On the representation of lexical geminates
Date: April 28, 2023
Location: SH 4430C/ PC Zoom Room
Speakers: Nina Topintzi (Aristotle University of Thessaloniki, Greece)
In languages where both singleton (C) and geminate (G) consonants are attested medially, all combinations of weight obtain, i.e. CVC light/heavy and CVG light/heavy. Two chief approaches have been entertained to account for these facts. The ‘weight’ approach treats geminates as underlyingly weightful (moraic) in contrast to weightless (non-moraic) singletons (e.g. Hayes 1989, Davis 1999, Topintzi 2010, among others). Singleton codas may acquire weight through Weight-by-Position (WbyP), but do not have to. Another approach views geminates as segmentally long, occupying two timing slots contra short singletons (e.g. Ringen & Vago 2011). Weight comes about as a result of WbyP, which, if applicable, affects codas (geminate and singleton) alike. None of the approaches captures all the patterns in (1). The former easily deals with (1.II-III) and the latter with (1.I-II), but both fail with regard to (1.IV): under the ‘weight’ approach, the weightlessness of CVG is paradoxical; under the ‘length’ approach, the asymmetrical weight across CVC and CVG is unexpected, given that codas should be uniformly heavy or light. To solve this problem, I present a model which argues that lexical geminates bear a mora underlyingly, whose association to higher and lower structure produces different effects. Integration under a syllable node produces weight; association to a lower node, i.e. a consonant, relates to phonetic length. Effectively, a consonant is interpreted as a geminate on the surface, due to its link to the underlying mora, which in turn, may end up prosodically integrated or unintegrated to a syllable, thus producing weightful and weightless geminates, respectively. Consonants that lack an underlying mora are interpreted as singletons. On the surface, these may acquire weight through WbyP. Importantly, such mora is an epenthetic one, a property visible to phonology, which can therefore always distinguish between singletons and geminates, as well as between underlying and derived-on-the-surface weight. I then build an OT analysis that captures the facts in (1) and briefly discuss the derived typology through it, as well as the co-existence of different types of geminates across different positions within the word.