Rejoinder to Pereltsvaig’s head movement in Hebrew nominals: A reply to Shlonsky [LINGUA Vol. 116 Issue 8 (2006) A1–A40]

Pereltsvaig claims that my phrasal pied piping analysis of Semitic NPs (Lingua 114:1465– 1526) is inadequate since it fails to account for (i) the distinct behavior of light and heavy adjectives, (ii) the position of DP and PP complements of the noun and (III) the correlation between agreement in definiteness with the position of the modifiers. She also finds my work ‘‘theoretically inelegant’’ and ‘‘non-minimalist’’. I’ll be brief and stick to the three substantive claims. The rhetorical ones require no response. Claim 1: Light adjectives appear in inverse order; heavy adjectives appear in canonical order. Shlonsky can’t account for this. Pereltsvaig can ‘‘in a more direct and elegant way’’. The claim is undermined by a fatal methodological error and benefits from no empirical support. Pereltsvaig’s own analysis is perhaps ‘elegant’ but insufficient. Consider the error. Pereltsvaig’s examples of DPs with heavy adjectives (her (12) and (13)) only show that a heavy adjective follows a light one. To show anything about adjective ordering, one needs examples which are independent of heaviness. In other words, one crucially needs a test case of a DP containing at least two heavy adjectives. Then, one could determine whether they are ordered ‘canonically’ or ‘inversely’. ‘‘Unfortunately,’’ writes Pereltsvaig, ‘‘this . . . cannot be tested, since strings with two heavy adjectives in Hebrew appear to be degraded, regardless of the order . . .’’. Unfortunate indeed. Relative clause modifiers are also heavy, must follow light adjectives and, unlike heavy adjectives, there may be more than one. Indeed, when there are two or more relative clauses in a DP, theyare freelyorderedamongthemselves.This looksrelevantbutPereltsvaig, rather thanattempting to state a generalization here, denies that one is possible, stating that relative clauses ‘‘have some independent reason for their placement’’. She never deigns to inform the reader what this reason is. www.elsevier.com/locate/lingua Lingua 116 (2006) 1195–1197