Call for Papers

Fifth North American Phonology Conference

Concordia University, Montreal

Theme: Phonology as Symbolic Computation

Invited speakers:
Andries Coetzee, Michigan
John Kingston, UMass
David Odden, Ohio State
Bridget Samuels, Harvard

Recent work in phonology has met with a number of recalcitrant problems.

1. Probabilistic and exemplar-based models of phonological learning and phonological computation have failed to deal with the same conceptual and empirical challenges that led to the demise of their empiricist, behaviorist forebears.

2. Objections against the computational complexity associated with derivations with multiple levels of representation have turned out to have been ill-grounded, and stubborn problems of analysis have forced 'two-level' theorists to allow complex derivations to sneak back in, as in the Stratal, Harmonic Serialism and Candidate Chains models of recent work in Optimality Theory.

3. The grounding of constraints in markedness 'theory' remains an elusive goal that fails on both logical and empirical grounds to provide explanations.

In this context, we invite papers on the prospects of future research in Good Old Fashioned Phonology (GOFP, an adaptation of Haugeland's Good Old Fashioned Artificial Intelligence). That is, we propose an exploration of phonology as a substance-free, symbolic computation system. Papers critiquing GOFP are also very welcome.

A substance-free theory considers the formal properties of a grammar without regard for transduction between symbols in the grammar and the input and output systems involved in language acquisition and use. Relevant sources for this position in phonology and elsewhere are Hjelmslev and Uldall (see Fudge 2006:88), Chomsky and Halle (1968), Kaplan 1987, Hale and Reiss (2000, 2008) and Pylyshyn (2003).

Formal topics might include the use of quantifiers or operator-variable structures, computation of locality, computational power of phonological grammars, and formal grammar and biolinguistic considerations.

Abstracts should be sent in pdf format to cogsci AT alcor.concordia.ca, up to 3 pages in length. Anonymity is not required.

Organizers:
Concordia Linguistics Program »
Concordia Linguistics Student Association »
Concordia Cognitive Science Group »
Computational Linguistics at Concordia »