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

Questions about INTJ #27

Open
muchang opened this issue Jul 20, 2021 · 6 comments
Open

Questions about INTJ #27

muchang opened this issue Jul 20, 2021 · 6 comments

Comments

@muchang
Copy link

muchang commented Jul 20, 2021

In UD definition, INTJ is defined as follows:

Note that words primarily belonging to another part of speech retains their original category when used in exclamations. For example, God is a NOUN even in exclamatory uses.

As a special case of interjections, we recognize feedback particles such as yes, no, uhuh, etc.

However, in GUM, there are some words that have their original category and are not feedback particles, while they are annotated as INTJ.

For example, in the following sentences, the "god"/"gosh"s are annotated as INTJ, which is conflicted with the definition.

# sent_id = GUM_vlog_lipstick-61
# s_type = decl
# speaker = AlyssaMarie
# text = Oh my god I had this back in college is when I ordered this, which is probably around twenty sixteen.
1	Oh	oh	INTJ	UH	_	13	discourse	13:discourse	Discourse=evaluation:79->80
2	my	my	PRON	PRP$	Number=Sing|Person=1|Poss=Yes|PronType=Prs	3	nmod:poss	3:nmod:poss	_
3	god	god	INTJ	UH	_	13	discourse	13:discourse	_
4	I	I	PRON	PRP	Case=Nom|Number=Sing|Person=1|PronType=Prs	5	nsubj	5:nsubj	Discourse=elaboration:80->78|Entity=(person-21)
5	had	have	VERB	VBD	Mood=Ind|Number=Sing|Person=1|Tense=Past|VerbForm=Fin	13	csubj	13:csubj	_
6	this	this	PRON	DT	Number=Sing|PronType=Dem	5	obj	5:obj	Entity=(object-47)
7	back	back	ADV	RB	Degree=Pos	9	advmod	9:advmod	_
8	in	in	ADP	IN	_	9	case	9:case	_
9	college	college	NOUN	NN	Number=Sing	5	obl	5:obl:in	Entity=(organization-48)
10	is	be	AUX	VBZ	Mood=Ind|Number=Sing|Person=3|Tense=Pres|VerbForm=Fin	13	cop	13:cop	Discourse=elaboration:81->80
11	when	when	SCONJ	WRB	PronType=Int	13	mark	13:mark	_
12	I	I	PRON	PRP	Case=Nom|Number=Sing|Person=1|PronType=Prs	13	nsubj	13:nsubj	Entity=(person-21)
13	ordered	order	VERB	VBD	Mood=Ind|Number=Sing|Person=1|Tense=Past|VerbForm=Fin	0	root	0:root|20:mark	_
14	this	this	PRON	DT	Number=Sing|PronType=Dem	13	obj	13:obj	Entity=(object-47)|SpaceAfter=No
15	,	,	PUNCT	,	_	20	punct	20:punct	_
16	which	which	SCONJ	WDT	PronType=Rel	20	mark	13:ref	Discourse=elaboration:82->81
17	is	be	AUX	VBZ	Mood=Ind|Number=Sing|Person=3|Tense=Pres|VerbForm=Fin	20	cop	20:cop	_
18	probably	probably	ADV	RB	Degree=Pos	20	advmod	20:advmod	_
19	around	around	ADP	IN	_	20	case	20:case	_
20	twenty	twenty	NUM	CD	NumForm=Word|NumType=Card	13	acl:relcl	13:acl:relcl	Entity=(time-49
21	sixteen	sixteen	NUM	CD	NumForm=Word|NumType=Card	20	goeswith	20:goeswith	Entity=time-49)|SpaceAfter=No
22	.	.	PUNCT	.	_	13	punct	13:punct	_
# sent_id = GUM_vlog_lipstick-94
# s_type = decl
# speaker = AlyssaMarie
# text = My gosh it's hard to think back.
1	My	my	PRON	PRP$	Number=Sing|Person=1|Poss=Yes|PronType=Prs	2	nmod:poss	2:nmod:poss	Discourse=evaluation:130->129
2	gosh	gosh	INTJ	UH	_	5	discourse	5:discourse	_
3-4	it's	_	_	_	_	_	_	_	_
3	it	it	PRON	PRP	Case=Nom|Gender=Neut|Number=Sing|Person=3|PronType=Prs	5	expl	5:expl	Entity=(abstract-68)
4	's	be	AUX	VBZ	Mood=Ind|Number=Sing|Person=3|Tense=Pres|VerbForm=Fin	5	cop	5:cop	_
5	hard	hard	ADJ	JJ	Degree=Pos	0	root	0:root	_
6	to	to	PART	TO	_	7	mark	7:mark	Entity=(abstract-68
7	think	think	VERB	VB	VerbForm=Inf	5	csubj	5:csubj	_
8	back	back	ADV	RB	Degree=Pos	7	advmod	7:advmod	Entity=abstract-68)|SpaceAfter=No
9	.	.	PUNCT	.	_	5	punct	5:punct	_

