana ‘made into mothers and fathers’ the population of the conquered town Nesa (Neu 1974: 11, l.9). This turn of phrase finds a striking counterpart in the Luwian version of KARATEPE 1 § 3: ‘Tarhunzas made me mother and father to á-*429-wa/i-’. By contrast, the Mesopotamian texts and their derivatives show the reverse order of parents. Thus a Sumerian hymn to the Sun-god states: ‘Utu, you are the father of the orphan, Utu, you are the mother of the widow’, while the Hittite adaptation of the same text renders this passage as ‘Sun-god, you are the father and mother of the oppressed and orphaned man’ (Metcalf 2011: 171–72). The same sentiments and order of parents percolate even into the Palaic invocation to the Sun-god: ‘Now, Faskhulassas Tiyaz, to tabarna the king you are indeed the father and the mother’ (Yakubovich 2005: 121). The goddess Ishtar says through a prophet to the New Assyrian king Asarhaddon: ‘I am your father and mother, I raised you between my wings’ (Cooper 2000: 441b). An extended variant of the same metaphor is found in the Phoenician inscription of Kulamuwa, king of Sam’al: ‘for some I was a father, for some a mother, for some a brother’ (Tropper 1993: 41–42). The Luwian form i-zi-ia-si, not understood in the editio princeps, receives satisfactory explanation in Rieken 2004 as the innovative third singular mediopassive preterit of izziya-, ‘to make’, used in the meaning ‘to become’ (cf. Yakubovich 2010a: 201–02). Consequently, there is no substantial difference between the structures of the Phoenician and Luwian clauses, and ‘father and mother’ in the Luwian text should be reconstructed in the nominative. § 6. The Phoenician and Luwian clauses are fully preserved and almost fully parallel, except that the Phoenician version names peoples, whereas the Luwian one mentions the corresponding lands. It is not necessary to assume that the content of this clause refers to a specific act of submission, since the mention of friendship with Assyria also occurs in the İNCİRLİ and HASSAN-BEYLİ inscriptions (cf. Lemaire 1983; Kaufman 2007). Curiously, this topos is absent in the text of KARATEPE, which makes one wonder if the regent Azatiwada did not share the pro-Assyrian sentiments of the ‘house of Mopsus’ or simply was not in direct contact with the Assyrians. The place of the friendship clauses within the narrative of the ÇİNEKÖY inscription suggests that it may have been used here as a mere rhetorical device, similar to the love of gods or material prosperity in other Neo-Hittite traditions. For the Semitic inscriptions from the kingdom of Sam’al, which celebrate or mention voluntary submission to the Assyrian Empire, see Lanfranchi 2009: 128– 29. The pro-Assyrian orientation of Que in the eighth century BC correlates with the use of Akkadian as the third language of the İNCİRLİ inscription. § 7. The Luwian and Phoenician versions clearly deviate from each other in that the Phoenician version refers to the construction of fortresses, whereas the Luwian version only mentions their destruction. This is clearly not an intended variation, but rather the translator’s error, which will be addressed later on in this paper. The attempt of the editio princeps to squeeze the mention of constructing new fortresses into the lacuna in the Luwian text is epigraphically unacceptable, as there is no such lacuna (cf. Tekoğlu, Lemaire 2000: 975, fig. 14). Note that the Phoenician version gives the total number of the constructed fortresses (7 + 8 = 15), whereas the Luwian text lacks this redundant information. 43 Downloaded from https://www.cambridge.org/core. Universitaetsbibliothek Marburg, on 16 Sep 2017 at 11:25:10, subject to the Cambridge Core terms of use, available at https://www.cambridge.org/core/terms. https://doi.org/10.1017/S0066154615000010 Anatolian Studies 2015 I lack a plausible interpretation for the Luwian sequence x-i?-ni? (cf. Tekoğlu, Lemaire 2000: 973, fig. 13). Nevertheless, the attempt of the editio princeps to restore here the phonetic equivalent of VERSUS-ia-na can be safely ruled out, because this post-position always occurs with the dative, never with the ablative, and the available traces are not compatible with any rendering of tawiyan, which is the established reading of VERSUS-ia-na. (CASTRUM)ha+ra/i-na-sà is the new spelling variant of (CASTRUM)ha+ra/i-ní-sà, ‘fortresses’, known from KARATEPE 1. This attestation allows me to propose an etymology for the Carian town Halicarnassus (Greek Ἁλικαρνᾱσσός), deriving it from Luwic *alV-xarnasa-, ‘high fortress’ (or ‘remote fortress’?). The etymologically compound character of the name of Halicarnassus is confirmed through its likely Carian spellings alos k̑arnos and alos-δ k̑arnos-δ (Adiego Lajara 2007: 68, 161). The town name is attested several times without initial aspiration in Athenian tribute lists, although the phonetic significance of this fact is debated (Zgusta 1984: 62). The spiritus asper in Ἁλικαρνᾱσσός may be due to a hypercorrect restitution of the Ionian psilosis or a folk etymology involving Greek ἅλς, ‘sea’ (Florian Sommer, personal communication). geographic considerations place Que squarely within the Neo-Hittite cultural sphere. If one has to choose between Phoenician migrations and Anatolian continuity, the second hypothesis has the allure of an easy default solution. Such arguments, however, are far from being compelling in addressing the formal issue of the relationship between the versions of the Cilician bilinguals. This question is essentially empirical and must be solved through linguistic analysis. Schmitz (2008: 6) appears to have been the first scholar to support the claim that the Phoenician versions of the Que bilingual inscriptions were primary and the Luwian ones secondary. In defence of his position, Schmitz refers the reader to some general observations taken from Payne 2006 and my doctoral dissertation, later published as Yakubovich 2010a. No linguistic evidence for such a hypothesis was, however, available at the time. Now I believe that the claim of Schmitz 2008 can be formally proven. In what follows I intend to advance a three-pronged set of linguistic and philological arguments pleading for the translated character of the Luwian versions of both the ÇİNEKÖY and KARATEPE 1 bilinguals. The most transparent evidence comes, in my opinion, from Luwian passages calquing the Phoenician word order. The placement of most predicative forms in Phoenician is ‘free’, i.e. governed by information structure constraints, whereas others are restricted to sentenceinitial position (Krahmalkov 2001: 290–95). By contrast, the unmarked word order in the Luwian original compositions of the Iron Age is subject–object–verb (SOV), and the preference for it is fairly strong. This is, for example, the only attested pattern in the long inscriptions of Katuwa, king of Carchemish (KARKAMIŠ A 2+3 11a and A 11b+c: Hawkins 2000: 1.94–100, 101–12) if one disregards the position of clitics and relative pronouns, whose placement is governed by special rules. The combined size of these inscriptions is 85 clauses, i.e. roughly similar to the ÇİNEKÖY and KARATEPE 1 inscriptions taken together. The influence of the Phoenician texts upon their Luwian translations manifests itself in the frequent use of Luwian verbal forms in positions other than SOV. Thus the Phoenician version of ÇİNEKÖY § 2 features the restored predicate [yrḥbt], ‘I extended’, between the relative pronoun ’š and the object phrase. The Luwian text converts the relative clause into an independent sentence by adding the subject group in lieu of the relative pronoun at the beginning. But the Luwian verb “[TER]RA”?