This is part of archived UD v1 documentation. See http://universaldependencies.org/ for the current version.
home ru/pos issue tracker

ADJ: adjective

Definition

Adjectives are words that typically modify nouns and specify their properties or attributes. They may also function as predicates, as in

Машина зеленая “The car is green.”

The ADJ tag is intended for ordinary adjectives only. See DET for determiners and NUM for cardinal numerals.

In accord with the UD approach, adjectival ordinal numerals (первый, седьмой, стопятидесятый)  are tagged as adjectives, although the traditional grammar classifies them as numerals. They behave like adjectives both morphologically and syntactically, with the exception that they cannot be compared and negated.

Most Russian adjectives inflect for ru-feat/Gender (большой – большое – большая) “big”, ru-feat/Number (большой – большие), ru-feat/Case (большой – большого – большому – большим – большом), ru-feat/Degree (большой – больше – наибольший) and Negation (большой – небольшой).

Examples

Border cases

Passive participles lie on the border between verbs and adjectives. Core participial forms (ending in consonant or short vowel) are tagged VERB. Long forms are participial adjectives and they are tagged ADJ. For example:

Only true participles (verbs) can be used to form the passive voice (but it may be sometimes difficult to distinguish from copula constructions, see AUX). On the other hand, the participial adjectives inflect for case and thus can modify nouns.

There is an analogy with some adjectives that preserved so called nominal (short) forms. And these adjectives are not derived from verbs. Example:

Here both groups are ADJ. The nominal forms are used in predication, the standard forms both in predication and to modify nouns.


Treebank Statistics (UD_Russian)

There are 3769 ADJ lemmas (20%), 6671 ADJ types (22%) and 12345 ADJ tokens (12%). Out of 16 observed tags, the rank of ADJ is: 3 in number of lemmas, 2 in number of types and 3 in number of tokens.

The 10 most frequent ADJ lemmas: ПЕРВЫЙ, Й, НОВЫЙ, ДРУГОЙ, ВТОРОЙ, БОЛЬШОЙ, САМЫЙ, ИЗВЕСТНЫЙ, РОССИЙСКИЙ, ОСНОВНОЙ

The 10 most frequent ADJ types: второй, й, 2008, 2010, х, других, 2004, первый, 2012, 1

The 10 most frequent ambiguous lemmas: ДРУГОЙ (ADJ 100, NOUN 10), РУССКИЙ (ADJ 57, NOUN 11), ВОЕННЫЙ (ADJ 52, NOUN 2), ПОСЛЕДНИЙ (ADJ 46, NOUN 8), 2008 (ADJ 41, NUM 2, ADV 1), ЛУЧШИЙ (ADJ 40, NOUN 1), 2010 (ADJ 39, ADV 1, NUM 1), НЕМЕЦКИЙ (ADJ 35, NOUN 1), 2012 (ADJ 34, NUM 1, ADV 1), 1 (NUM 43, ADJ 33, ADV 19)

The 10 most frequent ambiguous types: 2008 (ADJ 41, NUM 2, ADV 1), 2010 (ADJ 39, NUM 1, ADV 1), х (ADJ 39, NUM 5, NOUN 1), 2012 (ADJ 34, ADV 1, NUM 1), 1 (NUM 43, ADJ 33, ADV 19), 2011 (ADJ 33, NUM 1), 2007 (ADJ 32, NUM 1), II (ADJ 30, NUM 3), 12 (ADJ 29, NUM 17, ADV 6), 2005 (ADJ 25, NUM 4)

Morphology

The form / lemma ratio of ADJ is 1.769966 (the average of all parts of speech is 1.591757).

The 1st highest number of forms (14) was observed with the lemma “ИЗВЕСТНЫЙ”: известен, известная, известно, известного, известное, известной, известному, известную, известны, известные, известный, известным, известными, известных.

The 2nd highest number of forms (13) was observed with the lemma “БОЛЬШОЙ”: Большом, большая, большие, большим, большими, больших, большого, большое, большой, большому, большую, велика, велико.

The 3rd highest number of forms (12) was observed with the lemma “ПЕРВЫЙ”: первая, первого, первое, первой, первом, первому, первую, первые, первый, первым, первыми, первых.

