A perspective analysis of English relative clause formation

The English relative that and 7ωh-relatives (who, which, etc.) are generally regarded as morphophonemic variants. Some arguments, however, have been put forth to claim that the relative that is not a pronoun but a oomplementizer. These arguments seem highly promising, but they lack consideration from a semantic viewpoint. Little work has been done to investigate the problem of whether there are any relationships between the meaning and the syntactic behavior of relative clauses and, if there are, what kind of relationships these might be Bolinger [3] is the only exception as far as I know. In this paper, I propose a somewhat speculative analysis that the underlying structure (not necessarily the deep structure) of the English relative clause is of the form : (i) [<NP>_N_I[S^-(COMP)/(±Wh)[s‥x_i‥･]]] where x is bound pro-name in terms of Keenan [12]. The analysis stands on the hypothesis that the relative that, as well as ωh-relatives, has its own lexical property and that it is the lexical properties that give rise to the syntactic differences between that- and ωh-relative clauses.