(-)latara/i-ha, ‘I extended’, is still placed before the object phrase, as was the case in the Phoenician original. In a similar fashion, the predicate is placed before its direct object in both the Phoenician and Luwian versions of ÇİNEKÖY § 7. In the instance of Phoenician bn ’nk, ‘I The primary language of the inscriptions At this point one can proceed to determining the primary language of the KARATEPE 1 and ÇİNEKÖY bilinguals. Many modern studies of both inscriptions assume the primacy of their Luwian versions, which would then be translated into Phoenician as a regional lingua franca. Annick Payne (2006: 130) concludes after some hesitation: ‘the evidence thus appears to be in favor of the two bilinguals being original Luwian inscriptions with Phoenician translations’. No hesitation can be seen in the assertion of Lanfranchi (2007: 186): ‘for the aims of this analysis, the Phoenician text of both inscriptions can be a priori considered totally dependent on the Luwian text, as a direct translation from Luwian’. According to Alexandra Daues (2008), the KARATEPE 1 inscription represents a literal translation from Luwian into Phoenician, while the Luwian and Phoenician texts of the ÇİNEKÖY inscription correspond to each other less precisely (here Daues apparently follows the Luwian restorations adopted in the editio princeps). Federico Giusfredi (2009: 143) remarks that ‘the Phoenician version of the [KARATEPE] bilingual is not an exact translation of the Luwian text’. The motivation for this earlier consensus appears to be primarily sociolinguistic. There are no reasons to believe that Phoenicians ever settled in Cilicia in large numbers, whereas the Luwian presence in the kingdom of Que can be supported through the analysis of the local onomastics. As mentioned at the beginning of this paper, historical and 44 Downloaded from https://www.cambridge.org/core. Universitaetsbibliothek Marburg, on 16 Sep 2017 at 11:25:10, subject to the Cambridge Core terms of use, available at https://www.cambridge.org/core/terms. https://doi.org/10.1017/S0066154615000010 Yakubovich built’, this is due to a syntactic constraint on the position of the infinitive absolute (Krahmalkov 2001: 294–95). By contrast, the placement of Luwian *274-li-ha, ‘I destroyed’, before its object (CASTRUM)ha+ra/i-na-sà, ‘fortresses’, would not be expected in an original composition; cf. a-wa/i pa-PITHOS(URBS)-ha (*274)ha-ta-liha, ‘I destroyed the town Pa-PITHOS, too’ (KARKAMIŞ A25a § 1). Although the meaning of the predicate was misinterpreted in the Luwian version, on which see below, its syntactic position was faithfully preserved. Since one of the ÇİNEKÖY clauses discussed above involves restoration, while the other one has been variously interpreted, it is necessary to turn to the syntax of the KARATEPE 1 inscription. The patterns verb– subject and verb–object are attested there 14 times in the Luwian version, namely in §§ 4, 5, 7, 25, 26, 37, 49, 51, 52, 63, 66, 72, 73, 74. In 11 of these 14 cases, the position of the verb in a clause mirrors the syntax of the Phoenician version. The exceptions are § 37, where the Phoenician version features a nominalisation instead of a verbal predicate, § 52, where the predicate is ellipted, and § 74, where the place of the verb is different in the Phoenician clause. In § 37 the Luwian verbal clause roughly preserves the order of meaningful elements in the relevant passage seen in the Phoenician nominal phrase, i.e. Luwian |BONUS+RA/I-ia-ma-la-ha-wa/i SOLIUM.MI-ta |á-*429wa/i-sá(URBS), ‘And at ease dwelt A-*429-wa’, vs Phoenician w-nḥt lb l-dnnym, ‘And peace of heart to the Adaneans’. The syntax of § 52 is deliberately made parallel to that of the preceding Luwian clause § 51. On the other hand, the final §§ 74–75 probably display a chiastic figure, the word order of § 74 representing a mirror image of §75 (Melchert 2006: 293–94). As the last two examples suggest, the syntactic similarity between the Phoenician and Luwian versions of the bilingual inscriptions need not be exaggerated, and the application of a reverse test reinforces the same conclusion. In the majority of cases the Luwian translators rendered the verb-initial and verb-medial Phoenician clauses with their verb-final counterparts in both KARATEPE 1 and ÇİNEKÖY. The instances of syntactic calques resulting in verb fronting account for no more than 15% of the Luwian textual corpus. Nevertheless, such a result is significant, because, as mentioned above, deviations from the unmarked SOV word order are quite rare in original Luwian compositions. Although the syntax of the translated Luwian texts does not slavishly follow their Phoenician originals, a sharp increase in nonstandard word order patterns must be described in terms of Phoenician interference. An appropriate parallel here would be the syntax of the Hittite translation of the Hurrian and Hattic texts (as discussed in Sideltsev 2002: 188). An independent instance of calquing Phoenician syntax is addressed by Anna Bauer (2014: 58–59). The phrases PORTA-la-na-ri+i zi-na, ‘from this gate’ (KARATEPE 1 § 63), “CASTRUM<”>-ní-si za-ti, ‘to this fortress’ (KARATEPE 1 § 65), and “PORTA”-la-na za-ia, ‘this gate’ (KARATEPE 1 § 66), show post-position of the proximal deictic pronoun za- to its head nouns, which runs contrary to the general tendency to place Luwian demonstrative pronouns before their syntactic heads. In Phoenician, on the other hand, the post-position of the demonstrative z is normal, and so the corresponding phrases b-š‘r z, ‘in this gate’ (KARATEPE 1 § 63), h-qrt z, ‘this town’ (KARATEPE 1 § 65), and h-š‘r z, ‘this gate’ (KARATEPE 1 § 66), show the expected word order. Bauer (2014) plausibly hypothesises that the close imitation of the Phoenician syntactic pattern in KARATEPE 1 §§ 63–66 may be due to the translator’s attempt to preserve the illocutionary force of the curse formula. The second piece of evidence comprises those cases where the Phoenician original appears to have been paraphrased in Luwian under the impact of language-specific constraints. This phenomenon can be illustrated through contrasting the two versions of ÇİNEKÖY §§ 3–4. The Luwian statement ‘I made horse on top of horse, and I made army on top of army’ could be literally translated into Phoenician as a combination of two clauses. By contrast, the Phoenician construction with gapping ‘I made horse on top of horse and army on top of army’ can hardly be rendered into Luwian without supplying the second predicate. The Luwian clauses are rigidly separated from each other by clitic chains, and therefore the ellipsis of a clause predicate would result in a syntactic fragment, for example †kwalanza=ba=wa=tta kwalani sarranta, ‘And army upon army’. Fragments of this type occasionally do occur in the Iron Age Luwian corpus (for example KARATEPE 1 § 72a), but there is a strong tendency to avoid them. Therefore, the translation strategy adopted in ÇİNEKÖY 1 §§ 3–4 can be regarded as optimal only on the assumption that Luwian was the target language. One could, of course, attempt to argue that the original Luwian construction was compressed in the Phoenician version for purely stylistic reasons. There is, however, no reason to believe that the repetition of identical predicates was systematically avoided in the Phoenician inscriptions from the Neo-Hittite milieu. One can compare here the repetition of Phoenician ns‘, ‘to tear out’, in KARATEPE 1 §§ 71–72 and particularly the repetition of Phoenician p‘l, ‘to do’, in the opening lines of Kulamuwa’s inscription from Sam’al, where the ancestors of King Kulamuwa are uncharitably described as habitually inactive (Tropper 1993: 31–33). 