1 On tough-movement* Milan Rezac, University ... - Multimania.co.uk
1 On tough-movement* Milan Rezac, University ... - Multimania.co.uk
1 On tough-movement* Milan Rezac, University ... - Multimania.co.uk
You also want an ePaper? Increase the reach of your titles
YUMPU automatically turns print PDFs into web optimized ePapers that Google loves.
. =No person is such that it seems like he is here.<br />
c. ≠It seems that no one is here. (o.k. if there are no people present)<br />
d. No one seems to be here. (o.k. on both readings) seem > no one, no one > seem<br />
The evidence of these paradigms is very specific: the non-thematic DP in both <strong>tough</strong>movement<br />
and <strong>co</strong>py-raising λ-binds a pronoun, not a <strong>co</strong>py. Copy-raising wears this on its sleeve;<br />
for <strong>tough</strong>-movement it follows if the OP/gap must be pronominal, as discussed.<br />
The syntax of <strong>tough</strong>-movement proposed here predicts a related anti-re<strong>co</strong>nstruction effect.<br />
The TM subject is base-generated in [Spec, TP] of the TM trigger. There should therefore be no<br />
re<strong>co</strong>nstruction position within T' (not just one below the TM trigger), unlike what is expected on<br />
Browning's (1989: chapter 2) proposal where the TM subject is base-generated sister to the<br />
maximal projection of the TM trigger (e.g. AP/NP). This prediction can be <strong>co</strong>nfirmed. Hey<strong>co</strong>ck<br />
(1994:262) observes that TM <strong>co</strong>nstructions seem like individual-level predication. This fits in<br />
with Lasnik and Fiengo's (1974:544ff.) development of Postal's (1971) observation that TM<br />
subjects must be generic rather than existential if indefinite, since this is characteristic of<br />
individual-level predication. Copy-raising subjects behave thusly as well (Hey<strong>co</strong>ck 1994:293f.).<br />
(36) TM indefinite subjects are only generic<br />
a. Beavers are/a beaver is hard to kill. (Lasnik and Fiengo 1974:546; generic only)<br />
b. It was a pleasure to each a bunch of bananas; there are their skins. (Lasnik and Fiengo<br />
1974:546)<br />
c. *A bunch of bananas was a pleasure to eat; there are their skins. (Lasnik and Fiengo<br />
1974:546)<br />
(37) Individual level predication subjects are only generic<br />
a. *A building/someone was tall. (Lasnik and Fiengo 1974:545)<br />
b. Beavers are fat (*they're over there). (Lasnik and Fiengo 1974:546)<br />
(38) Copy-raising subjects are only generic<br />
a. #Snow sounds/seems like it's falling on the mountain. [generic reading made salient]<br />
b. Snow seems to be falling on the mountain. [existential reading possible]<br />
c. Snow sounds like it must be a strange thing: hexagonal ice crystals falling out of the<br />
sky? [generic reading made salient] (Hey<strong>co</strong>ck 1994:293f.)<br />
Kratzer (1995) and Diesing (1992: chapter 2) argue that the availability of an existential<br />
reading for such indefinites depends on binding by the existential closure operator of clausal<br />
partition theories. This operator is located somewhere between T' and the VP. The absence of an<br />
existential reading for individual level predicates is taken as evidence that their subjects, unlike<br />
those of stage-level predicates, are base-generated in [Spec, TP]. The result transfers to <strong>tough</strong>movement<br />
and <strong>co</strong>py-raising: there is no <strong>co</strong>py of the matrix subject where it <strong>co</strong>uld be bound by<br />
the generic operator, as there is in movement. Kratzer and Diesing also argue that basegenerating<br />
the subject in [Spec, TP] rather than lower is what gives rise to the other properties of<br />
individual level predication.<br />
The individual-level nature of TM predication may help in ac<strong>co</strong>unting for some hitherto<br />
rather mysterious restrictions on the interpretive relation between the TM subject and the OP<br />
13