Mike_Maxwell@sil.org
Thu, 15 Dec 1994 22:43 -0500 (EST)
Anne Abeille (Anne.Abeille@linguist.jussieu.fr) wrote s.t. that got me thinking: > ...if you look at so-called tough constructions in English you have > an adjective, say "difficult", which can take as cplt an infinitival > missing its complement, say "to see". what plays the role as a > filler can occur... as the subject of a copular construction: > This ant is difficult to see I believe Chomsky, in "On wh-movement", was the first to suggest that tough-constructions should be conflated with "ordinary" wh-movement constructions (including complementizerless relative clauses, which also lack an overt operator). But periodically people have noticed that, unlike "ordinary" wh-movement constructions, tough-movement constructions become less acceptable if the embedding is at all deep: ?My aunt is difficult to try to persuade anyone to visit. --and downright unacceptable if the gap is embedded in a tensed S inside the infinitival: ??Flying saucers are difficult to believe that anyone has seen. This is, of course, a problem for any theory that conflates these constructions, whether GBish or HPSGish. I wonder if the problem is semantic: that "difficult to see" is somehow a reasonable property of s.t., whereas "difficult to try to persuade anyone to visit" is less like a property. I guess there are a lot of factors, some of which don't have anything to do with tensed vs. infinitival ("this problem is difficult to try to understand" seems better than "...difficult to seem to understand", for instance). And maybe we should approach it the other way: what makes long-distance dependencies good? Ertischik-Shiir (sp?) had some to say on that, I think. But enough speculation on my part. Surely someone has written about the reasons for acceptability, or lack therof, of tough-movement constructions?
This archive was generated by hypermail 2.0b3 on Fri Dec 18 1998 - 20:31:40 PST