Ethnic groups and language contact in Lycia (I): the "Maritime Interface"

There can be little doubt that the contact with the Near East reflected in the letters from Ugarit has left some traces in the Lycian language. First of all, one can expect a number of oriental borrowings connected with trade be present in Lycian, such as terms for goods of Eastern origin, terminology for measures, vessels, possibly also some sea-ship terminology. In the extant Lycian corpus there is only one word which can be identified as a borrowing from the Levant, namely sixli-. The word is attested in two texts: in the Letoon Trilingual (TL 320: 22), where the form sixlas corresponds to 5uo SpaXM-bg of the Greek version (l. 20; no corresponding part in the Aramaic version), and in TL 57 as sixli (l. 5) and sixla (l. 6). The contexts make it clear that sixli- refers to a coin, probably the most common Lycian coin type (`stater' weighting 8,3-8,6 g, see Frei 1977: 70-71), which, however, in no way excludes that it was used also as a weight measure. The word obviously comes from a Semitic source, just as Greek огуЛод/огкЛод does, reflecting a measure name based on the root sql `weight'. There are no special reasons to think that it has been brought to Lycia by the Persians, also because Persian siglos seems to correspond in weight to Lycian ada (5,5-5,6 g, cf. Frei 1977: 69-70). Both vocalization and historical considerations allow Hebrew (seqel) to be excluded, and a direct borrowing form Akkadian (seqlu) does not seem probable either. There remain Ugaritic tql and Punic (and consequently also Phoenician) sql, whose vocalizations are unclear. In view of the evidence adduced above, the first option seems to be preferable, and it is quite possible that the borrowing goes as far back as Late Bronze Age There is no place here to go into discussion of the exact values of Lycian sibilants (s and z), which might affect the question, but probably rather not, since there is quite wide variation in renderings of Iranian and Greek names in Lycian, and both s and t of the Semitic source could probably be reflected as s in Lycian..

Due to a rather modest volume of the Lycian corpus and its genre specifics, many other potential Oriental borrowings are simply not yet attested (or identified). However, it would hardly be far-fetched to assume that the majority if not all words identifiable in Greek as early borrowings from an Oriental source were present in Lycian as well. Given the trade contacts, it is quite possible that other terms for measures were also present in Lycian, such as, for instance, *m(V)na, cf. Greek qva < poss. from Ugar. mn (cf. Akk. manu etc.), as well as some technical terms from this domain, cf. Greek appaptov `caution-money' connected with Ugar. `rbn `guarantor, surety', Phoen. rb `to guarantee' etc. For these and further examples and a general discussion of the words of Near-Eastern origin in Greek see, first of all, a sober treatment in Masson 1967; cf. a more recent discussion by Rosol 2013 which claims many more oriental borrowings in Greek. One can also hardly doubt that the Ly- cians knew a term corresponding to Greek xltu>v/kl0u>v (Myc. ki-to) `linen, linen tunic', which is connected with Ugar. ktn, Phoen. ktn `linen, linen tunic', Akk. kitu, kitinnu `linen', and other comparable terms for specific types of fine cloths (as puoaog or orv5d>v); or a word for `sesame' corresponding to Greek opoapov (Myc. sa-sa-ma), Phoen., Ugar. ssmn, Akk. samassammu and Hitt. sapsama, as well as other comparable terms (as Kupivov); or terms for oriental aromatic substances as `myrrh', cf. Greek puppa which goes back though Phoen. or Ugar. mr and to Arab. murr (cf. also paAoapov and Aipavog).

On the other hand, one should note that the Ugaritic evidence pointed out above suggests that Lycia might also have been a likely place for language contact between Greek and the Levantine languages, on par with the Levantine coast itself or Cyprus.

To the case of sixli- discussed above a further word can be added, which represents a more interesting example of an Oriental influence in Lycia, presenting, moreover, a clear clue that the Lycians were rather `reasonable people' already in Late Bronze Age. In a recent discussion of the Xanthos trilingual (Oreshko forthcoming) I presented arguments for interpretation of a part of the text on side A (TL 44a: 41-55) containing a recurrent term hatahe and apparently describing the victories won by the author as a summary of a longer text which has been incised on a different monument similar. This monument is referred by the term prulija in l. 41, which is probably the Lycian word for `pillar monument'. It is furthermore probable that the text part immediately preceding the hatahe-passage (ll. 31-40) describes other parts of the same monument, beginning with `bovine protomes' (wawadra, l. 32) on the cornice and proceeding to the reliefs with different scenes (as archery/hunting) in the upper part of the column, closely corresponds to the decoration of the pillar monument containing TL 44 as testified by the archaeological finds.. Now, the two lines immediately preceding the hatahe-passage (ll. 38-40) have a peculiar structure containing two practically identical clauses:

³⁸tupelija: Trmmilis[. (.)] 39[.. qa]Kadunimi: puwejehn: tupelija: sl/m[(.) ] 40[... ]: qaKadunimi: puwejehn:

An interpretation of the lines as referring to a certain `QaKadunimi son of Puweje''<sup>34</sup> makes little sense, since it does not explain the unusual structure of the passage. Instead, one may note that the word tupelija is strikingly reminiscent of the HLuw. *tupaliya- (SCRIBA-li-ia-) `writing, script' and that the structure of the Lycian passage finds a curious parallel in KARKAMIS A15b, §19, a passage describing the writing skills of Yariri:

..,]URBS-si-ia-ti ISCRIBA-li-ia-ti Su+ra/i-wa/i-ni-ti(URBS) ISCRIBA-li-ia-ti-i A-su+ra/i(REGIO)-wa/i-na-ti(URBS) ISCRIBA-li-ia-ti-i Ta-i-ma-ni-ti-ha(URBS) SCRIBA-li-ti

