FUNCTIONALISM, THE COMPUTER MODEL OF THE MIND, AND CAUSAL CONNECTIONS.
1. Functionalism and the Computer Model of the Mind
The basic idea of functionalism is that what makes a state mental is its playing a causal-functional role of a certain pattern within a system, and, specifically, what makes a mental state the mental state that it is (e.g., its being a belief that it's raining) is its causal connections to perceptual inputs, behavioral outputs, and other mental states. Mental properties, on this approach, are causal-functional properties. The basic idea of the computer model of the mind is that the mind is the software of the body (or the brain), or, in other words, that the mind is to the brain as the program is to the hardware. (1) On this model, physical properties and mental properties are properties of the same entity (say a person), it's just that--like software properties of a computer --mental properties belong to a more abstract level. Although not all functionalists employ the terminology of computers, hardware, and software, an important element of the computer model characterizes all versions of the basic functionalist idea, namely the conception of mental properties as abstract high-order properties. Typically, functionalists take mental properties to be abstract properties of physical systems or physical states, namely of brains and brain states. (2)
Another typical characteristic of functionalist theories is that they allow the possibility of multiple realizations of mental states and properties. That is, according to most versions of functionalism, a mental state of any given type (e.g., feeling pain) may be realized by states of various physical types. (3) Further, in principle, a mental state of any given type need not even be realized physically, and thus the very idea of functionalism is compatible with dualism. As a matter of fact, though, most functionalists believe that mental properties are functional properties of the brain, and according to many of them the importance of this theory lies in its ability to account for mental properties and their allegedly unique features in the framework of physicalist ontology. (4) The multiple realizability of mental properties is reminiscent of the multiple realizability of software properties, namely of the fact that the same software can be realized by computers with different hardwares. This fact too highlights the strong linkage that obtains between functionalism and the computer model of the mind.
2. Two Quandaries
But now a quandary arises. As noted, functionalism characterizes mental states in terms of their causal connections with perceptual inputs, behavioral outputs, and other mental states. On the other hand, the computer model of the mind does not appeal to causation. Rather, it appeals to abstract characteristics. It seems that the claim that mental properties are abstract properties is incompatible with the claim that mental properties are constituted by causal connections, which connections are as concrete as may be. The first quandary, then, is this: how, if at all, can we reconcile the basic tenet of the computer model of the mind--its taking mental properties to be abstract--with the basic idea of functionalism--its taking mental properties to be constituted by causal connections?
Another quandary appears to threaten functionalism. On this theory, mental duplicates (that is, two creatures who share with each other all of their mental properties) share those of their causal powers that are relevant to their mental states. When we combine this feature of functionalism with its allowing for the multiple realizability of mental states, we get the conclusion that utterly different physical objects--for example, my brain and the silicon-based brain of a Martian creature--may be identical in their mentally relevant causal powers. This, however, seems mysterious, for it seems that it is the physical properties of physical objects that are responsible for their causal powers. If so, the question arises of how it is that objects with very different physical makeups may share causal powers. As far as the mental is concerned, the question is: how can mental states of the same type (e.g., the desire to drink water) be identical to each other in causal powers, when their physical realizations are utterly different from each other? (5)
3. Resolving the Quandaries
Let us start with this second quandary. It is the computer model of the mind that provides the answer to it. For according to this model, physically different objects are identical in some of the causal connections that they instantiate if they realize the same software. We know that computers with physically different makeups--that is, computers with different hardwares--may be identical in the causal connections that they instantiate; they are identical in the causal connections that they instantiate when they realize the same software. So the assumption that the mind is the software of the brain explains the claim, to which functionalism is committed, that some different physical objects--the physically different brains of mental duplicates (such as an Earthling and a Martian)--and those of their physical states that are the realizations of their mental states, are identical in causal powers. Thus, this functionalist claim is explained by the idea that the mind is the software of the brain, since if the mind is the software of the brain then those physically different states that realize the mental states of duplicates are realizations of the same software; and they are identical in causal powers exactly as computers with different hardwares that realize the same software are identical in causal powers. In this respect, the functionalist view and the computer model of the mind are well integrated.
We should note that the claim that mental duplicates share those of their causal powers that are relevant to their mental states should be accepted by everyone. For this claim does not depend upon the idea that mental states are characterized in terms of their causal powers. Rather, it follows from the weaker and quite uncontroversial claim that mental states possess typical causal powers. (6) So philosophers who accept the thesis of the multiple realizability of the mental are challenged by the second quandary even if they are not functionalists. Now as we saw, functionalists do have a response to this quandary, a response that is given in terms of the computer model of the mind. In other words, explicating functionalism in terms of this model reveals functionalism's explanatory advantage, one which may be taken to form a significant (though certainly not conclusive) consideration in favor of this view.
