Colloquium

We discussed Chapter 1 of Aspects. I claimed that the competence-performance distinction was orthogonal to Marr's levels. On my reading, the competence-performance distinction is fundamentally about investigating a phenomenon which is the interaction of multiple causal powers. Separating competence from performance is isolating a single causal power to study in the blooming and buzzing confusion, whereas Marr's levels tell us that we need to understand a single causal power at multiple levels of description.

Our next scheduled meeting time is easter, and so we will meet the week thereafter. Please have read the Miller and Chomsky paper, and have come up with 5 questions/comments etc.

We discussed Marr's 1977 paper. Next time we will read chapter 2 of Chomsky's Aspects.

During the discussion, we talked about levels in linguistics, and wondered whether linguistic theorizing really conforms to Marr's computational level. In the course of this discussion, I mentioned a recent paper by Mark Johnson, which argues that simplicity of high level computational theories (i.e. Chomsky's Merge) does not necessarily correspond to simplicity of their implementation, and therefore the evolutionary plausibility of Chomsky's intuition that Merge emerged in a single evolutionary step is low. Here is the link to that paper:

• Mark Johnson (2017) Marr's Levels and the Minimalist Program

We have decided to read papers on the relationship between grammar and parser. Provisionally, we will read the following:

• Marr (1977) Artificial Intelligence - A Personal View
• Chomsky (1965) Aspects of the theory of syntax chapter 1
• Chomsky & Miller (1963) Finitary Models of Language Users
• Bresnan & Kaplan (1982) Grammars as Mental Representations of Language
  • reply by Steedman
• Berwick & Weinberg (1984) The Grammatical Basis of Linguistic Performance chapter 2
  • reply by Stabler
  • reply to reply
• Phillips (2012) Parser Grammar Relations