In the following sentences, the words tagged as INTJ are not feedback particles and have the original category, although their annotations are consistent in GUM.

 # sent_id = GUM_vlog_lipstick-63
# s_type = decl
# speaker = AlyssaMarie
# text = Everything has like rubbed off of it.
1	Everything	everything	NOUN	NN	Number=Sing	4	nsubj	4:nsubj	Discourse=justify:84->85|Entity=(abstract-51)
2	has	have	AUX	VBZ	Mood=Ind|Number=Sing|Person=3|Tense=Pres|VerbForm=Fin	4	aux	4:aux	_
3	like	like	INTJ	UH	_	4	discourse	4:discourse	_
4	rubbed	rub	VERB	VBN	Tense=Past|VerbForm=Part	0	root	0:root	_
5	off	off	ADP	IN	_	7	case	7:case	_
6	of	of	ADP	IN	_	5	fixed	5:fixed	_
7	it	it	PRON	PRP	Case=Acc|Gender=Neut|Number=Sing|Person=3|PronType=Prs	4	obl	4:obl:off_of	Entity=(object-47)|SpaceAfter=No
8	.	.	PUNCT	.	_	4	punct	4:punct	_
# sent_id = GUM_voyage_cuba-13
# s_type = imp
# text = Please note that if departing from Europe (this may apply to other countries), you will require to have the visa before boarding the plane.
1	Please	please	INTJ	UH	_	2	discourse	2:discourse	Discourse=condition:33->35
2	note	note	VERB	VB	Mood=Imp|Person=2|VerbForm=Fin	0	root	0:root	_
3	that	that	SCONJ	IN	_	19	mark	19:mark	_
4	if	if	SCONJ	IN	_	5	mark	5:mark	_
5	departing	depart	VERB	VBG	VerbForm=Ger	19	advcl	19:advcl:if	_
6	from	from	ADP	IN	_	7	case	7:case	_
7	Europe	Europe	PROPN	NNP	Number=Sing	5	obl	5:obl:from	Entity=(place-47-Europe)
8	(	(	PUNCT	-LRB-	_	11	punct	11:punct	Discourse=concession:34->33|SpaceAfter=No
9	this	this	PRON	DT	Number=Sing|PronType=Dem	11	nsubj	11:nsubj	Entity=(event-48)
10	may	may	AUX	MD	VerbForm=Fin	11	aux	11:aux	_
11	apply	apply	VERB	VB	VerbForm=Inf	5	parataxis	5:parataxis	_
12	to	to	ADP	IN	_	14	case	14:case	_
13	other	other	ADJ	JJ	Degree=Pos	14	amod	14:amod	Entity=(place-49
14	countries	country	NOUN	NNS	Number=Plur	11	obl	11:obl:to	Entity=place-49)|SpaceAfter=No
15	)	)	PUNCT	-RRB-	_	11	punct	11:punct	SpaceAfter=No
16	,	,	PUNCT	,	_	5	punct	5:punct	_
17	you	you	PRON	PRP	Case=Nom|Number=Sing|Person=2|PronType=Prs	19	nsubj	19:nsubj|21:nsubj:xsubj	Discourse=joint:35->7|Entity=(event-48(person-12)
18	will	will	AUX	MD	VerbForm=Fin	19	aux	19:aux	_
19	require	require	VERB	VB	VerbForm=Inf	2	ccomp	2:ccomp	_
20	to	to	PART	TO	_	21	mark	21:mark	_
21	have	have	VERB	VB	VerbForm=Inf	19	xcomp	19:xcomp	_
22	the	the	DET	DT	Definite=Def|PronType=Art	23	det	23:det	Entity=(abstract-5
23	visa	visa	NOUN	NN	Number=Sing	21	obj	21:obj	Entity=abstract-5)
24	before	before	SCONJ	IN	_	25	mark	25:mark	Discourse=circumstance:36->35
25	boarding	board	VERB	VBG	VerbForm=Ger	21	advcl	21:advcl:before	_
26	the	the	DET	DT	Definite=Def|PronType=Art	27	det	27:det	Entity=(object-50
27	plane	plane	NOUN	NN	Number=Sing	25	obj	25:obj	Entity=event-48)object-50)|SpaceAfter=No
28	.	.	PUNCT	.	_	2	punct	2:punct	_
# sent_id = GUM_fiction_pixies-47
# addressee = Jenna
# s_type = intj
# speaker = Robert
# text = Welcome to New York.”
1	Welcome	welcome	INTJ	UH	_	0	root	0:root	Discourse=evaluation:79->74
2	to	to	ADP	IN	_	3	case	3:case	_
3	New	New	PROPN	NNP	Number=Sing	1	obl	1:obl:to	Entity=(place-90-New_York
4	York	York	PROPN	NNP	Number=Sing	3	flat	3:flat	Entity=place-90-New_York)|SpaceAfter=No
5	.	.	PUNCT	.	_	1	punct	1:punct	SpaceAfter=No
6	”	''	PUNCT	''	_	1	punct	1:punct	_

