Swati and Kikuyu reduplication evidence against exhaustive copy

This paper addresses the central question of deriving shape-invariance of partial reduplication in two competing models of reduplication: Selective Copy [McCarthy and Prince 1986, 1988, and 1990] and Exhaustive Copy [Steriade 1988]. I show that the core difference lies in the fact that whereas Selective Copy favors prosodic templates, Exhaustive Copy selects parameters defined in terms of prosodic units. This distinction is examined against Swati diminutive reduplication which shows the base-independent vowel a. Crucially, to insert a requires access to segmentally unspecified prosodic units. Under Selective Copy, prosodic templates provide exactly the structures for defining insertion. In contrast, Exhaustive Copy cannot provide an internally consistent mechanism to express a insertion, with its reliance on parameters. This incapability reveals a larger problem confronting Exhaustive Copy: how to account for reduplication involving base expansion rather than base reduction. Unless a mechanism is found, Exhaustive Copy, with its parameter approach, is unable to explain base-expansion cases in reduplication.