Re: ARG-S and Binding

About this list Date view Thread view Subject view Author view

nerbonne@let.rug.nl
Mon, 19 Oct 1998 09:39:37 +0200 (METDST)


Hear, hear. >> Date: Sun, 18 Oct 1998 18:32:34 -0500 >> From: Gert Webelhuth <webelhut@mindspring.com> >> Subject: Re: ARG-S and Binding >> >> [...] that argument >> attraction changes HPSG from a monostratal theory of grammar to a >> multistratal theory, i.e. it introduces something akin to a difference >> between D-structure and S-structure into the analysis of signs and allows >> these structures to be different. This has always been my question about ARG-S, too. At least for "relation-changing" rules (i.e., rules which change or are sensitive to grammatical relations), ARG-S reintroduces an additional stratum which CG, GPSG, and early HPSG explicitly postulated to be superfluous, referring mostly to David Dowty's work. To be a little more restricted in the analogy to earlier theories, since arbitrary tree structures are not allowed on ARG-S, only linear orders, it's not exactly D-structure. Since it's a limited (but very rich) range of information, it also looks like a richer f-structure (richer because realized as a sequence of HPSG synsems, which are richer than LFG f-structure arguments). Another distinction, of course, is that LFG has been explicit about what's handled in f-structure, while HPSG seems to be putting ARG-S up for grabs. Maybe that's because things haven't settled. The similarity might not be surprising, if one kept in mind that f-structure was invented to restrict D-structure to the minimum needed. --John Nerbonne


About this list Date view Thread view Subject view Author view

This archive was generated by hypermail 2.0b3 on Fri Dec 18 1998 - 20:38:22 PST