ABSTRACT

Nonrestrictive relatives are usually conceived of as a unitary type of relative clause (semantically and syntactically opposed to both restrictive and “amount”, or “third type”, relatives). In the literature, they have been analysed either as a sentence grammar phenomenon, specifi cally as clauses internal to the nominal projection that also contains the Head, like restrictive and “amount” relatives (see, among others, Smith 1964, Jackendo 1977, chapter 7; Huot 1978; Perzanowski 1980; Cornilescu 1981; Kayne 1994, chapter 8; Bianchi 1999, chapter 5; Kempson 2003; Arnold 2007), or as a discourse grammar phenomenon, i.e., as sentences generated independently of the sentence containing the Head, whose pronouns relate to the Head much like (E-type) pronouns relate to an antecedent across discourse (see, for instance, Ross 1967, 434 ; Aissen 1972; Emonds 1979; Stuurman 1983; Sells 1985; Haegeman 1988; Fabb 1990; Espinal 1991; Peterson 2004; Grosu 2005).1