NUM
: numeral
Definition
A numeral is a word, functioning most typically as a determiner, adjective or pronoun, that expresses a number and a relation to the number, such as quantity, sequence, frequency or fraction.
Note that cardinal numerals are covered by NUM
whether they are used
as determiners or not (as in Windows 7) and whether they
are expressed as words (четыре), digits (4) or Roman numerals
(IV).
Russian grammar distinguishes several subclasses of pronominal numerals (quantifiers): interrogative and relative (сколько “how many”); demonstrative (столько “this many”); indefinite (несколько “several”). These words behave similarly to (most) cardinal numbers, e.g. they require that the counted noun phrase be in Genitive. They are not similar to adjectives (unlike their English counterparts).
In addition, several types of (non-pronominal) numerals, such as ordinal numerals and multiplicative numerals, are tagged ADJ or ADV, based on their syntactic and morphological behavior.
Examples
- 0, 1, 2, 3, 4, 5, 2014, 1000000, 3.14159265359
- I, II, III, IV, V, MMXIV
- один, два, три, четыре, пять, семьдесят “one, two, three, four, five, seventy”
- половина, треть, четверть “one-half, one third, quarter”: denominators of fractions constitute a separate class of cardinal numerals.
- двое, трое, четверо, пятеро “four, five”: collective numerals (see specific-syntax on their morphosyntactic behavior).
- сколько, столько, предостаточно “how many, this many, more than enough”: pronominal quantifiers of imprecise quantity.
Counterexamples
- первый, второй, третий “first, second, third”: adjectival ordinal numerals. They are tagged ADJ, and the ru-feat/NumType feature reveals their semantic relation to numbers.
- впервые “for the first time”: adverbial ordinal numerals. They are tagged ADV, and the ru-feat/NumType feature reveals their semantic relation to numbers.
- однажды, дважды, трижды “once, twice, three times”: multiplicative numerals. They are tagged ADV, and the ru-feat/NumType feature reveals their semantic relation to numbers.
- пара, тройка, четверка “pair, triplet, foursome”: n-tuples (n-tice) are not considered numerals in the Russian grammar. They are tagged NOUN.
- единица, двойка, тройка, четверка, пятерка “number one, number two, number three, number four, number five”: names of numbers, or of objects identified by the number (e.g. of a bus route). They are not considered numerals and they are tagged NOUN.
Border cases
- тысяча, миллион, миллиард, триллион “thousand, million, billion, trillion”: words for large quantities are ambiguous between cardinal numerals (tagged
NUM
) and nouns. If they inflect as nouns, they are tagged NOUN; but the borderline is fuzzy. For instance, in phrases like тысячи людей вышли на улицы (“thousands of people went on the streets”), тысячи is a noun. In numeric expressions, e.g. 110 тысяч долларов (“110 thousand dollars”), it is a cardinal numeral. - много, мало, немного, немало, несколько, достаточно “many, few, not many, a lot, several, enough”: pronominal quantifiers are ambiguous between cardinal numerals (tagged
NUM
when they refer to imprecise quantities) and adverbs (tagged ADV when they refer to degree/intensity). As a rule, the latter have verbs, adjectives, and adverbs as their head (e.g. я был несколько груб “I was a bit rude”). Note that the words более, больше, менее, меньше “more than, less than” are considered comparative forms of the numerals много and мало when they are used in constructions with cardinal numerals, e.g. более пяти студентов “more than five students” (see specific-syntax).
Treebank Statistics (UD_Russian)
There are 683 NUM
lemmas (4%), 728 NUM
types (2%) and 2028 NUM
tokens (2%).
Out of 16 observed tags, the rank of NUM
is: 6 in number of lemmas, 6 in number of types and 9 in number of tokens.
The 10 most frequent NUM
lemmas: ОДИН, ДВА, НЕСКОЛЬКО, ТРИ, 2, 1, 10, ЧЕТЫРЕ, 4, 3
The 10 most frequent NUM
types: 2, два, один, несколько, 1, двух, 10, 4, три, 3
The 10 most frequent ambiguous lemmas: ОДИН (NUM 185, ADV 1), НЕСКОЛЬКО (NUM 68, ADV 5), ТРИ (NUM 58, ADV 1), 2 (NUM 55, ADV 23, ADJ 9), 1 (NUM 43, ADJ 33, ADV 19), 10 (NUM 40, ADJ 14, ADV 8), 4 (NUM 35, ADJ 14, ADV 13), 3 (NUM 31, ADV 13, ADJ 8), 5 (NUM 29, ADJ 9, ADV 5), МНОГО (NUM 29, ADV 9)
The 10 most frequent ambiguous types: 2 (NUM 55, ADV 23, ADJ 9), один (NUM 42, ADV 1), несколько (NUM 41, ADV 5), 1 (NUM 43, ADJ 33, ADV 19), 10 (NUM 40, ADJ 14, ADV 8), 4 (NUM 35, ADJ 14, ADV 13), три (NUM 29, ADV 1), 3 (NUM 30, ADV 13, ADJ 8), 5 (NUM 29, ADJ 9, ADV 5), 20 (NUM 24, ADJ 12, ADV 11)
- 2
- один
- несколько
- 1
- 10
- 4
- три
- 3
- 5
- 20
Morphology
The form / lemma ratio of NUM
is 1.065886 (the average of all parts of speech is 1.591757).
