Debate (p. 88) [via PaperHive@docloop]

[docloop[bot]]

Annotation imported from PaperHive.

Regarding this part:

Hence, this book is a contribution to the general debate about argument structure constructions.

Remi van Trijp wrote:

Yes! I agree that you make an important contribution to the debate, and I appreciate the faire conclusion you draw from it. If I can suggest anything at all, it is to clarify that you investigated constructions as phrase-structural rules, but that e.g. Cognitive CxG (Goldberg...) do not assume an ASC is phrase-structural. Obviously, it is up the the CxG proponents to offer the same formal precision to show how that would work.

Link to original comment. About docloop.

[docloop[bot]]

Annotation imported from PaperHive.

Stefan Müller wrote:

Maybe I am missing something crucial but I think that the phrasal construction in your and in Adele's sense are basically the same as phrase structure rules but without the assumption of continuity. Embodied CxG has this and HPSG has this as well. I think I understand the formalities behind all this pretty well and I think what I said here applies to this type of phrasal constructions as well. FCG also assumes this kind of discontinuous constituents. I discussed this in Müller (2017). I worked with discontinuous constituents for almost a decade (Müller, 1995). I think what I described here would apply to models with discontinuous constituents as well. The problem is that a benefactive construction contributes semantics and a resultative construction contributes semantics and this has to be integrated somehow. One of the contribution embeds the other. You cannot do this by just unifying stuff (as you seem to suggest in another comment). This does not work. You have to have embedding and in a correct order with the scope taking element embedding the other one. --- * Müller, Stefan. 1996. The Babel-System: An HPSG Fragment for German, a Parser, and a Dialogue Component. In Proceedings of the Fourth International Conference on the Practical Application of Prolog, 263–277. London. * Müller, Stefan. 2017. Head-Driven Phrase Structure Grammar, Sign-Based Construction Grammar, and Fluid Construction Grammar: Commonalities and Differences. Constructions and Frames 9(1). 139–174. DOI:10.1075/cf.9.1.05mul.

Link to original comment. About docloop.

[docloop[bot]]

Annotation imported from PaperHive.

Remi van Trijp wrote:

No, it's not discontinuous phrase structure I have in mind. Goldberg's ASCs make no reference to phrase-structural relations; one alternative she suggested at one point is to use dependency trees, but even about that I don't think she would make strong claims. Argument structure and functional structure can be decoupled from phrase structure, and an ASC can just map between those two structures without caring about phrase structure. In fact, an ASC approach to argument structure can work with both a continuous and discontinuous phrase-structure. As for FCG, the formalism itself does not assume any kind of structure, it's me who has often used discontinuous phrase structure :), but I use it alongside other "views" on sentence structure.

Link to original comment. About docloop.

[docloop[bot]]

Annotation imported from PaperHive.

Stefan Müller wrote:

You have two or more things combined in a structure. This could be [Subj Verb Obj], right? This is like specifying an HPSG schema with three daughters without giving the mother a name.

Link to original comment. About docloop.

[docloop[bot]]

Annotation imported from PaperHive.

Remi van Trijp wrote:

If the functional structure of Goldberg's approach is a dependency tree, then the Verb is the root and the subject and object its dependents. Is there work on doing dependency structures in HPSG? I know the Alpino parser for Dutch is a dependency parser, but they say it uses HPSG theory. I have always wondered what they mean with that.

Link to original comment. About docloop.

[docloop[bot]]

Annotation imported from PaperHive.

Stefan Müller wrote:

Well, HPSG has much of the dependencies encoded in valence features. The verb is the root and the elements in SPR and COMPS are the dependents. One has to care about the adjuncts in both approaches.

Link to original comment. About docloop.

[docloop[bot]]

Annotation imported from PaperHive.

Remi van Trijp wrote:

Unless I am mistaken, the values of SPR and COMPS in HPSG are phrasal nodes such as NP, which is different from dependencies that directly point to words. Is such information represented in HPSG?

Link to original comment. About docloop.

[docloop[bot]]

Annotation imported from PaperHive.

Stefan Müller wrote:

OK, yes. There is an object for the whole phrase. This is a difference between (some versions of) DG and HPSG. But CxG and HPSG are more similar then they are similar to DG. CxG does not have word dependencies or does it?

Link to original comment. About docloop.

[docloop[bot]]

Annotation imported from PaperHive.

Remi van Trijp wrote:

Depends which construction grammarian you talk to :) The grammar for English that I am implementing does, it should be online by the end of April

Link to original comment. About docloop.