The Architecture of Cognition: Rethinking Fodor and Pylyshyn's Systematicity Challenge

1. Systematicity: An Overview

Introduction In 1988, Jerry Fodor and Zenon Pylyshyn published “Connectionism and Cognitive Architecture: A Critical Analysis.” Their article presented a forceful and highly influential criticism of the explanatory relevance of neural network models of cognition. At the time, connectionism was reemerging as a popular and exciting new field of research, but according to Fodor and Pylyshyn, the approach rested on a flawed model of the human mind. Connectionism is the view that the mind can be understood in terms of an interconnected network of simple mechanisms. Its proponents contend that cognitive and behavioral properties can be modeled and explained in terms of their emergence from the collective behavior of simple interacting and adaptive mechanisms. According to Fodor and Pylyshyn, connectionist approaches neglect an essential feature of thought—its systematic nature. On their view, the basic psychological fact that thoughts are intrinsically related to other thoughts in systematic ways becomes inexplicable if one denies that representations are structured in a syntactically and semantically classical combinatorial manner.1 Connectionism, they argued, inevitably fails to provide a meaningful explanation of cognition insofar as it confuses the intrinsically systematic nature of thought with a system of associations.2 Connectionism might shed some light on the way that cognitive architectures happen to be implemented in brains, but the explanation of cognition does not take place at the level of biology or hardware. A cognitive architecture must be systematic to the core in order to shed light on the intrinsically systematic character of cognition. One prominent message of their article, that a cognitive architecture must explain systematicity in order to explain human cognition, came to be called the systematicity challenge. The meaning and implications of 1 Systematicity: An Overview John Symons and Paco Calvo 4 John Symons and Paco Calvo the challenge can be interpreted in more than one way, and it quickly generated a vigorous response in philosophy and cognitive science. The ensuing debate resulted in an enormous literature from a variety of perspectives .3 Philosophers and scientists on all sides of the issue generally agree that the paper helped to sharpen central questions concerning the nature of explanation in cognitive science and that Fodor and Pylyshyn have encouraged the scientific community to think carefully about what the goals of cognitive science should be. Nevertheless, in cognitive science and philosophy, opinion is sharply divided concerning the role of systematicity in a mature science of mind. The criticism of connectionism in the 1988 paper harks back to Fodor’s earlier arguments for the idea that cognition should be understood by analogy with classical computational architectures as a system of rules and representations. In the 1970s, Fodor had argued that the language of thought hypothesis explained the systematic features of thought. On his view, all thoughts that have propositional content are representational in nature, and these representations have syntactical and semantic features that are organized in a way that is similar to the transformational rules of natural languages. Insofar as tokens of mental representation figure in thoughts and insofar as we can judge those thoughts to be true or false, they must be organized in a language-like way (Fodor 1975). Fodor presents the language of thought hypothesis as the best way to account for a range of features of our psychology. The three basic explananda that Fodor highlights in his work are: (a) The productivity of thought: we have an ability to think and understand new thoughts and previously unheard sentences. (b) The systematicity of thought: to genuinely understand a thought is to understand other related thoughts. (c) The principle of compositionality: the meaning of sentences results from the meanings of their lexical parts. Fodor and Pylyshyn’s criticism of connectionism is shaped by Fodor’s early articulation of the language of thought hypothesis and by their view that competing explanatory strategies miss what is distinctively cognitive about cognition. Their 1988 article applied a challenging philosophical argument to a lively and ongoing scientific controversy. By any measure, their paper has served as a focal point for one of the most active debates in the philosophy of cognitive science over the past twenty-five years. In our view, the scientific landscape has changed in ways that call for a fresh look at this influential set of arguments. Most obviously, the Systematicity: An Overview 5 quarter-century since Fodor and Pylyshyn’s paper has seen the development of new approaches to connectionism that depart in a...