Syntax Comments on Pesetsky (1987)

Syntax Comments on Pesetsky (1987) Comments on Rudin (1988) This is a very famous paper which is sort about wh-movement, and sort of about the existence of LF movement. His idea is to show evidence of covert wh-movement for whin-situ (e.g., any wh-word in Japanese, or the second wh-word in English) by showing that some wh-words don’t “move at LF” and that they act differently from those that do. Rudin’s (1988) paper is a very prominent analysis of languages like Bulgarian in which all of the wh-phrases in a multiple-wh-question move to the front. She considers evidence that seems to differentiate two types of multiple wh-fronting languages, one exemplified by Bulgarian, the other exemplified by Serbo-Croatian. One of the first ideas in the paper is that you can get the meaning of a question in one of two ways. You need to have an operator in the CP projection and a bound variable in the base position of the wh-word. This is often accomplished by moving the wh-word (the head of the chain is the wh-operator, the bottom of the chain is the trace, serving as the bound variable). The idea is that we need a representation like: For what x. John bought x? Another possibility, though is to just base generate an operator in CP and let the whword serve as a variable without moving. This is the proposal of Baker (1970). This paper is not highly technical, but it is a nice, interesting paper—I think it should be pretty much accessible given Syntax I and what we’ve done this semester. (1) (2) WHATi did John buy ti ? Qi John bought WHATi? In (2), the wh-word doesn’t have to move. So we don’t expect the relation between Q and WHAT to obey island constraints. Pesetsky proposes that certain wh-words are “D-linked” (Discourse linked) in that both conversational participants know and have in mind a small, specific set of alternatives from which the answer to the wh-phrase is drawn. D-linked wh-words don’t have to move. Non-D-linked wh-words do have to move. Wh-words can be “aggressively non-Dlinked” by adding “…in the world” or “…the hell”—So what the hell will always have to move, and what the hell turns out (as predicted) to be bad in islands (something we can actually only test in Japanese due to some orthogonal[?] restrictions on placement of …the hell in English questions). Comments on Cole and Hermon (1998) This one is optional just because it seemed like a lot of reading without it. Cole and Hermon look at some interesting facts about wh-movement in Malay. Malay seems to have a wh-in-situ possibility and a wh-movement possibility. What’s perhaps weirder is that when you would have to do successive-cyclic movement (when multiple clauses are involved), you can move the wh-word partway to its scope position and stop. Overt wh-movement in Malay respects islands (CNPC, adjunct islands). However, wh-insitu is allowed inside islands (and in fact this is common among wh-in-situ languages like Japanese, Korean, Chinese). There is a diagnostic for movement—A “transitivity prefix” meng- disappears when an object moves over it. And this also tells us that wh-in-situ doesn’t move, and fronted whwords do (that is, meng- disappears only for moved wh-phrases). Here’s the weird thing: Partially moved wh-phrases are sensitive to islands—even if they haven’t moved out of them yet… If your wh-word is inside an island you can leave it in situ (Subjacency is fine), but if you move the wh-word either partially (keeping it inside the island) or all the way out of the island, you appear to violate Subjacency. Mengdeletion is sensitive only to the overt movement part. Stuff for you to do: • Do the readings. • Do a summary of Pesetsky (1987) in the same style as previous summaries. Next time: Discussion of wh-movement, student presentation of Pesetsky (1987).