Urartian language

Urartian or Vannic is an extinct Hurro-Urartian language which was spoken by the inhabitants of the ancient kingdom of Urartu, which was centered on the region around Lake Van and had its capital, Tushpa, near the site of the modern town of Van in the Armenian highlands, now in the Eastern Anatolia region of Turkey. Its past prevalence is unknown. While some believe it was probably dominant around Lake Van and in the areas along the upper Zab valley, others believe it was spoken by a relatively small population who comprised a ruling class.
First attested in the 9th century BCE, Urartian ceased to be written after the fall of the Urartian state in 585 BCE and presumably became extinct due to the fall of Urartu. It must have had long contact with, and been gradually totally replaced by, an early form of Armenian.

Classification

Urartian is an ergative, agglutinative language, which belongs to the Hurro-Urartian family, whose only other known member is Hurrian. It survives in many cuneiform inscriptions found in the territory of the Kingdom of Urartu. There have been claims of a separate autochthonous script of "Urartian hieroglyphs" but they remain unsubstantiated.
Urartian is closely related to Hurrian, a somewhat better documented language attested for an earlier, non-overlapping period, approximately from 2000 BCE to 1200 BCE, written by native speakers until about 1350 BCE. The two languages must have developed quite independently from approximately 2000 BCE onwards. Although Urartian is not a direct continuation of any of the attested dialects of Hurrian, many of its features are best explained as innovative developments with respect to Hurrian as it is known from the preceding millennium. The closeness holds especially true of the so-called Old Hurrian dialect, known above all from Hurro-Hittite bilingual texts.
The external connections of the Hurro-Urartian languages are disputed. There exist various proposals for a genetic relationship to other language families, e.g. Northeast Caucasian languages, Indo-European languages, or Kartvelian languages, but none of these are generally accepted.
Indo-European, namely Armenian and Anatolian, as well as Iranian and possibly Paleo-Balkan, etymologies have been proposed for many Urartian personal and topographic names, such as the names of kings Arame and Argishti, regions such as Diauehi and Uelikulqi, cities such as Arzashkun, geographical features like the Arșania River, as well as some Urartian vocabulary and grammar. Surviving texts of the language are written in a variant of the cuneiform script called Neo-Assyrian. Comparison:

Urartian	Hurrian	Meaning
esi	eše	Location
šuri	šauri	Weapon
mane	mane	3. Sg. Pers.
-ḫi	-ḫi	Affiliation affix
-še	-š	Ergative
-di	-tta	1. Sg. Abs.
ag-	ag-	to lead
ar-	ar-	to give
man-	mann-	to be
nun-	un-	to come
-di	-da	Directive
-u-	-o-	Transitivity marking
qiura	eše	Earth
lutu	ašte	Transitivity marking

Decipherment

The German scholar Friedrich Eduard Schulz, who discovered the Urartian inscriptions of the Lake Van region in 1826, made copies of several cuneiform inscriptions at Tushpa, but made no attempt at decipherment. Schulz's drawings, published posthumously in 1840 in the Journal Asiatique, were crucial in forwarding the decipherment of Mesopotamian cuneiform by Edward Hincks.
After the decipherment of Assyrian cuneiform in the 1850s, Schulz's drawings became the basis of the decipherment of the Urartian language. It soon became clear that it was unrelated to any known language, and attempts at decipherment based on known languages of the region failed. The script was deciphered in 1882 by A. H. Sayce. The oldest of these inscriptions is from the time of Sarduri I of Urartu. Decipherment only made progress after World War I, with the discovery of Urartian-Assyrian bilingual inscriptions at Kelišin and Topzawä. In 1963, a grammar of Urartian was published by G. A. Melikishvili in Russian, appearing in German translation in 1971. In the 1970s, the genetic relation with Hurrian was established by I. M. Diakonoff.
In 1962, after completing his military service, kurdish man Mehmet Kuşman, started to work as a guard in the Çavuştepe castle, built by the Urartians between 764–735 BC. He became fascinated with the mysterious language when an excavation team found an inscription at the site, but did not know how to decipher it. A quote reads: "I used to think that professors know everything in this world, but it wasn't like that." Despite being a primary school graduate, he started to learn the Urartian script from the archeologists and linguists who came to the region for excavation work and the books they gave, later traveling to study Urartian-Assyrian bilingual inscriptions. He reached a point where he needed an alphabet, traveling to Iran, Syria, and Armenia illegally without a passport for field study. In Armenia he got a tip of a village near the Georgia border where a man lived who could help him. After 3 years he learnt the Urartian script, becoming one of a handful of people who could could read and write it fluently. Despite various lucrative offers, Kuşman did not pursue an academic career abroad, instead teaching on location of his old job.