45 Downloaded from https://www.cambridge.org/core. Universitaetsbibliothek Marburg, on 16 Sep 2017 at 11:25:10, subject to the Cambridge Core terms of use, available at https://www.cambridge.org/core/terms. https://doi.org/10.1017/S0066154615000010 Anatolian Studies 2015 Instances where additional predicates are supplied in the Luwian version of the KARATEPE inscription include §§ 9, 10, 37, 50, 52, 75. In all these cases the Phoenician version features complex noun phrases coordinated with elements of a larger sentence: ‘and shield upon shield’, ‘an army upon army by the grace of Baal and the gods’, ‘and peace of heart to the Danuneans and to all the plain of Adana’, ‘and powerful strength above every king’ (×2), ‘like the name of the Sun and the Moon’. If one operates with the hypothesis of translation from Luwian into Phoenician, it will imply that the translator systematically complicated the syntax of the text for no apparent reason. By contrast, chunking large Phoenician sentences into smaller Luwian units receives a satisfactory explanation in terms of adjusting the discourse to language-specific constraints. The constructions discussed above reflect successful examples of paraphrases in Luwian. An instance where the Luwian translator failed to find a suitable equivalent of a difficult Phoenician construction is KARATEPE 1 §§ 59– 61, which contain the beginning of the curse formula. The Phoenician version lists three categories of potential perpetrators: ‘if a king among kings, or a prince among princes, or a man, whose name is “man” ...’. Halet Çambel (1999: 60b) came to a plausible conclusion that this list refers to three distinct social orders, ending with the common man. She also cites a parallel Old Babylonian formula from Mari, which would support the existence of such an idiom in Semitic literary tradition. In dealing with these three phrases, the Luwian translator faced the familiar challenge of turning each of them into a well-formed clause but failed to find a uniform solution. He rendered the passage in a rather uncouth fashion as: ‘If anyone rules as king, or if he is a prince, and to him there is a princely name ...’. The first clause here apparently contains the denominative verb hantawatta-, ‘to rule as king’ (the alternative preferred by Hawkins, who interprets hantawattadi as an ablative ‘from kings’, is syntactically more problematic). The second clause is provided with a nominal predicate, which destroys the syntactic parallelism seen in the Phoenician original. But the third Luwian clause differs from the matching Phoenician phrase even in terms of meaning, since the perpetrator listed there has the same reference as the subject of the second Luwian clause, i.e. a prince and not a common man. This brings us to the final cluster of supporting examples, which comprises the passages where the Luwian texts appear to have distorted the meaning of their Phoenician originals. Thus the likely reason for the discrepancy between the mention of building and destroying fortresses in ÇİNEKÖY § 7 is the confusion between Phoenician bn, ‘to build’, and ‘n, ‘to destroy’ (Hawkins 2005: 156, n.24). One must, however, specify that this confusion was caused not by the Phoenician scribe, who obviously knew the difference between these two verbs, but rather by his Luwian colleague. The misunderstanding of the Phoenician text may have happened either by negligence or because the translator had to deal with a damaged copy of the Phoenician text. Such a conclusion is all the more likely since the remainder of the ÇİNEKÖY inscription (§§ 8–11) is devoted to building activities rather than warfare. Another likely mistake in the Luwian version of ÇİNEKÖY is the replacement of the designation’špḥ mpš, ‘descendant of Mopsus’, with its inaccurate specification muksas(sa) hamsis, ‘grandson of Muksa’ in § 1. Since the Anatolian rendering of his name as Muksa goes back to the time when the Greek labiovelar stops were still in place, this Mopsus could hardly be the grandfather of the late eighth-century King Waraika. If one wishes to maintain the accuracy of the Luwian version in this case, one would have to assume that Waraika’s grandfather had the same name as the founder of his dynasty. By contrast, the hypothesis that the Phoenician title is the accurate one does not require additional historical assumptions. The omission of the appendix ‘and there were (altogether) 15’ in the Luwian version of ÇİNEKÖY § 7 can be best taken as suppression of redundant information. This argument is not particularly strong, since such an epexegetic clause could, in principle, be secondarily added to the Phoenician text. But turning to KARATEPE 1, one finds cases where the Luwian translator negligently omitted or hedged the clauses that convey new information or contribute to textual cohesion. If the Luwian version were original, one would then have to assume that the author of the inscription made negligent mistakes, whereas the translator edited it for content, which is, of course, a rather complicated and unlikely hypothesis. Thus KARATEPE 1 (Luwian Ho) lacks altogether § 17 ‘I established peace with every king’, the counterpart of which is extant in all three versions of the Phoenician inscription. Although the relevant slab of the Luwian Hu copy is not completely preserved, the lacuna comprising the end of § 16 and the beginning of § 18 is too small to accommodate an additional clause (cf. Çambel 1999: pl. 62). So one has to assume that KARATEPE 1 § 17 never existed in the draft of the Luwian text. Furthermore, the Luwian version of KARATEPE 1 § 32 appears to replace the relevant information of the Phoenician text with a clause recycled from the earlier material. The Phoenician text is telling us that the Adaneans were present on all the frontiers of the kingdom, which represents a logical transition from the previous § 31 referring to their resettlement to the subsequent §§ 33–35 describing the resulting improvement of security within the kingdom. By contrast, the Luwian version of § 31 reverts to the topic of extending the frontiers of the kingdom, which is already covered in § 5. This repetition does not trigger cohesion with either the preceding or following information. 