Perhaps we should refine the definition?

Commit: d38df82

@amir-zeldes
Copy link
Contributor

Thanks for raising this issue - this is another one of those cases where UD guidelines conflict with the PTB guidelines. Since GUM's native POS tags are PTB-based, and the UPOS column is just an automatic conversion output, there are some mapping issues. In all of these cases, I believe UH is the correct xpos tag, as in OntoNotes we get:

  • god/UH (by itself, in "oh god" and in "oh my god")
  • please/UH
  • welcome/UH (split next to VBP, which I honestly can't understand, since if anything etymologically it's VBN)
  • gosh/UH

The example of 'gosh' in particular questions the guideline for "god" being a noun in this context, since they are interchangeable, but "gosh" is not a noun (it's a euphemism for the word "god" only in this use). Personally I'm comfortable with these things having UPOS as INTJ and it seems more misleading to me to try to find some etymology-based category for them, when synchronically they are not really nouns or verbs, as far as I can tell. Do you have an opinion on this @nschneid ?

@nschneid
Copy link

This is partly a multiword expression issue: The phrase "my god" functions like an interjection but I see no reason why internally it shouldn't be a standard noun phrase. Same with "my gosh".

I would argue that there are lexical items that are pure interjections, like "yay", and there are other words/phrases that get repurposed with interjection pragmatics and discourse attachment. Not exactly sure where the line should go at the UPOS level.

@nschneid
Copy link

Another example of a phrase with transparent internal structure used as an interjection: "Son of a bitch!"

@amir-zeldes
Copy link
Contributor

I think the latter would probably not be tagged as UH (there is one instance in ON, tagged literally with NN). Still for "gosh" I find it odd, since it can't actually occur as a normal noun - only as an interjection. And maybe leaving "god/gosh" aside, I think "please" and "welcome" make very little sense as VERB synchronically. We don't go by etymology for a great many things that are lexicalized, and I don't see the motivation to do so here.

@nschneid
Copy link

What about ADJ for "welcome" in "You are welcome"?

@amir-zeldes
Copy link
Contributor

That seems like a normal adjective to me, and indeed no one is trying to say that's a passive participle there (the VBP tagging cases of "Welcome!" seems really bizarre to me)

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Labels
None yet
Projects
None yet
Development

No branches or pull requests

3 participants