Projects at the Language Production Lab

Language Production Lab logo

Please see the Publications page for a more complete list of papers, sorted by year and categorized by project.

SP Structural Priming in Language Production

Structural priming is a tendency to recreate a recently uttered syntactic structure in different words. It can be seen even when words and thematic roles change from one utterance to another. We are exploring an explanation of this kind of priming in terms of implicit learning, using experimental methods and computational models in our work.

AP Agreement Processes in Language Production

Agreement is a paradigm case of what syntax does. Relying on features like number and gender, agreement ties together different words (e.g., verbs and their subjects; pronouns and their antecedents) that reflect the same thought. We are developing an account of how agreement is implemented in spontaneous speech. The primary goals are to explain normal agreement and agreement errors within a uniform framework, to explain differences in the agreement properties of verbs and pronouns, and to explain cross-language and cross-dialect differences in how agreement works.

LP Lexical Processes in Language Production

The mental lexicon is the repository of much of what speakers know about their language, and a lot of our work is concerned directly or indirectly with how the lexicon is accessed and used for relating meanings to sounds in language production. Here we list a small sample of papers relating to topics of interest in this area.

ET Eyetracking During Language Production

Psycholinguistic interest in language production has often been stymied by the difficulty of getting access to the input to production processes. To begin to attack this problem, we are examining the relationship between eye movements over scenes and the features of language used to convey information about the scenes.

PE Phonological Encoding in Language Production

To produce a word, one must retrieve its pronunication from the mental lexicon and adjust that pronunication so that it fits its context. This process of retrieval and adjustment of speech sounds during production is phonological encoding. We study phonological encoding by analyzing everyday speech errors, by carrying out psycholinguistic experiments, and by developing connectionist models that mimic the findings of the analyses and experiments.

AL Aphasia and Language Production

Aphasic patients have a great deal of difficulty retrieving words when they are speaking. The particular errors that patients make (called "paraphasias") are quite similar to the errors that nonaphasic individuals make when speaking; the patients' errors are just far more numerous. We have been applying connectionist models of normal language production to paraphasias. By "lesioning" the parameters of models of normal lexical retrieval we can reproduce the error patterns exhibited by fluent patients. The "aphasia modelling project" link (under construction) allows researchers to apply our models to aphasic picture naming and repetition error data.

OG Other/General
Webmaster: Matt Rambert <rambert@uiuc.edu>

Last modified: July 2004