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

home en/pos issue tracker

PUNCT: punctuation

The English PUNCT covers PTB tags:

• ``
• ’’
• -LRB-
• -RRB-
• ,
• .
• :
• HYPH
• Some uses of NFP (for lines of hyphens, asterisks or tildes)

---

Treebank Statistics (UD_English)

There are 125 PUNCT lemmas (1%), 125 PUNCT types (1%) and 29064 PUNCT tokens (12%). Out of 17 observed tags, the rank of PUNCT is: 9 in number of lemmas, 9 in number of types and 3 in number of tokens.

The 10 most frequent PUNCT lemmas: ., ,, “, -, ?, ), (, :, !, …

The 10 most frequent PUNCT types: ., ,, “, -, ?, ), (, :, !, …

The 10 most frequent ambiguous lemmas: ” (PUNCT 1619, NOUN 6), - (PUNCT 1404, SYM 116, X 11), ) (PUNCT 1044, X 6), ( (PUNCT 1007, X 6), : (PUNCT 768, SYM 1), … (PUNCT 319, SYM 5), / (PUNCT 241, SYM 32, X 2), ’ (PUNCT 240, PART 38, NOUN 7, VERB 1), «/em> (PUNCT 68, SYM 1), > (PUNCT 68, SYM 3)

The 10 most frequent ambiguous types: ” (PUNCT 1619, NOUN 6), - (PUNCT 1404, SYM 116, X 11), ) (PUNCT 1044, X 6), ( (PUNCT 1007, X 6), : (PUNCT 768, SYM 1), … (PUNCT 319, SYM 5), / (PUNCT 241, SYM 32, X 2), ’ (PUNCT 240, PART 38, NOUN 7), «/em> (PUNCT 68, SYM 1), > (PUNCT 68, SYM 3)

• ”
  • PUNCT 1619: To pander to the mythical ” Arab street ” , of course .
  • NOUN 6: Not sure if I am going to buy 17 ” or 16 ” wheels for the winter .
• -
  • PUNCT 1404: TEHRAN ( AFP ) -
  • SYM 116: Intercept : - 0.3931 ( 0.0076 )
  • X 11: « Compaq.com - notebook.url »
• )
  • PUNCT 1044: TEHRAN ( AFP ) -
  • X 6: - ENRON-CPS ( GISB rev1 ) .doc
• (
  • PUNCT 1007: TEHRAN ( AFP ) -
  • X 6: - ENRON-CPS ( GISB rev1 ) .doc
• :
  • PUNCT 768: From the AP comes this story :
  • SYM 1: Now post your amended article to at least 200 news groups . :
• …
  • PUNCT 319: … Now comes the fun part ….
  • SYM 5: …
• /
  • PUNCT 241: Rudwell Johnson / ENRON@enronXgate
  • SYM 32: No way to read / relax .
  • X 2: Hop onto a Hop On / Hop Off bus .
• ’
  • PUNCT 240: He did comment on what he meant by the phrase . ’
  • PART 38: The latest spot for a real Hackney ‘s is Printers ’ Row :
  • NOUN 7: If you Do n’t want to grow out of it quickly get a 14 ’ 2 ..
• «/em>
  • PUNCT 68: ” Les Spahnn “ «/b> spahnn@hnks.com ></em>
  • SYM 1: IF you have NOT begun the Application process’s NOW «/b> you are about 2 to 3 years away from being approved by Immigration Canada .</em>
• >
  • PUNCT 68: ” Les Spahnn “ < spahnn@hnks.com >
  • SYM 3: > 100,000 Newsgroups

Morphology

The form / lemma ratio of PUNCT is 1.000000 (the average of all parts of speech is 1.173735).

The 1st highest number of forms (1) was observed with the lemma “!”: !.

The 2nd highest number of forms (1) was observed with the lemma “!!”: !!.

The 3rd highest number of forms (1) was observed with the lemma “!!!”: !!!.

PUNCT does not occur with any features.