`.in the script of the city [= Hieroglyphic Luwian], in the script of Tyre [= Phoenician alphabet], in the script of Assyria [= cuneiform], in the script of *Taima-\ The interpretation of tupelija as `writing' is further supported by the possibility to recognize in puwejehn a derivative of the Lycian root puwe- `write'. The distinction between tupelija and *puweja- possibly consists in that the latter refers to script, while the former to the physical writings/text incised in stone. Consequently, the passage should refer to writing the text in two scripts/languages³⁵. Cf. Melchert 2004: s.v.v. Kadunimi and Puweje and Schьrr 2009: 161-163. Neumann (2006: s.v.v. Kadunimi and puweje-) also takes Kadunimi for a personal name, but suggests no definitive interpretation for puweje-. The interpretation of the passage has quite important consequences for identification of the second non- Greek language of the Xanthos Trilingual, which will be addressed in detail in the second part of the present contribution.

In addition to tupelija, one can identify in the Lycian corpus two further words which are likely connected to it. The first is tupelezije (poss. dat.sg.) found twice in the Xanthos Trilingual (44b: 63-64) and possibly in TL 35: 5 in an erroneous spelling tupazalije (standing for *tupalazije). The word is apparently a derivative with the agentive suffix -z-, found also in maraza- `commander' or prnnezi(je)- `household member', and can be interpreted as `scribe'. This interpretation agrees well with the context of TL 44b: 63-64, which follows a passage mentioning Persian kings Darius (Ntarijeus) and Xerxes (Ertaxssiraza), as one can readily identify in the combination tupelezije: xntawatije: a Lycian counterpart of yQaqqaxiaxqg parnAqiog `royal scribe', mentioned, for instance, by Herodotus in the context of the Persian court at Sardis (Hdt. 3.128). The second word is tupa found several lines before tupelija (TL 44a: 36), where it is followed by a clause esbedi: hmmenedi: Trmmil[i]je The reading of the HLuw. title is quite certain given the semantics of the word and the usage of the sign SCRIBA with a phonetic value <TU>, cf. Hawkins 2000: 33. For a discussion of tuppalanura- see Tischler 1991-1994: s.v. and Yakubovich 2017: 41-43. For Hittite evidence see Tischler 1991-1994: s.v. tuppi-.di: se Medezedi which can be interpreted as `with shooting/hunting on the horse-back in the Lycian and Median (Persian) style'. Given that the passage likely describes a scene depicted on the pillar monument, tupa can be interpreted as `image', `relief' or the like.

Identification of this word set has quite important implications for the question of linguistic and cultural contact between Lycia and the East. The HLuw. *tupaliya- (SCRIBA-li-ia-) is based on *tup(p)ala- `scribe' standing behind the common HLuw. title SCRIBA-la- `scribe' and attested in full phonetic form in the cuneiform title tup(p)alanura- `chief scribe' (< *tup(p)ala(n) + ura- `big, great')³⁶. The word *tup(p)ala- is based in its turn on Luw. *tup(p)a/i- corresponding to Hitt. tuppi- `(clay) tablet', both of which finally go back, through Akkadian and Hurrian intermediary, to Sum. dub `clay tablet'³⁷. Both *tup(p)a- and *tup(p)ala- represent thus important Bronze Age terms associated with the Ancient Near Eastern cultural sphere and scribal tradition, and their presence in Lycian demonstrates that both the art of writing and the media for it -- quite probably wooden rather than clay tablets The writing on wood in Anatolia (and elsewhere) cf. Waal 2011 with further refs. -- were well known to the Lukka people. It is noteworthy that this linguistic evidence confirms, once again, the extraordinary ability of the Homeric text to encapsulate historical reality in small details which may seem insignificant or accidental on the first glance. The only mention of writing in the Iliad (Hom. Il. 6.168) -- the `baleful signs incised in a folded tablet' (oqqaxa Auypa ypa^ag ev uivaKi nxuKxq>) given to Bellerophont by Proitos -- is notoriously associated, albeit indirectly, with Lycia, and it is quite possible that this fact reflects memories of the early Lycian literacy testified by Mycenaean merchants coming there to trade metal and other Oriental goods. Also, the discovery of the wooden diptych tablet in the cargo of the Uluburun ship sunken close to the Lycian coast (about 10 km from Ka§/Antiphellos) is probably not as accidental as it may seem, even if the final destination of the ship and the kind of script used to write on the tablet remain quite un- certain The usual assumption is that the ship sailed to the Aegean, cf., e.g., Bachhuber 2006, Cline-Yasur-Landau 2007 or Goren 2013 with further refs. This is indeed quite possible, but by far not certain. In fact, the evidence of the letters from Ugarit adduced above may well suggest that the ship was sailing to -- or at least intended to visit -- Lycia.. It is not impossible that the script the tablet was most frequently exposed to was not the Ugaritic cuneiform alphabet or Linear B -- which might seem to be the most straightforward assumptions -- but a script which could be understood in Lycia and which there are good reasons to identify as Hieroglyphic Luwian The evidence of Hieroglyphic Luwian in southern Anatolia is extremely scarce, which is due probably first of all to the fact that is was written on perishable media, such as wood, and that the practice of sealing was less. The exact source of the borrowing of the scribal terms into Lycian is not entirely clear. In any case, there are no special reasons to connect it with the Hittite military involvement in the region, attested first of all by the YALBURT inscription (cf. above), which has probably never lead to the establishment of a Hittite administration in Lycia, as demonstrated, inter alia, by the fact that in the reign of Tudhaliya `III/IV' (ca. 1227-1209 BC) the region was considered as lying outside the Empirespread than in Central Anatolia. There is, however, one curious piece of evidence: a seal coming from a Late Mycenaean tomb at Ialysos on Rhodos (cf. Boardman 1966: 47-48 with fie. 2). I was able to examine the seal in the British Museum in July 2013, for which I am greatly indebted to Alexandra Villing and Andrew Shapland. Unlike some seals found in the Aegean (as the Perati seal), which appear to be simply coarse emulations of Luwian writing without any actual meaning, the epigraph of the seal makes an impression of being a genuine Luwian example. However, reading of the epigraph is not obvious. One can immediately identify only two signs on Side B: the title URCEUS and the last sign of the name, which is <ni>. The sign above <ni> is probably <wa/i>, despite the odd oblique position of its central element. The reading of the central sign is particularly difficult, since the shape does not exactly correspond to any attested sign. The two theoretical possibilities would be to see in it either a schematic representation of an animal head or a hand, although no attested HLuw. `hand'-signs have a triangular element in the upper part. Given the reading of the last two signs, one may tentatively propose to identify the sign as a very schematic form of CANIS.ZU(WA), which stands for zu(wa)na/i- (for the combination cf. Oreshko 2013: 413-416). Lastly, the upper sign might be simplified form of BOS = <u>. Accordingly, the name can be read (u-)CANIS.ZU(WA)-wa/i-ni. While Uzuwani remains a possibility, a reading Zuwan(n)i is more sensible, since such a name is indeed well attested (cf. Laroche 1966: s.v.v. Zuwanna, Zuwanni, Zu(w)ania; Euevvectlc, attested in later sources (cf. recently Simon 2019), is clearly its extension). On the Side A, in the central field there seems to be only one sign, which does not correspond to any HLuw. sign, but is quite reminiscent of a representation of a ship. This makes good sense in the geographical context of the find, and there are good chances that we are dealing with a seal which once belonged to a Lukka man.

