Lyn Frazier was a plenary speaker at the 34th DGfS Meeting at Frankfurt University on March 9th. The title of her talk was "Processing Ellipsis: Explorations at the Edge of Grammar." (An abstract follows.) Former student Maribel Romero was another plenary speaker, her talk was titled "From Form to Meaning."
In addition, Anisa Schardl gave a talk entitled "Finnish - hAn and the QUD" and former students Maria Beizma and Kyle Rawlins gave a talk entitled "Or What?"
For more information, see http://dgfs.uni-frankfurt.de/dgfs/dgfs_de.html
Processing Ellipsis: Explorations at the Edge of Grammar
It will be suggested that stringent grammatical conditions on ellipsis should be maintained, requiring syntactic matching of antecedent and elided constituent in the case of Verb Phrase Ellipsis, certain morphological features aside, (Sag, 1977 a.o.), and requiring movement of the ‘answer’ to a Focus projection with ellipsis of the TP in the case of fragment answers to questions (Merchant, 2004). However, speakers at times erroneously produce utterances that violate these stringent conditions, e.g., producing a syntactic blend of an antecedent from one utterance and an elided constituent appropriate for a closely related (competitor) utterance. These ungrammatical utterances can be repaired under the same conditions as garden-path misanalyses, namely, when few repair operations are needed and there is lots of evidence for those operations. The acceptability of such utterances depends on the difficulty of the repair and on whether the input sounds like a familiar form, possibly an error, that might be produced by normal sentence production mechanisms. Evidence supporting this account suggests that, in addition to the general form-meaning pairing characterized by the grammar, there is also a performance-based pairing of forms and meanings that is token-based, rather than type-based; this performance pairing results from implicit knowledge of the performance systems as well as grammar. The performance pairing of form and meaning is not tied to ellipsis per se, and thus its existence has implications not only for our understanding of ellipsis but more generally including, for example, the interpretation of quantifiers. The account suggests that certain problematic examples do not require us to complicate our grammatical theory but instead are appropriately explained in the realm of systematic language performance.
