Hypotheses of interference between Greek and the languages of Ancient Anatolia: the case of patronymics

Ðàçìåùåíî íà http://allbest.ru

20

Universidad Complutense de Madrid

Hypotheses of interference between Greek and the languages of Ancient Anatolia: the case of patronymics

Bartomeu Obrador-Cursach



Abstract

Following an overview of how the different languages attested in Anatolia during the Iron Age express patronymics, this paper explores the alleged interferences among the strategies found in these languages. Particular focus is placed on the possible interactions between Greek and the Anatolian languages in the use of genitive patronymics with or without a noun for `son' or `daughter' (following prior studies by Merlin and Pisaniello 2019 and Rutherford 2002) and on the claim of a Lydian origin for Greek patronymics in -i5aç/-i5r|ç (Dardano 2011), for which an internal Greek development is accepted after the inclusion of relevant data from Phrygian. All in all, very few local interactions are sustained as being valid.

Keywords: patronymics; Anatolian languages; Phrygian language; Greek language; language contact.



Introduction

A patronymic is a complement to the personal name based on the given name of father's name.

They were used to identify a person in many ancient cultures before surnames replaced them, although many modern surnames are fossilized patronymics, for example, Scottish Gaelic MacNéill `the son of Niall', English Peterson `the son of Peter', Spanish

Martinez `the son of Martin' and Portuguese Henriques `the son of Henrique'. Ancient traditions are still in use in places like Iceland, where there are no surnames.

For instance, the name of the current Icelandic president is Guöni Thorlacius Jôhannesson: Guöni Thorlacius is his first name and Jôhannes- son derives from his father's name, Johannes Sæmundsson, through the addition of the suffix -son `son' (as in many current Germanic surnames).

Because of their social relevance and their relationship with personal identity and ethnicity, patronymics can be used as a source of cultural information on a given community and can provide details about linguistic contact.

Earlier scholars have suggested some cases of contact between Anatolian languages and Greek on the basis of shared features with regard to the formation of patronymics. However, few of the proposals in the literature are based on updated knowledge of the Anatolian material and most avoid providing an overview of the subject.

Moreover, in the last decades, the ancient languages of Anatolia (crucially for this paper, those of the Anatolian branch of the Indo-European family and Phrygian) have been subject to more systematic scientific study and the data, when available, are now more abundant and more reliable.

Therefore, we have a good opportunity to provide an updated overview of patronymics in the languages attested during the Iron Age (§ 1) There are no claims of contact between Greek and Anatolian languages based on patronymics during the Bronze Age. Therefore, I leave aside the intricate question of the contact between Greek (attested only in the Mycenean dialect) and Hittite and Luwian. and to review prior proposals of contacts in the light of our current knowledge (§ 2, 3 and 4). The Greek patronymic suffix -iSaç / -iSqç is one of the most important points here, since it has been identified as a possible borrowing from Lydian (Dardano 2011). Finally, an account of Phrygian variation is given (§ 5), since this language has been largely overlooked in the linguistic map of Anatolia, despite the central position it occupies.



1. Overview of patronymics attested in Anatolia during the Iron Age

There were different ways of expressing the patronymics in Anatolia during the Iron Age, and in some cases several of these are attested in a single language. As we will see, Hieroglyphic Luwian, Lycian, Phrygian and Greek use different strategies to form patronymics, while Carian, Lydian, Sidetic, Phoenician, Aramaic and Urartian have only one patronymic form. It may be significant that of the languages in this group, it is the Anatolian ones that remain more cryptic (Lydian and Carian) or present an extremely fragmentary corpus (Sidetic). Note that Carian has personal names that can be considered to have a patronymic meaning, if mno- is accepted as meaning `son' (see fn. 9): cf. 'EKaxôpvmç, Carian ktmno-, ktmno- (Adiego 2007: 375, 378), as adduced by Yakubovich in his discussion about the origin of MeQ-pv-dh-eç (2017: 289).

The most widespread patronymic form in Anatolia is the creation of a “possessive” or “genitival” adjective, which derives from the father's names through a suffix. There are two variants in the different languages: those which place the adjective in agreement with the nouns for `son' and `daughter' and those in which the adjective stands alone. In the first group of languages, which use the noun, we find Hieroglyphic Luwian, which uses adjectives derived through the suffix -assa/i- (1), and Lycian, which uses genitival-adjectives derived through -h- (when governed by nominative)/-he (by locative) / -hn (accusative). I follow here the analysis by Adiego 2010. This work is not considered in Merlin and Pisaniello 2019, who assumed a derivation through the Lycian suffix -ahi-/-ehi-. This last suffix commonly creates adjectives, but never patronymics. For an overview of the possessive adjective in the Luwic languages in place of the genitive case, see Melchert 2012.

1. Hieroglyphic Luwian: Bohça § 1

[ ! ?]ku+ra/i-ti-i-sâ \â-<sa-hwa/i-si>-sa4 !

HEROS-li-i-sa ! (“INFANS”)ni-mu-wa/i-za-sa

Kurtis-n.sg.C Ashwis-adj.n.sg.C. of the hero-n.sg.C. son-n.sg.C.

`Kurtis, the hero Ashwis' son'

2. Lycian: TL 36, 2 (Xanthos)

ahqqadi: pizibideh: tideimi

Ahqqadi-nom.sg.C. Pizibide-adj.nom.sg.C. son-nom.sg.C.

`Ahqqadi, son of Pizibide'

3. Lycian: TL 25a, 5-6 (Tlos)

tikeukëprë... urtaquijahn kbatru

Tikeukepre-acc.sg.C. Urtaqija-adj.acc.sg.C. daughter-acc.sg.C.

`Tikeukepre, daughter of Urtaqija'

In the second group, those languages in which the adjective stands alone can be classed as Lydian, which builds an adjective through the suffix -li- (4), Lycian (5), Phrygian through -evais- (6) and, perhaps, -(e)yo- (7), Aeolic Greek through -yo- (8) and Urartian through -ha (9). Carian could be added to this list, if forms in -s are not considered as genitives (see fn. 10).

4. Lydian: LW 001, 4 (Sardes)

manelid kumlilid silukalid

Manes-adj.nom-acc.sg.N.

Kuml(i)-adj.nom-acc.sg.N. Seleukia-adj.nom-acc.sg.N.