⁴¹ Cf. the evidence of §10 the `Tudhaliya Instructions for Lords, Princes and Courtiers' (CTH 255.1) which mentions the `frontier posts' (auri-) between Hatti and Lukka, see Miller 2013: 286-287.. A most likely place for such a borrowing appears to be Plain Cilicia (Kizzuwadna), although a port city of the northern Levant, such as Ugarit where Luwian and Hittite scribes were certainly present, cannot be excluded either.

§6. The Aegean

Geographically, the Aegean lies even closer to Lycia than the Levant, and, if Attarima- Awpupa was indeed one the regions covered by the term `Lukka lands' (cf. above), the early Lycian ethnolinguistic area practically overlapped in part with the Aegean cultural sphere. There are no special reasons to doubt that Lycians visited the Aegean at least sporadically; the description of the venture of the author of the Xanthos trilingual (TL 44a: 53-55) into the region of Samos and Mykale is merely one example of such a visit. The material culture of the Classical Lycia demonstrates quite a few instances of Greek influence, and there can be no doubt that the Greek regularly visited Lycia and possibly even settled there from at least 800 BC (cf. below). However, Greek presence in Lycia dates to an even earlier time, and it seems that it went beyond simple contacts.

Greek literary tradition connected the very origin of the Lycians with the Aegean. There were two strands in this tradition^{42 For} details see Bryce 1986: 11-41, cf. also Keen 1998: 22-26.. The first one, reflected in the Iliad (6.152-205), associates Greek settlement in Lycia with Bellerophon, son of Glaukos and grandson of Sisyphos, whose homeland was Ephyra/Corinth. Sarpedon and Glaukos, the Lycian leaders in the Trojan war, were his descendants. The other strand of the tradition, known to later authors (Hdt. 1.173,

Str. 12.8.5, Paus. 7.3.7, Apollod. Bibl. 3.1.1-2), saw in Sarpedon a brother of Minos and, accordingly, connected the origin of the Lycians with Crete. Since Lycian is an Anatolian language, one cannot take this tradition quite literally: it is clear that a significant or even major part of the Lycian population in the 1st millennium BC, as well as its culture in general had local roots. However, it would be equally unwise to simply dismiss this Greek tradition as pure fantasy, as sometimes alleged See, for instance, Keen 1998: 26 with further refs.. In fact, accounts of settlement of Lycia from the Aegean agree rather well with the phenomenon of sea-born migrations from the Aegean to the East at the end of the 2nd millennium BC which can be glimpsed both from the epigraphic and historical record. These migrations resulted in Aegean settlement in Rhodos, Pamphylia, Cyprus, Plain Cilicia (Cilician Ahhiyawa), the Amuq Plain (Palastina/i) and even southern Levant (the Philistines) For the Aegean element Cilicia and the Levant cf. in general Singer 2013 and Oreshko 2018a with further refs.. In this context, it seems very likely that Lycia indeed received some Aegean ethnic element in this period, even if this has not resulted -- in contrast with Rhodos, Pamphylia or Cyprus -- in the establishment of Greek as the main idiom. Rather, the opposite process was the case: the Aegean settlers eventually switched to Lycian, becoming a part of the Lycian ethnos as we know it, a scenario which has parallels in Cilicia and the Amuq Plain Cf. Oreshko 2018a and, for Plain Cilicia, Yakubovich 2015.. From a so- ciolinguistic perspective, it appears very likely, almost inevitable, that this process left some traces in the Lycian language.

