A draft by
Nigel J.T. Thomas
To email me, click here
California State University Los Angeles.
This was originally written and presented at the National Endowment for the Humanities Summer Seminar for College Teachers on Folk Psychology vs. Mental Simulation: How Minds Understand Minds, run by Robert Gordon at the University of Missouri - St. Louis, June-July 1999. It has been only lightly revised since, and should be considered a rough draft. Needless to say, the ideas herein owe a lot to what I learned at the seminar from Robert Gordon and the other participants, particularly Jim Garson. However, any errors are my responsibility alone.
One reason that many people found Robert Gordon's (1986) "simulation" theory of folk psychology particularly interesting, when it first appeared, was that it seemed to undermine a very influential argument for eliminative materialism: the argument of the Churchlands (e.g. P.M. Churchland, 1979, 1981; P.S. Churchland, 1988; and cf. Feyerabend, 1963) to the effect that folk psychology (on which our common sense mentalistic ontology allegedly depends) is a theory of mind that is very long in the tooth, and that is thus ripe for wholesale replacement by a more up-to-date scientific theory in a Kuhn style scientific revolution. It seemed that, if Gordon was right about folk psychology, then it was not a theory at all, and thus not a suitable candidate for theoretical replacement. However, an important article by Stich & Ravenscroft (1994) brought this idea into question. They distinguished several conceptions of "folk psychology" which, they claimed, were often conflated, and on this basis they concluded (amongst other things) that the perceived conflict between simulation theory and eliminativism was more apparent than real. Although this argument has been very influential, simulation theory, though it remains a minority view, has continued to flourish and arouse interest for its other merits (empirical and conceptual), and indeed, has by now divided into several variants. In what follows I shall argue that when we consider one of these variants, the "radical simulationism" being developed by Gordon himself, and when we distinguish between two different, and perhaps only contingently linked, aspects of the eliminativist program, then the conclusions of Stich & Ravenscroft are no longer secure. Radical simulation theory is not, after all, compatible with out-and-out ontological eliminativism. However, it sits rather well with other aspects of the eliminativist program, and thereby shifts the terrain of the mind-body problem itself into largely unexplored, but perhaps very promising, areas.
Stich & Ravenscroft (1994) list six possible "theory-theory" interpretations of the concept of "folk psychology". The first two are what they call "external accounts", meaning, it would seem, views about the sorts of things that ordinary people are inclined to say, or assent to, about the mind. Thus "folk psychology" might be understood as denoting:
 "The set of folk psychological 'platitudes' that people readily recognize and assent to,"or
 "A theory that systematizes the folk psychological platitudes in a perspicuous way."The remainder are examples of what they call "internal accounts", which they explain as construing "folk psychology" as referring to "An internally represented knowledge structure used by the cognitive mechanism underlying our folk psychological capacities". Under this heading they distinguish versions of theory-theory according to how the information embodying the putative theory might be stored:
 "In a mental model or a connectionist network that maps onto a set of propositions in a unique and well motivated way,"
 "In a system consisting of rules and sentence-like generalizations,"
 "In a system consisting only of rules,"or
 "In a mental model or connectionist network that does not map onto a set of propositions in a unique and well motivated way."According to Stich & Ravenscroft, eliminativist arguments, to the effect that folk psychology is a radically false theory, cannot be coherently applied to  and , because these putative embodiments of folk psychology make no assertions and thus cannot turn out to make false assertions.
Of course, there are several brands of eliminativism. Some eliminativists, for example Rorty (1980), Dennett (1988, 1991), and Wilkes (1988, 1995), are principally concerned to eliminate phenomenal states, or the notion of consciousness, from our scientific ontology. However, that brand does not seem to be the most relevant here. Stich & Ravenscroft seem concerned with those eliminativists (most notably the Churchlands) who mainly seek to persuade us that a proper explanation of the human mind does not require us to make a commitment to the real existence of propositional attitude states. As the Churchlands see things, present day external folk psychology (i.e., the platitudes, the things the folk say about the mind) is committed to the reality of the attitudes, but the prospects for a materialist reduction of the attitudes, of somehow integrating them into the ontological scheme of physical science, seems remote. Therefore, in order to vindicate physicalism, they seek to argue that such folk psychology is (in this respect, at least) a radically false theory that is ripe for replacement by some new theory of the mind that, they hope, will be both more accurate and less ontologically problematic. Such eliminativism is directed, in the first place, at "external" folk theories of type  or . This is where the big philosophical payoff would be: in effect, a solution to the mind-body problem (or a large part of it(1)).
