The historical science was aware of the existence of Lydians and Lydian language because of numerous mentioning about them in ancient sources. Herodotus describe some features of Lydian, mainly personal and place names. Strabo and Xenophon also depicted Lydia and its inhabitants, naming their language. But only archaeological research helped to find written signs of the language. The identification of inscriptions showed that the Indo-European and Anatolian origins of the language was evident. Nowadays Lydian language studies are maintained in Europe, the USA and widely in Russia, where the famous linguists Ivanov, Shevoroshkin and Bayun have written a significant number of important works.
 

 § 1. The History of the Lydian Language.

It is known that Lydian and Carian (and maybe Misian) derived directly from the Hittite language. Lydians were those citizens of the central parts of the Hittite Empire, who decided to move westwards saving from invasions of Assyrian, Aramaean, Syrian armies and tribes of Sea Peoples, which appeared frequently since the 13th century BC. Palaic speakers are believed to have been assimilated by substratum and superstratum languages, western Luwians went to southern and south-western regions of Anatolia and there were called as Lycians and Sydetics. And Hittites who used to live in central and eastern Anatolia, either had to migrate west or to mix with conquerors and aboriginal tribes.

In the 7th century BC no Greek author mentions Hittites, and Lydia is widely known in the ancient world. The kingdom ruled by the dynasty of the Mermnadae was developing successfully and was broadening its borders steadily. Lydian kings gained a decisive victory over Phrygians and their kingdom, conquered Misia and Troy, repelled the intrusion of Iranian speakers nomadic Cymmerians and constantly pressed the Ionian Greek city-states on the shores of the Aegean Sea. The Lydian language was flourishing at that time, spoken also in Misia, Phrygia and Caria; Misian was obviously assimilated very early.

Greek cultural influence was very significant since the 6th century, and many Lydian inscriptions were written in both Greek and Lydian languages (bilingual texts). Greeks created their own transliteration of Lydian, which is of great help to modern linguists in defining the true Lydian pronunciation. But this Greek influence could not cause great changes in the language until in the 4th century BC Lydia, at that time already a province of the Persian Empire, was conquered by Alexander the Great. Since then, many Greek colonists started arriving and settling in Lydia, its capital Sardis and its shores, and since then, Lydia never gained independence again, being always under the Greek rule. Lydian language was no longer official, and in three or four centuries it became really extinct. The last glosses written in Lydian date back to the 1st century BC.
 
 
 § 2. The Lydian Phonetics.
 
Lydian inherited the phonetic system from its ancestor Hittite, but developed many peculiar traits which were unknown in Hittite, Luwian or Palaic. One of the main new features was the system of Lydian vowels. Hittite had just 5 of them: short a, i, u and long á and í. Lydian not only developed the sound o, but also acquired two nasal vowels, which will be written here as â and ê. In all, the Lydian vowel set was the following:

a < Common Anatolian a, and also o before labials. In Hittite it was also a.
    e.g. da (to give), Hittite da, Luwian da (to take)
â < Hittite long á, also from any vowel + n.
    e.g. amâs' (loving), Anatolian *amans
e < Hittite e, i
ê < Common Anatolian long e, also from e + n
    e.g. qên (to kill), Hittite kwen-
i < Common Hitto-Luwian nasal é, also from e, i
    e.g. kis'- (to comb), Hittite and Luwian kiš-, Common Hitto-Luwian kés
o < Hittite au, uwa. Hittite lacked this sound, but it was revived here in Lydian
    e.g. s'oy- (to fill), Hittite suwai-
u < Hittite u, uwa
    e.g. kud (where), Hittite kuwad
y < Hittite long i

Nasal vowels was not only the Lydian invention. Sydetic and especially Lycian also used them, and there were even four nasal vowels in Lycian. It was caused mainly by the reduction of the nasal m, n sound after vowels, and nasalizations of long vowels. We cannot exclude the Lycian influence as the possible sources of vowel nasalizations in Lydian. Sometimes the nasal ê also replaced the old long a.

The four syllable sounds m, n, r, l were not preserved in Lydian, as well as in Hittite and other Anatolian languages. Lydian turned them into the consonant + vowel combination: Hitto-Luwian tr > Lydian tor, tro (to speak).