There is no need to argue for the importance of the Greek factor already in pre- Hellenistic Lycia: the Greek influence clearly manifests itself in Lycian art and architecture, as well as in the existence of Greek-Lycian bilinguals and the presence of Greek names in Lycian inscriptions Cf., for instance, Keen 1998: 66-69 with further refs.. However, we have next to no historical evidence bearing on the sociolinguistic framework within which the Lycian-Greek interaction took place, and reconstructing the details of this process is definitely not a trivial task. As a matter of fact, after the conquest by the Persian general Harpagos around 546/45 BC, Lycia stayed for almost two centuries under more or less strong Iranian/Achaemenid influence, even if it still enjoyed a great deal of political autonomy For the political history Lycia see in general Keen 1998, esp. 61-70 on the Iranian and Greek cultural influence in Lycia.. After the campaigns of Kimon in south-western Anatolia around 470 BC, Lycian cities joined the Delian League, but by the beginning of the Peloponnesian War (431 BC) most of them (except Tel(e)messos) defected from the Athenians, subsequently siding with the Persians and the Peloponnesian League, as is reflected inter alia in the Xanthos Trilingual composed around 400 BC. After ca. 360 BC Lycia came under the control of the rulers of the Carian Hekatomnid dynasty, who remained under Persian authority only nominally, leading a conscious politics of Hellenization, an important aspect of which was the usage of Greek as the official written language.

While contact with the Greeks and acquaintance with Greek culture is beyond doubt in pre-Hellenistic Lycia, it is far from obvious what effect this contact could have had on the Ly- cian language. Neither the participation of the Lycian cities in the Delian League, nor their dealings with the Peloponnesians should necessarily have led to any perceptible Hellenization of Lycians in a linguistic sense, although this interaction certainly increased the awareness of Greek in Lycia Thus contra Rurtherford 2002: 201-202 and Colvin 2004: 51-53. Rurtherford assumes usage of Greek as an `imperial language' in Lycia already during the Lycian alliance with the Delian League and counts with an increasing presence of bilingual speakers during the Dynastic and the Carian period. Such a scenario is quite dubious especially for the 5th century, and virtually refuted by the evidence of the Xanthos Trilingual which has 138 lines in Lycian, 103 lines in `Lycian B' and only a short (12 line long) Greek poetic part written probably by an invited versifier (cf. below). These Greek verses give also quite a clear idea of the level of mastery and perception of Greek in Lycia at this time, since it contains over 12 lines at least 8 mistakes (letter omissions).. The influence of the Hellenizing policy of the Hekatomnid dynasty, testified by several extensive monuments in Lycia written in Greek, including the Letoon Trilingual, might have been somewhat stronger, and in any case in the 4th century BC Greek was already on the way of becoming a lingua franca of the Eastern Mediterranean. However, the attitude of the Lycians to the Greek of the Carian administration should not necessarily be entirely positive, and again it is not clear how deep its influence on the Lycian communities in general might have been. On the other hand, we have very little evidence about the time and circumstances of Greek settlement in Lycia in the 5th and 4th century BC, and whether the Greeks formed in Lycia a strong linguistic community before the Macedonian conquest in 334/333 BC. The evidence of bilingual inscriptions is ambiguous, since it testifies only to the existence of a linguistic community, but ultimately says nothing about its size and language attitudes. Ironically, the presence of bilinguals may indicate that bilingualism was just not that widespread in the community. Greek names in Lycian inscriptions, which constitute less than 10% of attested names, do not constitute evidence for a massive presence of the Greeks in Lycia Cf. Colvin 2004: 51-52. In fact, the actual percentage may be closer to 5%, since a part of Greek names are attested in the Xanthos Trilingual, which has nothing to do with settlement of the Greeks in Lycia. It is not clear what exactly stands behind Colvin's (2004: 51) statement `In the fourth century approximately 66 per cent of the recorded names are still Lycian'. Most importantly, it is not clear what part of these inscriptions dates after 333 BC. Moreover, the percentage of Greek names should not exactly reflect the percentage of Greek speakers, since some Lycians might have adopted Greek names out of considerations of prestige.. In fact, judging from the general historical context, one would rather think that settlement of the Greeks in Lycia before ca. 350 BC was most probably a matter of private business, connected first of all with trade activities and professional domains, such as architecture, sculpture and possibly pottery production.

In support of this two pieces of epigraphic evidence may be adduced here. The first is TL 25, a text associated with statue bases intended as a dedication to Apollo in Tlos Judging from the letter forms of the Greek text, the monument belongs to the 4th century BC, possibly ca. 380-350.. The text combines a Lycian and a Greek part, which closely, although not quite exactly, correspond to each other The Greek text lacks words referring to the object of dedication, which is ebeis tikedris `these statues'

(acc.pl.) in Lycian, and a verb, which is tuwete `set up' (pret.3.pl.) in Lycian. The Lycian text, as we have it, lacks the name of a deity to which the statues are dedicated, which is AnoAAwvi (dat.sg.) in Greek. The addressee of a dedication is, however, a very important, even crucial element of a dedicatory text, and there is every reason to think that precisely this element is lost in the gap at the end of line 1 of the Lycian text. Kalinka (TAM 1: 24) tentatively reads the damaged letter at the edge of the gap as M. However, one would rather suggest that the letter is N, and the name is Natri, the Lycian counterpart of Apollo. Accordingly, the first clause of the text can be restored as ebeis: tukedris: N[atri=ti] tuwete: lit. `These (are) the statues which to Apollo dedicated...' (cf. already Oreshko forthc., §3, fn. 61).. The dedication is made by a Lycian from Tlos, Xssbeze (= nopua^ in the Greek part), on behalf of himself and his family (wife, daughter and a niece). However, the statues were created by a Greek, which is indicated in the Greek epigraph found separately on the next block: ©eoStoQog AGqvaTog euoqae. It is very likely that Theodoros had been specially invited by Xssbeze to work on the statues in Tlos, and this was probably a more or less established practice, judging from the Greek artistic influences in other Lycian tombs. It is not clear whether Theodoros stayed in Lycia or returned to Athens, but it appears quite probable that at least some Greek artists decided to stay and work in Lycia, attracted by the local demand.

