NUM

This is part of archived UD v1 documentation. See http://universaldependencies.org/ for the current version.

home it/pos issue tracker

`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.

Cardinal numerals are covered by NUM whether they are used as determiners or not (as in Windows Seven) and whether they are expressed as words (four), or digits (4).

There are numerals belonging to other parts of speech in the universal tagging scheme, based mainly on syntactic criteria: e.g. ordinal numerals are tagged as adjectives or pronouns (first, second, third, …). In Italian there are no adverbial numerals like [cs] poprvé “for the first time” nor multiplicative numerals such as [en] (once, twice).

Note that, since Roman numerals (IV) are only used as ordinal numerals in Italian (Enciclopedia Treccani), are tagged ADJ for coherence with other ordinal numerals. In this case we specify NumType=Ord as morphological feature. A similar treatment is reserved to a cardinal with a suffix such as a, o, esimo, esima, or the symbol ° (similar to the English use of the suffix “th”).

Other words functioning as determiners (including quantifiers such as many and few) are tagged DET.

Examples

Tagged as NUM:

0, 1, 2, 2014, 1000000, 3,14159265359
zero, uno, due, settantasette

Tagged as ADJ:

XX secolo
48esimo
primo, secondo, ultima
20°

Tagged as PRON:

il terzo ad arrivare “the third to arrive”

References

Enciclopedia Treccani

Treebank Statistics (UD_Italian)

There are 1117 NUM lemmas (6%), 1128 NUM types (4%) and 5041 NUM tokens (2%). Out of 17 observed tags, the rank of NUM is: 5 in number of lemmas, 5 in number of types and 12 in number of tokens.

The 10 most frequent NUM lemmas: due, 1, 2, tre, 3, cinque, 4, mila, quattro, 5

The 10 most frequent NUM types: due, 1, 2, tre, 3, cinque, 4, mila, quattro, 5

The 10 most frequent ambiguous lemmas: 1 (NUM 178, PROPN 1), 4 (NUM 69, PROPN 1), 5 (NUM 58, PROPN 1), uno (DET 4549, PRON 215, NUM 42, ADJ 2, PROPN 1), 7 (NUM 34, PROPN 1), 1994 (NUM 31, NOUN 1), dieci (NUM 27, PROPN 2), 22 (NUM 16, PROPN 1), mille (NUM 11, PROPN 1), zero (NUM 6, PROPN 2, NOUN 1)

The 10 most frequent ambiguous types: 1 (NUM 178, PROPN 1), 4 (NUM 69, PROPN 1), 5 (NUM 58, PROPN 1), sei (NUM 35, VERB 5, AUX 4), 7 (NUM 34, PROPN 1), 1994 (NUM 31, NOUN 1), dieci (NUM 26, PROPN 2), venti (NUM 20, NOUN 4), un (DET 2359, NUM 16, PRON 1), 22 (NUM 16, PROPN 1)

1
- NUM 178: 1 ) tre metri per gli alberi di alto fusto .
- PROPN 1: La top ten prosegue con 10 chilometri di coda , sempre in il Veronese , fra Peschiera e Sommacampagna e altrettanti in l’ innesto tra l’ A / 1 e l’ A / 4 in il Bolognese .
4
- NUM 69: Emilio Fede , direttore Tg 4 , quasi s’ arrabbia :
- PROPN 1: La top ten prosegue con 10 chilometri di coda , sempre in il Veronese , fra Peschiera e Sommacampagna e altrettanti in l’ innesto tra l’ A / 1 e l’ A / 4 in il Bolognese .
5
- NUM 58: Il carabiniere rischia da 5 a 15 anni di condanna .
- PROPN 1: Il leader di il Movimento 5 Stelle , Beppe Grillo , non parla a i giornalisti .
sei
- NUM 35: Il libro era di cento pagine e costava sei lire .
- VERB 5: ” Tu che sei così sensibile , una compagna … .
- AUX 4: Ti sei sbagliato , riconosci lo “ .
7
- NUM 34: Cm 60x60 e uno spessore di mm 7 .
- PROPN 1: Quanto era grande il “ Grande Boccone “ che si vendeva in i negozi 7 - Eleven in il 1980 ?
1994
- NUM 31: Chi era il presidente di gli Stati Uniti in il 1994 ?
- NOUN 1: Quando si sono tenute le elezioni europee di il 1994 ?
dieci
- NUM 26: a i 6.405 “ dieci “ 116.000 .
- PROPN 2: Luciano Ligabue ha fatto anche due film : Radiofreccia di il 1998 e Da zero a dieci di il 2002 .
venti
- NUM 20: A volte i dettagli complessi su i luoghi si estendono da le quindici a le venti pagine .
- NOUN 4: A inguaiare Minetto , che si è sempre proclamato innocente , sono state soprattutto le amicizie con parecchi neofascisti , come Marcello Soffiati , coinvolto in l’ inchiesta su la “ Rosa di i venti “ .
un
- DET 2359: gusto un po’ facile e appeal sicuro .
- NUM 16: A Baldas il gravissimo reato d’ opinione costerà un anno di purgatorio :
- PRON 1: Secondo James Shapiro studioso di Shakespeare , in il Giulio Cesare “ i vari elementi di politica , personaggio , introspezione , eventi contemporanei , persino le stesse riflessioni di Shakespeare su la scrittura , iniziano a sovrappor si l’ un l’ altro .
22
- NUM 16: Come si sono leggermente ferite 22 persone in l’ Empire State Building ?
- PROPN 1: Code - Il record d’ incolonnamento spetta a la A / 22 , l’ autostrada di il Brennero :