`(belongs) to Manes (the son) of Kumli- (the grandson of) Seleukia'

5. Lycian: TL 105, 2 (Limyra)

esete muleseh

Esete-n.sg.C Mulese-adj.nom.sg.C `Esete (son) of Mulese'

6. Phrygian: M-01a (Yazilikaya)

ates arkievais akenanogavos

Ates-n.sg.M. Archias-adj.n.sg.M. akenanogavos-n.sg.M.

`Ates (son) of Archias the holder of the akenan'

7. Phrygian: G-183 (Gordion)

tiveia imeneia

Tiveia-n.sg.F. Iman-adj.n.sg.F. ?

`Tiveia (the daughter of) Iman' ?

8. Greek: Buck 24 (Kebrene)

'ϲ E0£V£Lai [...] TÖ NLKLaLOL

prep Stheneias-dat.sg.M. Nikias-adj.dat.sg.M.

`upon Stheneias (the son) of Nikias'

9. Urartian: CTU 1 A 08-01, 2-3 (Van)

mar-gi-is-ti-[se] mmi-nu-u-a-hi-ni-se

Argisti-erg.sg. Minua-adj.erg.sg.

`Argisti (the son) of Minua'

It is worth recalling here that Greek has different patronymic adjectives but only the Aeolic patronymic in -loç 5 is found in Anatolia instead of the widespread genitive. Other Greek types are reported by the Tékhnë Grammatike (GG I, 1: 25-26):

naxpwvupiKdv pev oùv ecttl to Kupiwç and naxpàç ecxppaxiopevov, KaxaxppoxiKmç 5e êà³ òî and npoyovwv, oiov npAei5pç, À³àê³áðä î AxiAAeùç. Tùnoi 5è xôv naxpwvupiKmv àpoeviKmv pev xpelç, î eiç 5pç, î eiç wv, î eiç à5þñ, oiov Axpei5pç, Axpeiwv, xpelç, î eiç 5pç, î eiç wv, î eiç à5þä, oiov Axpei5pç, Axpeiwv, êà³ î xwv AioAéwv l5loç xunoç Tppa5ioç- "Yppa yàp nalç î Ïüòòàêîä. 0pAuKà>v 5e oi lool xpelç, î eiç lç, oiov npiapiç, êà³ î eiç aç, oiov neAiaç, êà³ î eiç vp, oiov A5ppoxivp. And 5e ppxepwv où oxppaxiCei naxpwvupiKàv ei5oç î "Opppoç, aAA' oi vemxepoi. There is also a variant in -eioç as seen in [Aql]otôUvoç BdKxeioç `Aristoxenos (the son) of Bakkhos' (Né-cropole de Myrina 116, 20). It is a fact that in Homer there are no examples of patronymics derived from any mother's name. However, this strategy is not unknown in archaic poetry. See, nepoeùç Aavaiôpç `Perseus the son of Danae' in Aspis 229.

`The patronymic is, properly, a form derived from [the name of] the father and by extension from [that of] the ancestors, such as Pelides or Aiacides said of Achilles. There are three types of masculine patronymics: the type ending in -5pç, that in -wv, and the type proper to the Aeolians, in -a5ioç, as respectively Atreides, Atreion, and Hyrradios. Pittacus was, in fact, the son of Hyrra. Similarly, there are three types of feminine forms: the type in -lç, that in -aç, and that in -vp, as Priamis, Pelias, and Adrestine. Homer does not derive the patronymic from [the name of] the mother, but the new poets do'. Translation by S. Merlin, who generously shared it with me.

The famous patronymics ending in -i5aç/-i5pç are restricted to literary sources (mainly after the influence of Homer), with the sole exception of some cases found in syllabic inscriptions from Cyprus (see below). It is true that the suffix -i5aç/-i5pç is found in some personal names from Anatolia but it is never used as a proper patronymic. No examples of patronymics ending in -iwv and -aSioç (both secondarily derived) are found in epigraphic sources, including, of course, the inscriptions from Anatolia.

A second way of expressing patronymics in the languages attested in Anatolia is the use of the genitive case. Again, the languages have two variants: the genitive can occur followed by the noun for `son' or `daughter' or it can stand alone. Languages that use the first variant include Hieroglyphic Luwian (10) and Greek (11). Carian may also be added to this list, if the identification of mno- as the word for `son' is accepted (12). l follow here the identification by Adiego (defended, e.g., in 2010b: 167-170). However, an alternative inter-pretation was given by Schürr (2013: 28-29) and Simon (2019: 299-302). One could also include Lycian in this list (13), if the debated “adesinential genitives” reflect the inherited genitive ending *-s > -0 (Adiego 2010: 5). There are few examples of such an alleged category (TL 5 4, 61 1, 77 2, 117 3, 127 1, N 315 2, with 0 in TL 69 2). Cf the critical remarks in Schürr 2010: 120-121.

10. Hieroglyphic Luwian: HAMA 4 Hama A1 § 1

u+ra/i-hi-li-na PRAE-tâ-sa !lNFANS.Nl-wa/i-za-sa

Urhilina-n.sg.C Paritta-gn.sg.C son-n.sg.C

`Urhilina the son of Paritta'

11. Greek: TAM l 117 (with Lycian TL 117, Limyra)

EiSapioç naQiaévovTOç uloç

Sidarios-n.sg.M Parmenon-gn.sg.M son-n.sg.M `Sidarios the son of Parmenon'

12. Carian: C.Ka 1 (Kaunos)

psusoAs mals: mnos

PsusoA-gn.sg.C Mal-gn.sg.C son-gn.sg.C

`of PsusoA, son of Mal' (Adiego 2010, 170, with caveats)

13. Lycian: TL 61 (Phellos)

sbikezijëi: mrexisa: tideimi

Sbikezijëi-n.sg.C. Mrexisa-gen.sg.C son-nom.sg.C.

