UD for Hittite 
Tokenization and Word Segmentation
- In HitTB, tokens are delimited using whitespace. This is reflected in the original writing, as Hittite cuneiform also used whitespace to delineate.
- Cuneiform is sometimes difficult to display. The treebank instead displays forms in a way that closely approximates the cuneiform writing, using narrow transcription, where each cuneiform character transliteration is separated with hyphens. Broad transcription, which more closely approximates the pronounciation, is used for lemmas.
- In Hittite, clitics do not stand as independent tokens. However, clitics most often function grammatically differently from the other word or words in that token. Therefore, in HitTB, clitics are separated out as unique words. For ease, clitics are always introduced by ‘=’.
- Other multiword tokens include Sumerian compound-word borrowings and words with determinatives. A determinative acts as a class for the noun it modifies, such as the determinative LÚ “man”, which is used for people. The determinative is treated as a separate word to preserve the determinatives in the treebank.
Morphology
Tags
- In HitTB, 15 part-of-speech tags are used.
- Only SYM and PUNCT remain unused.
- AUX is used for the copula ēš- (which, like in English, also can function as a true auxiliary), as well as the auxiliary verbs ḫar(k)- and dai-. A final word tagged AUX is NUGÁL, which functions as a negative copula.
- DET is used for the demonstratives kā-, apā-, and aši (before nouns), the interrogative kuiš “which” (before nouns), and the word mekki “many”.
- PART is used for nu and -(m)a (when they clearly do not act as coordinating conjunctions), the clitics -za, -ašta, -pát, -kán, and -šan, and the negatives lē and natta.
Nominal Features
For nouns, the following features are possible:
- Case, which has 10 possibilities (Nom, Acc, Gen, Dat, Abl, Ins, All, Erg, Voc, Abs)
- Definite, which is used to model the Akkadian construct form and has 1 possibility (Cons)
- Gender, which has 4 possibilities (Com, Neut, Masc, Fem)
- Number, which has 2 possibilities (Sing, Plur)
The Case, Gender, and Number features can also be used with verbs (in participle form), pronouns, numerals, adjectives, and determiners.
Verbal Features
Verbs (not including participles) in HitTB can have the following features:
- Person (1, 2, 3)
- Mood (Ind, Imp)
- Tense (Pres, Past)
- VerbForm (Fin, Inf, Part, Sup, Vnoun)
- Voice (Act, Mid)
Language-Specific Features
HitTB contains one language-specific feature, called Language. Since Hittite uses forms from both Akkadian and Sumerian, it is important to know which of these two languages the word originated from, especially in the event that the word displays Akkadian or Sumerian syntax. The possible values of Language are Language=Akk and Language=Sum
Syntax
- Hittite is most often a SOV language.
Instruction: Give criteria for identifying core arguments (subjects and objects), and describe the range of copula constructions in nonverbal clauses. List all subtype relations used. Include links to language-specific relations definitions if any.
Treebanks
There is 1 Hittite UD treebank:
Instruction: Treebank-specific pages are generated automatically from the README file in the treebank repository and
from the data in the latest release. Link to the respective *-index.html page in the treebanks folder, using the language code
and the treebank code in the file name.