46 Downloaded from https://www.cambridge.org/core. Universitaetsbibliothek Marburg, on 16 Sep 2017 at 11:25:10, subject to the Cambridge Core terms of use, available at https://www.cambridge.org/core/terms. https://doi.org/10.1017/S0066154615000010 Yakubovich Turning to the lexicon, the direction of translation can be inferred from those cases when two different Phoenician terms belonging to the same semantic field have one and the same Luwian counterpart in the same text. For example, Phoenician qrt, ‘town’, and ḥmt (plural ḥmyt), ‘walled fortress’ (vel sim.), have the identical translation harnisa-, ‘fortress’, in parallel Luwian contexts (cf. KARATEPE § 23 vs § 38). The assumption of creative variation in the Phoenician translation of the Luwian word for ‘fortress’ would obviously be more complicated. The rendering of the toponym Que in the bilingual texts illustrates essentially the same phenomenon. The Phoenician version of KARATEPE uses the ethnic term dnnym, ‘Danuneans’, and the geographic term ‘mq ’dn, ‘plain of Adana’. The Luwian verson normally translates both terms with the same proper noun á-*429-wa/i-, which was presumably perceived as both the ethnic and the geographic entity. The only exception (§ 37) concerns the passage where the ethnic and geographic terms occur side by side in the Phoenician version, and therefore the Luwian translator had to render the second one with the calque á-ta-na-wa/i-za(URBS) TERRA+, ‘Adanean plain’. On the other hand, it is to be expected that the same Phoenician technical terms could be provided with different counterparts by the translators of KARATEPE and ÇİNEKÖY, who were probably distinct individuals. A case in point is the Luwian rendering of the royal epithet hbrk b‘l. We have seen in the commentary to the ÇİNEKÖY inscription that it was understood as ‘servant of his lord’ (vel sim.) in KARATEPE 1 § 1 but as a theophoric epithet in ÇİNEKÖY § 1. In the absence of supplementary evidence I find it impossible to say which of the two translations is closer to the meaning implied by the author of the inscriptions. It is, however, almost certain that at least in one of the two cases a misunderstanding of the Phoenician technical term hbrk b‘l must have been involved. A hypothesis that the same Phoenician phrase could be used as a cover term for the two unrelated Luwian epithets would be highly artificial. Another case of a similar kind has been noticed by Craig Melchert (personal communication). The Phoenician versions of both KARATEPE 1 § 3 and ÇİNEKÖY § 5 introduce the unisex parental metaphor using the Semitic order of elements ‘father and mother’. The Luwian version of the great KARATEPE inscription replaces it with the opposite order ‘mother and father’, which finds a direct counterpart in the Hittite Anitta text. In the ÇİNEKÖY inscription, however, the order of parents is the same in both versions (cf. the commentary). Presumably, the translator of KARATEPE 1 used the idiomatic Anatolian equivalent of the Semitic metaphor, whereas his colleague responsible for the Luwian version of ÇİNEKÖY merely calqued the Phoenician original. Differences in shades of lexical meaning are harder to trace and interpret given our insufficient knowledge of either the Phoenician or the Luwian lexicon. But the correspondence between Phoenician b‘l ’gddm, ‘masters of gangs’, and Luwian usallinzi, ‘thieves’ (vel sim.), in KARATEPE 1 § 20 is at least suggestive. The Phoenician noun ’gdd is cognate with Hebrew gedūd, which is commonly used in the Bible with the meaning ‘marauding band’ (Krahmalkov 2001: 32). Its Luwian counterpart usall(i)- is derived from usa-, ‘to bring, carry’ (cf. Hawkins 2000: 1.61a), either directly or via an unattested action noun. Both terms apply to the evil-doers who previously did not obey ‘the house of Mopsus’ but were reduced to obedience by the regent Azatiwada. The Phoenician term appears to be entirely appropriate for the context as a reference to warlords who did not accept royal authority. On the other hand, Luwian usall(i)- does not seem to convey connotations of either violent or organised crime and appears to represent a generic term for trespassers against property whose suppression would not necessarily require direct royal intervention. One is tempted to conclude that it was used as a vague hyperonym in the absence of an exact Luwian equivalent for the Phoenician phrase b‘l ’gddm. Another likely instance of a Luwian hyperonym corresponding to a more specific term in Phoenician is Luwian harrall(i)-, ‘weapon’ (vel sim.), rendering Phoenician mgn, ‘shield’. The meaning of the Phoenician noun is reasonably secure through its northwestern Semitic cognates, including Hebrew māgēn, ‘shield’, but its Luwian counterpart is semantically more problematic. Hawkins (2000: 1.59–60) appears to assign the meaning ‘shield’ to Luwian hara/i-li- and provide its determinative *272 with the Latin transliteration SCUTUM merely because this noun functions as an equivalent of the Phoenician mgn in KARATEPE 1 §§ 9–10. Nevertheless, the shape of the sign *272 does not resemble that of a shield, but bears a vague resemblance to those of signs *273 to *275, which all predetermine various lexemes pertaining to warfare. From the etymological viewpoint, I find it difficult to separate this noun from the Hittite and Luwian verb harra-, ‘smash, crush’ (Melchert 1993: 57; Hawkins 2000: 2.460a), hence the proposed interpretative transliteration harrall(i)-. This connection is supported by the etymological figure (“*272”)ha+ra/i-ti-ha-wa/i-mu |hara/i-li-na (ASSUR letter d § 8; cf. Hawkins 2000: 2.546), where the sign *272 predetermines a different derivative of harra-. Presumably, the derivation of harrall(i)- parallels that of usall(i)-, ‘thief’, discussed in the previous paragraph. The proposed etymology is compatible with harrall(i)- either having the generic meaning ‘weapon’ or referring to a specific type of offensive weapon, but the KARATEPE 1 context tips the scales in favour of the former solution. It is easy to 47 Downloaded from https://www.cambridge.org/core. Universitaetsbibliothek Marburg, on 16 Sep 2017 at 11:25:10, subject to the Cambridge Core terms of use, available at https://www.cambridge.org/core/terms. https://doi.org/10.1017/S0066154615000010 Anatolian Studies 2015 envisage how the Phoenician metaphoric description ‘(I made) shield on top of shield’ could be paraphrased in Luwian with a broader metaphor ‘I made weapon on top of weapon’, but a paraphrase in the opposite direction would be less felicitous. It is possible that individual arguments among those advanced above will be rendered invalid or relativised through subsequent philological research. I doubt, however, that it will affect the final conclusion, unless one succeeds in assembling a matching set of arguments pleading in favour of the opposite scenario. Strong convergent evidence upholds Phoenician as the primary language of both the ÇİNEKÖY and KARATEPE 1 bilinguals. We have seen thus far that the rulers of Que commissioned inscriptions in Phoenician and provided them with Luwian translations. Sceptics might question the metaphoric significance of this procedure, arguing that the preference for Phoenician originals was motivated by purely technical considerations. In order to preempt this potential objection, it is appropriate to turn to the general design of the respective inscriptions, which has already been properly discussed in Payne 2006: 128–29. The fortress built by Azatiwada was provided with three copies of the Phoenician text of the KARATEPE 1 bilingual (Phu/A, Pho/B and PhSt), but with only two copies of the respective Luwian text (Hu and Ho). The slightly divergent version of the Phoenician text PhSt, which lacks the Luwian counterpart, was inscribed on a colossal statue of the