`Sbikezijëi son of Mrexisa'

The second group of languages with this structure uses a sole genitive to express the patronymic, without using the noun for `son' or `daughter'. Here we include Carian (14, with caution), Carian forms in -s could be also interpreted as adjectives, since it is difficult to differentiate them from a possible genitive (Adiego, personal communication). Sidetic (15), Phrygian (16) The Phrygian patronymic or papponymic manitos in manes iyungidas manitos (B-07) can be added to this list, although the form iyungidas remains unclear (see below). It should be noted that, in a recent lecture, Rostislav Oreshko defended the identification of masa urgitos as the Luwian name Masaurahisas (cf. ]ma-sa-MAGNUS+ra/i-hi- sà-sâ in Porsuk 1 § 4). and Greek (17). ln addition, despite being attested only in lmperial Roman times, Pisidian exhibits the same feature: e.g. 5 Mouopxa Tag `Museta (the son) of Ta' (a man is depicted on the stele where the inscription occurs). Again, Lycian could be added to this list (18), if the “adesinential genitive” is accepted. Note, however, that the only Lycian inscription containing this kind of filiation formula is TL 145, and unfortunately it shows some gaps. As can be seen, Carian, Greek and, perhaps, Lydian provide examples of both variants (with and without the noun for `son' or `daughter').

14. Carian: C.Hy 1, 4 (Hyllarima)

mane : usoAs

Manes-n.sg.C UsoA-gn.sg.C or `Manes (the son) of UsoA'

15. Sidetic: S1, 1 (Side)

artmon Oanpijs

Artemon-n.sg.C Athenobios-gn.sg.C `Artemon (the son) of Athenobios'

16. Phrygian: K-01 (Kerkenes Dag)

masa urgitos

Masa-n.sg.M Urgi(s)-gn.sg.M `Masa (the son) of *Urgi(s)'

17. Greek: TAM II 232, 1-2 (Sidyma)

Xpuainnoç Zwoipou

Chrysippos-n.sg.M Zosimos-gn.sg.M `Chrysippos (the son) of Zosimos'

18. Lycian: TL 145, 1 (Limyra)

hla nterubila

Hla-n.sg.C Nterubila-gen.sg.C `Hla (the son) of Nterubila'

Finally, the two Semitic languages attested in Anatolia, Phoenician and Aramaic, use their particular, inherited system: construct from of the word for `son' followed by the father's name (19 and 20).

19. Phoenician: KAI 24 (Sam?al)

klmw br hy?

Kilamuwa son-cons.sg. Haya `Kilamuwa the son of Haya'

20. Aramaic KAI 318 (Daskyleion)

?lnp br ?sy

Elnaf son-cons.sg. Asay `Elnaf son of Asay'

	Genitive alone	Genitive + `son/daughter'	Possessive adjective	`son/daughter' construct + name
Hieroglyphic Luwian	-	+	+	-
Lycian	±	±	+	-
Carian	±	±	±	-
Lydian	-	-	+	-
Sidetic	+	-	-	-
Phrygian	+	-	+	-
Greek	+	+	+	-
Phoenician	-	-	-	+
Aramaic	-	-	-	+
Urartian	-	-	+	-

2. The Aeolic patronymic adjective as influenced by the Anatolian languages

A well-known feature of Greek as recorded in Aeolis and, mainly, in Lesbos is the use of the adjectival patronymic ending in -ioç instead of the widespread genitive. This has sometimes been considered an influence of those Anatolian languages which use a patronymic adjective, such as Lydian or Luwian (Hawkins 2001: 58-59). A problem derives from the fact that the Aeolic forms can be considered an archaism because the suffix *-iyo- forms possessive adjectives in a variety of Indo-European languages (Yakubovich 2010: 148-149). The question is difficult to evaluate, since we do not have a morpheme borrowing in Greek (see below) and the patronymic use of *-iyo- > -io- is not so far removed from other occurrences in Greek: it forms an adjective to express belonging or relation (àypoç `field, country' >> aypioç `living in the fields; wild').

Therefore, there are three possible scenarios: an archaism in Aeolic, an internal development, or a development triggered by the contact with the Anatolian language. Although far from certain, it is possible that a feature - the possibility of creating an adjective from a personal name - was promoted following the strategies of other neighboring languages, as a kind of influence.

However, Mycenaean shows the very same patronymic adjective for a Greek dialect during the Bronze Age. For example, in a-re-ku-tu-ru-wo e-te-wo-ke-re-we-i-jo Alektruon Etewoklewe- hiyos `Alectryon (the son) of Eteocles' (PY An 654.8-9), the suffix -i-jo /ijos/ (< *-iyo-) is used to build a patronymic after the personal name found in alphabetic Greek as TxepoKÀépqç (classical 'ExæKÀqç).13 Of course, it is feasible that both Mycenaean and Aeolic developed this strategy of building the patronymic adjective from an inherited adjective independently. But the simplest scenario is to assume that Mycenaean and Aeolic patronymics in *-iyo- are a common, inherited feature.

It is possible that, unlike the other Greek alphabetic dialects, Aeolic preserved this kind of patronymic because of its similarity to the analogous feature in neighboring languages, but this is difficult to prove; as an archaism, however, it does not need more substantiation than already provided by the current data. Note also that this feature could be treated as a typical archaism of a peripheral area. The same conclusion is reached by Hajnal (2018: 2046-2047), who adduced three points: 1. “io-adjectives that denominate belonging or possession of objects are not attested in Lesbian”. 2. In the Luwic languages, */-io-/ was affected by the i-mutation, which does not occur in Greek io-patronymics. 3. As in Mycenaean, Lesbian shows that the possessive adjectives in /-io-/ interfered on “material adjectives” in /-e(i)o-/. It is possible that Phrygian, the closest language to Greek in genetic terms, also has the suffix *-eyo-, used to form patronymics, if tiveia imeneia `Tiveya (the daughter) of Iman' (G-183b) is, in fact, a feminine personal name followed by a possessive adjective derived from iman (Obrador-Cursach 2020: 254) and not a “material adjective”: `those things belonging to Ti(v)es Iman' (vel. sim.). Leaving aside this example, Phrygian -eyo- expresses ethnics (x£|XQoy£ioç 1.1 = 48, derived from the name of the river Thymbris), matar kubeleya (B-01) ~ kubileya (W-04) `Cybele' (literally, `the mother from Kybelon') and seems to derive adjectives from nouns like mireyun (B-05) from meros (B-07)/qipoç (MPhr-01) `?'. In any case, it seems that, contrary to Aeolic Greek, in Phrygian the suffix -eyo- influenced the possessive adjectives in *-eyo-. One can add that these patronymics also occur in some names with the ending -i-jo in agreement with the nouns ko-wo /korwos/ `boy' and u-jo, i-*65 and i-jo `son', so they could be very productive in Mycenaean. Never-theless, these endings can be also interpreted as a variant of the thematic genitive in -i-jo-jo (cf. Duhoux 2008: 357-359).