Relations

PUNCT nodes are attached to their parents using 11 different relations: en-dep/punct (28843; 99% instances), en-dep/cc (119; 0% instances), en-dep/root (86; 0% instances), en-dep/case (6; 0% instances), en-dep/discourse (3; 0% instances), en-dep/dep (2; 0% instances), en-dep/conj (1; 0% instances), en-dep/det:predet (1; 0% instances), en-dep/list (1; 0% instances), en-dep/mwe (1; 0% instances), en-dep/nmod (1; 0% instances)

Parents of PUNCT nodes belong to 18 different parts of speech: VERB (14791; 51% instances), NOUN (6916; 24% instances), ADJ (2892; 10% instances), PROPN (2643; 9% instances), NUM (390; 1% instances), ADV (376; 1% instances), X (331; 1% instances), PRON (234; 1% instances), INTJ (145; 0% instances), ROOT (86; 0% instances), SYM (63; 0% instances), PART (56; 0% instances), DET (54; 0% instances), CONJ (38; 0% instances), AUX (23; 0% instances), ADP (13; 0% instances), PUNCT (10; 0% instances), SCONJ (3; 0% instances)

29015 (100%) PUNCT nodes are leaves.

44 (0%) PUNCT nodes have one child.

4 (0%) PUNCT nodes have two children.

1 (0%) PUNCT nodes have three or more children.

The highest child degree of a PUNCT node is 3.

Children of PUNCT nodes are attached using 9 different relations: en-dep/nummod (32; 58% instances), en-dep/punct (8; 15% instances), en-dep/conj (4; 7% instances), en-dep/dep (4; 7% instances), en-dep/goeswith (3; 5% instances), en-dep/advmod (1; 2% instances), en-dep/appos (1; 2% instances), en-dep/cc (1; 2% instances), en-dep/remnant (1; 2% instances)

Children of PUNCT nodes belong to 7 different parts of speech: X (34; 62% instances), PUNCT (10; 18% instances), NOUN (6; 11% instances), CONJ (2; 4% instances), ADV (1; 2% instances), NUM (1; 2% instances), PROPN (1; 2% instances)

---

Treebank Statistics (UD_English-ESL)

There are 1 PUNCT lemmas (6%), 1 PUNCT types (6%) and 9580 PUNCT tokens (10%). Out of 17 observed tags, the rank of PUNCT is: 13 in number of lemmas, 13 in number of types and 5 in number of tokens.

The 10 most frequent PUNCT lemmas: _

The 10 most frequent PUNCT types: _

The 10 most frequent ambiguous lemmas: _ (NOUN 15635, VERB 15080, PRON 10618, DET 10057, PUNCT 9580, ADP 8546, ADJ 5857, ADV 5704, AUX 4533, PART 3531, CONJ 3198, SCONJ 2516, PROPN 1795, NUM 844, INTJ 80, X 68, SYM 39)

The 10 most frequent ambiguous types: _ (NOUN 15635, VERB 15080, PRON 10618, DET 10057, PUNCT 9580, ADP 8546, ADJ 5857, ADV 5704, AUX 4533, PART 3531, CONJ 3198, SCONJ 2516, PROPN 1795, NUM 844, INTJ 80, X 68, SYM 39)

• _
  • NOUN 15635: _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
  • VERB 15080: _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
  • PRON 10618: _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
  • DET 10057: _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
  • PUNCT 9580: _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
  • ADP 8546: _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
  • ADJ 5857: _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
  • ADV 5704: _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
  • AUX 4533: _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
  • PART 3531: _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
  • CONJ 3198: _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
  • SCONJ 2516: _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
  • PROPN 1795: _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
  • NUM 844: _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
  • INTJ 80: _ _ _ _ _ _ _ _ _ _ _ _
  • X 68: _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
  • SYM 39: _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

Morphology

The form / lemma ratio of PUNCT is 1.000000 (the average of all parts of speech is 1.000000).

The 1st highest number of forms (1) was observed with the lemma “_”: _.

PUNCT does not occur with any features.