The next, even more curious piece of evidence is preserved in a monument (a statue base) of Arbinas, a Lycian dynast who ruled in the early 4th century BC (see Bousquet in Metzger et al. 1992: 155-165). The stone contains two rather long Greek poems, one of which is preserved almost completely, and its two last lines give information about the composer of the poem(s). The author is Euqqaxog EuqqSeog neAAaveug a `blameless seer' (qavxic; d[|xuqu>v]), thus a Greek from the Achaean neAAqvq (Dor. neAAava; less likely from the Spartan neAAava, which was probably too small at this time to produce a seer). As suggested by Bousquet (in Metzger et al. 1992: 162), the presence of the Pellenian poet in Lycia may be connected with the fact that his native city was a part of the Peloponnesian League, and thus the ships from Pel- lene might have been involved in the Peloponnesian activities in the region of Lycia at the end of the 5th century BC. While one can only guess about the details of the peculiar poetic career of Symmachos, his poems offer very interesting insight into the sociolinguistics of Greek in Lycia. On the one hand, the very presence of the poems clearly testifies in favor of an interest of the Lycian elite in the Greek literary culture, and a certain prestige associated with it. It also shows that in the early 4th century BC there were people in Lycia who were able to read and appreciate Greek. This is precisely what one might assume by looking at the Lycian artistic monuments of this period, such as the Nereid Monument (ca. 390 BC), or the somewhat later funerary monument of Pajawa (ca. 370-360 BC). On the other hand, the poetic skill of the `blameless seer' is arguably one of a rather technical character, hardly much higher than that of an average educated Greek well-read in Homer, and the fact that it was necessary to bring him to Lycia from the far-away Pellene -- which is due probably more to circumstances rather than to actual premeditation -- seems to indicate that the Greek community in Lycia was still not very strong and the formation of the Greek culture in Lycia was only at the initial phase. Keeping in mind these sociolinguistic considerations, we may now revisit the evidence adduced so far in the discussion of Greek-Lycian language contact For the data see primarily Rutherford 2002 and Melchert 2014, which should be consulted for further refs..

Lexical Borrowings. There are two substantives attested in Lycian corpus which can be readily recognized as Greek words. The first is trijere, which is attested in the Xanthos Trilingual (TL 44b: 22 and 23) in the context of what appears to be a sea battle. The word apparently corresponds to the common Greek xQiqpqc; `trireme' (< xqlc; `thrice' + ЈQЈxqc; `rower'), which is originally an adjective used with vauc; `ship'. One should note that the context of the attestation is quite specific: the passage seems to refer to `Chian trireme(s)' (trijere Kijeze), thus Greek ships. It is not clear whether Lycian ships could also be called that way; in other words, trijere- may be a foreign word in Lycian, i.e. a not fully embedded term connected to a specific cultural phenomenon (just like trireme in modern English). The second word is sttala, well attested in the corpus (six attestations, cf. Neumann 2007: s.v.), which reflects Greek oxaAa `stele', being borrowed either from Doric or from another dialect preserving a. What is curious is that Ly- cian also has the root on which sttala is based, stta-, which seems to have the basic meaning `stand'. The connection between sttala and stta- was clearly felt by the Lycians themselves, since both appear as a figura etymologica in 44c: 5 and 7. The case is ambiguous and interesting. On the one hand, there are no words in Lycian beginning with the cluster st(t)- which can be doubtlessly defined as inherited, From a formal point of view, it is not impossible to derive the verb from PIE *steh<sub>2</sub>- (see Melchert 2018a: 3132). However, the fact is that no other Anatolian language suggests the presence of the initial s- in the root, cf. Luw. ta- `stand' ((CRUS)ta-) and Hitt. titti- and tittanu-. It is not excluded that the prothetic s- is a later feature, introduced in some IE dialects after the split of the Anatolian branch. and the immediate association of sttala and stta- suggests that both originate from the same source, i.e. stta- is a borrowing of the Greek Loxqqi/ioxaw (see Schurr 2014 [2016] with further refs., cf. Melchert 2016: 31). On the other hand, the verb stta- has a very basic meaning and its use was clearly not confined to combinations with sttala or similar objects (cf. sttati in 44b: 35 in connection with erbbi `battle' and sttati=ti in damaged context in TL 93: 2), and it seems odd that this verb might have been borrowed from Greek (cf. Neumann 2007: s.v.; for a possible solution of the paradox see below).

