Collaborations with John Tienson
The
Spindel Conference 1987: Connectionism and the Philosophy of Mind, Southern Journal of Philosophy, Supplement (with John Tienson), 1988.
Settling into a New Paradigm (with J. Tienson), Spindel Conference 1987: Connectionism and the Philosophy of Mind, Southern Journal of Philosophy 26, Supplement (1988), 97-114. Reprinted in T. Horgan and J. Tienson (eds.), Connectionism and the Philosophy of Mind (Kluwer, 1991).
We argue (1) that an adequate account of cognition requires complex, syntactically structured representations that are processed in ways sensitive to content and structure; and (2) that cognitive processing is neither driven by, nor describable by, exceptionless rules. Standard cognitive science rightly assumes point (1), but wrongly assumes that the requisite processing must involve exceptionless, programmable rules. Connectionists have not yet produced models with the features we describe, but there is reason to hope that a version of connectionism can occupy this neglected region in the logical space of views about cognition.
Representations without Rules (with J. Tienson), Philosophical Topics 17 (1989), 27-43.
We argue (1) that in order for a cognitive system to deal successfully with its environment (as organisms do), it needs syntactically structured mental representations that interact with each other in ways that depend upon their structure and content; and (2) that the relevant kind of processing is not describable by exceptionless rules statable over the representations themselves. Connectionism, we argue, appears to have the potential for showing how there can be cognitive processing that is structure sensitive and content sensitive without being rule governed.
Connectionism and the Kuhnian Crisis in Cognitive Science (with J. Tienson), Acta Analytica 6 (1990), 5-17.
Soft Laws (with J. Tienson),
According to the standard philosophical conception of scientific law, proper laws permit no exceptions within their own domain. We argue, to the contrary, that proper laws in intentional psychology virtually always have exceptions within their own domain--psychological exceptions, as opposed to "had a stroke" or "hit by a bus" exceptions. These soft laws, as we call them, have ineliminable ceteris paribus clauses. The argument draws on a critique of classical cognitive science, which embodies the standard conception of law. We also argue that, given the proper understanding of law, psychological explanation requires laws.
Connectionism and the Philosophy of Mind (with J. Tienson). Kluwer, 1991.
Cognitive Systems as Dynamical Systems (with J. Tienson), Topoi 11 (1992), 27-43.
Connectionist networks are naturally viewed, conceptually and mathematically, as physical realizations of high-dimensional dynamical systems. Drawing on this mathematical framework for connectionism, we set forth a conception of cognitive systems as physically realizable dynamical systems. We focus on ways of conceptualizing and incorporating, within the dynamical systems framework, mental representations with complex syntactic structure, and cognitive processing that is sensitive to that structure. We also address some important ways in which this conception of cognitive systems differs from the conception embodied in classical cognitive science.
Structured
Representations in Connectionist Systems? (with J. Tienson), in S. Davis., ed., Connectionism: Theory and Practice (
Levels of Description in Nonclassical Cognitive Science (with J. Tienson), Philosophy 34 (1993), Royal Institute of Philosophy Supplement, 159-88.
Representations Don't Need Rules: Reply to James Garson (with J. Tienson). Mind and Language 9 (1994), 38-55. Invited reply to Garson's critique of item 35.
We have argued that cognitive systems require syntactically structured representations, but that these representations do not undergo processes which conform to programmable rules. Garson argues that nothing can determine the content of a physical state other than the rules to which it conforms. We grant that representations must have content appropriate causal roles, but argue that causal roles do imply regularities or rules. Representations of the precise locations of nearby objects are never exactly repeated, and hence do not participate in any regularities. Furthermore, any given such representation might help determine indefinitely many different bodily motions (depending upon current aims)--rendering programmable rules governing such representations impossible.
A Nonclassical Framework for Cognitive Science (with J. Tienson), Synthese 101 (1994), 305-345. Issue on philosophy and connectionism.