Indeed, trying to eliminate "internal" folk psychological theories (, , , ) through philosophical argument would seem to be a quixotic and quite forlorn project(2). Thus, when eliminativists concern themselves with views about "internal" folk theories, they generally, and quite properly, focus not so much on the content represented in the putative knowledge structures (as Stich & Ravenscroft seem to think), but on the form of whatever is represented. It is not that eliminativists suspect that "internally represented" folk psychologies might contain or imply or embody false generalizations about human behavior. Indeed, they would not be satisfied if any such false generalizations could somehow be replaced with true ones (even ones couched in neurological terms). Rather, (Churchland style) eliminativists object to scientific hypotheses that posit knowledge structures of certain forms. They object to scientific theories that posit discrete representations of propositional attitudes, and they do so primarily because they plausibly think that these, if shown to be true, will serve to vindicate the ontological commitment to the reality of attitudes allegedly displayed in current "external" folk psychologies ( or ).(3)
Stich & Ravenscroft would seem to be right (if not altogether for the right reasons) that simulation theory does not directly threaten the eliminativist project in the way that was once thought. As they read it, simulationism is an "internalist" theory - a theory of how we actually predict and explain each others behavior - and leaves the "external" question, the question of the truth of what the folk say about each others psychological processes, quite open. They argue that, although simulationism undercuts internalist eliminativism, it leaves the way open for externalist eliminativism, which, as I have argued, is actually the most philosophically radical version of the doctrine. Indeed, simulationism might even seem to give some quite strong support to externalist eliminativism. If what we really do to predict and understand people's behavior is to simulate, yet (as has been standardly assumed) the folk talk about the mind in terms of practical reasonings about propositional attitudes, then the eliminativists case is (at least partly) made: the "external" folk theory of the mind (as standardly understood) is false. Admittedly this is not quite so clear in the case of a simulationist like Goldman (1992), since he appears to countenance the view that the decision making system itself, the system that is run "offline" when we simulate, works by practical reasoning applied to propositional attitudes. However, the more radical version of simulation developed in the more recent writings of Gordon (1995, 1996, 2000) [and Heal?] seems to deny this picture, and promotes a sort of irrealism about the attitudes that appears far from incompatible with the stance taken towards them by outright eliminativists like the Churchlands.
This is most obvious in Gordon's recent work on the explanation of action (Gordon, 200), which directly challenges the canonical Davidsonian (Davidson,1963) story that identifies reasons for action with causally efficacious internal propositional attitude states (beliefs, desires, etc.). If Davidson was wrong about this, a major motivation for the reification of the attitudes is removed. However, Gordon's earlier work (1995, 1996) on "ascent routines" points in the same direction. The "ascent routine" story is offered as an account of belief ascription (to self and to others) that provides an alternative both to theory theory accounts (where beliefs are portrayed as posits of folk theory) and to introspectionist accounts such as Goldman's (1993). Both of these latter stories encourage reification: on the one hand it is now widely accepted that theoretical entities with causal powers should be construed realistically (when the theory is taken to be true, at any rate); on the other hand, if our propositional attitudes can be perceived (introspected) through some sort of inner sense then it would seem perverse to deny them at least the same degree of reality that we grant to the things (tables, trees, football matches) that we perceive with our outer senses. The "ascent routine" story, by contrast, gives no such hostages to realism. It explains how we could reliably report what we believe, and even (through simulation) what other people are likely to believe, without making any appeal to reified belief states. According to Gordon, we do not look within, into our belief boxes, to find out what we believe; rather we look outward, reporting how we take the world to be. Of course, we can, if we want, express our findings by talking about our beliefs and desires, and sometimes it is convenient to do so, but such grammatical nominalization should not be taken to imply that beliefs etc. are things, with causal powers, any more than, say, orbits, or habits. Belief talk, on this view, is derivative from the 'real' underlying processes of ascent routine and simulation.(4)
But, although they both resist the reification of the attitudes, there are very important differences between radical simulationism and radical scientistic eliminativism in the Churchland style. Gordon certainly does not look forward to a day when the mentalistic terms of ordinary "common sense" psychology will be replaced wholesale by the terminology of an advanced neuroscience (or cognitive science, or whatever). This is not, I think, a mere matter of temperament or style, for although (in its radical version) it has little use for reified beliefs, desires, and other propositional attitudes, simulation theory (radical and otherwise) in fact makes constant appeal to another group of "folk" mental notions, such as: perspective taking, seeing as, pretending, and, perhaps most centrally, imagination. Arguably, these play at least as large a role in actual folk talk about the mind as do the attitudes(5), and, like the attitudes, they are thoroughly intentionalistic notions. It does not look as if simulation theory can do without them (or some equally intentionalistic alternatives).
