You are here: Home Events MPI Colloquium series - Folder MPI Colloquium Series - 2014 Ray Jackendoff, Tuesday, February 18

This content is archived, it could be outdated.

Ray Jackendoff, Tuesday, February 18

Aligning linguistics and psycholinguistics: The Parallel Architecture and theories of language processing


Ray Jackendoff
Center for Cognitive Studies, Department of Philosophy at Tufts University
http://ase.tufts.edu/cogstud/jackendoff/index.html


The Parallel Architecture is a framework for linguistic theory that I have been developing over the last 20 years, in which phonology, syntax, and semantics are independent generative systems that communicate via interface rules.  Among its fundamental questions is one familiar to psycholinguists:  What aspects of language are stored (and in what form), and what aspects are built on line (and, again, in what form)?  This framing of knowledge of language leads to a view in which there is no strong distinction between words, idioms, meaningful constructions, and phrase structure rules:  they form a continuum in the lexicon.  It also lends itself to a natural view of incremental language processing.

I am spending a month at the MPI in the hope of strengthening the ties between the Parallel Architecture and theories of language processing.  This talk will lay out some of the questions on which I seek your wisdom (if you know the answer) or your collaboration (if you don't).  For instance:  

1.  Many linguistic theories, including the Parallel Architecture, make essential use of the notion of inheritance, though which a more specific stored item inherits structure from a more general schema.  I would like to understand this notion better, in particular figuring out how inheritance is to be formalized and what good it does in processing, storage, and acquisition.

2.  In addition to the traditional links between phonology and syntax and between syntax and semantics, the Parallel Architecture also permits phonology and semantics to be linked directly, without syntactic intervention.  Such an interface can link linear word order and prosody to meaning, using principles like "Agent First" and "Prosodic constituent = Semantic constituent."  I would like to explore the possibility that such principles are active in ordinary language comprehension, giving rise to what have been called "perceptual strategies" and "good-enough parsing."  

Where and when:
15:45-17:00 Feb 18, 2014
Donders Centre for Cognitive Neuroimaging (Colloquium Room), Kapittelweg 29, 6525 EN Nijmegen

Street address
Wundtlaan 1
6525 XD Nijmegen
The Netherlands


Mailing address
P.O. Box 310
6500 AH Nijmegen
The Netherlands

Phone:   +31-24-3521911
Fax:        +31-24-3521213
E-mail:   


Public Outreach Officer
Charlotte Horn