Internal head constraint

Relative clauses in English (and in an overwhelming number of languages) obey a `no internal head' constraint. This constraint is exemplified in the contrast between ((226)) and ((227)).

(225)0(225

(226)

the person [who_i [$\epsilon$_C Muriel likes $\epsilon$_i]]  (226)0(226

(227)

*the person [[which person]_i [$\epsilon$_C Muriel likes $\epsilon$_i]] 

We know of no good way to rule ((227)) out, while still ruling ((228)) in.

(227)0(227

(228)

the person [[whose mother]_i [$\epsilon$_C Muriel likes $\epsilon$_i]] 

Dayal (1996) suggests that `full' NPs such as which person and whose mother are R-expressions while who and whose are pronouns. R-expressions, unlike pronouns, are subject to Condition C. ((226)) is, then, ruled out as a violation of Condition C since the person and which person are co-indexed and the person c-commands which person. If we accept Dayal's argument, we have a principled reason for allowing overgeneration of relative clauses that violate the internal head constraint, the reason being that the XTAG grammar does generate binding theory violations.