Relations

PUNCT nodes are attached to their parents using 1 different relations: en-dep/punct (9580; 100% instances)

Parents of PUNCT nodes belong to 15 different parts of speech: VERB (6244; 65% instances), NOUN (1590; 17% instances), ADJ (1283; 13% instances), PROPN (233; 2% instances), ADV (76; 1% instances), NUM (52; 1% instances), PRON (46; 0% instances), DET (11; 0% instances), INTJ (11; 0% instances), SYM (10; 0% instances), X (10; 0% instances), AUX (9; 0% instances), ADP (2; 0% instances), CONJ (2; 0% instances), PART (1; 0% instances)

9579 (100%) PUNCT nodes are leaves.

1 (0%) PUNCT nodes have one child.

The highest child degree of a PUNCT node is 1.

Children of PUNCT nodes are attached using 1 different relations: en-dep/nummod (1; 100% instances)

Children of PUNCT nodes belong to 1 different parts of speech: NUM (1; 100% instances)

---

Treebank Statistics (UD_English-LinES)

There are 1 PUNCT lemmas (6%), 15 PUNCT types (0%) and 10025 PUNCT tokens (12%). Out of 17 observed tags, the rank of PUNCT is: 13 in number of lemmas, 14 in number of types and 3 in number of tokens.

The 10 most frequent PUNCT lemmas: _

The 10 most frequent PUNCT types: ,, ., –, “, ?, ;, ), (, :, …

The 10 most frequent ambiguous lemmas: _ (NOUN 14939, VERB 11076, PUNCT 10025, ADP 8281, DET 7865, PRON 7793, ADJ 5305, ADV 4610, AUX 3168, PROPN 2792, CONJ 2535, PART 2131, SCONJ 1512, NUM 581, INTJ 159, X 43, SYM 6)

The 10 most frequent ambiguous types:

Morphology

The form / lemma ratio of PUNCT is 15.000000 (the average of all parts of speech is 597.705882).

The 1st highest number of forms (15) was observed with the lemma “_”: !, “, ‘, (, ), ,, ., .., …, :, ;, ?, [, ], –.

PUNCT does not occur with any features.

Relations

PUNCT nodes are attached to their parents using 1 different relations: en-dep/punct (10025; 100% instances)

Parents of PUNCT nodes belong to 17 different parts of speech: VERB (6389; 64% instances), NOUN (1629; 16% instances), ADJ (667; 7% instances), ADV (376; 4% instances), AUX (221; 2% instances), PROPN (212; 2% instances), PRON (206; 2% instances), INTJ (95; 1% instances), ADP (62; 1% instances), NUM (59; 1% instances), X (39; 0% instances), CONJ (31; 0% instances), PUNCT (15; 0% instances), SCONJ (10; 0% instances), DET (6; 0% instances), PART (4; 0% instances), SYM (4; 0% instances)

9984 (100%) PUNCT nodes are leaves.

35 (0%) PUNCT nodes have one child.

5 (0%) PUNCT nodes have two children.

1 (0%) PUNCT nodes have three or more children.

The highest child degree of a PUNCT node is 3.

Children of PUNCT nodes are attached using 17 different relations: en-dep/punct (15; 31% instances), en-dep/det (4; 8% instances), en-dep/nmod (4; 8% instances), en-dep/nsubj (4; 8% instances), en-dep/advmod (3; 6% instances), en-dep/amod (3; 6% instances), en-dep/mark (3; 6% instances), en-dep/compound (2; 4% instances), en-dep/discourse (2; 4% instances), en-dep/acl (1; 2% instances), en-dep/aux (1; 2% instances), en-dep/compound:prt (1; 2% instances), en-dep/conj (1; 2% instances), en-dep/dobj (1; 2% instances), en-dep/neg (1; 2% instances), en-dep/vocative (1; 2% instances), en-dep/xcomp (1; 2% instances)

Children of PUNCT nodes belong to 13 different parts of speech: PUNCT (15; 31% instances), NOUN (5; 10% instances), PRON (5; 10% instances), ADV (4; 8% instances), DET (4; 8% instances), PROPN (4; 8% instances), ADJ (2; 4% instances), INTJ (2; 4% instances), PART (2; 4% instances), SCONJ (2; 4% instances), ADP (1; 2% instances), AUX (1; 2% instances), VERB (1; 2% instances)

---

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

© 2014 Universal Dependencies contributors. Site powered by Annodoc and brat

.