3. Greco-Anatolian influences in the use of `son/daughter' after genitive

for the patronymic formula?

In a recent paper, Merlin and Pisaniello (2019: 97-98) raised the question of whether the Greek pattern genitive + uloç `son' is triggered by Lycian genitive + tideimi `id.' in the bilingual inscriptions from Lycia. Note that they operate with the traditional assumption that Lycian patronymics in -h are mere genitive forms (as in eDiAna, but against Adiego 2010). Rutherford (2002: 210-212) stated earlier that the occurrence of uloç `son' following the genitive of the father's name in the Greek part of the bilingual inscriptions is a calque of the Lycian formula father's alleged genitive + tideimi. After consulting the bilinguals and the uses in the monolingual Greek inscriptions from Lycia, Merlin and Pisaniello observed that in the eight useful bilinguals, three texts show the Greek formula genitive + uloç for the patronymic where the Lycian part has genitive + tideimi (TL 72, On this text, see the new edition and commentaries by Christiansen (2019: 83-84). Note that Lycian also has a patronymic with a zero ending (Melchert 2012: 275). In most examples, the genitive depends on the word tideimeli- `son' (TL 5 4, TL 61 1, TL 77 2, TL 117 3, TL 127 1 and N 315 2) or kbatra- (TL 87 5), although there is an example without these words (Hla: Nterubila `Hlas (the son) of Nterubila-' TL 145 1). TL 117, N 320 [2x]), three other texts lack the word for `son' in both versions (TL 6 [2x], TL 45 A, N 312), and in only two instances (TL 25a, TL 56) divergences occur (in the Lycian text tideimi is used after the genitive while the Greek counterpart lacks the word uloç).15 They also observed that the “the occurrence of uloç with the father's name is quite sporadic” in the monolingual inscriptions from Lycia and, in fact, the only occurrence is found in “an inscription from Limyra (H ii 34, iv/iii c. BC), whose structure fully reproduces that of the Lycian sepulchral inscriptions, with a topicalized object, followed by the verb, the subject and the filiation formula with the father's name in genitive case + the noun `son' (both personal names are Anatolian), and the indirect object (the builder himself, his wife, and their sons)” (Merlin and Pisaniello 2019: 98). The inscription reads as follows: toûto to pvppa êàòºñòêºèàñòàòî Kevbaç Àîñòà uL[oç] ºàèòô <ò>º êà³ òð³ yuvaiKL àèòîé êà³ tol[ç] tekvolç. `This monument, Kendas son of Assas built (it) for himself, his wife, and (their) children'. All in all, one can easily conclude that the use of uloç in the Greek inscriptions from Lycia (almost all confined to the bilingual documents) can be attributed to the influence of the Lycian formula. However, Merlin and Pisaniello (2019: 98) acknowledge that the possible interference “does not operate against the Greek rules or create something new; it rather expands an uncommon epigraphic use already existing in the Greek”. Note also that a similar process occurred in Roman times with the Latin influence of the formula genitive + filius (usually abbreviated as F.) on Greek (Adams 2003: 670-677). See, for example, the inscription Klazomenai 11, l. 3-4 (Ionia, 41/54 AD): Tl(Péqloç) Kàau SLOç MevàvSpou uloç `Tiberius Claudius the son of Menandros'. A second point studied by the same authors (Merlin and Pisaniello 2019: 98-99) addresses the reverse scenario: the lack of tideimi in the Lycian inscriptions as a calque of the Greek formula, as suggested by Rutherford (2002: 212). They quote the interesting bilingual inscription N 312 (Xanthos), The patronymic sequences read in the bilingual inscriptions are Greek ˺ðîêË[º³]5ã^ 0e[q]Pectioç ˳^ðÅÑ^ `Demoklides (the son) of Therbesis, from Limyra' (l. 1-2) and Lycian ntemuxlida krbbe[s]eh zemuris `Ntemuxlida (the son) of Krbbese, from Limyra' (l. 4-5). where the Greek text precedes the Lycian and seems to be primary because of the layout and the content, as a Lycian text lacking tideimi likely influenced by the Greek formula. Three more cases of the twenty examples available are found in bilingual inscriptions, where the feature can be considered a Greek influence. However, this is difficult to state with any certainty because of the small number of instances with explicit Greek data and the same use in other surrounding Anatolian languages. patronymic greek language generic name

A similar situation with regard to possible interference with Greek can be found in other Anatolian languages, although these other languages are more fragmentary than Lycian and data are scarce or, in the worst cases, not consistently interpreted. One such case is Carian. According to Adiego (e.g. 2007: 291, 383), the word mno- means `son' and occurs 11 times C.Eu 1, C.Ka 5, C.Ka 1, C.Kr 1. E.Me 10, E.Me 12, E.Me 16, E.Me 27, E.Me 39, E.Me 43b, and E.Me 47. in contexts similar to the abovementioned C.Ka 1: snis : sôisas : psusoAs mats: mnos (C.Ka 1) `this (is) the tomb of PsusoA, son of Mal'. Note, however, that this point has been analysed in other terms by Schrürr and Simon (see fn. 9). In any case, “the most typical Carian onomastic formula” consists of individual name + father's name with the genitive ending s, sometimes followed by a postclitic ki (Adiego 2007: 265-266).

As far as we know from the few identified inscriptions, the Sidetic formula is always twofold: name followed by the father's name in genitive without any word for `son' or `daughter': artmon Qanpijs `Artemon (the son) of Athenobios' S1, 1 (Side). The case of Phrygian is also interesting: it has some suffixes to create patronymic adjectives but also the genitive without any noun for `son' or `daughter': masa urgitos `Masa (the son) of Urgis' (K-01, Kerkenes Dag). Therefore, western Anatolia provides examples of languages using the genitive without any noun to express patronymics. It is true that Caria was one of the earliest Hellenized territories in Anatolia, leaving apart the Greek colonies proper. Likewise, Sidetic was surrounded by Pamphylian Greek, and Phrygian is so close to Greek that this feature could be inherited. It is likely, then, that the lack of the noun for `son' or `daughter' after the patronymic genitive was triggered by Greek and/or Phrygian in the Anatolian languages. An areal feature could be a more feasible explanation. As in the case of Lycian, however, a simple elision of the noun as an independent innovation (as occurred in Greek and Phrygian) is very likely, and such an influence cannot be considered a fact.