Unfortunately, the way these notions have been deployed by simulationists has been extremely vague and metaphorical. Indeed, to my mind this(6) is the weakest link in current expositions of all versions of simulation theory. Simulationists are much in need of some account of these mentalistic notions that is at least as clear as current accounts of propositional attitudes. This might appear a large task, because at first sight they are a heterogeneous bunch, and imagination in particular is often taken to be a polysemous term of dubious coherence. However, the close relation between the notions of pretending and imagining has long been recognized (Ryle, 1949), and, as I have argued at length elsewhere, if we return to the traditional view that a suitably understood concept of mental imagery (or quasi-perceptual experience) is at the core of our concept of imagination, we can attain an understanding of the latter that is reasonably unitary and coherent, and that also brings out its intimate relationship with the notion of seeing as (which is itself, surely, closely allied to perspective taking(7)) (Thomas, 1997a, 1999). Although the particular way these concepts are used in simulation theory remains to be clarified, the task of doing so may not be so daunting as it might otherwise appear.
In any case, when we realize that it makes essential appeal to these "folk" intentional concepts, radical simulationism does not look so very radical. (Neither, incidentally, does it look much like an "internalist" theory, in the sense that Stich & Ravenscroft (1994) seem to understand the notion(8).) Although it certainly challenges current philosophical orthodoxy, that is, the prevalent, attitude based, philosophical caricature of folk psychological theory, it may not actually be nearly so much at odds with real common sense intentionalistic talk about the mind. Neither, I think, is it out of line with our subjective experience of our minds. For my own part, at least, (and pace Goldman, Searle, and their allies) I am unaware of having any direct phenomenal experience of any specific propositional attitudes. True, I sometimes silently "hear" or "say" English sentences in my head, but these "verbal images" are surely, at best, expressions of my beliefs, desires etc., in same way that they would be if they were spoken out loud; they are not the very things themselves. Just the opposite seems to be true of my mental imagery: the fact that I sometimes have quasi-perceptual experience (leave aside, for now, what causes it) seems to me to be an incontrovertible fact of experience(9). It is surely not an accident that there is no counterpart, for the attitudes, to the well established scientific research tradition concerned with the degree of subjective intensity or vividness with which we experience our imagery (e.g. Galton, 1880; Marks, 1973, 1999; Ahsen,1985). Unlike reified propositional attitudes (which may well be explanantia, theoretical posits that are only as real as the positing theory is true) quasi-perceptual experiences are basic explananda for a science of the mind.
But inasmuch as radical simulationism does not reject the reality of imagery, and perhaps relies upon it (since it relies on imagination) it does not offer us a direct solution to the mind-body problem in the way that eliminativism does. Nevertheless, it has shifted the ground in an interesting and possibly very fruitful way. Attempts to give a direct naturalistic account of the intentionality of the attitudes have been notably unsuccessful, but, like eliminativism, radical simulation theory lets us off this hook. However, the question now becomes how to naturalize the intentionality of imagination.
