Keynote Address

The Adaptable Speaker: Speech Errors Reveal Implicit Learning in the Language Production System

Gary Dell

University of Illinois at Urbana-Champaign, USA

Abstract 
Joe Biden and Sigmund Freud are famous for slips of the tongue, Biden because people believe he often makes them, and Freud because he thought people like Biden are revealing their unconscious state when they slip. In thinking this, Freud was right in at least one respect: Speech errors do reveal something about the speaker's unconscious - their implicit knowledge of the sound patterns of their language. For example, Biden and all English speakers implicitly know that you cannot begin a syllable with the "ng" sound. And one's slips "know" this fact, too. An English speaker might say "hangmang" instead of "hangman," but would never say "ngangman." Because other languages do allow "ng" syllable onsets, we can conclude that the English speaker's belief that "ng" cannot be an onset is the result of learning.

This talk shows that this kind of learning can be altered in the laboratory and that slip behavior will change accordingly. If speakers produce syllables in the lab in which /f/ only occurs in syllable onsets, their slips will follow this new rule. The speakers have implicitly learned a new sound pattern. I will review a number of phenomena related to this finding and suggest a new theory of implicit learning in language production. In this respect, the talk will introduce the audience to the many upcoming sessions of the Psychonomic Society meeting that focus on psycholinguistics and the nearly as large number of sessions that address implicit learning. (It is unfortunate that these two kinds of sessions are sometimes in complementary distribution at the meeting.) The theory will also touch on issues discussed in sessions on animal learning, motor behavior, developmental psychology, and memory consolidation. 

About Gary

Gary Dell is Emeritus Professor of Psychology at the University of Illinois at Urbana-Champaign. He obtained his PhD in Psychology at the University of Toronto in 1980, and subsequently held academic positions at Dartmouth College and the University of Rochester, before coming to Illinois in 1989. He received the APA Early Career Award in 1987 and was elected to the American Academy of Arts and Sciences and the National Academy of Sciences in 2015.

Dell's research deals with how people produce, comprehend, and learn language, and how these processes can be modeled using neural networks. He has developed models that make predictions about the properties of everyday “slips of the tongue”. These predictions are tested using procedures in which participants produce words and sentences under controlled conditions. Also, by changing the model’s processing characteristics, it can produce the speech error patterns of various types of aphasic patients. Recent work has used voxel-based lesion-symptom mapping in aphasia to link model parameters to brain regions. Other recent research topics include: structural priming in production, the effect of driving on language use, monitoring and attention in production, optical imaging of the neural signal during sentence processing, prediction in sentence processing, memory for word co-occurrences, and production mechanisms for timing and intonation. In his most recent work, he developed a theory of learning within the production system that attempts to link accounts of speech errors, implicit sequential learning, memory consolidation, and critical periods for language learning.

View a list of past keynote speakers.

Questions?
Contact Member Services at info@psychonomic.org or +1 847-375-3696.
Member Services is open Monday - Friday from 8:00 a.m. - 6:00 p.m., U.S. Central Time.