4. Greek -LSaç/-LSnç and variants and the alleged Lydian borrowing

As noted by Masson (1965: 222-227, also in Dardano 2011: 42), Greek alphabetic inscriptions lack any Greek patronymic in -i5âç/-i5qç;. Only syllabic Cypriot inscriptions attest four instances of this kind of patronymic: o-na-sa-to a-ra-wa-ti-ta-u Ovàoa(v)xo(ç) ÄppaxiSau `of Onasas, son of Arwatos', e-ke-ti-mo te-o-to-ki-ta-u 'Åõ£õ1ðÎ 0£î5îê³5àè `of Echetimos, son of Theodokos', sa-ta-si-wo-se to-pa-po-pa-si-le-wo-se sa-ta-si-pi-li-ta-u Exàoipoç òî ÏàôÎ ßaoiAqpoc; Åõàî¿ô³Ë³áàè `of Stasis, king of Paphos, son of Stasiphilos' and o-na-si-ti-mo-se ti-wi-so-ni-ta-se Ovaoixiqoç; Aipiowviôaç `Onasitimos, son of Diwison' (see Keurentjes 1997: 385-386). In fact, the suffix -iSâç is found for the first time in some personal names read in the Mycenean tablets. However, they are used as plain personal names, not as patronymics: see the suitable example of ko-ni-da-jo /konnidajos/ (KN AS 1516.7, quoted by Dardano 2011: 42). The Mycenean patronymic suffix is -ijos <(i)-jo> (Bartonëk 2003: 422), the relational suffix, also found forming the patronymic as -loç in the Aeolic dialect and in some Homeric instances (e.g. NqAqiq» uli Il. 2.20). From this last suffix, the variant -iwv derives through the agglutination of the individualizing suffix -wv. The patronymic use of -i5âç/-i5qç; seems to be an innovation of the literary tradition applied to significant characters which spread with Homer's influence. As Dardano summarizes (2011: 43, improving on prior statements by Keurentjes 1997), the function of the suffix was not to form patronymics in the strict sense; rather, it denotes a vague relation with an ancestor, as seems to be apparent through its use in the names of the demi (AiGaAiaSai, WvlSai, etc.), the names of the Attic ôèÀà³ (cf. AiavxiSqç), the names of relevant families (e.g., BayxinSai, Ku^eAlSai, neiaiaxpaxiSai, even AyaipeviSai for the same dynasty found in Old Persian as Haxamanisiya- `Achaemenid'). It may also denote a relation with a place, for example in EùpiuiSqç (derived from Eùpiuoç) and BpaaiSaç (from *Bpàaioç). In the light of these different functions, Dardano (2011: 43) concludes: “le forme in esame non sono pa- tronimici, sono piuttosto aggettivi relazionali derivati da un antroponimo o da un toponimo”.

A problem arises from the origin of -iSâç. After considering the difficulties to substantiate prior explanations for the etymology of this suffix, Dardano (2011: 48-58) suggests that Lydian -da- < *-ida- < *-iyo- (Gérard 2005: 89, with a common syncope) is the origin for Greek -iSâç. In fact, this suffix is used to create the ethnics in the light of sfar- `Sardis' >> *sfar-ida- > sfarda- `Sardian' (in LW 22 alternatively sfardëti- `Sardian') and, if accepted, *luda- `Lydian' (Greek ËuS6ç) < luwiya- `Luwian' (see Yakubovich 2017: 287-288, with references). Dardano (2011: 55) alleged other formations that now are considered to contain a clitic = dav: the alleged taac- `votive offering' >> taacda- `relative to the votive offering' became taac=dav `the podium' + `from'; sfen(i)- `rela-tive ?' (sfen(i)- for her, I use the current transliteration) >> sfênda- `?' became sfen=dav. The same clitic is also found in amu=dav `Ã (LW 023 and 024, see Yakubovich 2017: 278), cidaAm=dav `?' (LW 022) and TeAm=dav `?'. Only, *mAwen- `of grave' (in *mAwens(i)-) >> mAwêndav `(burial) installation ?' remains as so. However, the hypothesis cannot be upheld for several reasons.

First, the existence of Mycenean personal names in /-idas/ invalidates the theory of a Lydian borrowing because they predate the Lydian documentation by several centuries. It is true that at least mo-ri-wo-do- /moliwdo-/ `lead' (DMic. I 457-458, alphabetic Greek poAupSoç, ou, î, p) is suggested to be a borrowing from Lydian in the light of the theonym mariwda- (LW 004a). First suggested by Melchert 2008 and followed by Simon 2018: 400 § 107 and Bianconi 2020: 139 fn. 5. In fact, Myc. /moliwdo-/ shows Lydian consonant treatments, if we assume the likely etymology *morkw-iyo-, a derivative from PIE*merkw- `dark'. But the possibility exists that it is simply a cultural loanword, which does not imply the close contact required to support the theory of suffix borrowing. Note, however, that we do not know if any other language of the area was affected by the same shifts. It also remains unclear where Proto-Lydian was spoken during the Bronze Age. This is also the case of Greek Kuavoç `enamel, lapis lazuli, blue copper carbonate', found in Mycenaean as ku-wa-no- `smalt' (DMic. I 415-416), and somehow related to Hittite ku(wa)nna(n)- `copper, ornamental stone' (Simon 2018: 396 § 85). The same can be said of Kupivov `cumin', Myc. ku-mi-no /kuminon/ (DMic. I 401), very likely a borrowing from a Semitic language (Rosol 2013: 55-56). It is, however, to be noted that, leaving aside the possibility of the patronymic suffix -ISâç, there are no other traces of Anatolian structural influences on Greek (see a summary in Simon 2018: 277-378).