Phonetics process among Lydian vowels were not too numerous, but maybe we just do not know much about them. The only vowel interchanges known in the language are the following:

a - ê (amu - I, me; êmis - my)
y - i (qy / qi - which, that)

The consonants of the Lydian language were of a greater number than in Hittite or Luwian. The most important process that took place during the transition from Hittite to Lydian was the palatalization of many consonants, mostly dental ones, fricatives, nasals and liquids. Some linguists believe even all consonants had their palatal equivalents in Lydian, but the orthography followed only a few of them. Below is the table of Lydian consonant system:
 

Consonant interchanges in the word were not frequent in Lydian, and the only ones that were noticed are the following: b - v - f (e.g. borli / forli - a year), and d - t at the end of the word (e.g. kod, kot - how?).

As a whole, Lydian phonetics shows many interesting internal processes that were unknown in most other Indo-European groups of languages. We should not forget about the mixture of Anatolian elements with Syrian, Aramaean and Hurrian languages, and about the influence of the Asian substratum of aboriginal languages. Further we will see how this non-Indo-European element influenced the morphology of the Lydian language.
 
 
 § 3. The Lydian Substantives.
 
All Indo-European languages inherited two types of nominal declension from the Proto-Indo-European language: that of nouns and pronouns. The declension of nouns is also used in adjectives, possessive pronouns and in verbal nouns (participles, gerundives, verbal adjectives, semiparticiples). The pronominal declension is mainly used with demonstrative, personal, reflexive and interrogative pronouns, it has some peculiar features which make it dissimilar somehow from the noun declension.

There can be a situation, when those two systems of declension borrow some endings from each other, mixing pronoun inflections in nouns and vice versa. That happens in many Indo-European languages: some pronoun endings are traced in Baltic, Slavic, Italic and Greek languages, few exist in Germanic and Celtic.

Anatolian languages are of the same sort, and Lydian nouns have quite an extensive use of pronominal endings. Many Indo-European noun inflections died in the language replaced by Hittite pronominal ones, and this misled scientists for a long time. In fact, the problem of shortage of materials for Lydian studies does not allow us to learn the nominal system in detail, but the common processes are known by us and described below here.

And, finally, there were 5 noun cases. In Proto-Indo-European they used to be eight, and in Common Hitto-Luwian, a unique case, even nine, like nowhere in the language family. The ninth case was a directional one with an -a ending, used only in singular. Lydian demonstrates the significant reduction of noun cases, preserving only 5 cases: nominative, accusative, genitive, ablative, and dative-locative.

The usage of these cases was similar to any other Indo-European language with only one important exception. The dative-locative case had the second usage as an object case, using with nouns at which the action was directed. Here is the table of Lydian noun declension for all the noun stems:
 

No information was found about the Lydian adjectives, their special traits. The only thing we can say is that they were declined quite like nouns. There is no sign about pronominal adjectives or relative adjectives. Attributes were placed either before the subject, as in es-l mru-l (this stele) in Lydian, or after it, as in ora-l is-l bakill-l (in month this Bacchus's). So there was no fixed word order for attributes, as in Celtic, Italic or Germanic languages. One of the Lydian adjectives: alas' (other).

Lydian pronominal system has been studied much more thoroughly, and materials are not too scarce, fortunately. At least we can follow personal, possessive, demonstrative, relative, indefinite, interrogative and enclytical pronouns. This set of pronouns is a usual Indo-European one, and the system of pronouns was a decisive factor which made every linguist be sure that Lydian is an Indo-European tongue. Different theories were discussed since in the first half of this century Lydian was discovered. There is one version of their origin, which states Lydian people were descendants of ancient Asia Minor population, related to Hatti and having no connection with Hittites who used to live far in the east of the peninsula. This can be confirmed by the fact that already in the 2nd millennium BC, when no one knew Lydians and the Hittite Empire was prospering, Egyptian sources mention some people from Asia Minor which were called "shardanna". This can be a Egyptian-pronounced name of the city which is known as Sardis, the capital of Lydia. But, to my, mind, no name of the city proves the language of the people. Hittites called themselves Nesians, but came to the land named Hatti and acquired a new name, and even Hattusa, the ancient pre-Indo-European name of their capital, was not changed. Sardis can easily be a substratum name, but the language of Lydians who lived in Sardis in the 8th century (Homeros's Maeonians) was certainly Indo-European.

