|Title:||Maximizing First Positions|
|Comment:||This paper was published in Caroline Féry, Antony Dubach Green & Ruben van de Vijver (eds.), Proceedings of HILP5, 1-19, University of Potsdam|
|Abstract:||In this paper I discuss some possible extensions of Beckman\'s (1998) idea of prominence maximization. I argue that prominence maximization can be the trigger for seemingly unrelated phenomena as certain types of metathesis, the existence of monosyllabic templates and certain types of vowel deletion.
Differently from Beckman\'s proposal, the prominence maximization constraints I use are defined on the output only. In this form they are more similar to the family of COINCIDE constraints proposed by Zoll (1996, 1998) to account for \'positional markedness”. Zoll\'s constraints require certain marked output structure to appear in specified prominent positions in the output and thus account for cases where marked structure seems to be restricted to prominent positions. I will maintain Zoll\'s constraint format, but replace \'marked output structure” with \'every element”:
Output oriented prominence maximization:
COINCIDE-P: every element of the output is in P
(P = some prominent position).
The paper shows that one member of this constraint family, COINCIDE-ONSET-S1, can trigger metathesis to a prominent position such as the onset of the first syllable, as is the case in the Sardinian dialect of Sestu Campidanian (Bolognesi 1998).
Best satisfaction of another member of the constraint family, COINCIDE-S1 (requiring all segments to surface in the first syllable) is achieved if only the first syllable and nothing else is realized: in this case every element of the output is in initial position, since there is nothing but an initial position. This happens when COINCIDE-S1 dominates the anti-deletion constraint MAX. Under the ranking COINCIDE-S1 >> MAX all material that does not fit into the first syllable (and thus violates COINCIDE-S1) is deleted. The effects of this ranking show up in templates. Thus, the advantage of a monosyllabic reduplicant is that it ensures that every segment of the reduplicant is realized in an initial position, i.e. the first syllable of a morpheme. Monosyllabic reduplicants therefore are the direct consequence of the constraint COINCIDE-S1, which here acts as a size-restrictor constraint. A similar analysis can account for monosyllabic templates in truncation. Yet more dramatic consequences of COINCIDE-S1 can be observed in input-output relations when the constraint dominates MAX-IO. In this case all morphemes of a language consist of a single (hence prominent) syllable. This is the case in the majority of native German roots (and affixes).
Finally, I argue that constraints such as COINCIDE-(ONSET)S1 can receive a psycholinguistic grounding: first syllables are important for word-recognition since they are perceived first and more material in initial positions guarantees faster recognition of a word (s. Bruner & O\' Dowd 1958, Fay & Cutler 1977, Nooteboom 1981, Nooteboom & Vermeulen 1997, among others).