The Kashmiri data is written in Perso-Arabic script, right to left; every sentence has # rtl = yes.
The treebank has two parts: ks_nehru_* (Kashmiri Wikipedia) and ks_grammar_* (example sentences from Koul, Kashmiri: A Grammatical Sketch).
Sentence-level comments
Comment
Part
Content
# text
both
The real text. A token not followed by a space has SpaceAfter=No.
# original_text
Nehru
The sentence exactly as written in the Wikipedia article.
# translit
both
Transliteration of # text (see below).
# gloss
grammar
English word-by-word gloss.
# translation
both
English translation.
Transliteration
# translit is generated from the Perso-Arabic text with the same rules for both parts, so it follows the spelling, not a separate phonetic transcription. Words with a missing space in the source are transliterated with the corrected spacing. Punctuation is written as ,;?..
Perso-Arabic
Latin
Perso-Arabic
Latin
َُِ
a i u
ٕ
ụ
ٔ / أ
ạ
ٲ
ạ̄
اآ (after a consonant)
ā
ٟ
ụ̄
یٖ
ī
وٗ
ū
ی inside a word / at the end
e / ī
و after a consonant / elsewhere
ō / v
ۆ
o
ۄ
ọ
ے
e (after a, ā: y)
ؠ
ʲ
ٹڈڑ
ṭ ḍ ṛ
ژ
ts
چ / چھ
ch / chh
ش
sh
قخغ
q kh gh
consonant + ھ
aspirated (kh, th, ṭh, bh …)
Tokenization and Word Segmentation
Tokens are separated by whitespace. Punctuation is always a separate token: ۔ (80), ، (54), " (30), : (9), ( and ) (6 each), ؛ (4), ' (2), ... (2).
Every sentence of the Nehru part ends with ۔, as in the source. Sentence-final punctuation is left out in the grammar part.
Digits are written as in the source: the Nehru part uses 1912 …, except ks_nehru_043 and ks_nehru_057, which have ۱۹۱۲ and ۱۹۱۵ in the source. The lemma always uses Western digits (1912, 1915).
Misspellings in the source are kept in FORM and corrected in MISC with CorrectForm; the lemma uses the correct spelling: اَكھ → اَکھ (ks_nehru_001), کێنٛہہ → کیٚنٛہہ (ks_nehru_036), مظهرُل → مظہرُل (ks_nehru_055), تٔحذیٖب → تٔہذیٖب (ks_nehru_056), ہیٛێنٛدؠن → ہیٛیٚنٛدؠن (ks_nehru_063), یێلہِ → ییٚلہِ and تێلہِ → تیٚلہِ (ks_nehru_066).
A missing space in the source between two words is marked with SpaceAfter=No|CorrectSpaceAfter=Yes: مَنٛزآزٲدی (ks_nehru_010), گوکھَلےموٗد (ks_nehru_057).
Case endings written as separate words are separate tokens, tagged ADP with case on the preceding word: اَس (16), یَس (8), وَن (3), یَن (2), ہَس, اُک, وَس, یُک, یُن (1 each), for example 1947 اَس مَنٛز, شا ہَس, گانٛدھی یُن “Gandhi’s”.
Genitive postpositionsسُنٛد, ہُنٛد, سٕنٛز, سٕنٛدِ … are separate words (ADP, case). کِس (8) is also written as a separate word and is treated in the same way. A pronoun with a genitive postposition is two words: تٔمؠ سُنٛد “his” (in the grammar part also تٔمؠ سٕنٛز, تٔمؠ سٕنٛدِ, اَکؠ سُنٛد).
The possessive pronouns تِہُنٛد “his/her/their (respectful)” and تُہُنٛد “your”, with their inflected forms, are single words (PRON, Poss=Yes).
Multiword token:کانٛگرٛیسکِس (ks_nehru_004) is written as one word in the source and split into کانٛگرٛیس + کِس.
The izafat compound جَنٛگہِآزٲدی “freedom struggle” (ks_nehru_003) is written as one word in the source and kept as one token. This is different from مَنٛزآزٲدی (ks_nehru_010), which is two independent words (a postposition and a noun) with a missing space.
Morphology
Tags
The data uses 14 UPOS tags. XPOS is not used.
