L322
Islands are constituents out of which embedded constitutes cannot move.
We will cover only CP islands here. If the CP is the complement of a NP,
nothing can be extracted out of it:
If the WH-phrase is not extracted, the related structures fail where the WH-phrase does not have an echo reading:
The vast majority of NPs do not subcategorize [+WH].Hence, the movement of the WH-phrase to the embedded CP crashes:
There are a few nouns that take a WH-complement. The WH-phrase moves and is adjoined to CP:
We have a case of a D-structure which has no grammatical derivation. If one wants to find out the identity of such an embedded object, he must resort to the following type of discourse connection:
Extraction is not possible out of a CP complement of an adjective:
There are certain verbs that take complements out of which extraction
is not possible or varies amongst speakers of English:
The fact that some verbs resist extraction of WH-phrases from their CP complements suggests that there is a general (universal) constraint against extraction from CP. Some languages permit no extraction at all. It seems reasonable to propose that some verbs in certain languages permit bridges. These bridges cross over the barrier of CP permitting extraction. Bridges cannot be built over barriers of NP and AP complements. This is interesting in that nouns and adjectives must be Case marked and cannot assign Case structurally (see ). Is there a relationship between Case and bridges? We leave this question for future pondering.
![]()
To return to the course outline for 322 click
here
![]()