However, the functionalist explanation in question cannot be considered complete. This explanation is supposed to account for the identity in causal powers of mental duplicates by relying on their identity in software characteristics. But as noted, software characteristics are abstract, whereas causal powers are concrete: they are responsible for bringing about changes in the world. We thus face the following question: how can identity in abstract characteristics explain identity in causal powers? We can now see the close linkage between the two quandaries. For the first quandary, recall, was how, if at all, we can reconcile the claim that mental properties are abstract with the claim that they are constituted by causal connections.
But there is a further complication. Imagine that I and my mental duplicate, whose brain is silicon-made, share the following mental states: the belief that it's raining, the belief that the only way for us (given our circumstances) to avoid getting wet it to open an umbrella, and the desire to avoid getting wet. These three mental states give rise in both my case and my duplicate's case to a further shared mental state, namely the decision to open an umbrella. But this decision is realized in both cases by physically different states. So in an important sense the three initial mental states are not identical in causal powers, and mental duplicates are not identical in those causal powers that are relevant to their mental states. However, we've just seen that there is good reason to accept the claim that mental duplicates share those of their causal powers that are relevant to their mental states (it should be accepted by everyone, recall). So what's the way out?
I believe that some clarification will show us the way out. Two physical systems are the realizations of the same software if the networks of their causal powers have the same structure. As a simple example of a structure of the relevant kind, consider a system that involves a move from states A and B to state C, a move from states C and D to states E and F, and a move from state F to state A. Such moves constitute a structure--a structure of transitions (the identity of the specific states is not part of the structure, and we can replace each of the A-F states with whatever other states, as long as the replacement is consistent, so that if, e.g., we replace state A by state Q in one move, A has to be replaced by Q in the other moves which involve A). Physical systems can certainly realize this structure. Further, even physically different physical systems can realize this structure; this happens in cases in which the states that stand for A-F in one system physically differ from the respective states in the other. In any realization of this structure, the moves among the states are causal (if miracles are out of the picture), and so it seems that there is a sense of "causal" (call it "the weak sense") in which the systems' identical structure is a causal structure, and the systems exemplify the same causal connections and may be said to be causally identical. But in another sense of causal ("the strong sense"), the structure in itself is not causal; for one thing, it involves no causal laws. In this sense, the physically different realizations of the structure do not exemplify the same causal connections, and thus their counterpart states do not share causal powers.
We can now come back to the two quandaries. We suggested to explain the identity in causal powers of mental duplicates by relying on their identity in software characteristics, but then we found ourselves facing the question of how identity in abstract characteristics such as software characteristics may explain identity in causal powers. The quandary is dissolved once we realize that the relevant causal identity, such as the one that obtains between physically different mental duplicates, is a causal identity only in the weak sense of "causal." That is, such properties are only identical in abstract properties, and there is no difficulty in explaining identity in abstract properties by appealing to abstract properties; how else can we explain it?
The first quandary, recall, was how, if at all, we can reconcile the claim of the computer model of the mind that mental properties are abstract, with the basic functionalist claim that mental properties are constituted by causal connections. Yet according to functionalism (or at any rate, according to plausible versions of functionalism) mental properties are constituted by causal connections only in the weak sense of "causal." It is the abstract structure and abstract connections that determine the identity of a mental state. Thus, the basic functionalist claim does not clash with the idea of the computer model of the mind. Rather, these two ideas are equivalent.
4. Causation in the Weak Sense and Causal Relevance
Although the very idea of functionalism is neutral with respect to the physicalism-dualism controversy, most functionalist theories are physicalist--they take mental properties to be functional properties of physical systems. Of course--since functionalism opposes physicalist reduction--those versions are versions of non-reductive physicalism. As such, they face the challenge of the "exclusion problem" of mental causation, namely the problem that since in a physicalist framework it's physical properties that are responsible for the causal role of mental states, mental properties cannot be causally efficacious in the production of behavior and other mental states. The specific form that this charge takes in the case of functionalism is that it's the realizing (physical) property that excludes the realized functional property from being causally efficacious. (7)
Now what is the implication of the picture that endorses those two kinds of causal connections--"causal connections in the strong sense" and "causal connections in the weak sense"--for the exclusion charge against functionalism? It seems that this picture illustrates that functionalism is subject to no causal exclusion. For clearly, since causal connections in the strong sense and causal connections in the weak sense are causal connections in different senses, there is no competition between them. None exclude the other. Those connections can coexist because they belong to different levels of abstractness (or concreteness); they belong to different level in the same respect in which physical properties and mental properties themselves belong to different levels in the functionalist picture.