Corpus

The oldest recorded texts originate from the reign of Sarduri I, from the late 9th century BCE. Texts were produced until the fall of the realm of Urartu, approximately 200 years later.
Approximately two hundred inscriptions written in the Urartian language, which adopted and modified the cuneiform script, have been discovered to date.

Writing

Cuneiform

Urartian cuneiform is a standardized simplification of Neo-Assyrian cuneiform.
Unlike in Assyrian, each sign only expresses a single sound value.
The sign gi has the special function of expressing a hiatus, e.g. u-gi-iš-ti for Uīšdi. A variant script with non-overlapping wedges was in use for rock inscriptions.

Hieroglyphs

Urartian was also rarely written in the "Anatolian hieroglyphs" used for the Luwian language. Evidence for this is restricted to Altıntepe.
There are suggestions that besides the Luwian hieroglyphic inscriptions, Urartu also had a native hieroglyphic script.
The inscription corpus is too sparse to substantiate the hypothesis. It remains unclear whether the symbols in question form a coherent writing system, or represent just a multiplicity of uncoordinated expressions of proto-writing or ad-hoc drawings.
What can be identified with a certain confidence are two symbols or "hieroglyphs" found on vessels, representing certain units of measurement: 100x33px for aqarqi and 100x33px for ṭerusi. This is known because some vessels were labelled both in cuneiform and with these symbols.

Phonology

Hachikian gives the following consonants for Urartian inferred both from Urartian writing as well as loans into neighboring languages, mainly Armenian:
The three-way laryngeal contrast for stops and affricates was faithfully represented in Urartian writing, except for the "emphatic' /pʼ/ which the Semitic-based cuneiform writing system did not have a distinct symbol for. Their values are confirmed by loans in Armenian. Urartian voiceless stops and affricates were loaned as voiceless aspirates in Armenian, while Urartian "emphatic" stops are found as unaspirated voiceless stops in Armenian. E.g., Urartian ul-ṭu 'camel' ↦ Armenian ուղտ ułt, Urartian ṣu-pa- 'Sophene ' ↦ Armenian Ծոփ- Copʰ-. Contrasting the last example with Urartian ṭu-uš-pa- ‘Tushpa ’ ↦ Armenian Տոսպ Tosp, Hachikian reconstructs an "emphasis" distinction in the bilabial position.
The cuneiform signs usually transliterated with ‹s, z, ṣ› were not fricatives, but affricates, as again shown by loans in Armenian. E.g., Urartian ṣa-ri ‘orchard’ ↦ Armenian ծառ caṙ ‘tree’, Urartian al-zi- ‘Arzanene ’ ↦ Armenian Աղձնի- Ałʒni-. Urartian ‹š› was loaned into Armenian as /s/: Urartian ša-ni ‘kettle’ ↦ Armenian սան san.
The precise phonetics of emphasis is not recoverable. It possibly may have been ejectivization or glottalization /pʼ, tʼ, t͡sʼ, kʼ/ as in Semitic languages of the time and the nearby endemic languages of the Caucasus, or just plain unaspirated as in Armenian, in either case, contrasting fully with the respective aspirated and voiced series. Near front vowels, /g/ was palatalized and probably merged with, or at least became perceptibly close to, /j/. A distinct /v/ is suggested by variant spellings alternating between and and by the toponym rendered in Armenian as Վան Van ‘Van’ and written bi-a-i-ni- in Urartian.
Hachikian also suggests and. For a phonemic distinct from, there is limited evidence from the Greek rendering of the toponym Κομμαγηνή Kommagēnḗ ‘Commagene’ for Urartian qu-ma-ḫa-; thus, and were not orthograpically distinguished.

Vowels

The script distinguishes the vowels a, e, i and u. Hachikian believes that there was an /o/ as well, as reflected in loans such as the rendition of Urartian ṭu-uš-pa- ‘Tushpa ’ as Armenian Տոսպ Tosp and Greek Θοσπ- Thosp-. There may have been phonemic vowel length, but it is not consistently expressed in the script. Word-finally, the distinction between e and i is not maintained, so many scholars transcribe the graphically vacillating vowel as a schwa: ə, while some preserve a non-reduced vowel. The full form of the vowel appears when suffixes are added to the word and the vowel is no longer in the last syllable: Argištə 'Argišti' - Argištešə 'by Argišti '. This vowel reduction also suggests that stress was commonly on the next-to-the-last syllable.
In the morphonology, various morpheme combinations trigger syncope: *ar-it-u-mə → artumə, *zaditumə → zatumə, *ebani-ne-lə → ebanelə, *turulyə → tulyə.