AUX has two lemmas:
آسُن “be”, with the forms چھُ, چھَ, چھِ, چھُس, چھُسَتھ, چھَے, چھُم, چھُنہٕ, چھَنہٕ, اوس, اوسُس, أس, أسؠ, ٲسؠ, آس, آسہِ, آسان, آسِوہِ.
یُن “come”, the passive auxiliary, with the forms آیہِ, آو, آمَژٕ, یِوان.
ہیکُن “can”, گَژُن “go”, بَنُن “become” and لَگُن are VERB.
PART is used for the negation نہٕ, نَہ, the question particle ما, the focus particle تہِ “also”, and for وول after a verb (see below).
DET and PRON: the demonstratives سُہ, سۄ, یہِ, یِم, ہُ, یَتھ, تَتھ are DET before a noun and PRON when they stand alone. The quantifiers سٲری, سارِوٕے and دۄشوؠن “all, both” are PRON with PronType=Tot when they stand alone (ks_nehru_046, ks_nehru_065, ks_grammar_056). Other DET lemmas: اَکھ, کیٚنٛہہ, آل, دَ (English “the” in a book title).
The relatives یُس, یۄس, یٚمہِ are DET with det before a noun (ks_grammar_001, 002, 036, 038, 045) and PRON otherwise.
وول after an infinitive or participle forms an agent or relative participle and is PART with mark (دِلہِ روزَن وٲلؠ کٲشِرؠ خاندانٕچ “of a Kashmiri family living in Delhi”, ks_nehru_060). After a noun it forms a noun “person of X” and is NOUN: کانٛگرٛیس والؠن “the Congress people” (ks_nehru_067).
کِژھ / کِژھِ “for” is ADP with case ("ہوم روٗل" اَس کِژھ, ks_nehru_058; پانَس کِژھ “for himself”, ks_grammar_092).
Numbers are NUM with NumType=Card, also number words: زٕ “two”, ژۄداہ “fourteen”, کاہ “eleven” (ks_nehru_026), ترُٛواہ “thirteen”, اَرٕترٕٛہ “thirty-eight”, ڈایَن ساسَن “two and a half thousand” (ks_nehru_050).
Features
UPOS
Features
PRON
PronType (Prs, Dem, Ind, Int, Rel, Tot), Person (1, 2, 3), Number (Sing, Plur), Gender (Masc, Fem), Poss=Yes, Reflex=Yes
DET
PronType (Dem, Ind, Tot, Art, Rel), Number=Plur
NUM
NumType=Card
VERB
VerbForm (Fin, Inf, Part, Conv), and on finite forms Mood (Ind, Imp) and Tense (Past, Fut)
Every VERB and AUX has at least VerbForm. Finite indicative forms have Mood=Ind and Tense where the form shows it.
Future forms have Tense=Fut: گَژھہِ “will go”, یِیہِ “will come”, آسہِ “will be” (ks_nehru_004), میلَن “will get” (ks_nehru_069b).
Commands have Mood=Imp: کَر “do!” (ks_grammar_048), یِزہِ “come!” (ks_grammar_058).
Reflex=Yes is used on پان, پَنُن and پانہِوانؠ. Polarity=Neg is used on چھُنہٕ and چھَنہٕ; the particles نہٕ, نَہ have no features.
Nouns, adjectives and proper nouns have no features. Case is not annotated.
Syntax
Core arguments
Ergative subjects such as تٔمؠ, نِہروٗ وَن are nsubj.
iobj is used for the recipient beside a direct object: لیٖچھ نِہروٗوَن شا ہَس چِٹھؠ “Nehru wrote a letter to Shaw” (ks_nehru_034).
A dative experiencer is obl: تٔمؠ سُنٛد نَژُن چھُ مےٚ پَسَنٛد “I like his dance” (ks_grammar_049).
A postposition is case on its noun.
Copula and auxiliaries
With a nominal or adjectival predicate, the predicate is the head and آسُن is cop: یہِ چھُ سؠٹھا بۆڈ کُل “this is a very big tree” (ks_grammar_003). With an infinitive clause as predicate, cop attaches to the infinitive (ییٚمیُٛک مَقصَد اوس … ژھانٛڈُن “whose aim was to find …”, ks_nehru_068).