ADJ occurs with 6 features: ru-feat/Animacy (12314; 100% instances), ru-feat/Case (12314; 100% instances), ru-feat/Number (12314; 100% instances), ru-feat/Variant (9636; 78% instances), ru-feat/Gender (9591; 78% instances), ru-feat/Degree (133; 1% instances)

ADJ occurs with 17 feature-value pairs: Animacy=Anim, Animacy=Inan, Case=Acc, Case=Dat, Case=Gen, Case=Ins, Case=Loc, Case=Nom, Degree=Cmp, Degree=Sup, Gender=Fem, Gender=Masc, Gender=Neut, Number=Plur, Number=Sing, Variant=Brev, Variant=Full

ADJ occurs with 114 feature combinations. The most frequent feature combination is Animacy=Inan|Case=Gen|Gender=Fem|Number=Sing|Variant=Full (871 tokens). Examples: российской, отечественной, великой, новой, мировой, железной, московской, гражданской, русской, Вологодской

Relations

ADJ nodes are attached to their parents using 23 different relations: ru-dep/amod (10637; 86% instances), ru-dep/conj (493; 4% instances), ru-dep/nmod (323; 3% instances), ru-dep/root (241; 2% instances), ru-dep/list (234; 2% instances), ru-dep/appos (136; 1% instances), ru-dep/acl (40; 0% instances), ru-dep/parataxis (40; 0% instances), ru-dep/nsubj (33; 0% instances), ru-dep/acl:relcl (32; 0% instances), ru-dep/remnant (26; 0% instances), ru-dep/ccomp (25; 0% instances), ru-dep/dobj (19; 0% instances), ru-dep/advcl (17; 0% instances), ru-dep/advmod (13; 0% instances), ru-dep/xcomp (10; 0% instances), ru-dep/goeswith (6; 0% instances), ru-dep/nummod (6; 0% instances), ru-dep/compound (4; 0% instances), ru-dep/nsubjpass (4; 0% instances), ru-dep/case (3; 0% instances), ru-dep/iobj (2; 0% instances), ru-dep/name (1; 0% instances)

Parents of ADJ nodes belong to 11 different parts of speech: NOUN (10760; 87% instances), ADJ (539; 4% instances), PROPN (443; 4% instances), VERB (280; 2% instances), ROOT (241; 2% instances), ADP (36; 0% instances), PRON (18; 0% instances), ADV (12; 0% instances), NUM (8; 0% instances), DET (6; 0% instances), PUNCT (2; 0% instances)

10399 (84%) ADJ nodes are leaves.

562 (5%) ADJ nodes have one child.

781 (6%) ADJ nodes have two children.

603 (5%) ADJ nodes have three or more children.

The highest child degree of a ADJ node is 16.

Children of ADJ nodes are attached using 33 different relations: ru-dep/punct (1224; 25% instances), ru-dep/goeswith (851; 17% instances), ru-dep/conj (487; 10% instances), ru-dep/nmod (415; 9% instances), ru-dep/cc (351; 7% instances), ru-dep/nsubj (325; 7% instances), ru-dep/advmod (269; 6% instances), ru-dep/cop (156; 3% instances), ru-dep/case (155; 3% instances), ru-dep/amod (75; 2% instances), ru-dep/xcomp (72; 1% instances), ru-dep/iobj (50; 1% instances), ru-dep/mark (49; 1% instances), ru-dep/parataxis (44; 1% instances), ru-dep/neg (43; 1% instances), ru-dep/discourse (41; 1% instances), ru-dep/appos (37; 1% instances), ru-dep/advcl (33; 1% instances), ru-dep/remnant (30; 1% instances), ru-dep/cc:preconj (18; 0% instances), ru-dep/dobj (17; 0% instances), ru-dep/nsubjpass (16; 0% instances), ru-dep/acl (15; 0% instances), ru-dep/aux (14; 0% instances), ru-dep/det (14; 0% instances), ru-dep/auxpass (12; 0% instances), ru-dep/list (11; 0% instances), ru-dep/ccomp (10; 0% instances), ru-dep/nummod (10; 0% instances), ru-dep/nummod:gov (9; 0% instances), ru-dep/mwe (6; 0% instances), ru-dep/acl:relcl (4; 0% instances), ru-dep/compound (1; 0% instances)