The 1st highest number of forms (10) was observed with the lemma “ОДИН”: один, одна, одним, одних, одно, одного, одной, одном, одному, одну.
The 2nd highest number of forms (5) was observed with the lemma “ДВА”: два, две, двум, двумя, двух.
The 3rd highest number of forms (5) was observed with the lemma “МНОГО”: более, больше, многим, многих, много.
NUM
occurs with 5 features: Case (2026; 100% instances), Animacy (1013; 50% instances), Gender (601; 30% instances), Number (316; 16% instances), Degree (2; 0% instances)
NUM
occurs with 14 feature-value pairs: Animacy=Anim
, Animacy=Inan
, Case=Acc
, Case=Dat
, Case=Gen
, Case=Ins
, Case=Loc
, Case=Nom
, Degree=Cmp
, Gender=Fem
, Gender=Masc
, Gender=Neut
, Number=Plur
, Number=Sing
NUM
occurs with 82 feature combinations.
The most frequent feature combination is Case=Nom
(470 tokens).
Examples: один, 1, 10, 2, 5, два, 0, несколько, 16, 4
Relations
NUM
nodes are attached to their parents using 20 different relations: nummod:gov (845; 42% instances), nummod (520; 26% instances), nummod:entity (125; 6% instances), nmod (92; 5% instances), root (83; 4% instances), list (68; 3% instances), appos (64; 3% instances), conj (57; 3% instances), compound (53; 3% instances), amod (26; 1% instances), nsubj (21; 1% instances), dobj (17; 1% instances), parataxis (15; 1% instances), advmod (11; 1% instances), goeswith (10; 0% instances), remnant (9; 0% instances), nsubjpass (5; 0% instances), iobj (4; 0% instances), acl (2; 0% instances), ccomp (1; 0% instances)
Parents of NUM
nodes belong to 11 different parts of speech: NOUN (1562; 77% instances), VERB (102; 5% instances), NUM (92; 5% instances), ROOT (83; 4% instances), SYM (83; 4% instances), PROPN (63; 3% instances), ADJ (28; 1% instances), ADP (5; 0% instances), ADV (4; 0% instances), PRON (3; 0% instances), PUNCT (3; 0% instances)
1593 (79%) NUM
nodes are leaves.
200 (10%) NUM
nodes have one child.
108 (5%) NUM
nodes have two children.
127 (6%) NUM
nodes have three or more children.
The highest child degree of a NUM
node is 10.
Children of NUM
nodes are attached using 24 different relations: punct (338; 36% instances), nmod (155; 17% instances), nsubj (82; 9% instances), case (69; 7% instances), conj (57; 6% instances), advmod (51; 5% instances), cc (27; 3% instances), cop (27; 3% instances), discourse (27; 3% instances), goeswith (25; 3% instances), appos (12; 1% instances), neg (12; 1% instances), remnant (11; 1% instances), list (9; 1% instances), parataxis (8; 1% instances), nummod (7; 1% instances), amod (4; 0% instances), compound (4; 0% instances), nummod:gov (3; 0% instances), advcl (2; 0% instances), det (2; 0% instances), acl (1; 0% instances), dobj (1; 0% instances), iobj (1; 0% instances)
Children of NUM
nodes belong to 14 different parts of speech: PUNCT (340; 36% instances), NOUN (223; 24% instances), NUM (92; 10% instances), ADP (71; 8% instances), ADV (66; 7% instances), VERB (35; 4% instances), CONJ (27; 3% instances), PART (27; 3% instances), PROPN (15; 2% instances), PRON (12; 1% instances), SYM (12; 1% instances), ADJ (8; 1% instances), DET (6; 1% instances), AUX (1; 0% instances)
Treebank Statistics (UD_Russian-SynTagRus)
There are 1220 NUM
lemmas (3%), 1314 NUM
types (1%) and 16014 NUM
tokens (2%).
Out of 17 observed tags, the rank of NUM
is: 6 in number of lemmas, 6 in number of types and 13 in number of tokens.