Second, Bronze-Age Greek speakers would need a large number of borrowed forms in -(i)da- to be able to recognize and use this suffix in a productive way in their own language (borrowing indirectly through complex loanwords containing this suffix) In fact, neither of the other two criteria established by Seifart (2015: 513) can be applied here: Greek does not have “a set of pairs of loanwords, one with and one without the affix” (e.g., profit-profitable) and, conse-quently, one cannot observe if “[w]ithin pairs of complex loanwords and corresponding simple loanwords, com-plex loanwords have a lower token frequency than the corresponding simplex loanwords”. but we lack any evidence for this (the documentation does not provide any more alleged Lydian words). Direct borrowing of the suffix is a possibility, but it implies a direct knowledge of the donor language, in this case Lydian. Such knowledge is not confirmed in the use of the suffix: Lydian -(i)das is used to form at least one ethnic (if not two, as said above) but never occurs in patronymics, which are built through the suffix -li- (as seen in example 4).

Moreover, it remains unclear why Mycenean adopted the Lydian suffix -ida- as -do- in mo- ri-wo-do- /moliwdo-/ but as -ida- in personal names such as ko-ki-da (MY Au 102.8). Finally, onomastics represents a special sphere of the language and the introduction of a suffix that is only operative in onomastic formulae would only be plausible in a context of strong contact and even with a linguistic subordination of Mycenean to Lydian. As discussed so far, it seems unlikely that Proto-Lydian exerted such an influence that Mycenaean could have borrowed any suffix. Therefore, it is time to move to another scenario. All the given approaches to Greek -iSâç ignore the recent Phrygian evidence, yet this evidence is relevant given the genetic relationship of the two languages. In fact, Phrygian is the closest language to Greek (Obrador-Cursach 2019) and the languages share very specific exclusive features such as the suffix *-eu-/*-eu- (Greek -£uç, -éwç, Epic -qoç, Phrygian -avos thematized) and a dental enlargement of the masculine i-stems in cases other than nominative and accusative. This last Phrygian feature was identified by Brixhe (2006: 40; see also Obrador-Cursach 2019: 236) through genitives: artimitos (B-05, the Greek goddess ÂQT£qiç, genitive ApréqiSoç, but Mycenean a-te-mi-to, âqt£^ltoç Alcm.54, ApTàqiToç CID 1.10.8, 12), manitos (B-07, genitive of manes) and, perhaps, urgitos (K-01). Importantly, manitos and urgitos are genitives used as patronymics: manes iyungidas manitos (if iyungida- is a patronymic, manitos is a papponymic) and masa urgitos. 23

After considering the existence of a shared dental enlargement, one might ask if the Phrygian feature can be used in discussing the origin of Greek patronymics in -iSâç. In fact, the Phrygian enlargement -it- may go back to a proto-form *-it- or *-id- (through the Phrygian de- voicing of the stops identified by Lubotsky 2004). Therefore, the presence of this feature in both close languages confirms the plausibility of an inherited treatment from a proto-Greco- Phrygian suffix nom. *-is, gen. -idos. This point is the key to ruling out a borrowing from Lydian: the older the form, the more unlikely the Lydian borrowing. Of course, this approach implies that Greek -iSâç is to be analysed as -iS-âç, where the second element is the individualizing suffix which goes back to *-efe+s, as found in some nouns such as v£aviaç `young man' or in Greek and Latin verbal governing compounds of the type paGuSivqç `deep-eddying' and agricola `farmer', respectively (Fellner and Gretenberger 2016). This analysis for Greek -iS-âç is not new and has been considered by prior scholars (Chantraine 1933: 339, 362; Schwyzer 1953: 510). What we can rule out definitively is that -iS-âç is the mere masculinization of feminine words in -iS-, since *-eh2- > -â+ç only individualizes the form, which is a suitable explanation for the creation of mere personal names, as seen in the Mycenean stage. Secondly, when used to qualify masculine personal names, it receives the analogical -s form nominatives of other stems, as is also the case of v£avia-ç.

Despite this segmentation, the origin of *-id- remains unexplained and has no parallels outside the Greco-Phrygian branch. 24 As such, a form inherited from Proto-Indo-European and only attested in these two languages is highly unlikely. As stated above, a borrowing from Lydian is difficult to substantiate because of the Lydian chronology and use. Therefore, only an internal innovation seems plausible. At this point it can be proposed that the Greek derivatives in -iç, -iSoç are in fact etymological derivatives in *-i- (see an earlier proposal in Chantraine 1933: 339, 362). If this is correct, it would explain the relatively limited presence of this suffix in Greek and the abundance of *-iS-. Note, however, that I suggested a possible continuity from Hurrian Urhi-, as in Urhi-Tesub, heteronym of Mursili III (Obrador-Cursach 2020: 142), see fn. 9. I leave aside the problem of the origin of some dental suffixes found in Luwian such as CLuwian -ahit- /-ahid/ < -eh₂+id- (according to Hajnal 1994: 137 fn. 4) or -it- /-id-/ (Starke 1990: 151-224), which goes beyond the scope of my knowledge.

Proto-Indo-European ã-stems are used to form nouns and adjective of all genders, as one can infer from some ancient Indo-European languages: see, e.g., Hittite palh-i- `broad', Vedic suc-i- `bright', Greek òðîô-³-ä `stout, large', Latin turpis `ugly' (< *mor-i-), Old Irish maith `good'. However, details on this possible category in PIE remain cryptic if not controversial, since exact formations in different branches are not found (see Balles 2009; Rau 2009: 72, 132; Grestenberger 2014: 94-95). Leaving aside the exact origin of i-stem adjectives in Indo- European languages, we can assume that those found in the prehistory of Greek are relational adjectives in *-i- `belonging, related to', with no gender distinction (as assumed for PIE and as still seen in Latin and in some Greek words), The same claim was made by Leukart 1994: 255-256, who stated “Das Suffix /-id-/ war ursprünglich genus-indifferent, vgl. ïàð³á-”. Note that ïà¿ä (Epic ïà¿ä), ïà³áîä `child' (masculine and feminine) goes back to *peh<sub>2</sub>u-. We can infer from *dh<sub>3</sub>-ti-m > totin `gift, offering' (Ligorio 2016) that the same accusative is preserved in the derivatives in -ti-, as in Greek (cf. the acc. áîî^ of áîî³ä, £ìä or ³îä). secondly individualized (*-eh2-) and masculinized (through an analogical *-s in the nominative) `he who is related to', as seen in the diversity of types of words that -³áàä derives: they can derived from a personal name, such as the gentilics derived from a mythical ancestor (ÀÞà˳áà³, 'ImviSai, Åêàððè^³áà³) and the names of families (Âàóõ³àáà³, ÊèôºË³áà¿, Ïàîþòðàò³áà³, etc.), from a place name (Åèð³ï³áïä, Âðàî³áàä) or even from nouns (like åÎïàòð³áðñ; or KOipavi6pç). In fact, the history of the patronymic in -iu>v is very similar, where after an adjective in *-yo- a patronymic was built by the addition of the individualizing -u>v. At some point in Greek pre-history after the creation of personal names in -³áàä, the suffix -³á- was identified as a feminine formation, since many words derived through *-i- (but not all) are found as feminine in historical times. Consequently, -³áàä was not the masculine of -³á-, as previously suggested.