Storm-god, which arguably made it the most prominent inscription in the fortress. In addition, the order of the orthostats covered with Anatolian hieroglyphs appears to be irregular, which creates an impression that the architects of the fortress either were not concerned about making the Luwian text legible or were unable to accomplish this goal. In the instance of the ÇİNEKÖY inscription, the Phoenician text was placed in the frontal space between two bulls carrying a statue of the Stormgod, while fragments of the Luwian text were squeezed between the feet of the animals, on the back of a carriage and along the base of the sculptural complex. One can conclude that the designers of both texts tried to give more prominence to their Phoenician versions. The convergence between linguistic and iconographic asymmetries supports the claim that neither of the two is accidental. culty is, of course, the discrepancy between the postulated socially-dominant language (Greek) and the attested primary written language (Phoenician). As mentioned above, there is no independent evidence for the social dominance of Phoenician newcomers in Early Iron Age Cilicia, but a number of arguments have been advanced for Greek resettlement to this region in the late second millennium BC. How then can one assume that the ‘House of Mopsus’ chose Phoenician instead of Greek for the purposes of writing? The answer to this question logically falls into two parts. On the one hand, one can look for typological parallels illustrating comparable sociolinguistic situations. Of relevance here are the cases when a new elite group avails itself of a foreign written tradition. On the other hand, one must investigate whether the reconstructed symbiosis between Greek population groups and Phoenician written culture might have left historical traces beyond Cilicia. Pursuing the typological line of argumentation, one can begin with a parallel from a rather different region. The First Turkic Khaghanate came into being through the efforts of the Ashinas clan on the steppes to the north of China in the mid sixth century AD. It was the first steppe empire that left us written texts, notably the Bugut inscription found in Mongolia. Although the founders of the Khaghanate were Turks, and its closest literate neighbours were the Chinese, the language that its rulers adopted for public display was neither Turkic nor Chinese, but Sogdian (de la Vaissière 2005: 202–03). The Sogdian native speakers inhabited the area of Samarkand and Bukhara, which came to be dependent on the Khaghans but remained on the periphery of their empire. The Sogdian principalities had no geopolitical significance, but Sogdian traders were active at the time along the eastern stretch of the Silk Road, reaching up to the Chinese capitals. Presumably, the decision to adopt the Iranian Sogdian language for official purposes in an empire of Altaic nomads was prompted by the easy availability of Sogdian scribes, the relative simplicity of the script and the desire to maintain a separate cultural identity from the neighbouring Chinese. All these factors, mutatis mutandis, might have been at work when the ‘House of Mopsus’ chose to engage Phoenician scribes for chancery purposes in the principality of Que. In searching for a parallel from the Neo-Hittite world, one may turn once again to the situation in the principality of Sam’al, the closest neighbour of Que. As mentioned in the introductory part of this article, this polity was a foundation of Aramaean tribes, which established themselves in the midst of Luwian-speaking population groups. Nonetheless, the earliest monumental inscription emanating from Sam’al is not in any form of The sociolinguistic situation in Que If the original character and iconographic prominence of the Phoenician texts can be taken as a metaphor for the sociolinguistic situation in Que, we have arrived at the basic linguistic correlate of the migrationist hypothesis. The principal official language of Que/Hiyawa was not Luwian, even though Luwian was historically spoken by the bulk of its population. The principal remaining diffi- 48 Downloaded from https://www.cambridge.org/core. Universitaetsbibliothek Marburg, on 16 Sep 2017 at 11:25:10, subject to the Cambridge Core terms of use, available at https://www.cambridge.org/core/terms. https://doi.org/10.1017/S0066154615000010 Yakubovich Aramaic but in Phoenician (Tropper 1993: 5). André Lemaire (2001: 189) hypothesises that the written use of Phoenician in the earlier part of King Kulamuwa’s reign was connected in some way with close political ties between Sam’al and Que. This hypothesis, however, is hard to maintain, given that Kulamuwa’s Phoenician inscription explicitly refers to the king of Que as his adversary (Tropper 1993: 37). As an alternative, one can propose that Phoenician at the time simply represented the default option for a king who ruled in a Neo-Hittite milieu but did not wish to use Anatolian hieroglyphs. Kulamuwa’s motivation here must have been more or less the same as that of the rulers of Que, namely an attempt to distance themselves from the neighbouring Neo-Hittite kings, who continued to embrace Anatolian traditions. Naturally, the genetic similarity among the northwestern Semitic languages and the simplicity of the Phoenician script may also have played a role in this decision. The written use of Phoenician was, however, abandoned as soon as the native Sam’alian dialect of Kulamuwa and his descendants underwent reduction to writing or literisation (Verschriftlichung). At the same time, we know that the rulers of Sam’al, just like those of Que, eventually came to a compromise with the conquered Luwian-speaking population. Thus the Phoenician incription of Kulamuwa makes it clear that the king extended his protection to the mškbm, understood as a reference to the local sedentary population groups, possibly of Luwian origin (Tropper 1993: 39–43). This statement correlates well not only with the Luwian origin of the Sam’al dynastic names Kulamuwa and Panamuwa, but also with the mingling of Luwian and Aramaean traditions reflected in the local art and architecture (Bryce 2012: 172 with references). The burial inscription of KTMW, recently found in the course of the University of Chicago excavations at Zincirli, displays a blend of Semitic and Luwian pantheons (Yakubovich 2010b: 296). Although no instances of assured bilingual texts are connected with this principality, a signet ring of the last Sam’alian king, Bar-Rakib, bears his name inscribed in Anatolian hieroglyphs (Hawkins 2000: 2.576). The difference in the extent to which Luwian was deployed as an ancillary written language in Sam’al compared to Que may have been related to the divergent dynamics of language coexistence in the two principalities. In the instance of Sam’al, the Semitic conquerors appear to have maintained their language throughout the existence of the state, judging at least by the onomastic data. One cannot even exclude that the general direction of language shift in Sam’al was from Luwian to Aramaic. By contrast, the arguments presented thus far hardly point to anything more than the Greek cultural heritage of the Que elites. They would be perfectly compatible with the assumption that the rulers of the ‘House of Mopsus’ were no longer Greek speakers, or only