Other cases of possible Greek borrowings are more dubious. Neumann (cf. 2007: s.v. with further refs.) suggested that the term ammama- (five attestations in the corpus) may correspond to Greek dpwpog `blameless'; Melchert, following Starke (for refs. see Melchert 2003: s.v.), interpreted it as `fine, penalty'. It is quite difficult to decide between these two options: the latter is more straightforward, but it is striking indeed that all contexts seem to associate ammama- with animals `paid' to certain deities `Cow' (wawa-/uwa-) in TL 111:4, TL 131: 4 and TL 149: 9; puwa in TL 102: 3, possibly `goat'; and kerut[i] in TL 111: 3, which appears to refer to a `horned' animal (cf. Neumann 2007: s.v.)., which matches rather well the usage of dpwpog as an epithet of sacrificial victims in Greek. It is not impossible that the word could be borrowed into Lycian as a specific ritual terminus technicus, but at present this remains only a possibility. Two further cases are even less certain One should mention that Melchert (2003: s.v. sttrat[ ]), adopting an old suggestion by Savelsberg (see ref. in Neumann 2007: s.v.), tried to see in the fragmentary sttrat[ in TL 44b: 18 a reflection of Greek OTpaTpyog. This is quite unlikely, since, in all probability, the word picks up the Greek name attested several lines above in the text (ll. 15-16) Stt[...]ani[...]he). Whether the restoration Stt[rat]ani[da]he) suggested by Stoltenberg (cf. refs. in Neumann 2007: s.v.) is correct or not, the name should be in any case a composite Greek name based on oxpaxog.. The word manaxine found on monument of Pajawa (cf. above §3 with fn. 24) represents either a title or a sort of ethnic connected with the Pam- phylian origin of Pajawa, and its connection with Greek povoyevqg is quite arbitrary. The suggestion that garai, attested in TL 44b: 62 (followed by zeusi), represents an adoption of Greek dyopa (Shevoroshkin 2011: 34, cf. Melchert 2014: 68) is equally problematic. The relevance of the attestation of neleze Tarqqnt- = Zeug dyopafog in the bilingual N324 is questionable, since it may be asked why one did not use neleze also in TL 44b, and it is difficult to either give a convincing Lycian interpretation of the morphology of the word or to explain the phonetic development of garai from dyopafog Formally, garai looks like comm.nom.pl. of an n-stem (cf. mahai `gods' or tahntai = oiKqpaxa, for which cf. below), and, despite the formal discrepancies, should probably be connected with the preceding ebeija `these' (nom.-acc.pl.neutr.), since independent usage of pronouns seems to be otherwise not attested in Lycian. It may be noted that the idea to interpret zeusi as a dat. form of Zeus strikes one as rather bizarre, despite the phonetic similarity of the words. As a matter of fact, the Lycians `translated' the name of gods, even when they were used in Greek contexts, cf. the epigraphs on the vessel with the scene of the Judgement of Paris (N307), which features Pedrita (Aphrodite), Alixssa[tra] (Paris-Alexander) and Mal[ija] (Athene); or Turaxssa[l]i: Natri in TL 44c: 47-48 who is Greek Apollo of the Mount Torax; or translation of the personal name ArcoAAoboxog as Natrbbijemi. To this should be added the odd dative form in -si, which would presuppose the perception of nom.sg. ending as part of the stem. It seems that the level of knowledge of Greek in Lycia around 400 BC was still higher than one which could allow for such a blunder. In sum, zeusi is probably not Zeus at all.. But even if garai zeusi is an odd phonetic rendering of Zeug dyopafog, it in no way suggests that *gara- was present in Lycian as an independent word, since in this context it would merely be an epithet of a foreign deity. In fact, both its bizarre phonetic form and the evidence of N324 which translates dyopafog with neleze plainly runs against the assumption that dyopa was present as a borrowing in Lycian.

On the other hand, there are two terms which can be identified as Lycian borrowings in Greek, although their usage was clearly localized. One is pivSig (and pevSfxqg, derived from it), which reflects Lyc. minti. The term likely referred to the local community as a whole, representing a close counterpart of Greek Sqpog, and not to `cemetery administration' as sometimes claimed (see in detail Oreshko 2019: 105-117 with further refs.). The other term is maxpa which is a feminine kinship term, possibly referring to `daughter-in-law' (see Schurr 1999). As argued by Schurr, the word may be based on the Anatolian root `to give' (piya-), thus designating `one who bears gifts = dowry' Thus contra Brixhe (1999: 89-91) who supported a connection with rcdxpa (`father's sister, aunt') presented in LSJ.. The fact that one adopted the Lycian term instead of using a Greek term, such as vuog or vuq^q, is due apparently to the specific local character of marriage regulations and the legal status of daughters-in-law.