In recent times, this has hardly been attempted. It has been assumed (when the issue has been considered at all) that any intentionality attaching to imagination (pretense, etc.) must be derivative from the intentionality of the propositional attitudes. On the one hand, either pretense and imagination (Nichols & Stich, 2000; White, 1990), or even imagery (Pylyshyn, 1981), are taken to be reducible to complexes of attitudes; on the other hand we are told that the intentionality of imagery can only derive from the description under which it is understood (Fodor, 1975; Tye, 1991). Such moves do not appear to be open to radical simulationism: if it is to be a naturalistic theory it will need to account for the intentionality of imagination, or of its core component, imagery, directly.
This project will be hopeless, however, if we retain the traditional and still very current (Kosslyn, 1980, 1994; Tye 1991) view that the quasi-perceptual experience commonly called mental imagery is to be explained as somehow the result of having pictures (or 'quasi-pictures") in one's head. The old story was that such pictures represent through their resemblance to what they depict, and that this form of representation forms the foundation for all others, including the meaningfulness of language. But although this depictive theory of intentionality has a long and distinguished history, twentieth century philosophical arguments have decisively shown it to be untenable(10). These arguments, indeed, have been a major motivation for the move to accounts of the mind in general, and imagination in particular, as being somehow linguistic (or quasi-linguistic) rather than perceptual at its foundations. However, the arguments loose their bite if imagery is not fundamentally pictorial. I have developed a non-pictorialist theory of imagery (Thomas, 1999, 2000) that is specifically designed to account for its intentionality without either appealing to resemblance or making it derivative from the intentionality of propositional or linguaform (mentalese) representations of the sort envisaged by theory theorists and other propositional attitude realists(11). Perhaps this is not the only possible theory of this type, and perhaps it is not the correct one, but I submit that the radical simulationist program needs some such theory if its full potential is to be realized(12). Like eliminativism, radical simulation theory allows us to avoid including reified proposition attitudes in our ontology, but without doing anything like such serious violence to our conception of ourselves as persons. Augmented, as I have suggested, with a non-pictorial theory of imagery, it might form the basis of a much more palatable and believable naturalistic account of the mind than eliminativism could ever provide.
1. Only part because consciousness and qualia are still there. These problems did not loom large when the Churchland eliminativist program got going. However, other eliminativists (as noted) would eliminate these, and other philosophers would (quite apart from any eliminativist commitments) reduce them to intentionality. Thus the program to eliminate attitudes plausibly retains its promise as an overall solution.
2. This is clearly true if it is held that the "internal" folk psychological theory is innate. The project of trying to "eliminate" (i.e. bring about a conceptual revolution in) the "internal" folk theory would perhaps make sense if it were held that the "internal" theory is caused by, and fairly directly reflects, whatever theory is embodied externally, in what people say: if one's "folk psychology" is learned at one's mother's knee, for example. In that case you might conceivably be able to bring about an internal theoretical revolution by bringing about a scientific/philosophical revolution in how people externally theorize the mind. However, eliminativists do not typically favor the sorts of "internal" stories (such as  and ) that would render such a relationship between the external and the internal coherent (though not mandatory). They have good reason not to do so, as will shortly emerge.
3. In this regard, eliminativists may be expected object most vehemently to theories of type . They may also find  objectionable since it seems to posit sentence like internal representations with intentional contents, and it is not at all clear that these would prove any more amenable to physicalistic reduction than the Attitudes.  is clearly acceptable to the eliminativist. It bolsters his case since the discovery that it is the correct sort of scientific story would result in a clear mismatch between the underlying neural story and any propositional attitude involving story that might be told at the "external" level. On the other hand,  may not be entirely unacceptable to the eliminativist, since it would seem that in this case, although it is possible to map the representation onto a set of propositions, it is not mandatory to do so. The ontological commitment of the science would be to the non-propositional data structure, not the propositions it maps to.