semi-speakers, by the eighth century BC, when the royal multilingual inscriptions were produced. Onomastic evidence from Cilicia would support a shift in the direction of the Luwian language, as is particularly clear in the instance of personal names in the seventh-century Phoenician inscription CEBELİREİS DAĞI (Younger 2003). Under such an assumption, the rulers of Que would have had few incentives to abandon the tradition of writing Phoenician in favour of Greek literacy, even though it was already cultivated in the Aegean and southern Italy in the same period. The same assumption also implies that the metaphoric significance of the use of Phoenician in Que could no longer be clear to the outsiders. This would explain why Phoenician came to be deployed alongside Luwian in the unpublished bilingual İVRİZ 2 from the inner Anatolian principality of Tabal (Lipiński 2004: 133–35). A historical argument that can support the reconstruction of Phoenician literacy in the Greek milieu is, in my opinion, the origin of the Greek alphabet. Recent discussions of this topic (Krebernik 2007; Lemaire 2008) converge in reaffirming the traditional view, according to which it was adapted from a Phoenician, not Aramaic prototype. Although the first attested Greek inscriptions postdate 800 BC, the analysis of their palaeography speaks for the Phoenician ductus of the late ninth century BC as a starting point for the development of Greek letters (Krebernik 2007: 123 with references). Another product of the adaptation of the Phoenician writing system is the Phrygian alphabet, which displays non-trivial common innovations with its Greek counterpart in the domain of vowel marking, in particular the creation of signs for e and o vowels out of the Phoenician consonantal letters he and ayin respectively. The earliest Phrygian inscriptions appear to date back to the late ninth century BC and thus predate the attested specimens of Greek writing, but there are ways to show that the Greeks did not borrow their alphabet from the Phrygians (Krebernik 2007: 116–17). We know that the adaptation of a particular script to a new language was rarely a spontaneous process in ancient societies. What preceded it was usually an extension of the same script in association with its original language to new communities whose native languages had lacked a written tradition. For example, ‘the earliest inscriptions of the Urartian kings are written in the Assyrian script and language ... though after a single generation the Urartian language, for most purposes, replaced the use of Assyrian’ (Wilhelm 2004: 119). The use of Akkadian as the main written language should probably be reconstructed for the early days of the kingdom of Hattusa before the reforms of King Telepinu introduced Hittite literacy on a larger scale (van den Hout 2010: 103–04). The written use of 49 Downloaded from https://www.cambridge.org/core. Universitaetsbibliothek Marburg, on 16 Sep 2017 at 11:25:10, subject to the Cambridge Core terms of use, available at https://www.cambridge.org/core/terms. https://doi.org/10.1017/S0066154615000010 Anatolian Studies 2015 official Aramaic in Achaemenid and Hellenistic Iran predated the adaptation of the Aramaic alphabet to the transmission of Middle Iranian (Skjaervø 1995). Therefore, as long as one accepts the Phoenician alphabet as the source of the Greek writing system, its historical use for writing purposes in certain parts of the Greek-speaking world emerges as the default hypothesis. What makes it even more likely is the Semitic origin of Greek names for individual letters (Krebernik 2007: 148– 61). While the northwestern Semitic letter names have transparent etymologies (such as aleph, ‘bull’, beth, ‘house’, etc.), which are further motivated in terms of their original acrophonic values, neither is the case for the Greek letter names, such as alpha and beta. This situation can be contrasted, for example, with abbreviating Greek letter names for the Roman alphabet or coining new letter names with Slavic etymologies for the Cyrillic alphabet. The transmission of writing together with elements of useless linguistic tradition speaks in favour of the literisation of Greek being an initiative of professional bilingual scribes acting under institutional patronage rather than a mere by-product of casual informal encounters between Greeks and literate Semites. The reform modifying a Semitic script for purposes of writing Greek could have hardly taken place in Cilicia, where Phoenician continued to be used in the official setting even in the seventh century BC. Although Lemaire (2008: 51–52) suggests that alphabetic writing may have spread to Phrygia via Cilicia, the shared innovations of the Greek and Phrygian alphabets would rather plead for a common reformed archetype, which probably came into being further west. In this connection it is interesting to observe that the control of Que may have extended westwards on occasions as far as Pamphylia. This follows from the etymology of the local town Aspendos, which was apparently written *Εστϝεδυς (genitive Εστϝεδιιυς) in the Pamphylian dialect of Greek (Brixhe 1976: 194–200). Stephen Durnford and Heiner Eichner remind me about its possible etymological connection with the name of the Que regent Azatiwada and his eastern foundation Azatiwadaya (cf. already Brixhe 1976: 80, 193). If the area of Aspendos marked the western limits of Que, it could be one possible point where the Greek population of Pamphylia might have come into contact with the Phoenician script, perhaps before the town of Aspendos was even founded. It goes without saying that this contact episode is not assured, and even if it were, its impact on the further development of the Greek alphabet would still remain a matter of speculation pending further material finds. But an independent piece of indirect evidence for the westward extension of Que is the CEBELİREİS DAĞI inscription found in the west of Rough Cilicia and featuring the dynastic name wryk (Younger 2003). No matter how one may venture to reconstruct further developments, the principality of Que emerges as a likely venue for the initial adoption of Phoenician for purposes of writing in a Greek-language community. Beside the Greek presence in Que, the only assumption that is required for the consistency of this hypothesis is the use of the Phoenician script in Que since at least the mid ninth century BC. Although no direct evidence supports such an early date, there is nothing unlikely about it either, given that this is the point in time when King Kulamuwa commissioned his Phoenician inscription in Sam’al. Kulamuwa’s inscription refers to Que as an organised state, and this information can be confirmed through contemporary Assyrian sources (Bryce 2012: 155). It seems reasonable to assume that the principality of Que had its own chancery, just as was the case in the other NeoHittite states. This chancery can be envisioned as a place where engaged Phoenician scribes rubbed shoulders and shared knowledge with their Greek disciples in the mid ninth century BC and probably even earlier. Concluding thoughts We have seen that the reconstruction of the ethnic landscape in Early Iron Age Cilicia can be approached from various angles. As usual in interdisciplinary research, the outcome of this reconstruction will depend not on