next

lecture one
"i thought of
what i called
'an automatic
sweetheart'"

 

 

William James at one point makes the remarkable thought experiment of imagining a human being who (undetectably) lacks all mental properties:

I thought of what I called an "automatic sweetheart," meaning a soulless body which should be absolutely indistinguishable from a spiritually animated maiden, laughing, talking, blushing, nursing us, and performing all feminine offices as tactfully and sweetly as if a soul were in her. Would anyone regard her as a full equivalent? Certainly not.¹

But James was not the first philosopher to have this worry. The philosopher who more than any other fixed the shape of our subsequent worries about mind and body, René Descartes, considers it too. In the Discourse on Method Descartes even proposed a kind of "Turing test":

For we can certainly conceive of a machine so constructed that it utters words, and even utters words which correspond to bodily actions causing a change in its organs (e.g., if you touch it in one spot it asks what you want of it, if you touch it in another it cries out that you are hurting it, and so on). But it is not conceivable that such a machine should produce different arrangements of words so as to give an appropriately meaningful answer to whatever is said in its presence, as the dullest of men can do.²

The worry certainly seems far-fetched. Yet the issue has arisen again in a recent debate between two of our leading philosophers of mind, Jaegwon Kim and Donald Davidson.

In the essays collected in his book Supervenience and Mind, Kim devotes an extraordinary amount of attention to the philosophy of mind of Donald Davidson.³ The reason is not hard to guess: Davidson's view is diametrically opposed to Kim's on the issue of what Kim calls "strong supervenience"⁴ and what I shall refer to for the time being simply as reductionism. (However, Kim is impressed by Davidson's arguments against the existence of the sorts of psychophysical laws required for strong supervenience, and even includes a paper that appears to defend them in Supervenience and Mind.)⁵ Although Davidson is a materialist to the extent of believing that each individual mental event is identical with an individual physical event, there are, according to Davidson no psychophysical laws. However (in most of the papers in Supervenience and Mind), Kim defends the possibility of finding such laws. As he puts it, even if certain arguments of mine⁶ show that we aren't going to find any unrestricted laws of the form P´M, where P is a physical property, M is a mental property, and the law quantifies over all physically possible organisms, still, there may well be "species-specific bridge laws," i.e., laws of the form S_iÆ(M´P_i) "which, relative to a species or structure S_i, specifies a physical state P_i as both necessary and sufficient for the occurrence of mental state M."⁷ Indeed, Kim's position in "The Myth of Nonreductive Physicalism" is that only if there are such laws can it be the case that mental properties have genuine causal efficacy.

First, here is Kim's description of Davidson's position:

Mental events, Davidson observed, enter into causal relations with physical events. But causal relations must be backed by laws; that is, causal relations between individual events must instantiate lawful regularities.⁸ Since [according to Davidson] there are no laws about the mental, either psychophysical or purely psychological, any causal relation involving a mental event must instantiate a physical law, from which it follows that the mental event has a physical description, or falls under a physical event kind. For an event is physical (or mental) if it falls under a physical event kind (or a mental event kind).

It follows then that all events are physical events—on the assumption that every event falls into at least one causal relation.⁹

And here is Kim's criticism:

Davidson's ontology recognizes individual events as spatiotemporal particulars. And the principal structure over these events is causal structure; the network of causal relations that interconnects events is what gives intelligible structure to this universe of events. What role does mentality play, on Davidson's anomalous monism, in shaping this structure? The answer: None whatever.

For anomalous monism entails this: the very same network of causal relations would obtain in Davidson's world if you were to redistribute mental properties over its events in any way you like; you would not disturb a single causal relation if you randomly and arbitrarily reassigned mental properties to events, or even removed mentality entirely from the world. Remember: on anomalous monism, events are causes or effects only as they instantiate physical laws, and this means that an event's mental properties make no causal difference.¹⁰

davidson defended

I shall begin by defending Davidson against Kim's argument.¹¹ It isn't, let me say at the outset, that I share Davidson's anomalous monism or Davidson's view on causality. On anomalous monism, my position is that mental events aren't "identical" or "not identical" with physical events; I don't believe that the notion of "identity" has been given a sense here.¹² And I also believe that Elizabeth Anscombe's "Causality and Determination"¹³ contains conclusive counterexamples to Davidson's claim that every true singular causal statement must be backed by a strict law.¹⁴ However, I believe that the above objection to Davidson's position can be answered, and that trying to answer it is an excellent way to come to grips with a number of central issues in the philosophy of mind.

