Philosophy Connectionism
by
Kenneth Aizawa
  • LAST REVIEWED: 02 September 2022
  • LAST MODIFIED: 11 January 2017
  • DOI: 10.1093/obo/9780195396577-0153

Introduction

“Connectionism” in its most generic sense describes theories that postulate interconnected networks of simple neuron-like information-processing elements (often called “nodes”) with modifiable interconnections (often called “weights”) to explain cognitive processes or their implementation. Although the term was used in the first half of the 20th century to describe theories of the neurobiological implementation of principles of associationism, most recent philosophical work on connectionism has focused on cognitive scientific research that uses computers to simulate the activities of artificial neural networks. Often this work goes under the rubric of “parallel distributed processing” (PDP). Typically, the neural network simulations consist of presenting a network with data in order to have the network modify the connections between the processing elements in such a way as to enable a network to compute a desired function. Although advocates of connectionism often insist on the “neurobiological inspiration” of their models, this does not always translate into models that are in all respects neurobiologically plausible. In many cases, the models are, in the first instance, developed with an eye to accounting for one or another psychological phenomenon. (There are, however, those who study connectionist models as a kind of computational apparatus of interest or value apart from psychology.) In these cases, advocates of connectionism often advance their models as alternatives to computational models of cognition. Since computational models typically invoke computer programs that manipulate syntactically and semantically combinatorial representations, much of the philosophical discussion of connectionism has focused on questions of representation, the structure of representations, the existence of (explicit) rules for the manipulation of representations, and the nature of computation. Although research on connectionism is an extremely active area of cognitive science, this article is largely, and somewhat artificially, limited to works by philosophers. Those wishing to conduct more serious research on connectionism will have to delve into the connectionist scientific literature.

General Overviews

Numerous general overviews of connectionism are available. These overviews vary in the level of technical detail they offer regarding the apparatus of nodes, weights, and weight change procedures. There are also differences in the audiences to which they are directed, such as psychologists, philosophers, or computer scientists. Bechtel and Abrahamsen 2002 provides the most robust attention to both technical detail and philosophical issues. McClelland and Rumelhart 1987 and Rumelhart and McClelland 1987 provide a technical introduction directed primarily toward psychologists. There is also a companion volume, McClelland and Rumelhart 1989, with software for running connectionist simulations. Ellis and Humphreys 1999 and Dawson 2005 provide extensive introductions suitable for an undergraduate or graduate psychology course in connectionism. Smolensky 1988, Rumelhart 1989, Goldblum 2001, and Garson 2007 offer article-length introductions to connectionism with divergent levels of technical detail.

  • Bechtel, William, and Adele A. Abrahamsen. Connectionism and the Mind: Parallel Processing, Dynamics, and Evolution in Networks. 2d ed. Malden, MA: Blackwell, 2002.

    A detailed, well-informed introduction to both the technical dimensions of connectionist modeling and some of the relevant philosophical issues that this modeling raises.

  • Dawson, Michael R. W. Connectionism: A Hands-On Approach. Malden, MA: Blackwell, 2005.

    DOI: 10.1002/9780470694077

    An extensive textbook covering familiar types of connectionist networks primarily from the 20th century.

  • Ellis, Rob, and Glyn W. Humphreys. Connectionist Psychology: A Text with Readings. Hove, UK: Psychology, 1999.

    A textbook with reading directed primarily to psychologists.

  • Garson, James W. “Connectionism.” In The Stanford Encyclopedia of Philosophy. Edited by Edward N. Zalta. Stanford, CA: Stanford University Press, 2007.

    A brief overview of the structure and capabilities of current connectionist networks, along with some discussion of connectionist representation, systematicity, state space semantics for connectionist networks, and the elimination of folk psychology.

  • Goldblum, Naomi. The Brain-Shaped Mind: What the Brain Can Tell Us about the Mind. Cambridge, UK: Cambridge University Press, 2001.

    DOI: 10.1017/CBO9780511612749

    A short, informal introduction that does not use any of the mathematical apparatus typical of introductions to connectionism.

  • McClelland, James L., and David E. Rumelhart, eds. Parallel Distributed Processing. Vol. 2, Psychological and Biological Models. Cambridge, MA: MIT, 1987.

    This is one-half of a two-volume collection of essays that constitute the principal source of scientific input into philosophical discussions of connectionism.

  • McClelland, James L., and David E. Rumelhart. Explorations in Parallel Distributed Processing: A Handbook of Models, Programs, and Exercises. Cambridge, MA: MIT, 1989.

    A companion volume to McClelland and Rumelhart 1987 and Rumelhart and McClelland 1987, explaining many of the technical details needed to run computer simulations of connectionist networks. It also includes computer software with which to run the simulations.

  • Rumelhart, David E. “The Architecture of Mind: A Connectionist Approach.” In Foundations of Cognitive Science. Edited by Michael Posner, 133–159. Cambridge, MA: MIT, 1989.

    A relatively brief overview of some of the principal ideas in the “parallel distributed processing” approach to connectionism.

  • Rumelhart, David E., and James L. McClelland, eds. Parallel Distributed Processing. Vol. 1, Foundations. Cambridge, MA: MIT, 1987.

    This is one-half of a two-volume collection of essays that constitute the principal source of scientific input into philosophical discussions of connectionism.

  • Smolensky, Paul. “On the Proper Treatment of Connectionism.” Behavioral and Brain Sciences 11.1 (1988): 1–23.

    DOI: 10.1017/S0140525X00052432

    A widely cited account of what connectionism is all about.

back to top

Users without a subscription are not able to see the full content on this page. Please subscribe or login.

How to Subscribe

Oxford Bibliographies Online is available by subscription and perpetual access to institutions. For more information or to contact an Oxford Sales Representative click here.

Article

Up

Down