Semantics

Lexical obsolescence is a topic whose study spans centuries, and yet it is not well-understood. Variously termed lexical mortality, lexical death, and lexical loss, among other names, the phenomenon has been described as both a product and a process, but the scholarship on how and why expressions go out of use has, until recently, been sporadic and sparse. The last few years have seen attempts to situate obsolescence among other processes of language change, but these have mostly focused on obsolescing constructions in modern languages. The present study, by contrast, investigates words that went obsolete in Late Middle English, suggesting a methodological approach designed to overcome the challenge of finding that which is no longer there, namely the consultation of a comprehensive online historical dictionary, and proposing an explanatory framework within the tradition of onomasiology and semasiology that positions obsolescence as a diachronic result of the habitual and contextually driven corporate deselection of linguistic constructions.

This experiment investigated the role of sentence meaning in auditory language comprehension. Tokens from a GOAT-COAT speech voicing continuum were embedded in carrier sentences that were biased toward either a "goat" or "coat" interpretation and presented to subjects for a word identification task. The identification function showed a boundary shift in favor of the biased context, and an interaction localized to the ambiguous boundary region. Response times were largest in the boundary region and the interaction between the two factors was localized to the boundary region and the voiced endpoint. There was also a response time advantage for context consistent responses specifically in the boundary region. These results and those of earlier research (Connine, 1987; Connine & Clifton, 1987) are described in terms of interactive activation of potential response categories by acoustic parameter and sentence context.

The aim of this project was to clarify the findings of Shapiro and Levine (1990) by exploring post-verb argument structure complexity effects. Three verb types, transitives, datives and obligatory three-place, were probed at four positions during an on-line sentence processing task that utilized cross-modal naming as the secondary task in a reaction time paradigm. No significant verb x probe interaction was found at any probe position with any of the three verb types. Two possible explanations are given for this pattern of results: (1) the choice of cross-modal naming as the secondary task; and, (2) the high variability of reaction times among subjects.

The base structures of Hutchins' deep case grammar are designed
in exclusively semantic terms. Part I of this study of the model
assesses its theoretical implications: (1) To compare the depth
of Hutchins' base structures with that of structures in other
models, a definition of grammatical depth is needed. The depth
of an abstract structure is defined in terms of the range of
similar surface structures derivable from it. (2) Hutchins
distinguishes some aspects of semantic study from the study of
general knowledge. Arguments for and against this dichotomy are
evaluated. (3) Hutchins' rules usually produce surface structures
so that early segments can be produced before the form of later
segments is fixed. Such part-by-part generation is seen as useful
in modeling performance. Part II attends to the grammar's partieulars.
A summary description is provided; critical comments are
made; and Hutchins' treatment of paraphrases involving causality
and incongruity ("although") is extended.