proving or disproving particular points but rather on weighing up the relative compatibilities of various scenarios with all the heterogeneous data at our disposal. I have tried to demonstrate that the existing historical hypothesis can be reconciled with the new linguistic results. The combination of both yields a coherent picture if one assumes that the written use of Phoenician in Que/Hiyawa developed as a contrastive statement of identity on the part the Greek colonists, as opposed to the indigenous Luwian population. I have also endeavoured to stress the relevance of the proposed reconstruction for tracing the emergence of the Greek alphabetic script. Acknowledgements My work on the topic of multilingualism in Que/Hiyawa and its sociolinguistic interpretation became possible through the award of a Humboldt Fellowship tenured at the Philipps-Universität Marburg and was further facilitated by participation in the project Egea i Lewant na przełomie epoki brązu i żelaza within the framework of the National Program of the Development of the Humanities of the Polish Ministry of Science and Higher Education. Earlier versions of this paper were presented at the Translation and Bilingualism in Ancient Near Eastern Texts workshop (Oxford, March 2013), the 32nd Deutscher Orientalistentag (Münster, September 2013), Journée Langues rares (Paris, November 2013) and The Aegean and the Levant 50 Downloaded from https://www.cambridge.org/core. Universitaetsbibliothek Marburg, on 16 Sep 2017 at 11:25:10, subject to the Cambridge Core terms of use, available at https://www.cambridge.org/core/terms. https://doi.org/10.1017/S0066154615000010 Yakubovich Arsuz (İskenderun): ARSUZ 1 and 2’ Anatolian Studies 65: 59–77 Gander, M. 2011: Die geographischen Beziehungen der Lukka-Länder. Heidelberg, Winter — 2012: ‘Ahhiyawa – Hiyawa – Que: Gibt es Evidenz für die Anwesenheit von Griechen in Kilikien am Übergan von der Bronze- zur Eisenzeit?’ Studi micenei ed egeoanatolici 54: 281–309 Giusfredi, F. 2009: ‘The problem of the Luwian title tarwanis’ Altorientalische Forschungen 36.1: 140–45 Goedegebuure, P. 1998: ‘The Hieroglyphic Luwian particle REL-i=pa’ in A. Süel (ed.), Acts of the Third International Congress of Hittitology. Ankara, Uyum Ajans: 233–45 — 2009: ‘hbrk b‘l or tiwadamis zidis, the steward of the king in Luwian and Phoenician society’. Paper presented at the Midwestern Branch of the American Oriental Society meeting (Bourbonnais, February 2009) Hajnal, I. 2003: Troia aus sprachwissenschaftlicher Sicht. Die Struktur einer Argumentation. Innsbruck, Institut für Sprachwissenschaft der Universität Innsbruck — 2011: ‘Namen und ihre Etymologien – als Beweisstücke nur bedingt tauglich?’ in C. Ulf, R. Rollinger (eds), Lag Troia in Kilikien? Der aktuelle Streit um Homers Ilias. Darmstadt, Wissenschaftliche Buchgesellschaft: 241–63 Hallo, W.W., Younger, K.L. 2003: The Context of Scripture (3 vols). Leiden, Brill Harrison, T.P. 2007: ‘Neo-Hittites in the Orontes valley. Recent excavations at Tell Ta’yinat’ Journal of the Canadian Society for Mesopotamian Studies 2: 59–68 Hawkins, J.D. 2000: Corpus of Hieroglyphic Luwian Inscriptions 1. Inscriptions of the Iron Age (parts 1, 2: texts; part 3: plates). Berlin and New York, de Gruyter — 2005: ‘Die Inschrift des Warikas von Hiyawa aus ÇİNEKÖY’ in M. Lichtenstein (ed.), Texte aus der Umwelt des Alten Testaments 2. Staatsverträge, Herrscherinschriften und andere Dokumente zur politischen Geschichte. Gütersloh, Gütersloher Verlagshaus: 155– 56 — 2011: ‘The inscriptions of the Aleppo Temple’ Anatolian Studies 61: 35–54 — forthcoming: ‘The ending of the ÇİNEKÖY inscription’ in a forthcoming Festschrift Hutter, M. 2003: ‘Aspects of Luwian religion’ in H.C. Melchert (ed.), The Luwians. Leiden, Brill: 211–80 Kahn, D. 2011: ‘The campaign of Ramesses III against Philistia’ Journal of Ancient Egyptian Interconnections 3.4: 1–11 Kaufman, S. 2007: ‘The Phoenician inscription of the Incirli trilingual: tentative reconstruction and translation’ Maarav 14: 7–26 at the Turn of Bronze and Iron Ages workshop (Warsaw, January 2014). I am grateful to the audiences of these gatherings for their constructive feedback. At a later stage the paper benefited from the helpful advice of Stephen Durnford (Brighton), Heiner Eichner (Vienna), H. Craig Melchert (Los Angeles), Norbert Oettinger (Erlangen), Elisabeth Rieken (Marburg), Florian Sommer (Zürich) and David Sasseville (Marburg). Among these, Stephen Durnford and Craig Melchert also made substantial contributions toward improving its style. Alexander Fantalkin (Tel-Aviv) and Rostislav Oreshko (Hamburg) guided me on the archaeological problems relevant to the topic, while Alexei Kassian (Moscow), Gianni Lanfranchi (Padova) and Philip Schmitz (Ypsilanti MI) helped me with the historical and philological bibliography. Yuri Koryakov (Moscow) kindly agreed to design a historical map that accompanies this paper. All the above-mentioned scholars deserve my heartfelt gratitude and none of them is to be blamed for my possible shortcomings. References Adams, M.J., Cohen, M.E. 2013: ‘The “Sea Peoples” in primary sources’ in A.E. Killebrew, G. Lehmann (eds), The Philistines and Other ‘Sea Peoples’ in Text and Archaeology. Atlanta, Society of Biblical Literature: 645–64 Adiego Lajara, I.-X. 2007: The Carian Language. Leiden, Brill Bauer, A. 2014: Morphosyntax of Noun Phrase in Hieroglyphic Luwian. Leiden, Brill Brixhe, C. 1976: Le dialecte grec de Pamphylie: documents et grammaire. Paris, Maisonneuve Bryce, T. 2012: The World of the Neo-Hittite Kingdoms: A Political and Military History. Oxford, Oxford University Press Çambel, H. 1999: Corpus of Hieroglyphic Luwian Inscriptions 2. Karatepe-Arslantaş. Berlin, de Gruyter Cooper, J. 2000: ‘Assyrian prophecies, the Assyrian tree, and the Mesopotamian origins of Jewish monotheism, Greek philosophy, Christian theology, gnosticism, and much more’ Journal of the American Oriental Society 120: 430–44 Daues, A. 2008: ‘Mehrsprachigkeit im vorchristlichen Anatolien: Sprachlicher Wandel durch Kontakt von Sprach- und Schriftsystemen’ in A. Casaretto, S. Kutscher (eds), Proceedings of the Second Graduate Workshop of the Center of Linguistics, University of Cologne, 23rd–24th of June 2006. Aachen, Shaker Verlag: 1–28 de la Vaissière, É. 2005: Sogdian Traders: A History. Leiden, Brill Dinçol, B., Dinçol, A., Hawkins, J.D., Peker, H., Öztan, A. (2015) ‘Two new inscribed Storm-god stelae from 51 Downloaded from https://www.cambridge.org/core. Universitaetsbibliothek Marburg, on 16 Sep 2017 at 11:25:10, subject to the Cambridge Core terms of use, available at https://www.cambridge.org/core/terms. https://doi.org/10.1017/S0066154615000010 Anatolian Studies 2015 Killebrew, A.E., Lehmann, G. 2013: ‘Introduction: the world of the Philistines and other “Sea Peoples”’ in A.E. Killebrew, G. Lehmann (eds), The Philistines and Other ‘Sea Peoples’ in Text and Archaeology. Atlanta, Society of Biblical Literature: 1–17 Krahmalkov, C.R. 2000: Phoenician-Punic Dictionary. Leuven, Peeters — 2001: A Phoenician-Punic Grammar. Leiden, Brill Krebernik, M. 2007: ‘Buchstabennamen, Lautwerte und Alphabetgeschichte’ in R. Rollinger, A. Luther, J. Wiesehöfer (eds), Getrennte Wege? Kommunikation, Raum und Wahrnehmung in der Alten Welt. Mainz, Verlag Antike: 108–75 KUB = Figulla, H.H. et al. (eds) 1921–: