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 (чотири), digits (4) or Roman numerals (IV).
Ukrainian grammar distinguishes several subclasses of pronominal numerals (quantifiers):
interrogative and relative (скільки “how many”);
demonstrative (стільки “this many”);
indefinite (кілька, багато, мало “several, many, few”).
These words behave similarly to (most) cardinal numbers, e.g. they require that the counted noun phrase be in genitive and have different forms depending on the case (with the exception of мало “little, few”. They are not similar to adjectives (unlike their English counterparts). However, in accord with the UD standard, they should be tagged DET, not NUM.
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, чотири, п’ять, сімдесят “one, two, three, four, five, seventy”
- половина, третина, четвертина (чверть) “one-half, one third, quarter”: denominators of fractions constitute a separate class of cardinal numerals. (They are not considered numerals in the Ukrainian grammar. They are tagged NOUN.)
- четверо, п’ятеро “four, five” (These are special forms, so-called generic numerals.)
- одні, двоє, троє “one set of, two sets of, three sets of”
Counterexamples
- перший, другий, третій “first, second, third”: adjectival ordinal numerals. They are tagged ADJ, and the NumType feature reveals their semantic relation to numbers.
- вперше, вдруге, втретє “for the first time, for the second time, for the third time”: adverbial ordinal numerals. They are tagged ADV, and the NumType feature reveals their semantic relation to numbers.
- двічі, тричі “twice, three times”: multiplicative numerals. They are tagged ADV, and the NumType feature reveals their semantic relation to numbers.
- двійня, дует, трійня, тріо, квартет “pair, triplet, foursome”: n-tuples (n-tice) are not considered numerals in the Ukrainian grammar. They are tagged NOUN.
- одиниця, двійка, трійка, четвірка, п’ятірка “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.
- тисяча, мільйон, мільярд, трильйон “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 тисячі людей взяли участь в демонстації (“thousands of people demonstrated in the streets”), тисячі is a noun. In numeric expressions, e.g. 132 тисячі доларів (“132 thousand dollars”), it is a cardinal numeral.
Treebank Statistics (UD_Ukrainian)
There are 19 NUM lemmas (3%), 23 NUM types (3%) and 37 NUM tokens (2%).
Out of 16 observed tags, the rank of NUM is: 9 in number of lemmas, 8 in number of types and 11 in number of tokens.
The 10 most frequent NUM lemmas: 50, мільйон, 3, 5, багато, 200, 8, два, п’ять, сім
The 10 most frequent NUM types: 50, мільйонів, 5, 200, 3, 8, П’ять, багатьма, дві, 1
The 10 most frequent ambiguous lemmas: мільйон (NUM 4, NOUN 1), багато (NUM 3, ADV 1)
The 10 most frequent ambiguous types:
Morphology
The form / lemma ratio of NUM is 1.210526 (the average of all parts of speech is 1.172859).
The 1st highest number of forms (2) was observed with the lemma “3”: 3, 3,2.
The 2nd highest number of forms (2) was observed with the lemma “багато”: багатьма, багатьох.
The 3rd highest number of forms (2) was observed with the lemma “сім”: Сім, Сімом.
NUM occurs with 5 features: NumType (37; 100% instances), Case (15; 41% instances), Number (11; 30% instances), Gender (4; 11% instances), PronType (3; 8% instances)
NUM occurs with 10 feature-value pairs: Case=Acc, Case=Dat, Case=Gen, Case=Ins, Case=Nom, Gender=Fem, Gender=Masc, NumType=Card, Number=Plur, PronType=Ind
NUM occurs with 10 feature combinations.
The most frequent feature combination is NumType=Card (22 tokens).
Examples: 50, мільйонів, 5, 200, 3, 8, 1, 14, 2016, 24
Relations
NUM nodes are attached to their parents using 6 different relations: uk-dep/nummod:gov (16; 43% instances), nummod (10; 27% instances), compound (6; 16% instances), iobj (2; 5% instances), root (2; 5% instances), det (1; 3% instances)
Parents of NUM nodes belong to 7 different parts of speech: NOUN (22; 59% instances), NUM (6; 16% instances), ADJ (3; 8% instances), PRON (2; 5% instances), ROOT (2; 5% instances), DET (1; 3% instances), VERB (1; 3% instances)
25 (68%) NUM nodes are leaves.
5 (14%) NUM nodes have one child.
5 (14%) NUM nodes have two children.
2 (5%) NUM nodes have three or more children.
The highest child degree of a NUM node is 6.
Children of NUM nodes are attached using 7 different relations: compound (8; 33% instances), punct (5; 21% instances), advmod (4; 17% instances), amod (3; 13% instances), nmod (2; 8% instances), case (1; 4% instances), nsubj (1; 4% instances)
Children of NUM nodes belong to 7 different parts of speech: NUM (6; 25% instances), PUNCT (5; 21% instances), ADV (4; 17% instances), NOUN (4; 17% instances), ADJ (3; 13% instances), ADP (1; 4% instances), SYM (1; 4% instances)
NUM in other languages: [bg] [cs] [de] [el] [en] [es] [eu] [fa] [fi] [fr] [ga] [he] [hu] [it] [ja] [ko] [sv] [u]