Now back to Lydian pronouns. We will follow their evolution from Hittite, Common Anatolian and Proto-Indo-European languages here step by step.

Personal pronouns had two persons in Proto-Indo-European, the third person ("he, she, it") did not exist, and certain demonstrative pronouns functioned instead of them. Anatolian as a very archaic tongue preserved this structure, though sometimes in Hittite and even stronger in Lydian and Lycian demonstratives were used in the meaning of personal third person pronouns. The only known Lydian personal pronoun is amu (I, me), which - surprisingly - was not declined. This resulted from Hittite, which lost personal pronoun declension due to non-Indo_European influence. Anyway, in Luwian they were declined, and the form known is ammedaz - ablative case of "I, me" ("from me"). But Lydian lost the declension, as well as the very inflections with amu. We can reconstruct Lydian 2nd person pronoun somehow like *tik, *tuk (note that in Hittite this was declined, unlike the previous one, and had a dative form), but this is just a version.

The demonstrative pronoun which was used to denote third person is bis' ("this, he") originated from Proto-Indo-European *ebho- via Hittite apa- ("that").

Possessive pronouns managed to keep their Indo-European structure. They are declined just like the nouns: remember, that noun declension had many pronominal signs. Here is everything found about them including the reflexive possessive pronoun lost in English but common to all ancient Indo-European tongues (Latin suus, Lithuanian savo, Russian svoj):
 

The demonstrative pronouns played an important part in Hittite, Luwian and Palaic. Five different stems for them are found in Hittite, and some additional ones are formed also with suffixes. All of them are Indo-European and can be traced in Latin, Ancient Greek and other languages. Lydian preserved only two of these pronouns, from Proto-Indo-European stems *s- and *to-. Remember Greek 'o, 'h, to < *so, *sa, *tod, or Slavic *siji, *ta, *to. Lydian demonstratives looked as follows:
 

We see that this pronoun is declined just like demonstrative ones, so they represent the same type of declension which is called "Hittite pronominal". The sound -i- which is included in the stem, is reduced before a palatal sound; this palatal consonant reflect the combination -in, -il, and supposedly the initial consonant q- is also softened a little.

Among other relative pronouns we can name nâqis - "whoever", and as for interrogative, kot - "how?, in what way?", and kud - "why?"; the last two lost labial sounds which were present in their Hittite ancestors kuwat and kuwad. This is common according to phonetic laws of the Lydian language described above (see).

The indefinite pronouns are just derivative from interrogatives: qi-, qesi-, qel-, niqesi- all mean the same "someone, something", the difference in their usage is unknown. The only one that is interesting is fisfid, a neuter pronoun which also means "something" but derived not from the stem *kwi- but directly from Hittite pippid, maybe an example of some dialectal process of turning the Proto-Indo-European *kw into p. Anyway, no other signs of such revertion is found.

And finally, the enclitic pronoun inserted into the word. We should mention that Lydian words consisted not only of the stem.  its suffixes and inflections, but also of a great number of particles and conjunctions right in the middle of the word. The pronoun could also be placed among them, and so did the pronouns -i-, -a- meaning "this, his" and being morphologically reduced demonstrative pronouns. The second of them, -a-, originated from the same Hittite form, and from Common Anatolian *-e-.

Certainly, there were other pronouns in the Lydian language, and the set shown above is not the full picture of the pronominal system. But we must be happy with what has been found by now, the present material is enough to judge about the language and its origin, and I am sure in the future the Lydian language discoveries will grant us more information.
 
 
 § 4. The Lydian Verb.
 
Lydian materials do not leave us enough material to be sure that the verb system of the language was exactly the way we now reconstruct it. Hittite and Luwian languages are known much better, and their verb structures look quite complete in modern Indo-European scientific research. As for late Anatolian languages, we can't sometimes point even the basic verb forms. Anyway, sorry for the lack of information.

