Claire Grover (grover@cogsci.edinburgh.ac.uk)
Mon, 19 Dec 94 09:59:02 GMT
As mentioned by Ivan Sag, in my thesis I present an HPSG account of tough and related constructions which rejects the UDC analysis. In my account the missing object in the tough complement is promoted by means of a lexical rule similar to passive and the sign for tough causes the promoted object to be structure-shared with the tough subject. Ivan Sag and Bob Levine gave the following examples to demonstrate that the tough construction is unbounded: (1) Getting herself arrested on purpose is hard for me to imagine Betsy being willing to consider __. (2) This new kind of box was hard for me to persuade Leslie to retool her factory to produce __ To account for such apparently unbounded examples I allow raising and control predicates and auxiliaries to inherit and pass on information about the promoted object. My claim is that rather than being UDCs, these examples involve sequences of local dependencies mediated by the raising/control/aux lexical signs. An advantage to this account is that it follows that tough gaps are always complements not subjects: (3) *Kim is hard for me to believe __ likes lime pickle and more generally it follows that any tough gap in a finite sentential complement is bad: (4) *Lime pickle is hard for me to believe that Kim likes __ As Mike Maxwell pointed out, these examples are hard for the UDC account to explain. For (3) standard HPSG simply stipulates that the NP in SLASH must be accusative and (4) isn't blocked at all. A further advantage to my approach is that it extends well to other European languages. In Dutch the tough construction is completely bounded and cannot therefore be treated as a UDC. In my account the Dutch tough construction is just like the English one except that control and raising predicates cannot mediate the tough relation. In Italian and Spanish, the tough construction is usually bounded but there are exceptions when restructuring verbs intervene between the tough adjective and its complement. My account of English can be adapted to Italian and Spanish with one modification: only the restructuring verbs are able to be tough mediators. Furthermore, the mechanisms I have developed for the tough construction lend themselves well to other usually local constructions where restructuring verbs permit objects to be promoted out of their licensing clause (clitic-climbing and object promotion in impersonal 'si'/'se' constructions). Bob Levine observed that an earlier version of my analysis (unpublished) did not deal with the fact that parasitic gaps may occur with the tough construction nor with the fact that the tough gap may occur inside an NP. In my thesis I have attempted to rectify this omission: I provide an extensive reanalysis of parasitic gaps and the object promotion mechanism is made more general to cover more cases. My thesis hasn't been examined yet but I hope that it will be available for distribution in the not too distant future. Claire Grover
This archive was generated by hypermail 2.0b3 on Fri Dec 18 1998 - 20:31:40 PST