Logic, Neural Networks, and Optimality Theory

Workshop Berlin, July 23-25, 2003
Jägerstraße 10-11, 10117 Berlin (Mitte), Room 006
Organizers: Anton Benz (Berlin), Reinhard Blutner (Amsterdam), Helen de Hoop

The primary aim of this workshop is to talk about foundational issues of OT including the connection to connectionism and some aspects of evolutionary grounding and stochastic OT. The secondary intention of this workshop is to collect and to discuss ideas for a bigger project proposal. The organization of this workshop was financially supported by the ZAS project group "Bidirectional Optimality Theory" (Berlin), the PIONIER Project "Case Cross-Linguistically" (University of Nijmegen), and the NWO preliminary grant "Logic, Neural Networks, and Optimality Theory" (University of Amsterdam).
Possible themes (this is not an exhaustive list):
(A)  Founding aspects of OT within a connectionist setting
  • Symbolism as high-level description of neural networks
  • Bidirection, recoverability, and compositionality
  • Localist interpretation of neuronal networks and their logic.
  • Towards a logic of neural networks (?). The binding problem
(B)  Founding aspects of OT within a evolutionary setting
  • Evolutionary grounding of OT constraints
  • Variants of Bidirection. Motivation and foundation (learning, evolution)
  • The evolution of discourse
(C)  Stochastic OT
  • Gradedness. Harmonic Grammar vs. stochastic OT vs. OT with partial constraint ordering.
  • Strictness of dominance. Apparent and real counterexamples.Motivation (learning? automaticity). Neural underpinning.
  • Stochastic OT, bi-directional learning, evolution.

Invited Participants:

David Beaver
Paul Boersma
Paola Escudero
Hartmut Fitz
Hans-Martin Gärtner
Peter beim Graben
Silke Hamann
Petra Hendriks
Gerhard Jäger
Geraldine Legendre
Jason Mattausch
Yukiko Morimoto
Ernst Odolphi
Robert van Rooy
Anette Rosenbach
Katrin Schulz
Paul Smolensky
Henriette de Swart
Ralf Vogel
Henk Zeevat