Lydian Language and Inscriptions

H. Craig Melchert


Lydian is attested in just over a hundred inscriptions, but less than thirty of these consist of more than a few words and are reasonably complete. Some coin legends, graffiti, and one or two inscriptions on objects date from the late seventh or early sixth century, but the vast majority of extant texts date from the fifth and fourth centuries. Most inscriptions were found in the Lydian capital, Sardis, or its vicinity, but there have been scattered finds elsewhere: Çobanisa (Hermos valley), several sites in the Kaistros valley (Falaka, Tire, Haliller), Manisa (Magnesia), Pergamon, Ephesus, Emre (Maeonia), Eğriköy (Aeolis), Aphrodisias (Caria), Smyrna, and even a graffito in Silsilis in Upper Egypt.

The Lydian script is derived from a variant of the Greek alphabet, and nearly all of its 26 signs (8 vowels, 18 consonants) have direct counterparts in Greek.1 Most letters either indicate the same sounds as the respective letters in Greek or have values that are easily derivable from those seen in Greek. Like Lycian, Lydian also uses a few Greek letters for sounds it does not have for others not existing in Greek. For example, it uses Greek consonant letters for its nasalized vowels ã and . One slightly confusing practice: ś represents ordinary /s/, while s stands for a palatal(ized) sibilant (Turkish ş).2 Lydian is regularly written right to left, but in the archaic period one finds also left to right and one example of boustrophedon (alternate lines left to right and right to left). Word divisions are usually marked by a space. Again in the Archaic period one also finds examples of continuous writing without spaces and occasional use of word dividers.3

The inscriptions on stone are mostly funerary, but there are also several decrees. For example, the text numbered 24 in Gusmani represents a mutual declaration by a priest named Mitradastas and the “chief temple authority” (serlis srmlis) in which each apparently bequeaths to the other whatever property he has (perhaps because it was felt to belong ultimately to the temple).4 Text 22 in Gusmani similarly contains mutual arrangements of some kind between the Sardians and a group of people designated by the word mλimna-, but we do not yet understand any of the details. Inscriptions on seals and other objects either indicate the owner (manel-im “I belong to Manes” on a seal,5 or are dedicatory: titiś-in ẽmν tiśardν fabil ataλ kitwaλ “Titis gave my contents(?) to Ata, (son) of Kitwa” (on a terracotta vase found in a Sardis tomb).6 Several of the inscriptions on stone from Sardis are, remarkably, apparently in verse, showing vowel assonance in the last syllable of each line (either a or o) and undoubtedly some kind of meter, although the nature of the latter remains much debated.

The basics of Lydian grammar are well understood, but knowledge of the lexicon remains very limited and tentative; hence, interpretation of the texts does as well. We have only two short Lydian-Greek bilingual texts and two Lydian-Aramaic ones, but in one of the latter nearly all the Aramaic text is missing, and in the other the Aramaic text raises problems of its own. We still await a Lydian-Greek bilingual of significant length. Pending its discovery, all analyses of Lydian must be regarded as provisional.7

Linguistic Origins and Features

Lydian is a member of what is now called the Anatolian sub-branch of the large Indo-European language family. It is thus most closely related to Hittite, Palaic, and Cuneiform Luvian of the second millennium BCE (attested in cuneiform documents from the Hittite capital Hattusha in central Anatolia), Hieroglyphic Luvian of the second and early-first millennia (from inscriptions found in Anatolia and northern Syria), and Lycian, Carian, Pisidian, and Sidetic of the first millennium (attested in inscriptions from the southwestern and southern coasts of Asia Minor). In spite of important cultural interaction between Lydians and Phrygians, the Lydian language has no special connections with Phrygian, which represents an Indo-European language entirely separate from the Anatolian family as defined above.

Against occasional claims to the contrary, most scholars maintain the long-held view that the Anatolian Indo-European languages are intrusive to Asia Minor. The prehistoric speakers entered from the north no later than the middle of, and possibly as early as the beginning of, the third millennium BCE. It remains indeterminate whether they followed an eastern route through the Caucasus or a western one across the straits from the Balkans. There are some arguments in favor of the western entry, but they cannot be viewed as decisive.

There is a widespread view that in the second millennium pre-Lydian speakers lived in the northwest, in the region of classical Bithynia, and that they entered the area of classical Lydia only later, mingling there with a Luvian-speaking population, possibly as a ruling elite. Several scholars have independently argued that the name Lydoi is in fact the Lydian outcome of the attested place name Luwiya (unattested Lydian *lūda-). By this scenario, the name of the Lydians would not be their own self-designation (which is unknown), but would reflect their name for their new homeland. However, even if the explanation of the name is valid, it does not prove a relatively late southward movement. Evidence for classical Lydia (centering on the valley of the Hermos) as Luvian-speaking is sparse.8

Linguistic features do support the idea that Lydian was relatively isolated from the other Anatolian languages, but only to a degree that is also compatible with its location in classical Lydia already in the second millennium (with, if necessary, assumption of some later northward movement to, for example, Daskyleion). Lydian does share some innovations with the other western Anatolian languages (see below), but was distinct from the southwestern dialect group of Luvian-Lycian-Carian that lay to its south. There are at present too many unknowns regarding peoples and languages of western Anatolia in the second millennium to decide on the location of the pre-Lydian speech community.