Kim's argument evidently turns on the following remarkable conditional:

Kim's conditional (assuming that Davidson's views are correct): You would not disturb a single causal relation if you randomly and arbitrarily reassigned mental properties to events, or even removed mentality entirely from the world.

An immediate consequence of this conditional is the following:

(AUTOMATA) (assuming that Davidson's views are correct): Even if certain people did not have any mental properties, as long as all of their physical properties were the same and their physical environments were the same, the same physical events would happen.

In effect, Kim has invited us to make the same thought experiment as James (assuming Davidson's philosophy of mind is correct): imagine that there
were a world in which some body just like mine (just like yours) were in exactly the same physical state and exactly the same physical environment as mine (as yours), but lacked all mental properties. Would it not behave exactly as mine (as yours) does? The body would be exactly like the body of James's imagined "automatic sweetheart," an automatic me or an automatic you. And from the fact (as Kim must claim it to be) that the automatic Hilary Putnam would turn on the water just as efficiently even though the mental event, the "decision to turn on the water for a bath," did not take place, Kim concludes that (if Davidson's philosophy of mind is correct) the mental event is without causal efficacy. Mental properties are mere "epiphenomena," just as Santayana supposed them to be.¹⁵

Of course, the argument can be blocked by denying that the counterfactual "makes sense." And, for the counterfactual not to make sense, Kim thinks that it would have to be the case that our mental properties are somehow reducible to our physical properties (or more precisely, "strongly supervenient" on them in some satisfactory way).¹⁶ But reduction (including "strong supervenience") requires psychophysical laws, which Davidson's position rules out. Thus Kim's conclusion is disjunctive: either Davidson is wrong, and there are psychophysical laws, or else our mental properties are ephiphenomenal. (I leave out a third alternative, that it is an illusion that we have mental properties, since neither Davidson nor Kim—nor I, for that matter—take that one seriously.)

setting the stage:
finding fault with jaegwon kim's counterfactual

Suppose I am running water for a bath. If someone asked me "Why is the water running in the bathroom?" I might reply, "I'm running the water in order to take a bath," but I might also say simply "I decided to take a bath." And no one would doubt that I have explained why water is running in the bathroom. To be pedantic about it, water is running in the bathroom because I decided to take a bath, and because I decided to take a bath in this particular bathtub at this particular time, and I decided to take a "real" bath as opposed to a shower, and I know full well that to fill the tub for a "real" bath I need to turn the faucet, and I know how to turn the faucet to the right position, and I acted on my decisions and my knowledge.

John Haldane has wisely remarked¹⁷ that "there are as many kinds of cause as there are senses of 'because.' " (He added that "Aristotle's doctrine of four causes was only a preliminary classification.") If the "because" in the preceding explanation is correct, then in some sense of "cause" my decision to turn that faucet to the middle position (between hot and cold) is the cause of water (of a certain comfortable temperature) running right now. Decisions (which are among the things that philosophers refer to as mental or psychological events) can be the causes of water runnings (which are among the things that philosophers refer to as physical events). And to suppose (as Santayana would have done) that the "because" is not correct—that is, to suppose that it isn't because of my decision that water of that comfortable temperature is now running, that indeed my decision was totally "inefficacious"would be to reject ways of talking that are essential to our whole conception of ourselves as beings in a world; ways of talking even Santayana (who constantly offers advice as to how to live, advice that would make no sense if decisions were causally inefficacious) was unable to do without or to offer a coherent alternative to. Epiphenomenalism is C-R-A-Z-Y.

Does it then follow that our mental properties must in some way be reducible to physical ones? Before we try to crack that chestnut, let us first look into the question of how we might defuse Kim's counterfactual.

Well, we don't want to simply accept the antecedent as a fully intelligible possibility (Certain people do not have any mental properties, but all of their physical properties are the same as if they did and their physical environments are the same as if they did) while denying the consequent (The same physical events would happen); i.e., we do not want to assert:

(NOT-AUTOMATA) If certain people did not have any mental properties, but all of their physical properties were the same as if they did and their physical environments were the same, DIFFERENT physical events would happen.¹⁸

(NOT-AUTOMATA) is not something that we find believable.

