Tibor Kiss (tibor@heidelbg.ibm.com)
Mon, 16 Oct 1995 16:24:42 +0200
Hi HPSGians, I agree that the representation on page 32 (Kim walks), where both "Kim" and "walks" are taken to be phrases (this idea is taken over in fn. 32 on the same page, and on p. 392, see below) may lead to the representation below (which looks awful to me). > S > / \ > NP VP > | | > N V > | | > Kim left But isn't it so that the representation on page 32 either is due to simplicity of presentation (the apprentice reader would possibly be confused if the first structure he encountered shows that subjects are words)? The grammar in PS 1994 actually licenses the structure below, where the phrasal status of the subject is without relevance (come on authorities, correct me ...), if we interpret the following quote as mentioned below: S / \ N VP | | Kim V | left PS 1994:346 write: " inasmuch as all phrases will now be [COMPS <>]. we might incorporate this specification into the definition of the sort _phrase_. In this case, we can simply say that nonheads (other then markers) are phrasal." I interpret this passage in the following way: A nonhead is required to bear the specification [COMPS <>]. This requirement is satisfied by [word PHON: <kIm>]. (Please note that the relationship between phrase and [COMPS <>] is simple implication and not bi-conditional.) So the interpretation should be: Each nonhead is saturated. This assumption is more natural than the one which says nonheads are phrasal (and it incorporates markers, if somebody wants to have them around). BTW: This assumption goes against one of the basic assumptions of my thesis, but see Feyerabend (1975) on contradiction and coherence in science. The other place where the phrasal status of the subject is addressed, is the table on page 392, where it is stipulated that subject daughters have Bar level 2. If we interpret this as an abbreviation for [COMPS <>], we may have a lexeme as a subject (or even as a complement daughter), as long as it is saturated. The projection of the verb is a different story, it's actually needed to combine the ID-grammar of PS 1994 with the LPs in PS 1987 (it seems that LP2 [PS 1987:181] really does not work because it allows sentential and nonfinite subjects to be realized after the finite verb, but that is irrelevant for the present concerns -- definitely not personal, Ivan). Best regards, Tibor (the impersonal 8->) =========================================================================== Dr. Tibor Kiss IBM Germany Institute for Logic and Linguistics Vangerowstr. 18 69115 Heidelberg 0049-6221-594483 (phone) 0049-6221-593200 (fax) kiss@vnet.ibm.com ============================================================================
This archive was generated by hypermail 2.0b3 on Fri Dec 18 1998 - 20:33:04 PST