Extraction and Inversion

$\langle$extracted$\rangle$, possible vales are +/- All sentential trees with extracted components, with the exception of relative clauses are marked S.b$\langle$extracted $\rangle = +$ at their top S node. The extracted element may be a wh-NP or a topicalized NP. The $\langle$extracted$\rangle$ feature is currently used to block embedded topicalizations as exemplified by the following example.

(518)0(518

(519)

* John wants [Bill_i [PRO to leave t_i]]  

$\langle$trace$\rangle$: this feature is not assigned any value and is used to co-index moved NPs and their traces which are marked by $\epsilon$. $\langle$wh$\rangle$: possible values are +/-
NPs like who, what etc. come marked from the lexicon with a value of + for the feature $\langle$wh$\rangle$. Non wh-NPs have - as the value of their $\langle$wh$\rangle$ feature. Note that $\langle$wh$\rangle$ = + NPs are not restricted to occurring in extracted positions, to allow for the correct treatment of echo questions. The $\langle$wh$\rangle$ feature is propagated up by possessives - e.g. the + $\langle$wh$\rangle$ feature of the determiner which in which boy is propagated up to the level of the NP so that the value of the $\langle$wh$\rangle$ feature of the entire NP is +$\langle$wh$\rangle$. This process is recursive e.g. which boy's mother, which boy's mother's sister. The $\langle$wh$\rangle$ feature is also propagated up PPs. Thus the PP to whom has + as the value of its $\langle$wh$\rangle$ feature. In trees with extracted NPs, the $\langle$wh$\rangle$ feature of the root node S node is equated with the $\langle$wh$\rangle$ feature of the extracted NPs. The $\langle$wh$\rangle$ feature is used to impose subcategorizational constraints. Certain verbs like wonder can only take interrogative complements, other verbs such as know can take both interrogative and non-interrogative complements, and yet other verbs like think can only take non-interrogative complements (cf. the $\langle$extracted$\rangle$ and $\langle$mode$\rangle$ features also play a role in imposing subcategorizational constraints). The $\langle$wh$\rangle$ feature is also used to get the correct inversion patterns.