The suffix -id- may have arisen as an analogical inflection after the Greco-Phrygian shift *-VT-s# > -Vs#. In the case of Greek, this shift can be seen in Attic veÔTqç `youth' (Doric, Aeolic veÔTâç, Latin nouitas `newness, novelty') < PIE *néuo-teh2t-s (Rix 1992: 143, §157), õàð³ä, Õàð³òîä `beauty, elegance; gratitude' < PIE *ghr-i-t- and Doric ïè>ä, ëþáîå; `foot' < PIE *pods (Attic ïîèä is problematic). For Phrygian, the shift occurs in Old Phrygian nevos `male descendant' < *nepot-s (acc. sg. nevotan, identified by Hämmig 2013) and the theonym Âàä < *bhôh2-t-/ *bhéh2-t- (acc. sg. batan, BaTav see Obrador-Cursach 2017). Importantly, there are problems and divergences in the inflection of i-stems in Greek (Chantraine 1933: 114; Beekes 1973: 241-245).

Alongisde the words which retained the inherited inflection (î³ä, î'þä î and q `sheep, ram') and variants (ï(ò)î˳ä, ïî˳îä or -ðîä / -ºøä `city'), there are words that follow two inflections, with and without the dental enlargement: ð^³ä `wrath' gen. ð^³îä or ð^³áîä, ððò³ä, q `wisdom, skill, craft' gen. ð^ò³îä and ð^ò³áîä, euvp;, î, q `reft of, bereaved of' gen. ºïó³áîä and ºïó³îä and nom.pl. £ÏÓ³á£ä and £ÏÓ³£ä, ïòºð³ä, ³áîä has the variant ïòºð³ä, £øä q `male fern, Aspidium Filix-mas', etc. There are also words with an accusative singular which shows an earlier stage: for instance, åð³ä, ³áîä `strife, quarrel, contention' (an original -i stem, EDG 459) has åðãó together with åð³áà. As far as we know, Phrygian i-stems only retain the original ending in the accusative: the personal name manes has a genitive manitos but an accusative manin.26

Therefore, it can be inferred that the inflection of some i-stem words merged the dental stems (which have a -Vs nominative), resulting in the creation of a new derivative suffix not found outside Greco-Phrygian or an analogy of the genitive from dental stems (-VTos) that prompted a reanalysis of such stems. The advantage of this hypothesis is that it would explain the presence of the dental in unexpected environments such as ïà¿ä, ïà³áîä, î, q `child; daughter or son' < *péh2u-i-s << *péh2u- `few, little' (seen in Attic ïàèä, cf. EDG 1142-1143) and even why this word can be used as both masculine and feminine. Just like -is, -is in the third Latin declination, in the early stages of Greek pre-history, the suffix *-is could have been indifferent to the masculine and feminine gender. Consequently, the formation -iS-âç can be considered the “Maskulinisierung des genusindifferenten Suffixes -l5-” (Meier 1975, § 66), during the specialization process of the suffix *-id- as forming feminine and diminutive words. A similar process can be found in Luwian. In this language the i-stems of foreign origin were adapted as neuter id-stems (Starke 1990: 210-226). This is, for example, the case of Hurrian loanwords: e.g., Hurrian abi- `sacrifical pit' is adapted in Luwian as abid- and erippi-, irimpi- `cedar' as irim- pid-, irippid-.27

If the above scenario is accepted for Greek, it is likely that *-id- can in fact be analysed as *°i-d-, since in many instances it is an innovation of inherited i-stems (e.g. *uelp- `expect' > èArc-tç, tSoç `hope, expectation', EDG 415). This new suffix -d- was also analogically added to other stems such as derivatives in *-iH- > *-z- (kAclç, kAciSoç `bar, bolt' < *kleh.2U-iH-/ EDG 711) and some non-inherited u-stems: £|ruç, -uSoç, r\ `fresh-water tortoise, esp. Emys lutaria' (also à|ruç, -uSoç, r), n^Aa^uç, uSoç, r\ `young tunny', xAa|auç `short mantle , cloak'. One can add to this list nàApvç, uSoç `king', a clear borrowing from Lydian qa\mu- `king'. The same process occurs in the inflection of foreign personal names, such as the eight Lydian names in genitive used as patronymic which were adapted in Greek as dental stems28 in the inscriptions of Ephesos IEphesos I.2 (= Ephesos 572):

AxaSoç (l. 24, 26, 54, 55), genitive of Axaç (LGPN V5a-45664-5, KPN § 119-2), Lydian ata- (LW 030, Sardis).

AxiSoç (l. 23, 43, 47, 56), genitive of Axiç (LGPN V5a-45671-6, KPN § 119-4).

BapaSoç (l. 50), genitive of Bapaç (LGPN V5a-45956 and 45962, KPN § 133-1).

IpiSoç (l. 55), genitive of Ipiç (not included in KPN nor in LGPN, only in literary sources).

KaSwSoç (l. 37, 50), genitive of KaSwç (LGPN V5a-49067-8, KPN § 500-10), Lydian katowa- (LW 010, Sardis).

KapouSoç (l. 16, 17, 18, 33, 45, 46, 48, 49, 54), genitive of Kapouç (LGPN V3b-28762-72, KPN § 542-2), Lydian karo- (e.g. LW 010, Sardis).

KovSaSoç (l. 33), genitive of KovSaç (LGPN V5a 49365, not included in KPN).

noxaSoç (l. 57), genitive of noxaç (LGPN V5a 52275-6, not included in KPN).