David Marr provided a useful framework for theorizing about cognition within classical AI-style cognitive science, interms of three levels of description: the levels of (1) cognitive function, (2) algorithm, and (3) physical implementation. We generalize this framework: (1) cognitive state transitions, (2) mathematical/functional design and (3) physical implementation or realization. Specifying the middle, design level to be the theory of dynamical systems yields a nonclassical, alternative framework that suits (but is not committed to) connectionsim. We consider how a brain's (or a network's) being a dynamical system might be the key both to its realizingh various essential features of cognition--productivity, systematicity, structure-sensitive processing, syntax--and also to a nonclassical solution of frame-type) problems plaguing classical cognitive science.
Connectionism and the Commitments of Folk Psychology (with J. Tienson),
Philosophical Perspectives 9 (1995),
127-52.
William Ramsey, Stephen Stich, and Joseph Garon argue that folk psychology (FP) is committed to a feature of mental states they call "modularity," that this feature is incompatible with certain connectionism models, and hence that those models support eliminativism. We offer three replies. First, FP is clearly committed only to a weak form of modularity that connectionism accommodates easily. Second, even if FP were committed to other forms of modularity not manifested in human cognition, this would not mean that FP is so radically false that there are no beliefs and desires. Third, connectionism actually can accommodate various other forms of modularity anyway.
Connectionism and the Philosophy of Psychology (with J. Tienson), Bradford/M.I.T., 1996.
Precis of Connectionism and the Philosophy of Psychology (with J. Tienson), Philosophical Psychology 10 (1997), 337-56. Symposium on T. Horgan and J. Tienson, Connectionism and the Philosophy of Psychology (MIT Press).
Connectionism was explicitly put forward as an alternative to classical cognitive science. The questions arise: How exactly does connectionism differ from classical cognitive science, and how is it potentially better? The classical "rules and representations" conception of cognition is that cognitive transitions are determined by exceptionless rules that apply to the syntactic structure of symbols. Many philosophers have seen connectionism as a basis for denying structured symbols. We, on the other hand, argue that cognition is too rich and flexible to be simutable by the exceptionless representation-level rules that classicism requires. However, this very richness of cognition requires syntactically structured representations--what philosophers call a language of thought. The natural mathematical characterization of neural networks comes from the theory of dynamical systems. We propose that the mathematics of dynamical systems, not the mathematics of algorithms, holds the key to how cognitive structure and cognitive processes can be realized in the physical structure and processes of a network.
Authors' Replies (with J. Tienson). Acta Analytica 22 (1998), 275-87. Issue containing proceedings from the 1997 conference on Horgan and Tienson's Connectionism and the Philosophy of Psychology.
Resisting the Tyranny of Terminology: The General Dynamical Hypothesis in Cognitive Science (with J. Tienson). Behavioral and Brain Sciences 21 (1998), 643. Invited commentary on T. van Gelder, “The Dynamical Hypothesis in Cognitive Science.”
Short Precis of Connectionism and
the Philosophy of Psychology (with J. Tienson). Acta Analytica 22 (1998): 9-22.
Rules (with J. Tienson). A Companion to Cognitive Science (Blackwell, 1998), 660-70.
Rules and Representations (with J. Tienson). The MIT Encyclopedia of Cognitive Science (MIT, 1999), 724-26.
Deconstructing
New Wave Materialism (with J. Tienson). In B. Loewer, ed., Physicalism and Its Discontents.
The Intentionality of Phenomenology and the Phenomenology of
Intentionality (with
J. Tienson). In D. Chalmers (ed.), Philosophy of Mind: Classical and Contemporary
The Phenomenology of Embodied Agency (with J. Tienson). In J. Joao Saagua
(ed.), Mind and Action III, in press.
(In Portugese translation)
Connectionism (with J. Tienson), invited for the Dictionary of Philosophy of Mind, online web resource, Washington University St. Louis.
Paper on the cognitive architecture debate in cognitive science (with J. Tienson),
invited for a Blackwell volume on philosophy of mind edited by S. Stich and T. Warfield.