Lydian shares some crucial defining characteristic innovations with the other Anatolian languages. The first person pronoun ‘I, me’ is amu, with u-vocalism in the second syllable that is unique to the Anatolian sub-branch: Hittite ammug, Hieroglyphic Luvian /amu/, Lycian amu/ẽmu. Lydian also shows various Anatolian changes involving the stop consonants. Word-initial voiced stops are devoiced: cf. Lydian teśaśti- vs. Latin dexter ‘right(-hand).’ Word-internal voiceless stops are under certain conditions voiced: cf., Lydian kaτared ‘protects’ with present third person singular verbal ending -d vs. Sanskrit ending -ti and Latin -t (in Lydian and Latin the original final vowel was eventually lost). Lydian also shows rather complex “chains” of sentence-initial conjunctions, particles, and pronouns, like the other Anatolian languages: fak=τ=ad (conjunction fak plus local particle -t- plus reflexive pronoun -ś- [spelled together as τ] plus pronoun ‘it’ in Manisa 1, No. 10). The precise nuances of the various particles are poorly understood.

As noted above, Lydian also shares particular features with the other western Anatolian languages: Luvian, Lycian, and Carian. The present first person singular verbal ending is -u *-wi (e.g., kantoru ‘I grant’). This matches the endings Luvian -wi and Lycian -u. Lydian also shares with those languages the peculiar feature known as “i-mutation” by which only the nominative and accusative of some nouns and adjectives add an -i- to the stem before the endings: thus, nominative singular śfardẽt-i-s, but dative plural śfardẽt-aν ‘Sardian’. Lydian, again like Luvian and Lycian, tends to mark possession not by the genitive case of a noun but by a derived adjective that then agrees with the noun in number, gender, and case. Lydian chooses for this purpose an -l- suffix: es anlola atraśtal śakardal “This monument (or similar) (is) of Atrastas, (son) of Sakardas” in the beginning of Manisa 1 (No. 10; fig. 1). Finally, Lydian reflects a syntactic feature seen also in Luvian by which a “nominal sentence” in the first or second person must contain a reflexive pronoun: the inscription on a seal manelim means ‘I (am) of Manes’ (i.e. ‘I belong to Manes’), reflecting *manel-mi ‘of Mane’ (adjective) plus first person reflexive. There is no expressed verb ‘am.’

Lydian also has some unique features that distinguish it from its closest relatives, at least superficially. Alone among the Anatolian languages, it preserves the adjective aλa- ‘other’ (cf., Latin alius, Greek állos, etc.). It has a word for ‘priest’ kawe- whose closest match is found in distant Avestan kauuā- (princely title) and Sanskrit kavi- ‘seer, poet.’Prehistoric yod becomes a kind of voiced continuant spelled «d»: thus Lydian pid- matches Luvian piya- ‘give’. Under conditions not yet understood, the prehistoric stop *p becomes f: the conjunction fa(k) cited above matches Cuneiform Luvian . Perhaps most strikingly, Lydian undergoes massive deletion of unaccented vowels, both internally (“syncope”) and word-finally (“apocope”). Several examples have already been seen above. In some cases this produces consonant sequences that are then broken up by newly inserted vowels: thus as per above, *manel-mi » *manelm » manelim. However, Lydian tolerates many consonant sequences that its relatives do not, some quite remarkable: note kśpλtaλ or dctdid (meanings unknown).

There is also limited linguistic evidence for cultural contact. A newly found inscription from Haliller (No. 11; fig. 2) contains the title śatrapaś ‘satrap’, obviously a loanword from Persian. The Greek poet Hipponax of Ephesus clearly knew some Lydian, and it has been tentatively suggested9 that his σκαπερδεῦσαι ‘to steal’ reflects indirectly Lydian kaprdokid in No. 10, whose context fak=τ=ad kaprdokid also points to a meaning ‘and steals/plunders it for himself’ and whose form resembles the Hittite word for ‘rat’ (compare ‘packrat’ and German mausen ‘steal’ from Maus ‘mouse’).

  • Fig. 1

    Stele of Atrastas, from Sardis, detail of inscription (©Archaeological Exploration of Sardis/President and Fellows of Harvard College)

  • Fig. 2

    Detail of stele from Haliller (©Archaeological Exploration of Sardis/President and Fellows of Harvard College)