Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

In consistency in the annotation of date expression #22

Open
Oneplus opened this issue May 27, 2017 · 4 comments
Open

In consistency in the annotation of date expression #22

Oneplus opened this issue May 27, 2017 · 4 comments
Labels
mischievous nominal See https://arxiv.org/abs/2108.12928

Comments

@Oneplus
Copy link

Oneplus commented May 27, 2017

# sent_id = weblog-blogspot.com_rigorousintuition_20050518101500_ENG_20050518_101500-0046
# text = Third ranking CIA officer, Briggs signed a declaration on February 3rd, 1983, that on November 8th, 1982, he had authorized a search of CIA records "for any material that in any way pertains to Edwin P. Wilson, or the various allegations concerning his activities after February 28th, 1971, when he retired from the CIA."
1	Third	third	ADV	RB	_	2	advmod	_	_
2	ranking	rank	VERB	VBG	VerbForm=Ger	4	amod	_	_
3	CIA	CIA	PROPN	NNP	Number=Sing	4	compound	_	_
4	officer	officer	NOUN	NN	Number=Sing	6	compound	_	SpaceAfter=No
5	,	,	PUNCT	,	_	6	punct	_	_
6	Briggs	Briggs	PROPN	NNP	Number=Sing	7	nsubj	_	_
7	signed	sign	VERB	VBD	Mood=Ind|Tense=Past|VerbForm=Fin	0	root	_	_
8	a	a	DET	DT	Definite=Ind|PronType=Art	9	det	_	_
9	declaration	declaration	NOUN	NN	Number=Sing	7	obj	_	_
10	on	on	ADP	IN	_	11	case	_	_
11	February	February	PROPN	NNP	Number=Sing	7	obl	_	_
*12	3rd	3rd	NOUN	NN	Number=Sing	11	nummod	_	SpaceAfter=No
13	,	,	PUNCT	,	_	12	punct	_	_
14	1983	1983	NUM	CD	NumType=Card	12	nummod	_	SpaceAfter=No
15	,	,	PUNCT	,	_	7	punct	_	_
16	that	that	SCONJ	IN	_	25	mark	_	_
17	on	on	ADP	IN	_	18	case	_	_
18	November	November	PROPN	NNP	Number=Sing	25	obl	_	_
*19	8th	8th	NOUN	NN	Number=Sing	18	nummod	_	SpaceAfter=No
20	,	,	PUNCT	,	_	19	punct	_	_
21	1982	1982	NUM	CD	NumType=Card	19	nummod	_	SpaceAfter=No
22	,	,	PUNCT	,	_	25	punct	_	_
...
44	,	,	PUNCT	,	_	48	punct	_	_
45	or	or	CCONJ	CC	_	48	cc	_	_
46	the	the	DET	DT	Definite=Def|PronType=Art	48	det	_	_
47	various	various	ADJ	JJ	Degree=Pos	48	amod	_	_
48	allegations	allegation	NOUN	NNS	Number=Plur	41	conj	_	_
49	concerning	concern	VERB	VBG	VerbForm=Ger	48	acl	_	_
50	his	he	PRON	PRP$	Gender=Masc|Number=Sing|Person=3|Poss=Yes|PronType=Prs	51	nmod:poss	_	_
51	activities	activity	NOUN	NNS	Number=Plur	49	obj	_	_
52	after	after	ADP	IN	_	54	case	_	_
53	February	February	PROPN	NNP	Number=Sing	54	compound	_	_
*54	28th	28th	NOUN	NN	Number=Sing	51	nmod	_	SpaceAfter=No
55	,	,	PUNCT	,	_	54	punct	_	_
56	1971	1971	NUM	CD	NumType=Card	54	nmod:tmod	_	SpaceAfter=No
...

I guess 28th should be nummod?

@nschneid
Copy link
Contributor

See UniversalDependencies/docs#455

@dan-zeman
Copy link
Member

We should discuss whether listing dates as an example of flat was a mistake or we really want to change the previous approach. We should also create a documentation page dedicated to dates. As a matter of fact, there has been a long discussion in UniversalDependencies/docs#113 and especially in UniversalDependencies/docs#210.

In this particular example from UD English, I am not sure whether nummod is better than amod, given that we have ordinal numbers here. But no matter what the preferred approach is, it should be applied consistently to all three occurrences in the sentence, hence there obviously is a bug. (Note that it is not just a nummod --> nmod typo. The dependency direction is reversed, there is compound(28th, February), 28th is analyzed as the head, and the whole date is attached as nmod to activities. This is wrong; there should be nmod(activities, February), and 28th should be attached as a child of February.

@nschneid nschneid added the mischievous nominal See https://arxiv.org/abs/2108.12928 label Jan 29, 2022
@nschneid
Copy link
Contributor

Just noting that one interesting date construction is "Sat. the 26th". appos in EWT, nmod:tmod in GUM

@amir-zeldes
Copy link
Contributor

I think appos is right, will fix the GUM error

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Labels
mischievous nominal See https://arxiv.org/abs/2108.12928
Projects
None yet
Development

No branches or pull requests

4 participants