Stephen M. Wechsler (wechsler@mail.utexas.edu)
Thu, 25 Jul 1996 11:08:58 -0500 (CDT)
> Now suppose I believe that LFG is the correct linguistic theory. When >designing a grammar for a new language I have to make sure that the >rules have some pre-defined format; I can use a certain set of >meta-variables such as up-arrow and down-arrow; I know what the >lexicon should look like. In short, it's possible to formally define >the structure of LFG grammars. > Coming back to HPSG, the situation is not that clear. I know what >the general architecture is: a grammar is specified using totally >well-typed feature structures over a signature that specifies the type >hierarchy and the appropriateness specification. But what do rules >look like? Let's assume that a grammar consists of a set of >principles, a set of rules, a set of ID schemata and a lexicon. Two >questions pop up: (1) What is the formal structure of each >constituent? (2) What other requirements have to hold for a >well-formed grammar in order for it to be an HPSG grammar? Joan Bresnan has always been careful to emphasize the distinction between the descriptive formalism of LFG (up and down arrows etc) and the theory stated in terms of that formalism. I assume the same applies to HPSG. The formal machinery doesn't rule out natural languages; 'you could do your income taxes with it' as we used to say. For example, you could write a grammar in either LFG or HPSG with the equivalent of 'affix-hopping', as is standard in GB/P&P (say, by putting a SLASH feature in the morphology). But both LFG and HPSG assume a strong kind of lexicalism that outlaws such moves. There's a list of 'leading ideas' of HPSG at: http://hpsg.stanford.edu/hpsg/leading-ideas.html --Steve o----------------------------------------------------------------o | Stephen Wechsler, Assistant Prof. of Linguistics. 403 Calhoun | | Spring 95 office hrs: Tu 2-4, F 11-12. MAIL: Linguistics Dept.,| | 501 Calhoun, U. of Texas, Austin, TX 78712-1196.(512) 471-1701 | | http://uts.cc.utexas.edu/~wechsler/ | o----------------------------------------------------------------o
This archive was generated by hypermail 2.0b3 on Fri Dec 18 1998 - 20:33:28 PST