David P. Baxter (dbaxter@uiuc.edu)
Mon, 22 Jun 1998 15:57:07 -0500 (CDT)
Dear fellow HPSGers, I would like to know if there is a reason why the feature ARG-ST cannot or should not be defined for headed phrases, as a HEAD feature inherited via the HFP from the lexical head of the phrase. I will welcome either emailed explanations or reference to relevant literature. I want to make ARG-ST a HEAD feature because of a fact about purpose infinitives that can be explained if the modified phrase has a list of the SYNSEM values of its arguments. The fact is the following: For a purpose infinitive sentence like (1) to be well formed, the gap in the infinitive must be coreferential with an argument in the modified phrase. (1) Dana bought War and Peace to give __ to Sandy. In (1), the gap is coreferential with "War and Peace". If you replace "bought War and Peace" with "did so", as in (2), the sentence is no longer acceptable. (2) *Dana did so to give __ to Sandy. I would like to account for this fact with a constraint on the purpose infinitive specifying that the ARG-ST list of the modified phrase contains a synsem object whose LOCAL value is shared with an element in the SLASH set of the infinitive VP. If anyone has any ideas about other ways to account for the data in (1-2), please let me know. Thanks, David Baxter ______________________________________________________________________ David P. Baxter Department of Linguistics dbaxter@uiuc.edu University of Illinois http://www.uiuc.edu/ph/www/dbaxter at Urbana-Champaign
This archive was generated by hypermail 2.0b3 on Fri Dec 18 1998 - 20:38:20 PST