4. The ascent routine theory has, so far, only been spelled out in any detail for the case of belief. It is not yet clear whether something like it can be successfully applied to the other attitudes. Radical simulationism is a research program that is still in the early stages of its development. However, its thrust is clearly away from the reification of any of the attitudes. If the ascent routine theory proves unequal to the task then radical simulationists will need to devise different non-reificatory theories of the other attitudes. If this cannot be done then the radical simulationist program will fail. However, there is no reason at present to think that such failure is inevitable.
5. To the best of my knowledge (and somewhat to my bemusement), there seems to have been little systematic study made of what sorts of things "the folk" actually do say about the mind, and the terms they actually do use explaining behavior. I would be prepared to wager that explicit talk of nominalized beliefs and desires makes up only a small proportion of such discourse, although, no doubt, much more of it could, with some ingenuity, be suitably construed so as to fit into the preconceived mold of propositional attitude theory.
6. Together with the very sketchy nature of the account Gordon gives of what is, for his "radical" version of ST, the key notion of "patching".
7. Cf. Wittgenstein (1958 pp. 207e, 213e) on imagery and seeing aspects.
8. At any rate, it does not make any direct appeal to the sort of sub-personal, intentional-representational data structures that characterize the internalist views listed by Stich & Ravenscroft, and I would argue that it need not make any such appeal, even implicitly (see Thomas, 2000). Such representations belong to the sort of cognitivism that radical simulationism is challenging.
9. True, some do deny it, but their reasons seem likely to be more a result of conceptual confusion coupled with theoretical bias, rather than real experiential difference (Thomas, 1989).
10. See Fodor (1975) for a quick version of the oft repeated arguments. For a quick summary of the arguments against imagery being foundational for linguistic meaning see Thomas (1997b).
11. See Ellis (1995; also Thomas 1997b) for some reasons to think that the project of grounding the semantics of language in imagery thus understood may not be so hopeless as is generally thought. Ellis arrived independently at a conception of imagery that parallels the theory developed by me (Thomas, 1999).
12. Another relevant advantage of the theory of Thomas (1999) is that, on independent grounds, it allows for the possibility of content-specific deficits in imaginative abilities (paralleled by similar perceptual deficits). This might be able to explain the folk psychological deficits seen in autistics in a way compatible with ST, since it would not appeal to deficits in their theory of mind. Relevant perceptual deficits are suggested by the work of Baron-Cohen (1995) (although Baron-Cohen himself seems to be inclined to theory theory).
Ahsen, A. (1985). "Unvividness paradox." Journal of Mental Imagery (9) 1-18.
Baron-Cohen S. (1995). Mindblindness: An Essay on Autism and Theory of Mind. Cambridge, MA: MIT Press.
Churchland P.M. (1979). Scientific Realism and the Plasticity of Mind. Cambridge University Press.
Churchland P.M. (1981). "Eliminative Materialism and the Propositional Attitudes." Journal of Philosophy (78) 67-90.
Churchland P.S. (1988). Neurophilosophy: Toward a Unified Philosophy of the Mind-brain. MIT Press. Cambridge MA.
Davidson, D. (1963). "Actions, Reasons, and Causes." Journal of Philosophy (60). Reprinted in his Essays on Actions and Events. Oxford University Press, (1980). pp. 3-19.
Dennett D.C. (1988). "Quining Qualia." In A. Marcel & E. Bisiach (eds.) Consciousness in Contemporary Science. New York: Oxford University Press. pp. 42-71.
Dennett D.C. (1991). Consciousness Explained. Little, Brown. Boston MA.
Ellis, R. D. (1995). Questioning consciousness: The interplay of imagery, cognition, and emotion in the human brain. Amsterdam: John Benjamins.
Feyerabend, P.K. (1963). "Mental Events and the Brain." Journal of Philosophy (60) 295-296.
Fodor, J. A. (1975). The Language of Thought. New York: Thomas Crowell.
Goldman A.I. (1992). "In Defense of the Simulation Theory." Mind and Language (7) 104-119.