Morphology

The form / lemma ratio of NUM is 1.009848 (the average of all parts of speech is 1.488836).

The 1st highest number of forms (3) was observed with the lemma “cinquanta”: cinquant’, cinquanta, cinquant’.

The 2nd highest number of forms (3) was observed with the lemma “uno”: un, una, uno.

The 3rd highest number of forms (3) was observed with the lemma “venti”: vent’, venti, vent’.

NUM occurs with 1 features: it-feat/NumType (5041; 100% instances)

NUM occurs with 3 feature-value pairs: NumType=Card, NumType=Gen, NumType=Range

NUM occurs with 3 feature combinations. The most frequent feature combination is NumType=Card (5016 tokens). Examples: due, 1, 2, tre, 3, cinque, 4, mila, quattro, 5

Relations

NUM nodes are attached to their parents using 15 different relations: it-dep/nummod (3854; 76% instances), it-dep/nmod (563; 11% instances), it-dep/conj (186; 4% instances), it-dep/compound (153; 3% instances), it-dep/root (90; 2% instances), it-dep/dobj (57; 1% instances), it-dep/name (56; 1% instances), it-dep/nsubj (45; 1% instances), it-dep/mwe (17; 0% instances), it-dep/nsubjpass (9; 0% instances), it-dep/acl:relcl (4; 0% instances), it-dep/xcomp (3; 0% instances), it-dep/advcl (2; 0% instances), it-dep/ccomp (1; 0% instances), it-dep/foreign (1; 0% instances)

Parents of NUM nodes belong to 13 different parts of speech: NOUN (3051; 61% instances), VERB (1098; 22% instances), NUM (364; 7% instances), PROPN (249; 5% instances), ROOT (90; 2% instances), SYM (84; 2% instances), ADJ (54; 1% instances), PRON (29; 1% instances), ADV (15; 0% instances), PUNCT (3; 0% instances), X (2; 0% instances), AUX (1; 0% instances), DET (1; 0% instances)

2673 (53%) NUM nodes are leaves.

693 (14%) NUM nodes have one child.

1225 (24%) NUM nodes have two children.

450 (9%) NUM nodes have three or more children.

The highest child degree of a NUM node is 22.

Children of NUM nodes are attached using 24 different relations: it-dep/punct (1696; 35% instances), it-dep/det (1060; 22% instances), it-dep/case (945; 19% instances), it-dep/nmod (384; 8% instances), it-dep/conj (232; 5% instances), it-dep/compound (154; 3% instances), it-dep/cc (139; 3% instances), it-dep/advmod (136; 3% instances), it-dep/nummod (36; 1% instances), it-dep/mwe (24; 0% instances), it-dep/cop (23; 0% instances), it-dep/amod (21; 0% instances), it-dep/nsubj (21; 0% instances), it-dep/advcl (11; 0% instances), it-dep/appos (6; 0% instances), it-dep/acl:relcl (5; 0% instances), it-dep/dobj (3; 0% instances), it-dep/parataxis (3; 0% instances), it-dep/mark (2; 0% instances), it-dep/aux (1; 0% instances), it-dep/ccomp (1; 0% instances), it-dep/det:poss (1; 0% instances), it-dep/det:predet (1; 0% instances), it-dep/neg (1; 0% instances)

Children of NUM nodes belong to 15 different parts of speech: PUNCT (1696; 35% instances), DET (1065; 22% instances), ADP (918; 19% instances), NOUN (405; 8% instances), NUM (364; 7% instances), ADV (165; 3% instances), CONJ (139; 3% instances), ADJ (65; 1% instances), VERB (45; 1% instances), PROPN (18; 0% instances), PRON (13; 0% instances), X (8; 0% instances), SCONJ (2; 0% instances), SYM (2; 0% instances), AUX (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]