Of course, the idea that those causal connections are causal connections in different senses--the idea that enables functionalism to rebuff the threat of causal competition and causal exclusion--does not enable functionalism to remove the related threat of type-epiphenomenalism, namely the objection that, according to it, mental states do not exert causal powers in virtue of their mental properties, or in other words, in virtue of falling under mental types. (8) Further, this idea may even be thought to call upon the threat of type-epiphenomenalism. That is, it is essential to this idea that mental properties, qua functional properties, are responsible for the obtaining of causal connections in the weak sense, and only in the weak sense. So mental properties qua functional properties can only be causally efficacious in a weak sense, not in the stronger sense in which physical properties are causally efficacious. Arguably, only causal connections in the strong sense are really causal connections, and so only physical properties are really causally efficacious. It might then seem that on the picture under discussion, although mental states are causally efficacious and thus token-epiphenomenalism is false, mental properties are not causally efficacious and so type-epiphenomenalism is true.
It might be--and many philosophers believe--that mental properties play a significant causal role vis-a-vis the production of behavior and other mental states that falls short of making these properties causally efficacious. Mental properties are supposed to be causally relevant. Whether or not various views that advance such a conception of mental causation should be referred to as "epiphenomenalist" is not an interesting question, but it is an interesting question whether mental properties qua functional properties may play the causal role that is as significant as the causal role that (we have reasons to believe) mental properties in fact play. This question takes a specific form when mental properties are to be characterized in terms of causal connections in the weak sense characterized above.
It is not the aim of this paper to discuss this question. What I'd like to stress is that such a functional characterization of mental properties is not subject to the charge of failing to reflect the true nature of mental causation more than the initial functionalist characterization. (9) As noted, the characterization in terms of causal connections in the weak sense is not threatened by the exclusion problem. Now if this problem is set aside, the reason why the functional characterization of mental properties has been charged for distorting the real nature of mental causation concerns these properties' being abstract or high-order properties of their realizers. (10) It is their being such that has been supposed to prevent them from being causally efficacious and playing any causal role. But my claim that the causal connections that figure in the functionalist characterization of mental properties can only be causal in a weak sense is concerned precisely with this nature of functional properties. This claims stems straightforwardly from the claim that functional properties are abstract high-order properties of their realizers: if mental properties are abstract high-order properties, then trivially the connections in which they figure cannot but be abstract connections; and to the extent that causal connections in the ordinary sense are concrete, those psychological connections can only be called "causal connections" by courtesy. That is, characterizing the "causal" connections in question as causal merely in the weak sense is a called upon elucidation of the very idea of functionalism. Thus, if this characterization is guilty of failing to respect the truth about mental causation, so is the very functionalist idea; and if the latter avoids this charge, so does the former. This characterization, then, is not subject to the charge of failing to reflect the true nature of mental causation more than the initial functionalist characterization.
Various accounts that aim to explain how functional properties can be causally relevant properties have been suggested. Let me close by mentioning one that seems to bear directly on the specific functionalist characterization under discussion. According to Jackson and Petit, what makes a non-efficacious property causally relevant for the causal production of an event is the fact that its realization "ensures--it would have been enough to have made it suitably probable--that a crucial productive property is realized and, in the circumstances, that the event, under a certain description, occurs" (Jackson and Petit 1990, p. 214). Jackson and Petit use the metaphor of programming in this context, and dub the causal explanations that are given in terms of causally relevant yet non-efficacious properties "program explanations." The realization of the causally non-productive but relevant property "programs for the appearance of the productive [realizing] property and, under a certain description, for the event produced... The analogy is with a computer program which ensures that certain things will happen--things satisfying certain descriptions--though all the work of producing those things goes on at a lower, mechanical level" (ibid.) (11)
Evidently, explanations in terms of properties that figure in causal connections in the weak ("software") sense are program explanations in Jackson's and Petit's sense. Thus, if mental properties are causally relevant in the sense suggested by this account, then functional properties that are characterized in terms of causal connections in the weak sense are causally relevant properties, and a functional theory that is characterized in such terms does not distort the nature of mental causation.
(1.) See Johnson-Laird (1988).
(2.) Prominent functionalists are Putnam (1960, 1967, 1973), Fodor (1968), and Lewis (1972, 1980). Putnam, who pointed out the analogy between minds and Turing machines, can be considered the founder of the computer model of the mind. Putnam later abandoned functionalism (see, e.g., Putnam 1988).
(3.) See Putnam 1967.
(4.) For what might be considered a dualist functionalism see Chalmers 1996.
(5.) Note that this puzzle is independent of whether our mental properties are indeed realized physically.
(6.) Note that the claim that mental duplicates share those of their causal powers that are relevant to their mental states should be accepted also by externalists. Externalism is in no conflict with the idea that mental states possess typical causal powers; it just sets unique requirements for mental identity.
