"Where is page 3?" unnecessarily ambiguous quantifier?

All of the parses for “Where is page 3” in erg-2018-osx-0.9.31 quantify page 3 using udef_q like variants of this (from the first ACE parse):

[ TOP: h0
RELS: < 
[udef_q__xhh LBL: h15 ARG0: x14 [ x PERS: 3 NUM: sg IND: + PT: notpro ] RSTR: h16 BODY: h17 ]
[ _page_n_1__x LBL: h18 ARG0: x14 [ x PERS: 3 NUM: sg IND: + PT: notpro ] ]
[ which_q__xhh LBL: h6 ARG0: x4 [ x PERS: 3 NUM: sg ] RSTR: h7 BODY: h8 ]
[ place_n__x LBL: h5 ARG0: x4 [ x PERS: 3 NUM: sg ] ]
[ number_q__xhh LBL: h9 ARG0: x3 [ x PERS: 3 NUM: sg ] RSTR: h10 BODY: h11 ]
[ compound__exx LBL: h12 ARG0: e13 [ e SF: prop TENSE: untensed MOOD: indicative PROG: - PERF: - ] ARG1: x3 ARG2: x14 ]
[ card__cxi LBL: h12 CARG: "3" ARG0: x3 [ x PERS: 3 NUM: sg ] ARG1: i20 ]
[ loc_nonsp__exx LBL: h1 ARG0: e2 [ e SF: ques TENSE: pres MOOD: indicative PROG: - PERF: - ] ARG1: x3 ARG2: x4 ]
HCONS: < h0 qeq h1 h7 qeq h5 h10 qeq h12 h16 qeq h18 > ]

It seems like this loses the fact that, in this case, page 3 is referring to a countable noun (page) that seems to have an unambiguous meaning of “a specific page” (ala “the”). It seems like there should be a quantifier in this case that says “the user didn’t specify a quantifier but the quantification semantic is like ‘the’”.

I understand why you’d get udef_q for “where is water?”, but not this case.

Are there cases where countable nouns used in a sentence like this really aren’t implying a semantic like “the”?

If not, is there a way to tell, from the parse, that the quantification semantic is referring to a specific page?

The analysis of the phrase “page 3” introduces two quantifiers, one to bind the ARG0 for “page” and one for “3”. The ERG treats “3” as the semantic (and syntactic) head of this phrase, and introduces the quantifier number_q to bind the ARG0 of “3”. This number_q is closely related to the proper_q introduced for names, and is to be treated as definite, like the_q, referring to some specific individual that exists in the relevant discourse. The “page” is interpreted as a kind of classifier of the individual whose name is “3”. This is analogous to the analysis of “professor Smith”, interpreted to refer to an individual named “Smith” who is in the set of individuals who are professors. So “chapter 3” is like “page 3” in that both refer to an individual whose name is “3”, but the first is in the set of things that are chapters, while the second is in the set of pages. One can of course dispute analyzing “page 3” as analogous to “professor Smith”, but as it is, you do get the definiteness you want via the number_q quantifier.

1 Like

Ahh, makes sense. Great. Thanks @Dan!

The fact that proper_q is treated as definite as well is something I haven’t implemented either. I’m sure that’ll fix some future bugs.