The optimal second position in Pashto

Introduction In a detailed study of Pashto clitics, Tegey (1977) argues that some interesting interactions between clitic placement and phonology in the language cannot be accommodated in a derivational model of grammar like (1), in which there is a strict separation between the syntactic and phonological components. (1) D-structure S-structure PF LF Rather, Tegey suggests that phonological rules must be interleaved among syntactic ones. Kaisse (1981; 1985: 132-143), on the other hand, observes that the Pashto data may be accounted for simply by regarding a small set of verb stems as polymorphemic, rather than as monomorphemic. The present paper brings additional evidence to support Kaisse's conclusion that Pashto may be accommodated by a grammar like (1), though from a different body of facts. After outlining the facts of Pashto second position clitic placement in sectionn 1, sectionn 2 will challenge the assumption of both Tegey and Kaisse that second position clitic placement in Pashto is a syntactic process. It will be shown that Pashto clitic placement in fact is best explained by a purely phonological account, a point suggested recently also by van der Leeuw (1995) and Hock (1996). The extent to which syntactic structure contributes to clitic placement is narrowly constrained by principles proposed by Selkirk (1984; 1986) for the mapping of syntactic structure to prosodic structure. Pursuing recent ideas of Anderson (1996) regarding second 2 position phenomena, it will be claimed that, once the prosodic structure of a sentence has been derived, principles of Optimality Theory (Prince and Smolensky 1993; McCarthy and Prince 1993b) select the output form. This analysis will be proposed in sectionn 3. Pashto clitics are particularly well suited to an analysis within Optimality Theory, since highly ranked prosodic constraints may compel clitics to move away from second position in either direction. Pashto has often been cited as a language that is recalcitrant to traditional models of grammatical analysis. However, the conclusion here—that clitic placement in Pashto is not syntactic—reveals that Pashto clitics do not pose a serious challenge to traditional ideas about grammar and clitics. The Pashto data remain as interesting as ever, though not necessarily for the reasons originally brought forth by Tegey (1977); rather, Pashto clitic placement is fascinating now because its general "second position" requirement is violated by other constraints in the language, suggesting that Anderson's (1996) application of Optimality Theory to second position clitics may yield considerable force in an …