Lexical and semantic calques. Several words have been suggested to represent Lycian literal translations (`calques') of Greek terms. Creation of calques is possible even in the situation of only relatively slight language contact, and the level of Greek-Lycian relationships was in any case enough for that. However, the evidence is not too impressive. Rurtherford (2002: 204-205) suggested that the term kumehe/i- found in the Letoon Trilingual (N320: 27) as a noun `sacrificial animal' -- as contrasted with its usual adjectival sense `sacred' -- is a calque of Greek lЈQЈlov `sacrificial animal' used in the Greek translation of the respective clause. This seems entirely possible in the context of the text, but it is not clear whether it is a real sociolin- guistic phenomenon, or simply an effect of translation of this particular text from Lycian into Greek. On the other hand, the idea is so trivial that one did not probably even need a Greek word to produce something similar in Lycian. Equally trivial is another pair tentatively adduced by Rutherford (2002: 205-206) as an example of calque: Lyc. prnnezi(je)- `household member' vs. Greek oiKЈioq `Household members' obviously existed in Lycia before the Greek settlement there, and one does not actually need the Greek word to explain the transparent semantic and morphological structure of the Lycian one. The third example comes again from the Letoon Trilingual and concerns Lyc. ahntai (N320: 17). This word was traditionally translated as `possessions' and connected with the Lyc. verb es-/ah- `to be', which automatically suggests a correspondence with Greek (xa) ovxa or ouoia (cf. Neumann 2007: s.v. and Rutherford 2002: 205). However, ahntai is most probably simply a phantom word. As pointed out by Schurr (2016: 125 with fn. 6) and Melchert (2018b, ad 11abc), there is no motivation for the presence of the enclitic particle -te having a locative semantics in the clause N320: 17-18, and the sequence setahntai can be interpreted simply as se-tahntai. Moreover, the Lycian word corresponds in Greek not to (xa) ovxa or ouoia, but to oiKqqaxa `dwellings', and there are thus no reasons to assume any interference between the two terms Schurr (2016: 125-126) suggested that the term tahnta- may be connected with 66e- (< *tahe) which in combination with kumezije `sacred, sacrificial' corresponds to Greek |3«pog `altar (with base), raised platform' in the Letoon Trilingual, and with Lycian B tasntuwadi. The latter connection is unverifiable, but the connection with 66e- looks plausible. It was convincingly argued by Schurr that there is no reason to derive either 66e- or Luw. tasa/i-, connected with it by Eichner (1983: 60-61), from PIE *d<sup>h</sup>eh<sub>1</sub>s- `holy, divine'. A connection with PIE *d^hh₁-, more positively assessed by Schurr, is possible, but is neither compelling. Luwian evidence implies rather that tasa/i- means simply `stone block' and may be an Anatolian areal term. The context of CEKKE §15 strongly suggests that FINES-ha+ra/i-ia ta-sa are `border stones', and the interpretation of tas(a/i)- as `stone block' well agrees with the context of KARKAMIS A6 §27-28 which features tasa/i- parallel to another term for `stone (block)', SCALPRUM-su(wa)- (= asu(wa)-). It is not impossible that (“*256”)td-sd- found in KULULU 2, §6 represents the same word, although the context does not unequivocally support this, and the difference in spelling (<ta> vs. <ta>) may be significant. It is noteworthy that neither attestation of the word features an ideogram hinting at an action (e.g. CRUS or PONERE), which implies that synchronically the term was not regarded as a derivative of a verb. Accordingly, tahnta- = oiKqpaxa probably just designated something built of stone blocks (rather than `installation')..

To these, one further piece of evidence may now be added. In a recent article (Oreshko 2019: 95-101, esp. 100) I suggested that the Lyc. term alaha- `concede' is based on the Luwic word for `place' (Lyc. ala(d)-) and means literally `to let somebody (into) a place', thus finding a close semantic correspondence in Greek anyx^ppaai `concede, let', which is based on Xd>pog. Given the striking correspondence in the underlying semantics of both terms built around the idea of `place/space' (ala(d)- and x^Qog), which is not that trivial, it would be natural to assume that one of the terms influenced the other. Since the meaning `concede, let' is normal for ouyxwQqaai, the direction should be Greek > Lycian, even if it may seem strange that a term connected with a specifically Lycian burial culture would have been influenced by Greek. Lastly, one should mention the idea of Seyer (2006: 727) that the common specification used in Lycian funerary inscriptions (hrppi:) atli: ehbi `for himself' is an attempt to render in Lycian Greek the word eauxq». The idea is interesting, but is difficult to verify. Contra Mel- chert (2014: 69), in the context of meticulous stipulations of the Lycian funerary inscriptions the usage of atli: ehbi does not seem especially redundant. The question is what one would expect to find in a Luwic inscription in a region outside the contact area with the Greeks.

Structural Influence. Besides borrowing and calques, Rutherford (2002) and Melchert (2014) have discussed several possible instances of a finer structural interference between Greek and Lycian. Doubtlessly the most intriguing case is the semantics of the Lycian adverbial element epi `upon' and the composite hrppi `upon' (< hri `above' + epi). The semantics of both Lycian adverbs seems to correspond to that of Greek eni rather precisely, but is far from the semantics of its presumable etymological counterparts, Luwian appi and Hittite appa, both of which mean `back(ward), again'. In the preserved texts, Lycian epi does not have the meaning `back' at all, although the very existence of the composite hrppi might imply that it was still present in the simple form epi. It is noteworthy that the Lycian usage of hrppi might in its turn have influenced the mode of usage of Greek eni (cf. Rutherford 2002: 206). The next feature concerns the Lycian usage of the connective se which quite exactly corresponds to that of Greek Kai, starkly contrasting with the exclusive preference of Luwian (-ha), Hittite (-a/-ya) and Lydian (-k) for enclitic conjunctions following the second member. Furthermore, this is not the only unusual feature of the Lycian clause architecture, since Lycian syntax is different from what one can usually see in Hittite and Luwian texts in other respects as well. This is particularly noticeable in the typologically rare OVS word order as seen in the common Lycian funerary formulas (cf. Rurtherford 2002: 214), and in the generally quite flexible clause structure seen especially in the longer texts, such as the Xanthos and the Letoon Trilinguals It is noteworthy that in a recent article Mouton-Yakubovich (2020) make an attempt to establish links between the unusual Lycian syntax and the proleptic construction found in Luwian. In particular, they suggest that Lycian OVS clause construction with a nasalized preterits (e.g., prnnawate) -- which probably contain an enclitic -(e)n (comm.acc.sg.) (for the interpretation see refs. there) -- should go back to a proleptic construction which would be directly comparable with the ones observed in Luwian. I am not convinced that such a reconstruction would be the only possibility for Lycian. However, the many interesting examples adduced by Mouton-Yakubovich demonstrated that Luwian syntax was flexible enough and that the tendencies which later lead to the Lycian OVS construction might have been present already on the early language stage..