Children of ADJ nodes belong to 15 different parts of speech: PUNCT (1646; 34% instances), ADV (706; 15% instances), NOUN (699; 14% instances), ADJ (539; 11% instances), CONJ (365; 8% instances), VERB (347; 7% instances), ADP (180; 4% instances), PRON (99; 2% instances), PART (77; 2% instances), PROPN (68; 1% instances), SCONJ (50; 1% instances), AUX (36; 1% instances), NUM (28; 1% instances), DET (16; 0% instances), SYM (8; 0% instances)


Treebank Statistics (UD_Russian-SynTagRus)

There are 8714 ADJ lemmas (20%), 27742 ADJ types (24%) and 115829 ADJ tokens (11%). Out of 17 observed tags, the rank of ADJ is: 3 in number of lemmas, 3 in number of types and 4 in number of tokens.

The 10 most frequent ADJ lemmas: который, другой, новый, самый, первый, сам, должен, российский, большой, нужный

The 10 most frequent ADJ types: которые, который, которых, которая, других, нужно, которой, другой, многие, должны

The 10 most frequent ambiguous lemmas: который (ADJ 5029, DET 54), первый (ADJ 1337, NOUN 1), такой (DET 1819, ADJ 719), тот (DET 1503, ADJ 710, NOUN 535, PROPN 4), один (NUM 1927, ADJ 575, NOUN 1), весь (DET 2645, ADJ 422), русский (ADJ 346, NOUN 52, PROPN 1), военный (ADJ 338, NOUN 52), простой (ADJ 327, NOUN 2), этот (DET 5073, ADJ 255, NOUN 6)

The 10 most frequent ambiguous types: которые (ADJ 1314, DET 23), который (ADJ 794, DET 17), которых (ADJ 570, DET 3), которая (ADJ 518, DET 7), которого (ADJ 307, DET 2), первый (ADJ 221, NOUN 1), которое (ADJ 261, DET 2), все (DET 907, NOUN 858, PART 337, ADJ 237), необходимо (ADJ 172, ADV 2), невозможно (ADJ 182, ADV 4)

Morphology

The form / lemma ratio of ADJ is 3.183613 (the average of all parts of speech is 2.665758).

The 1st highest number of forms (31) was observed with the lemma “важный”: важен, важна, важная, важнее, важней, важнейшая, важнейшего, важнейшее, важнейшей, важнейшем, важнейшие, важнейший, важнейшим, важнейшими, важнейших, важнейшую, важно, важного, важное, важной, важном, важному, важную, важны, важные, важный, важным, важными, важных, наиважнейших, поважнее.

The 2nd highest number of forms (30) was observed with the lemma “сильный”: посильнее, силен, сильна, сильная, сильнее, сильней, сильнейшая, сильнейшего, сильнейшей, сильнейшем, сильнейшему, сильнейшие, сильнейший, сильнейшим, сильнейшими, сильнейших, сильнейшую, сильного, сильное, сильной, сильном, сильному, сильную, сильны, сильные, сильный, сильным, сильными, сильных, силён.

The 3rd highest number of forms (28) was observed with the lemma “близкий”: Поближе, ближайшее, ближайшей, ближайшем, ближайшему, ближайшие, ближайший, ближайшим, ближайшими, ближайших, ближайшую, ближе, близка, близкая, близки, близкие, близкий, близким, близкими, близких, близко, близкого, близкое, близкой, близком, близкому, близкую, близок.

ADJ occurs with 6 features: ru-feat/Degree (115829; 100% instances), ru-feat/Number (113958; 98% instances), ru-feat/Case (105119; 91% instances), ru-feat/Gender (76038; 66% instances), ru-feat/Animacy (10546; 9% instances), ru-feat/Variant (8841; 8% instances)

ADJ occurs with 17 feature-value pairs: Animacy=Anim, Animacy=Inan, Case=Acc, Case=Dat, Case=Gen, Case=Ins, Case=Loc, Case=Nom, Degree=Cmp, Degree=Pos, Degree=Sup, Gender=Fem, Gender=Masc, Gender=Neut, Number=Plur, Number=Sing, Variant=Brev

