Abstract
This paper argues that, in Twi (Akan), the shape of a reduplicant, and the resulting segmental processes and reduplicated output are largely determined by its input's inherent tone and the ‘strength’ of its syllables; an argument in support of Mophological Doubling Theory (MDT) (Inkelas and Zoll 2005) within which the reduplicant like its base has access to the input. This observation led to the classification of disyllabic verb inputs into two: a low-toned group with a contrastive, and a non-contrastive, syllable; and, a low-high-toned group with two contrastive syllables. Non-contrastive syllables are banned from the reduplicant, which is why reduplicants of the former group are monosyllabic, and, respect *[−High] (McCarthy and Prince 1995) just like monosyllabic verbs' reduplicants. Their final segment(s) are unparsed, and are deleted without compensatory lengthening, except in cases where the final segment becomes homorganic with the following consonant. The latter group's reduplicants are disyllabic with every segment parsed; homorganic assimilation of the reduplicant-final segment applies optionally; and, with the loss of (a) reduplicant-final segment(s), vowel lengthening applies to avoid a monosyllabic reduplicant and, consequently, *[−High]. *[−High] applies to distinguish outputs of the former group from outputs of monosyllabic/disyllabic adjectives. The ranking argument, *Contour ≫ Align LH ≫ Tone Polarity ≫ Tone Spread, derives output tones.
©[2013] by Walter de Gruyter Berlin Boston
