Head-specifier phrase, HOOK, and agreement

This is regarding a learner grammar of Spanish that I am working on, which needs to model agreement mismatches between specifiers and nouns:

vemos   muchos       torres
see.1pl many.masc.pl tower:fem.pl
`We see many towers.' [spa]

The original head-specifier phrase has the following apparently important identity:

HEAD-DTR.SYNSEM.LOCAL [ AGR #agr,
                          CONT.HOOK #hdhook & [ INDEX #agr ] ],
  NON-HEAD-DTR.SYNSEM.LOCAL.CAT.VAL.SPEC < [LOCAL.CONT.HOOK #hdhook ] > 

I have to take that out if I want to allow mismatches in agreement between the specifier and the noun but how do I keep the MRS together in such a case? It seems important that the hook is identified between the head daughter and the non-head daughter. I’ve made attempts but so far am not succeeding so I think I better consult people in terms of what I should be doing here.

I see two options:

  1. Leave the HOOK identity (#hdhook) but remove the AGR/INDEX identity.
  2. Create a version of muchos that is fem.pl — and marked as a mal entry.
1 Like

Thanks, @ebender !

I now have a mal-version of headed and binary-headed phrases which do not insist on agreement, and this way I am able to do what you suggest in (1), which is leave the HOOK identity.

It’s interesting because we sort of have to hypothesize whose agreement values are more important: the determiner’s or the noun’s. I think it is reasonable to assume that the determiner’s agreement values are clear to the learner while the noun’s are not necessarily internalized and so it is the noun that is assigned a wrong value.

Right now I am getting the opposite in the MRS but it is probably possible to get there :). At least this MRS seems properly connected.

Linguistically speaking, the agreement values are inherent to the noun — and the determiner is just reflecting them, I would say. So I think that is the MRS that you want.

2 Likes

You can take a look in the PorGram too.

% ace -g porgram.dat -Tf1
SENT: O artista vendeu muitos gatos
[ LTOP: h0
INDEX: e2 [ e SF: prop-or-ques E.TENSE: past E.ASPECT: aspect E.MOOD: indicative ]
RELS: < [ "_o_q_rel"<-1:-1> LBL: h4 ARG0: x3 [ x SPECI: bool COG-ST: cog-st PNG.PER: 3rd PNG.NUM: singular PNG.GEND: masculine ] RSTR: h5 BODY: h6 ]
 [ "_artista_n_rel"<-1:-1> LBL: h7 ARG0: x3 ]
 [ "_vender_v_rel"<-1:-1> LBL: h1 ARG0: e2 ARG1: x3 ARG2: x8 [ x SPECI: bool COG-ST: cog-st PNG.PER: person PNG.NUM: plural PNG.GEND: masculine ] ]
 [ "_muito_q_rel"<-1:-1> LBL: h9 ARG0: x8 RSTR: h10 BODY: h11 ]
 [ "_gato_n_rel"<-1:-1> LBL: h12 ARG0: x8 ] >
HCONS: < h0 qeq h1 h5 qeq h7 h10 qeq h12 >
ICONS: < > ]

% ace -g porgram.dat -Tf1 
As artistas venderam muitas ratazanas.
SENT: As artistas venderam muitas ratazanas.
[ LTOP: h0
INDEX: e2 [ e SF: prop E.TENSE: past E.ASPECT: aspect E.MOOD: indicative ]
RELS: < [ "_o_q_rel"<-1:-1> LBL: h4 ARG0: x3 [ x SPECI: bool COG-ST: cog-st PNG.PER: 3rd PNG.NUM: plural PNG.GEND: feminine ] RSTR: h5 BODY: h6 ]
 [ "_artista_n_rel"<-1:-1> LBL: h7 ARG0: x3 ]
 [ "_vender_v_rel"<-1:-1> LBL: h1 ARG0: e2 ARG1: x3 ARG2: x8 [ x SPECI: bool COG-ST: cog-st PNG.PER: person PNG.NUM: plural PNG.GEND: feminine ] ]
 [ "_muito_q_rel"<-1:-1> LBL: h9 ARG0: x8 RSTR: h10 BODY: h11 ]
 [ "_ratazana_n_rel"<-1:-1> LBL: h12 ARG0: x8 ] >
HCONS: < h0 qeq h1 h5 qeq h7 h10 qeq h12 >
ICONS: < > ]

Yeah but maybe in this case I want the noun to switch the gender value then… Perhaps that would be clearer. Not sure.