ADJ occurs with 60 feature combinations. The most frequent feature combination is Case=Gen|Degree=Pos|Number=Plur (12548 tokens). Examples: которых, других, новых, самых, российских, многих, разных, научных, различных, политических

Relations

ADJ nodes are attached to their parents using 20 different relations: ru-dep/amod (87754; 76% instances), ru-dep/root (6271; 5% instances), ru-dep/conj (5965; 5% instances), ru-dep/nmod (5606; 5% instances), ru-dep/nsubj (3396; 3% instances), ru-dep/advcl (1344; 1% instances), ru-dep/parataxis (1172; 1% instances), ru-dep/dobj (1145; 1% instances), ru-dep/acl (1117; 1% instances), ru-dep/compound (578; 0% instances), ru-dep/acl:relcl (523; 0% instances), ru-dep/nsubjpass (366; 0% instances), ru-dep/advmod (259; 0% instances), ru-dep/dep (148; 0% instances), ru-dep/appos (119; 0% instances), ru-dep/nmod:agent (25; 0% instances), ru-dep/iobj (20; 0% instances), ru-dep/mwe (18; 0% instances), ru-dep/name (2; 0% instances), ru-dep/nummod (1; 0% instances)

Parents of ADJ nodes belong to 14 different parts of speech: NOUN (86202; 74% instances), VERB (11510; 10% instances), ADJ (7814; 7% instances), ROOT (6271; 5% instances), PROPN (2228; 2% instances), PRON (601; 1% instances), ADV (461; 0% instances), NUM (411; 0% instances), SCONJ (171; 0% instances), CONJ (58; 0% instances), SYM (40; 0% instances), PART (38; 0% instances), X (22; 0% instances), INTJ (2; 0% instances)

83340 (72%) ADJ nodes are leaves.

14733 (13%) ADJ nodes have one child.

7048 (6%) ADJ nodes have two children.

10708 (9%) ADJ nodes have three or more children.

The highest child degree of a ADJ node is 18.

Children of ADJ nodes are attached using 31 different relations: ru-dep/punct (19458; 26% instances), ru-dep/advmod (10185; 14% instances), ru-dep/nsubj (9062; 12% instances), ru-dep/nmod (6519; 9% instances), ru-dep/conj (6288; 9% instances), ru-dep/cc (5748; 8% instances), ru-dep/case (3014; 4% instances), ru-dep/parataxis (2442; 3% instances), ru-dep/cop (2272; 3% instances), ru-dep/dep (1688; 2% instances), ru-dep/amod (1635; 2% instances), ru-dep/advcl (1598; 2% instances), ru-dep/mark (1357; 2% instances), ru-dep/neg (1011; 1% instances), ru-dep/compound (570; 1% instances), ru-dep/acl:relcl (221; 0% instances), ru-dep/aux (164; 0% instances), ru-dep/nummod (119; 0% instances), ru-dep/iobj (106; 0% instances), ru-dep/acl (86; 0% instances), ru-dep/appos (77; 0% instances), ru-dep/nummod:gov (75; 0% instances), ru-dep/mwe (47; 0% instances), ru-dep/foreign (22; 0% instances), ru-dep/nmod:agent (9; 0% instances), ru-dep/dobj (8; 0% instances), ru-dep/discourse (4; 0% instances), ru-dep/vocative (3; 0% instances), ru-dep/expl (2; 0% instances), ru-dep/auxpass (1; 0% instances), ru-dep/name (1; 0% instances)

Children of ADJ nodes belong to 16 different parts of speech: PUNCT (19458; 26% instances), NOUN (11873; 16% instances), ADV (8192; 11% instances), ADJ (7814; 11% instances), VERB (7269; 10% instances), CONJ (5000; 7% instances), PART (3434; 5% instances), ADP (3014; 4% instances), PRON (2313; 3% instances), AUX (2238; 3% instances), SCONJ (2194; 3% instances), PROPN (729; 1% instances), NUM (211; 0% instances), SYM (34; 0% instances), X (15; 0% instances), INTJ (4; 0% instances)


ADJ in other languages: [bg] [cs] [de] [el] [en] [es] [eu] [fa] [fi] [fr] [ga] [he] [hu] [it] [ja] [ko] [sv] [u]