With a verb, آسُن is aux: شُر چھُ تێز دوران “the child runs fast” (ks_grammar_004).
In passives, یُن with an infinitive is aux:pass, the subject is nsubj:pass and the “by” phrase is obl:agent: پَلَو آیہِ نہٕ مینِنہِ ذَرِیہِ چھَلنہِ “the clothes were not washed by Meena” (ks_grammar_016). See also ks_nehru_007, ks_grammar_076, ks_grammar_077.
The tag question چھَ نَہ “isn’t it?” is parataxis of the main predicate (ks_grammar_023).
Light verbs, compounds and reduplication
A noun combined with a light verb is compound of the verb:
with دِیُن “give”: دیٛان; with کَڈُن: کٕٹھ; with گَژُن: شۆروٗ
A noun or proper noun that modifies a following noun is compound of it: نِہروٗ خاندان (ks_nehru_017).
In reduplication, the second copy is compound:redup of the first: زورہِ زورہِ (ks_grammar_005), کُس کُس (ks_grammar_029), کؠا کؠا (ks_grammar_030), پَکان پَکان (ks_grammar_054), پانہٕ پانہٕ (ks_nehru_064).
Names
Names of persons and places are PROPN; the first word is the head and the following words are flat: جَواہَر لال نِہروٗ (ks_nehru_001), اِللّٰه اَباد (ks_nehru_024, ks_nehru_040).
Names of organizations and titles that have a regular internal structure are annotated with regular relations, following the UD guidelines for flat: اِنڈیَن اِنسٹیٖچیٛوٗٹ آف ٹیکنالَجی (ks_nehru_013), The Discovery of Indiaدَ ڈِسکَوَری آف اِنڈیا (ks_nehru_009), اِنَر ٹیمپلَس (ks_nehru_038), مُسلِم لیٖگ (ks_nehru_055, ks_nehru_063).
Nominal modifiers
A genitive is nmod, with the genitive postposition as case.
An izafat phrase is nmod of its head: تٔحریٖکہِ آزٲدی “freedom movement” (ks_nehru_002, ks_nehru_005).
A participle with وول is acl of the noun that follows it: بَننہٕ وول ہِندوستٲنؠ قومَس “the Indian nation that was to be formed” (ks_nehru_006).
Clauses with the relatives یُس, یۄس, یَتھ, ییٚمؠ, ییٚمیُٛک, یٚمہِ are acl:relcl. In یُس … سُہ correlatives, the relative clause is attached to the correlative word: یُس کوٹھ ژٚ ہؠۆتُتھ سُہ (ks_grammar_001).
An apposition follows its head and is appos. A bracket that comes before the word it explains is parataxis, because appos cannot precede its head (ks_nehru_058).
Clauses
ccomp clauses are introduced by زِ, کہِ or اَگَر.
csubj is used for a clause that is the subject: with لَگُن (ks_nehru_041, 044, 049) and with the predicates مَژَر, مُمکِن and پَسَنٛد.
Infinitives are annotated by their function:
xcomp when the infinitive is a complement required by the verb: “want to see” (ks_grammar_033), “forget to give” (ks_grammar_035), “let me eat” (ks_grammar_074), “told him to read” (ks_grammar_051, ks_grammar_092).
advcl for purpose, time and cause: “went to see a play” (ks_grammar_050), “will go to pluck flowers” (ks_grammar_073), “marched … to do satyagraha” (ks_nehru_050), “before he comes” (ks_grammar_046), “after his departure” (ks_grammar_047), “because of the arrival” (ks_grammar_056).
csubj as subject, acl as noun modifier (ks_nehru_051, ks_nehru_058).
A postposition after an infinitive is mark: خاطرہِ, پؠٹھ, وَرٲے, بروہ, پَتہٕ, سٟتؠ, کہِ. After a noun it is case.
Other adverbial clauses are advcl, for example those with ییٚلہِ, تِکیٛازِ, اَگَر, اَگَرچہِ, and converbs in -ِتھ.
Coordination: the coordinators تہٕ, مَگَر, یا, کِنہِ and نَتہِ are cc on the following conjunct; every later conjunct is conj of the first.
parataxis is used in ks_nehru_035, ks_nehru_051, ks_nehru_058, ks_nehru_066 and ks_grammar_023.