Goldman A.I. (1993). "The Psychology of Folk Psychology." Behavioral and Brain Sciences (16) 15-28. [Preprint available at http://cogsci.soton.ac.uk/~harnad/Papers/Py104/goldman.psyc.html]
Gordon, R.M. (1986). "Folk Psychology as Simulation." Mind and Language (1) 158-171.
Gordon, R.M. (1995). "Simulation Without Introspection or Inference from Me to You." In M. Davies & T. Stone (Eds.) Mental Simulation. Oxford: Blackwell. pp. 53-67.
Gordon, R.M. (1996). "'Radical' Simulationism." In P. Carruthers & P.K. Smith (Eds.) Theories of Theories of Mind. Cambridge: Cambridge Universty Press. pp.11-21.
Gordon, R.M. (2000). "Simulation and the Explanation of Action." In H.H. Kögler & K.R. Stueber. Empathy and Agency: The Problem of Understanding in the Human Sciences. Boulder, CO: Westview Press. [Draft available at http://www.umsl.edu/~philo/Mind_Seminar/New%20Pages/papers/Gordon/sim&action.htm
Kosslyn S. M. (1980). Image and Mind. Cambridge, MA: Harvard University Press.
Kosslyn S. M. (1994). Image and Brain: The Resolution of the Imagery Debate. Cambridge, MA: MIT Press.
Marks, D.F. (1973). "Visual imagery differences in the recall of pictures." British Journal of Psychology (64) 17-24.
Marks D.F. (1999). "Consciousness, Mental Imagery and Action." British Journal of Psychology (90) 567-585.
Nichols S. & Stich S. (2000). "A Cognitive Theory of Pretense." Cognition (74) 115-147. [Drafts available at http://ruccs.rutgers.edu/ArchiveFolder/Research%20Group/Publications/ACTOP/actop.html and http://www.umsl.edu/~philo/Mind_Seminar/New%20Pages/papers/Stich/SN%20Pretense%20doc.pdf]
Pylyshyn, Z. W. (1981). The imagery debate: Analogue media versus tacit knowledge. Psychological Review (88) 16-45.
Rorty R. (1980). Philosophy and the Mirror of Nature. Blackwell. Oxford.
Ryle, G. (1949). The concept of mind. London: Hutchinson.
Stich, S.P. & Ravenscroft, I. (1994). "What is Folk Psychology?" Cognition (50) 447-468.
Thomas, N. J. T. (1989). "Experience and theory as determinants of attitudes toward mental representation: The case of Knight Dunlap and the vanishing images of J. B. Watson." American Journal of Psychology, 102, 395-412.
Thomas, N. J. T. (1997a). "Imagery and the coherence of imagination: A critique of White." Journal of Philosophical Research, 22, 95-127.
Thomas, N. J. T. (1997b). "A Stimulus to the Imagination." Psyche (3) [Online serial] URL http://www.theassc.org/files/assc/2287.pdf [PDF download]
Thomas, N.J.T. (1999). "Are Theories of Imagery Theories of Imagination? An Active Perception Approach to Conscious Mental Content." Cognitive Science (23) 207-245.
Thomas, N.J.T. (2000). "A Non-Symbolic Theory of Conscious Content: Imagery and Activity." Presented at Tucson 2000 (Toward a Science of Consciousness) Conference, Tucson AZ, April 14th 2000.
Tye, M. (1991). The Imagery Debate. Cambridge, MA: MIT Press.
White, A. R. (1990). The language of imagination. Oxford: Blackwell.
Wilkes, K. V. (1988). "__, Yishi, Duh, Um, and Consciousness," in: A. J. Marcel & E. Bisiach (eds.). Consciousness in Contemporary Science (pp. 16-41). Oxford University Press.
Wilkes, K. V. (1995). "Losing Consciousness," in: T. Metzinger (ed.). Conscious Experience. Schöningh/Imprint Academic. Paderborn/Thorverton. pp. 97-106.
Wittgenstein L. (1958). Philosophical Investigations. (2nd edn. trans. G.E.M. Anscombe.) Blackwell. Oxford.
Return to Home Page:
Imagination, Mental Imagery, Consciousness, Cognition:
Science, Philosophy & History.