Lastly, it has been suggested that the formation of some Lycian personal names might have been influenced by Greek names. In particular, this might be the case with Lycian names containing a participle in -mi- in the second part of the composite, such as Natr-bbije-mi `Given- by-Natr(i)' which corresponds to AnoAAoSoxog in the Letoon Trilingual. It seems that names of this structure were indeed largely restricted to Lycia with sporadic irradiation to Pisidia (cf. also Melchert 2013: 41-42), while in other Luwic areas it was enough to use a simple verbal root, either in the first or second part of the name. To this a further possible example of an onomastic influence may be now added. In the discussion of the hatahe-passge of the Xanthos trilingual, I have argued that Herikle mentioned in TL 44a: 50 has nothing to do with the mythical Herakles, as is usually claimed, but refers to a real person, in all probability a late 5th BC century governor (sehaxlaza) of Kaunos (Oreshko forthcoming, §9). The name might correspond to HpaKAag sporadically attested in the Classical period, but there is actually no real necessity to interpret it in this way, especially given that Herikle was a governor installed by the Persian administration. The same is probably true also for another Lycian name of a comparable structure, Perikle, who is well known as an early 4th century BC dynast of Limyra: it is difficult to suspect pro-Athenian sentiments in a Lycian dynast, still wholly in the sphere of the Persian influence; even more difficult to see in Perikle an admirer of Thucydides whom the Athenian statesman nepiKAqg arguably owes a great deal of his present fame. In fact, both names can well be genuine Lycian composite names: the first parts heri- and peri- may be easily explained as Anatolian adverbial elements (cf. Lyc. hri `up, on (top)' and *peri = Luw. pari `beyond, exceedingly)'), and the root kle- might well be present in Lycian as well Cf. hrkkeledi (instr.) in N324: 11); muni-klei-me in TL 107a:2, as contrasted with muneite in TL 127: 2 and muneita in TL 44b: 20; and Lycian B kllei-me (kllei-ma in 44c: 45, kllei-me in 44d: 61 and kllei-me-di in 44c: 49 and 60).. However, interpretation of -kle as `fame' (= Greek kACog) makes a good sense: Heri-kle and Peri-kle can be interpreted as `Upmost-Fame' and `Exceeding-Fame' respectively. It is not quite impossible that Lycian could preserve a native reflex of PIE *kleu- (with a loss of u in the syllable-/word-final position), but in the areal context it seems likelier that the popular Greek names in -KAqg played a role in the introduction of the pattern of names in -kle in Lycia.

Summarizing the evidence, one can note the following principal points:

1. The number of Greek lexical borrowings in Lycian is in fact very low. Both trijere (if it was indeed embedded in Lycian) and sttala are cultural terms, and their adoption does not presuppose any intensive language contact, let alone bilingualism. If one accepts the connection of ammama- with aqwqog, this would provide an interesting glimpse of an influence of Greek ritual terminology in Lycia. In view of this, the case of stta- `stand' looks quite strange.

2. The number of possible calques is not much higher. The closeness of morpho-semantic structure of kumehe/i- = lepeiov, prnnezi(je)- = o'lxeiog and alaha- = auyxwpqaai is certainly notable, but ultimately it demonstrates similarity in thinking rather than a straightforward linguistic influence of Greek.

3. The structural similarities are more impressive and intriguing. Even if the number of arguable cases is still not too high, it would be fair to say that from a structural point of view, Lycian is closer to Greek than, for instance, Luwian or Hittite.

Seen from a sociolinguistic perspective, the picture is quite puzzling For the typology and scale of contact-induced language changes see Thomason-Kaufman (1988: 74-95), cf. Thomason 2001: 59-98, esp. 70-71. Cf. Oreshko 2018b: 95-102 for general observations on the sociolinguistics of Greek-Anatolian language contact.. On the one hand, neither the actually attested lexical borrowings, which are the clearest and most basic indicators of the language interaction at its initial phases, nor the general sociolinguistic situation as it can be reconstructed for the 5th and the early 4th century BC Lycia hint at a significant level of Greek-Lycian bilingualism. Greeks were clearly present in Lycian cities as merchants and artisans, but the Greek community was probably still rather slim, and there is no question of a `Greek-Lycian' ethnocultural merge at this time. On the other hand, the structural similarities between Greek and Lycian suggests a very high level of bilingualism in the whole community, when two languages begin literally to intertwine and to align their structures in the minds and on the tongues of the speakers.

The paradox can be explained if one goes beyond the chronological framework imposed by the definition `Greek-Lycian contact' -- 6th-4th centuries BC -- and brings into the picture the early migration to Lycia from the Aegean reflected in the Greek legendary tradition. One may suggest a different model of the `Lycian-Aegean' ethnolinguistic contact than applied hitherto: one assuming two essentially different phases and associated phenomena. The first phase is connected with the migration from the Aegean, for which the associated events in other parts of Eastern Mediterranean (Rhodos, Pamphylia, Cyprus, Plain Cilicia etc.) suggest a rough dating to the 12^th-11th centuries BC. This migration has probably resulted in the settlement in the Xanthos valley of a more or less substantial group of Aegean migrants, who were not necessarily exclusively Greeks, and their subsequent intermingling with the local Anatolian population with the formation of a largely bilingual community. For reasons that remain unclear, the community eventually switched completely to Lycian, but the process left a number of traces in the language, primarily on the structural and conceptual level. It is noteworthy that the time depth assumed for the phenomenon -- 500-600 years before the actual epi- graphic attestation -- agrees rather well with the apparent embedded-ness of the `Aegean features' in Lycian.