(7.) On the exclusion problem see Kim 1989, 1993, 1998, 2005, and O'Connor and Churchill 2010.
(8.) On type-epiphenomenalism see McLaughlin 1989. In Horowitz 1999 I called into question the view that type-epiphenomenalism is problematic.
(9.) This might be the place to mention that the charge under discussion is meant to apply to role functionalism, that identifies mental properties with higher-order functional properties, rather than to realize functionalism (or filler functionalism), that identifies mental properties with the lower-order properties that realize the functional properties. On this distinction see McLaughlin 2006.
(10.) There is another reason that led philosophers to take the functional characterization of mental properties to distort the real nature of mental causation, one that is rooted in the idea of "metaphysically necessitated effects" (Ludwig 1998, Rupert 2006, Bennett 2007). I cannot discuss this problem here, but clearly the distinction between weak and strong senses of causal connections is irrelevant to it.
(11.) See also Jackson and Petit 1988, 2004a, 2004b.
Bennett, K. (2007), "Mental Causation," Philosophy Compass 2: 316-337.
Chalmers, D. (1996), The Conscious Mind. Oxford: Oxford University Press.
Fodor, J. (1968), Psychological Explanation. New York: Random House.
Horowitz, A. (1999), "Is There a Problem in Physicalist Epiphenomenalism?," Philosophy and Phenomenological Research 59: 421-434.
Jackson, F., and Petit, P. (1988), "Functionalism and Broad Content," Mind 97: 381-400.
----- (1990), "Program Explanation: A General Perspective," Analysis 50: 107-117.
----- (2004a), "Causation in the Philosophy of Mind," in F. Jackson, P. Pettit and M. Smith (eds.), Mind, Morality and Explanation. Oxford: Oxford University Press, 45-69.
----- (2004b), "Structural Explanation and Social Theory," in F. Jackson, P. Pettit and M. Smith, Mind, Morality and Explanation. Oxford: Oxford University Press, 131-163.
Johnson-Laird, Philip N. (1988), The Computer and the Mind: An Introduction to Cognitive Science. Cambridge, MA: Harvard University Press.
Kim, J. (1989), "Mechanism, Purpose, and Explanatory Exclusion," Philosophical Perspectives 3: 77-108.
---- (1993), "The Non-Reductivist's Troubles with Mental Causation", in Heil, J. and Mele, A. (eds.), Mental Causation, Oxford: Clarendon Press, 89-210.
---- (1998), Mind in a Physical World. Cambridge, MA: MIT Press.
---- (2005), Physicalism, or Something Near Enough. Princeton, NJ: Princeton University Press.
Lewis, D. (1972), "Psychophysical and Theoretical Identifications," Australasian Journal of Philosophy 50: 249-258.
---- (1980), "Mad Pain and Martian Pain," in Block, N. (ed.), Readings in the Philosophy of Psychology. Cambridge, MA: Harvard University Press, 216-222.
Ludwig, K. (1998), "Functionalism, Causation, and Causal Relevance," Psyche 4(3).
McLaughlin, B. P. (1989), "Type Epiphenomenalism, Type Dualism, and the Causal Priority of the Physical," Philosophical Perspectives 3: 109-135.
Mclaughlin, B. P. (2006), "Is Role-Functionalism Committed to Epiphenomenalism?," Journal of Consciousness Studies 13: 39-66.
O'Connor, T., and Churchill, J. R. (2010), "Is Non-reductive Physicalism Viable within a Causal Powers Metaphysic?," in C. Macdonald and G. Macdonald (eds.), Emergence in Mind. Oxford: Oxford University Press, 43-60.
Putnam, H. (1960), "Minds and Machines," reprinted in Putnam 1975, 362-385.
----(1967), "The Nature of Mental States," reprinted in Putnam 1975, 429-440.
----(1973), "Philosophy and our Mental Life," reprinted in Putnam 1975, 291-303.
----(1975), Mind, Language, and Reality. Cambridge: Cambridge University Press.
----(1988), Representation and Reality. Cambridge, MA: MIT Press.
Rupert, R. D. (2006), "Functionalism, Mental Causation, and the Problem of Metaphysically Necessary Effects," Nous 40: 256-283.
The Open University of Israel
Received 25 August 2014 * Received in revised form 26 October 2014
Accepted 27 October 2014 * Available online 1 November 2015
|Printer friendly Cite/link Email Feedback|
|Publication:||Analysis and Metaphysics|
|Date:||Jan 1, 2015|
|Previous Article:||THE UNIVERSITY IN THE EPOCH OF DIGITAL REASON: FAST KNOWLEDGE IN THE CIRCUITS OF CYBERNETIC CAPITALISM.|
|Next Article:||A REPRESENTATIONAL ACCOUNT OF FICTIONAL CHARACTERS.|