On parsers and grammars
Carlos Acuña-Fariña · 2023
This chapter focuses on the all-important issue of whether grammars and parsers are two different things, or just the same thing in different time frames. It starts by arguing that this issue can make sense only if one assumes the merits of psychological adequacy (basically the idea that our theories in linguistics must refer to ‘real’ processes in the mind). This is where well-known ‘divergencies’ in the world of linguistics come into play: Chomsky, Jackendoff, Bever, language-for-thought, language-for-communication, isolable syntax, syntax first, processing heuristics before grammar, grammar before processing heuristics, both simultaneously and opportunistically, etc. The chapter touches on important developments in the cognitive science world of the 1970s (the derivational theory of complexity ‘failure’ and the emergence of the notion of processing heuristics) and discusses current work on islands (where we seem to see grammar being obeyed in the face of stringent processing challenges) and grammatical illusions (where we seem to see grammar being ignored, often when facing easy processing problems). It also discusses currently fashionable models of good-enough processing and related goal-directed proposals, such as surprisal theory, predictive coding models and noisy channel models.