The 10 most frequent NUM
lemmas: один, два, несколько, три, 1, 10, 20, 2, пять, четыре
The 10 most frequent NUM
types: один, несколько, два, три, 1, одной, 10, двух, 20, 2
The 10 most frequent ambiguous lemmas: один (NUM 1927, ADJ 575, NOUN 1), несколько (NUM 739, ADV 90), 6 (NUM 145, NOUN 1), пол (NUM 79, NOUN 69, PROPN 16), 2005 (NUM 66, NOUN 1), i (NUM 22, PROPN 5, X 3), 2012 (NUM 21, NOUN 1), x (NUM 9, PUNCT 2, PROPN 2), v (NUM 5, PROPN 1)
The 10 most frequent ambiguous types: один (NUM 493, ADJ 98), несколько (NUM 531, ADV 82), одной (NUM 321, ADJ 79), 10 (NUM 324, ADJ 1), одного (NUM 192, ADJ 55), одна (NUM 154, ADJ 56), одно (NUM 131, ADJ 50), одним (NUM 125, ADJ 22), одну (NUM 109, ADJ 18), одном (NUM 112, ADJ 23)
- один
- несколько
- одной
- 10
- одного
- одна
- одно
- одним
- одну
- одном
Morphology
The form / lemma ratio of NUM
is 1.077049 (the average of all parts of speech is 2.665758).
The 1st highest number of forms (11) was observed with the lemma “один”: один, одна, одни, одним, одними, одно, одного, одной, одном, одному, одну.
The 2nd highest number of forms (8) was observed with the lemma “оба”: оба, обе, обеим, обеими, обеих, обоим, обоими, обоих.
The 3rd highest number of forms (6) was observed with the lemma “три”: трем, тремя, трех, три, трём, трёх.
NUM
occurs with 3 features: Case (6012; 38% instances), Gender (2875; 18% instances), Animacy (1420; 9% instances)
NUM
occurs with 11 feature-value pairs: Animacy=Anim
, Animacy=Inan
, Case=Acc
, Case=Dat
, Case=Gen
, Case=Ins
, Case=Loc
, Case=Nom
, Gender=Fem
, Gender=Masc
, Gender=Neut
NUM
occurs with 31 feature combinations.
The most frequent feature combination is _
(10002 tokens).
Examples: 1, 10, 20, 2, 15, 5, 3, 30, 4, 100
Relations
NUM
nodes are attached to their parents using 10 different relations: nummod (11751; 73% instances), nummod:gov (3536; 22% instances), nsubj (346; 2% instances), root (177; 1% instances), conj (138; 1% instances), nsubjpass (42; 0% instances), dep (10; 0% instances), advmod (9; 0% instances), advcl (4; 0% instances), acl (1; 0% instances)
Parents of NUM
nodes belong to 13 different parts of speech: NOUN (11519; 72% instances), VERB (1480; 9% instances), NUM (1153; 7% instances), SYM (833; 5% instances), PROPN (564; 4% instances), ADJ (211; 1% instances), ROOT (177; 1% instances), ADV (48; 0% instances), PRON (14; 0% instances), SCONJ (7; 0% instances), X (4; 0% instances), CONJ (2; 0% instances), PART (2; 0% instances)
10887 (68%) NUM
nodes are leaves.
2873 (18%) NUM
nodes have one child.
1692 (11%) NUM
nodes have two children.
562 (4%) NUM
nodes have three or more children.
The highest child degree of a NUM
node is 18.
Children of NUM
nodes are attached using 22 different relations: punct (2443; 30% instances), nmod (1820; 22% instances), advmod (987; 12% instances), nummod (937; 11% instances), case (786; 10% instances), amod (334; 4% instances), conj (197; 2% instances), cc (162; 2% instances), nsubj (132; 2% instances), parataxis (131; 2% instances), nummod:gov (128; 2% instances), neg (49; 1% instances), appos (47; 1% instances), acl:relcl (19; 0% instances), mark (15; 0% instances), foreign (7; 0% instances), advcl (5; 0% instances), acl (2; 0% instances), compound (2; 0% instances), cop (2; 0% instances), mwe (2; 0% instances), iobj (1; 0% instances)
Children of NUM
nodes belong to 14 different parts of speech: PUNCT (2443; 30% instances), NOUN (1878; 23% instances), NUM (1153; 14% instances), ADP (785; 10% instances), ADV (611; 7% instances), PART (454; 6% instances), ADJ (411; 5% instances), CONJ (158; 2% instances), VERB (121; 1% instances), PRON (90; 1% instances), PROPN (64; 1% instances), SCONJ (22; 0% instances), SYM (17; 0% instances), X (1; 0% instances)
NUM in other languages: [bg] [cs] [de] [el] [en] [es] [eu] [fa] [fi] [fr] [ga] [he] [hu] [it] [ja] [ko] [sv] [u]