This short list shows how three different patterns were adapted29 in a similar way: we have names with a nominative in -ouç (Kapouç), in -is (Axiç and Ipiç) and in -as (Axaç, Bapaç, KovSaç and noxaç). KaSwç, KaSwSoç is a special case, since it shows the evolution of a name in -as with the vowel contraction occurring after the loss of /w/: Lydian katowa- > KaSoaç (KPN § 500-7) > KaSwç. The dental adaptation for names in -is is not surprising in light of observations made above. The adaptation of Kapouç can be explained by the inflection of nouç, noSoç. The names in -as follow the inflection of some Greek nouns in -àç -àSoç (e.g. ô£uyàç, ô£uyàSoç `exile'), which remain problematic. It is common to read that the suffixes -iS-, -uS- and -aS- are variants of a Pre-Greek suffix (see Keurentjes 1997: 397). However, no traces of Pre-Greek have been identified in Phrygian, the language which shares at least the dental ex- I provide these examples according to Yakubovich 2015. For a possible instance of a similar development in an inherited Luwian lexeme, cf. Yakubovich 2016: 481. In fact, it seems that at some stage of this language almost all the neuter i-stems were eliminated, some by way of transfer to the id-class through a proportional analogy of a similar kind (Yakubovich in p.c.). See Kearns 1994, who defended the presence of the dental adaptation as a “transformation of the Lydian possessive adjectives in -lis”. This statement is unnecessary, since they are names in genitive, not adjectives. The inscription contains a death sentence passed against Sardian men for sacrilege against Artemis Ephesos. Note that the adaptations of these names do not always follow the same pattern. In Pisidia we find the da-tive Axa (MAMA 4.245) of Axaç without any trace of a dental inflection. The same is seen with Baßa, the genitive of Bapaç found in Bithynia and Phrygia. tension for the ã-stems. If we add to this the possible internal development of the suffix as argued here, an alternative origin for -aS- must be found. To develop the hypothesis of analogy suggested for -iS-, it would be necessary to find a kind of nominative which could be the source of such a stem. Here, one could suggest some as-stem neuters. In fact, they have multiple origins: they can be “primary s-stems derived from a root in final *-h2” like yqpaç `old age' or yépaç `prize of honor' < *gerh.2-s-, from “secondary s-stem derivatives from *h2-stems” (Nikolaev 2010a: 191) like Sépaç `bodily frame' (see Nikolaev 2010b) and from a heteroclitic *r/n-stem.

The latter is the case of only two Greek words, “xépaç `marvel' and neipaç/népaç `limit', which may continue *kwer-r and *per-m” (Nikolaev 2010a: 190). Note that the first has two inflections, Epic xépaç, xépaoç vs. the dental extension in the most common xépaç, xépaxoç, and the second also has variants, Epic and Lyric nripap and nâpaç, but it is commonly inflected as a f-stem: népaç, népaxoç. However, a complex process of stem reconversion and a gender shift is unlikely as an explanation for the origin of the words containing the suffix -aS- because it requires the assumption of two processes that cannot be proved.

The process can be explained in other terms, including a prior neuter stage. In fact, the words with -aS- may derive from ancient collective neuters in *-h2 with the ending for animate nouns in *-s, a kind of “singulative” found, e.g., in Spüç, Spuoç `tree, oak' (feminine) < *druh2-s `a single tree' << *dru-h2 `wood' << *do/eru `tree' (Janda 1997: 141-145). A similar process has been considered for the masculine Greek noun Aâaç, Aâoç by Nikolaev (2010a: 192-193): *leh2- e/os- `stone' >> *leh2-es-h2 `mass of stones' >> Proto-Greek laha-s `stone'. Thus, we can hypothesize that a singulative like v^àç àSoç, q `snowflake; snowstorm' has the following prehistory: PIE *sneigwh- `to snow' >> *nigwh-h2 `the snow' >> *nigwh-h2-s `snowflake'. The same applies to ôèóàç, àSoç, î, q `one who flees, fugitive, exile': PIE *bheug- `to flee' >> *bhugh `exile' >> *bhug- h2-s `a single exile'. The parallel process can be hypothesized for Ai0àç, àSoç, q `stone', Aaprcàç, àSoç, q `torch' and words with the same suffix. After the creation of *s-stem nonneuter nouns and the shift *-VT-s# > -Vs# with the rise of the new suffix -S-, the nominative resulting from *-h2-s, with a strange short-vowel nominative and *s-stem inflection was reanalysed as *-a-d-s > *-a-ç and a new stem *-a-d- considered to be a suffix. Following this process, the new suffix -a-S- was also used as an allomorph of -iS- after -i-, as in ''IAiov/'IAioç >> 'IAiàç, àSoç, q `the Troad; a Trojan woman; The Iliad (the poem)'. It also occurs in the patronymics built after this suffix such as Aé^ioç >> A£^iàSqç (Il. 7.15).

To sum up, the main idea is that the nominatives of some ancient stems (mainly the relational adjective *-is and the singulative of an ancient collective *-h2-s > -as) were re-analysed after the shift *-VT-s# > -Vs# and, as a result, a new suffix *-V(S)- was created in Greek (and Phrygian). On the basis of the resulting suffix *-id-, singulatives in *-id-eh2-s were created and became used as patronymics in Homer (and in Cyprus, according to the syllabic inscriptions).

5. Variation in Phrygian patronymics

Similarly to Lycian and Greek, Old Phrygian shows different ways of expressing patronymics. To the best of our knowledge, Phrygian has patronymics built with a plain genitive without any noun for `son' or `daughter'. There are two different structures of this formula. The first is the personal name followed by the father's name in genitive: The identification of -oy, written both <oi> and <oy>, as the thematic genitive is a recent proposal by Oreshko (fthc.), who also adds the Middle Phrygian inscription W-11 viKooTQaxoç KA£upaxoi `Nikostratos (the

W-08: ates agomoi `Ates (the son) of Agomos' (according to Oreshko fthc.).

B-05: atriyas davoi `Atriya (the son) of Davos' (according to Oreshko fthc.).

K-01: masa urgitos `Masa (the son) of Urgis'

HP-102: midas aiasay `Midas (the son) of Aiasa' ?

Conversely, there are several examples of a possible genitive preceding the name, which is usually the case of normal genitives in Indo-European languages but does not occur with patronymics:

M-06: davoi iman `Iman (the son) of Davos' (according to Oreshko fthc.).

G-136: tadoy . iman `Iman (the son) of Tados' ?? dative?

...