The Lydian verb categories were similar to those of Hittite. Only two tenses are found: the past (preterite) and the present. No perfect or any other complex past tense was known either in Hittite or any other Anatolian language. It looks as if the Anatolian past denoted both perfective and imperfective meanings, and that's no analytic forms for other past tenses existed either. The present is believe to act also for the future action, otherwise it is unclear how to express future in Lydian.

Hittite and Lydian had 3 persons, as usual in Indo-European tongues, and the endings for those persons are quite Indo-European, too, as we will see below. Verbs had 2 moods in Lydian - indicative ("I go") and imperative ("go!"), and so do Hittite and Luwian. The main difference lies in the category of voice. Hittite, Luwian, Palaic had two voices - active and mediopassive, repeating the Proto-Indo-European voice division. But Lydian inscriptions do not show any medium or passive forms, and they are not found also in Lycian, Sydetic and Carian. This can hardly be just a coincidence: perhaps, all late Anatolian languages reduced mediopassive forms and replaced them somehow with this or that active form. But the shortage in information can be the reason for our ignorance in this question.

Now let us come to the verb conjugation. We must say that the conjugation was simplified greatly in Lydian, and there was no longer such a strict observance of verb endings as in Hittite, inflections could be interchanged in different numbers, persons and even tenses (e.g. an ending of preterite could be used also in the present). These are the basic endings:

    The Present tense indicative mood.
1st person sg.     -u < Hitto-Luwian -wi, -wa (used only after consonants)
                          -f / -v < Hitto-Luwian -wi, -wa (used only after vowels, represents just the reduced form of the -u ending)
                          -n (palatal n) < the preterite ending
Examples: kon (I find), kantoru (I trust), fakorfif (I do harm)

2rd person sg.    -t < Hittite -ti / -si (later it coincided with the 3rd person sg. in Lydian)
Example: kot (thou find)

3rd person sg.    -t / -d < Hittite -zi, Luwian -ti (the two forms vary due to the previous vowel or some other reasons)
                          -nt < Hitto-Luwian -nti (from the 3rd person plural)
Examples: kot (finds), silavad (takes care), int (does, they do)

1st person pl.    -vn < Hittite -weni (here -ni turned to the palatal sound)
Example: karttivn (we abolish)

3rd person pl.   -nt < Hittite -nti
Example: inânt

    The Past tense indicative mood.
1st person sg.   -n < Hittite -un / -nun, -hun (the reason for palatalization is unknown)
                        -idn / -idn (the Lydian neologism)
Examples: dân (I gave), vidn (I built)

3rd person       -l < Hittite participle form (used both for singular and plural in Lydian)
Example: kardal  (he / they made)

    The Imperative mood.
3rd person       -u, -v, -f < Hittite -u (both singular and plural)
Example: s'of

So we can see that though a lot of verb forms were clearly derived from their Hittite ancestors, the system of the Lydian verb conjugation changed greatly. The phonetic processes in the language turned many inflections into new forms, dissimilar to Hittite, some new ones were developed. The preterite 3rd person used a participle ending (compare with Common Slavic participle in -l, which was also turned into the past tense ending: Lydian katanil = Slavic postavil "he placed, constructed"). This is the common process in Indo-European languages, but in Lydian it meant the analytization of the language - the number of endings was reducing more and more.

Verbal nouns included two main categories: the participle and the infinitive. Hittite used more of them: there we about three or four kinds of participles, the infinitive, the supine and also some verbal nouns with unclear functions. Lydian preserved only active forms of the participles, and two different infinitives which could be interchanged in all situations.

The participle examples: amâs' (loving), êtos'rs' (encircling), lalens' (speaking). All of them represent the Proto-Indo-European participle in -ns which existed practically in all Indo-European branches: Latin amans (loving), Lithuanian mylinti (loving), Common Slavic ljubint (loving), etc. Phonetic mutations caused the original ending to turn into -rs' or -âs'.

The infinitive examples: sanvas'tal (to possess), un (to order, to prescribe). The ending of the last one comes from Hittite -na, -una. In fact, all Indo-European languages form infinitives from different verbal nouns, their case forms. There was no infinitive in Proto-Indo-European, that is why forms vary so much in different groups.