Se-Kyung Kim
Murmur transfer and Extension of the morphological boundary>This paper deals with root-internal murmur transfer and extension of the
>root boundaries in Classical Sanskrit. The murmur in codas transfers to an
>onset, both locally and non-locally. The paper argues that the surface
>position of the murmur is determined by the ranking between the constraint
>that demands the murmur to be aligned to the right edge of the root
>(ALIGN(murmur, root, R)), and the morphological constraints on the domain of
>roots (I-ANCHOR-R(F)) and the contiguity of the root string (ALIGN(R, root,
>L, suffix)). I-ANCHOR-R(F) demands that the rightmost root feature be
>anchored to the rightmost root segment in the input and the output.
>Consequently, any suffix segments with a root feature are considered to be a
>part of the root. The local murmur transfer to the suffix is possible by
>this extension of the root boundary. When local process is impossible,
>non-local murmur transfer is derived by MAX(murmur) and ALIGN(R, root, L,
>suffix) that punishes interruption of the root string by a suffix segment.
>Dominance of ALIGN(R, root, L, suffix) over MAX(murmur) accounts for the
>murmur deletion when the non-local transfer would involve skipping a suffix.
>The proposed constraint ranking will account for the murmur transfer
>regardless of the number of murmur features present in the input or the
>linking status of the murmur.Graham Horwood
A Different Approach to Optional OptimaApproaches to optionality in OT have met with a number of conceptual challenges: constraint ties are notoriously prone to computational difficulties; adducement of optionality to equivalence in violation profiles hinges upon knowledge of the entirety of CON or results in a disturbing fragility of analysis; and approaches which refute optionality entirely place an undersiably large explanatory burden on the input. The present study will present a new attack on the optional optima problem, arguing that a differentiation constraint -- a constraint which renders candidates anharmonic with respect to one another -- may engender an n-ary split in the candidate set, allowing pure optionality without any of the aforementioned conceptual difficulties. The utility of this type of constriant will be demonstrated in analysis of various problems in syntax and phonology.
John Hale
Derivational and Non-derivational Approaches to German Scope-MarkingQuestion to be answered: What should the OT Syntax of German Scope-Marking look like? Should it be harmonic serialism [local optimization] or harmonic parallelism [global optimization]? Are these two actually notational variants?
Lisa Davidson
Hidden strata and graded performance in Optimality TheoryIn this study, I look at the performance of native English speakers on Polish word initial consonant clusters. Results show that the minimal sonority distance hypothesis cannot explain why some clusters have significantly better production rates than others. Instead, other factors which increase the markedness of a cluster, such as voicing and place of articulation, must be considered. This variable performance can be easily accounted for with an Optimality Theoretic ranking of "hidden" constraints.
Matt Goldrick
The Emergence of the Unmarked in Gender Agreement Systems
The typology of gender agreement systems as analyzed in Corbett (1991) suggests that within a language common forces are at work in syncretism and gender resolution. An OT analysis which utilizes a single markedness hierarchy provides a framework to understand this phenomena which is more constrained than previously rule based analyses. The approach is illustrated through an analysis of the crossed gender.
Markus Hiller
FEATURE APPROPRIATENESS AND TRANSCATEGORICAL PLACE ASSIMILATION
Division of labor between ranking and feature geometryWhile vowels may assimilate liprounding from the labial place of consonants (plain labial as opposed to liprounding) (1), the same kind of labial consonants may be perfectly transparent to liprounding harmony (2). Cases like (1) are known as _transcategorical assimilations.
(1) Tulu epenth. [I]~[u] in: a. kappu ``black'' b. kaTTI ``(?)''
(2) Turk.rd.harmony: a. mum+u ``candle-ACC'' b. kebab+I ``kabob-ACC''
The problems that these apparently contradictory pieces of evidence pose for autosegmental theory are resolved by Clements (1993) and Hume (1991) by representing [round] and [labial] by privative autosegments on the same tier (3), interpreted as [round] for each association from [V-Place] and as [labial] for each association directly from [Place].
(3) Place
| \
| V-Place
| /
labialSimilar relations were found for the pairs [front]/[coronal], [back]/[dorsal] and [RTR]/[radical]. Now, one important argument that was advanced in support of feature
geometric representation was _feature appropriateness_, e.g. [+/-anterior] is represented as dependent on [coronal] to capture that it only meaningfully subdistinguishes coronal place but not between any two noncoronal segments (Sagey 1986). In particular, it doesn't make any interpretable distinction among [front] vowels, which is a problem for the Clements/Hume strategy of representing [coronal] and [front] by like autosegments on the same tier. Hume (1994) argues that this actually was a virtue of this kind of representation and that universally fixing sub-place features of [V-Place|coronal] (``[front]'') to [-anterior] etc. explains changes of these subfeatures in coronal consonats before front vowels as assimilations. This conflicts with Hume's argument that vowel frontness can be assimilated even from [+anterior] coronal place of consonants e.g. in Maltese (4).
(4) /jV/- a. before cor.obs.: ji+drab b. otherwise: jo+ktor
(4a) does not become *[ji+dZrab] even though they (qua assimilation) share the autosegment that specifies [i] as [front] and [d] as [coronal], and [i] is not [+anterior]. That conflict can be resolved by representing [coronal] and [front] (as well as the other pairs of cognate place features) on two separate tiers interconnected by an autosegmental plane, transcategorical assimliation between them as driven by feature licensing constraints such as (5).
(5) LICENSE-PLACE: Place of articulation features are either realized in or released into [V-Place].
This correctly captures that transcategorical assimilations (a) are only C-Place to V-Place, not conversely (NiChiosa'in and Padgett 1993), (b) lack a universal directionality (leftward/rightward) and (c) do not universally require adjacency. By the same argument, change of coronal subfeatures before front vowels cannot be analyzed as driven by feature licensing but rather must be analyzed as driven by contextual markedness constraints that need not necessarily be referring to phonological features at all (cf. Pater 1996 for the same argument with obstruent voicing and sonorant voicing in *NC effects): The relevant characteristics of these effects are (a) that they are only triggered by coronal consonants before, not after, front vowels, (b) that the two segments involved have to be adjacent and (c) that there are other repair strategies such as backing the vowel while leaving the coronal consonant unchanged (6).
(6) /si/ --> [su] in Touhoku dialects of Japanese
In sum, transcategorical assimilations of place features are most insightfully analyzed as featurally driven, which requires a representational link in each of the pairs [round]/[labial], [front]/[coronal], [back]/[dorsal] and [RTR]/[radical]. However, this link does not involve representation of both features in each pair on the same tier, because the consonant place feature in some of these pairs has dependent features whereas the vowel features in these pairs don't. Hume's argument to the contrary, more concretely, that front vowels must have coronal sub-features because coronal consonants seem to assimilate them from vowels turns out to be unwarranted, since the constraints driving coronal consonant backing before front vowels do not even have to refer to features at all.
CLEMENTS, George N., 1993. ``Lieu d'articulation des consonnes et des voyelles: une the'orie unifie'e''. In: Laks, Bernard and Rialland, Annie (eds.), _Architecture des Repre'sentations Phonologiques_ , pp. 101--145. Paris: CNRS.
HUME, Elizabeth V., 1994. _Front Vowels, Coronal Consonants and their Interaction in Nonlinear Phonology_ . New York: Garland.
NiCHIOSA'IN, Maire and PADGETT, Jaye, 1993. Paper presented at HILP 1, Leiden.PATER, Joe, 1996. ``*NC''. NELS 27.