Keilschrifturkunden aus Boghazköi. Berlin, Akademie Verlag Lanfranchi, G.B. 2005: ‘The Luwian-Phoenician bilingual of Cineköy and the annexation of Cilicia to the Assyrian empire’ in R. Rollinger (ed.), Von Sumer bis Homer. Festschrift für Manfred Schretter zum 60. Geburtstag am 25. Februar 2004. Münster, Ugarit-Verlag: 481–96 — 2007: ‘The Luwian-Phoenician bilinguals of Cineköy and Karatepe: an ideological dialogue’ in R. Rollinger, A. Luther, J. Wiesehöfer (eds), Getrennte Wege? Kommunikation, Raum und Wahrnehmung in der alten Welt. Mainz, Verlag Antike: 179–217 — 2009: ‘A happy son of the king of Assyria: Warikas and the ÇİNEKÖY bilingual (Cilicia)’ in M. Luukko, S. Svärd, R. Mattila (eds), Of God(s), Trees, Kings, and Scholars: Neo-Assyrian and Related Studies in Honour of Simo Parpola. Helsinki, Finnish Oriental Society: 127–50 Latacz, J. 2004: Troy and Homer: Toward a Solution of an Old Mystery. Oxford, Oxford University Press Lehmann, G. 2007: ‘Decorated pottery styles in the northern Levant during the Early Iron Age and their relationship with Cyprus and the Aegean’ UgaritForschungen 39: 487–550 Lemaire, A. 1983: ‘L’inscription phénicienne de HassanBeyli reconsiderée’ Rivista di Studi Fenici 11: 9–19 — 2001: ‘Les langues due royaume de Sam’al aux IXe– VIIIe s. av. J.-C. et leurs relations avec le royaume de Qué’ in É. Jean, A.M. Dinçol, S. Durugönül (eds), La Cilicie: Espaces et pouvoirs locaux (2e millénaire av. J.-C. – 4e siècle ap. J.-C.). Istanbul, Institut français d’études anatoliennes Georges Dumézil: 187–92 — 2008: ‘The spread of alphabetic scripts (c. 1700–500 BCE)’ Diogenes 218 (55.2): 44–57 Lipiński, E. 2004: Itineraria Phoenicia. Leuven, Peeters Melchert, H.C. 1987: ‘PIE velars in Luvian’ in C. Watkins (ed.), Studies in Memory of Warren Cowgill (1929– 1985). Berlin, de Gruyter: 182–204 — 1993: Cuneiform Luwian Lexicon. Chapel Hill, selfpublished — 2002: ‘Hieroglyphic Luwian REL-i-pa “indeed, certainly”’ in M.R.V. Southern (ed.), Indo-European Perspectives. Washington, Institute for the Study of Man: 223–32 — 2006: ‘Indo-European verbal art in Luwian’ in G.-J. Pinault, D. Petit (eds), La langue poétique indo-européenne. Leuven, Peeters: 291–98 Metcalf, C. 2011: ‘New parallels in Hittite and Sumerian praise of the sun’ Welt des Orients 41: 168–76 Moran, W.L. (ed.) 1992: The Amarna Letters. Baltimore, Johns Hopkins University Press Neu, E. 1974: Der Anitta-Text. Wiesbaden, Harrassowitz Oettinger, N. 2008: ‘The seer Mopsos (Muksas) as a historical figure’ in M. Bachvarova, B.-J. Collins, I. Rutherford (eds), Anatolian Interfaces: Hittites, Greeks, and their Neighbours in Ancient Anatolia. Oxford, Oxbow: 45–56 — 2010: ‘Seevölker und Etrusker’ in Y. Cohen, A. Gilan, J. Miller (eds), Pax Hethitica. Studies on the Hittites and their Neighbours in Honour of Itamar Singer. Wiesbaden, Harrassowiz: 233–46 Oreshko, R. 2013: ‘The Achаean hides caged in yonder beams: the value of Hieroglyphic Luwian sign *429 reconsidered and a new light on the Cilician Ahhiyawa’ Kadmos 52.1: 19–33 Payne, A. 2006: ‘Multilingual inscriptions and their audiences: Cilicia and Lycia’ in S. Sanders (ed.), Margins of Writing, Origins of Cultures. Chicago, Oriental Institute of the University of Chicago: 121–36 Rieken, E. 2004: ‘Das Präteritum des Medio-Passivs im Hieroglyphen-Luvischen’ Historische Sprachforschung 117.2: 179–88 — 2010: ‘Das Zeichen <sà> im Hieroglyphen-luwischen’ in A. Süel (ed.), Acts of the VIIth International Congress of Hittitology, Çorum, August 25–31, 2008 2. Ankara, Anıt Matbaa: 651–60 Rieken, E., Yakubovich, I. 2010: ‘The new values of Luwian signs L 319 and L 172’ in I. Singer (ed.), Ipamati kistamati pari tumatimis: Luwian and Hittite Studies Presented to J. David Hawkins on the Occasion of his 70th Birthday. Tel-Aviv, Institute of Archaeology: 199–219 Rollinger, R. 2006: ‘The terms “Assyria” and “Syria” again’ Journal of Near Eastern Studies 65: 283–87 Schmitz, P.C. 2008: ‘Archaic Greek words in Phoenician script from Karatepe’ Newsletter of the American Society of Greek and Latin Epigraphy 12.1: 5–9 — 2009: ‘Phoenician KRNTRYŠ, Archaic Greek *ΚΟΡΥΝΗΤΗΡΙΟΣ and the Storm-god of Aleppo’ KUSATU 10: 119–60 Sideltsev, A.V. 2002: ‘Inverted word order in Middle Hittite’ in V. Shevoroshkin, P. Sidwell (eds), Anatolian Languages. Canberra, Association for the History of Language: 137–88 52 Downloaded from https://www.cambridge.org/core. Universitaetsbibliothek Marburg, on 16 Sep 2017 at 11:25:10, subject to the Cambridge Core terms of use, available at https://www.cambridge.org/core/terms. https://doi.org/10.1017/S0066154615000010 Yakubovich Simon, Z. 2011: ‘The identification of Qode: reconsidering the evidence’ in J. Mynářová (ed), Egypt and the Near East: The Cross-roads: Proceedings of an International Conference on the Relations between Egypt and the Near East in the Bronze Age. Prague, Charles University: 249–70 — 2014: ‘Awarikus und Warikas: Zwei Könige von Hiyawa’ Zeitschrift für Assyriologie 114.1: 91–103 Singer, I. 2006: ‘Ships bound for Lukka: a new interpretation of the companion letters RS 94.2530 and RS 94.2523’ Altorientalische Forschungen 33.2: 242–62 — 2013: ‘“Old country” ethnonyms in “new countries” of the “Sea Peoples” diaspora’ in R. Koehl (ed.), Amilla: The Quest for Excellence. Studies Presented to Guenter Kopcke in Celebration of his 75th Birthday. Philadelphia, INSTAP Academic Press: 321–33 Skjaervø, P.O. 1995: ‘Aramaic in Iran’ Aram 7: 283–318 Tekoğlu, R., Lemaire, A. 2000: ‘La bilingue royale louvito-phénicien de ÇİNEKÖY’ Comptes Rendus de L’Académie des Inscriptions 2000: 961–1006 Tropper, J. 1993: Die Inschriften von Zincirli. Neue Edition und vergleichende Grammatik von des phönizischen, sam’alischen und aramäischen Textkorpus. Münster, Ugarit-Verlag van den Hout, T. 2010: ‘The rise and fall of cuneiform script in Hittite Anatolia’ in C. Woods (ed.), Visible Language: Inventions of Writing in the Near East and Beyond. Chicago, Oriental Institute of the University of Chicago: 99–114 Vanschoonwinkel, J. 1990: ‘Mopsus: légendes et réalité’ Hethitica 10: 195–211 Ventris, M., Chadwick, J. 1973: Documents in Mycenaean Greek (2nd edition). Cambridge, Cambridge University Press West, M.L. 1997: The East Face of Helicon: West Asiatic Elements in Greek Poetry and Myth. Oxford, Oxford University Press Wilhelm, G. 2004: ‘Urartian’ in R.D. Woodard (ed.), The Cambridge Encyclopedia of the World’s Ancient Languages. Cambridge, Cambridge University Press: 119–37 Yakubovich, I. 2005: ‘Were Hittite kings divinely anointed?’ Journal of Ancient Near Eastern Religions 5: 109–39 — 2010a: Sociolinguistics of the Luvian Language. Leiden, Brill — 2010b: ‘West Semitic god El in Anatolian Hieroglyphic transmission’ in Y. Cohen, A. Gilan, J. Miller (eds), Pax Hethitica. Studies on the Hittites and their Neighbours in Honour of Itamar Singer. Wiesbaden, Harrassowitz: 385–98 Younger, K.L. Jr 2003: ‘Cebel İres Dağı’ in W.W. Hallo, K.L. Younger Jr (eds), The Context of Scripture 3: Archival Documents from the Biblical World. Leiden, Brill: 137–39 Zgusta, L. 1984: Kleinasiatische Ortsnamen. Heidelberg, Winter 53 Downloaded from https://www.cambridge.org/core. Universitaetsbibliothek Marburg, on 16 Sep 2017 at 11:25:10, subject to the Cambridge Core terms of use, available at https://www.cambridge.org/core/terms. https://doi.org/10.1017/S0066154615000010