Of course, the fact that we do not find (NOT-AUTOMATA) believable has everything to do with the fact that we no longer find strong Cartesian dualism ("interactionism") believable. For (NOT-AUTOMATA) is exactly what Descartes himself believed: he believed that the mind, conceived of as an immaterial entity that "bears" all of our mental properties, causes our bodies to act in a way in which they would not act if the mind were taken away.

In these lectures I shall be arguing that neither the classical problems in the philosophy of mind nor the "philosophical positions" they give rise to are completely intelligible. I think this whole idea of the mind as an immaterial object "interacting" with the body is an excellent example of an unintelligible position in the philosophy of mind, but this is not the customary criticism of interactionism.

The customary criticism was beautifully and trenchantly stated by Bertrand Russell.¹⁹ If dualist interactionism is true, then my body must have a trajectory that is different from the trajectory that would be predicted for it by the laws of physics on the basis of the totality of the physical forces acting upon it. Descartes himself was aware of this, and since even he did not want to postulate that when human bodies are involved in physical interactions such basic laws as the Conservation of Momentum are violated, he posited that although the mind can alter the direction in which the body moves (by acting on the body in the region of the pineal gland), it cannot alter the "quantity of motion" (this is more or less what we today call "total scalar momentum"). But shortly after Descartes's time it was realized that total momentum in each direction in space is also a conserved quantity. Thus, if interactionism is true, some conservation laws of physics are violated when humans act on the basis of decisions and other thoughts. In short, interactionism implies that human bodies behave in ways that violate the laws of physics. Since there is not a shred of evidence that this is the case, we are compelled to reject interactionism just as we reject vitalism and other outmoded views that postulated various phenomena ("life" was a popular example in the nineteenth century) to be "exceptions to the laws of physics." As Kim himself puts a less detailed form of the same argument,

There is a further assumption that I believe any physicalist would grant [and Kim is evidently a "physicalist" in this respect], namely "the causal closure of the physical domain." Roughly, it says this: any physical event that has a cause at time t has a physical cause at t. This is the assumption that if we trace the causal ancestry of a physical event, we need never go outside the physical domain. To reject this assumption is to accept the Cartesian idea that some physical events need nonphysical causes, and if this is true there can in principle be no complete and self-sufficient physical theory of the physical domain. If the causal closure failed, our physics would need to refer in an essential way to nonphysical causal agents, perhaps Cartesian souls and their psychic properties, if it is to give a complete account of the physical world. I think most physicalists would find that picture unacceptable.²⁰

Russell's (and Kim's) argument is an empirical argument. It assumes that interactionism is fully intelligible (contrary to the position that I will argue for) and argues that the empirical evidence does not support it. But whether we reject interactionism on empirical grounds, as Russell did, or on the grounds that we cannot make sufficient sense of it, as I would argue we should, the fact is that scientifically minded people today do reject interactionism. So if we wish also to reject (AUTOMATA) it cannot be by simply asserting (NOT-AUTOMATA).

There are, in fact, quite a number of philosophical positions that imply that (AUTOMATA) is not true but that do not imply (NOT-AUTOMATA)for example, reductionist physicalism, logical behaviorism, and verificationism. Philosophers who defend these positions would attack (AUTOMATA) by finding one or another thing wrong with the antecedent, which I reformulated²¹ above as Certain people do not have any mental properties, but all of their physical properties are the same as if they did and their physical environments are the same. More precisely, such philosophers would agree that the conjunction of the denial of interactionism with the antecedent of (AUTOMATA) suffers from one or another defect—unintelligibility (according to verificationists) or self-contradiction (according to logical behaviorists) or metaphysical impossibility (according to some reductionist physicalistsas we shall see other reductionist physicalists would argue that the antecedent of (AUTOMATA) is irrelevant if we are interested in what happens in the actual world).

To see why this is the case, observe that if interactionism is false, and if the antecedent of (AUTOMATA) describes a state of affairs that could conceivably hold even when interactionism is false, then (AUTOMATA) must be true. And that would mean that James's "automatic sweetheart"—the person who has no mental properties, no "consciousness," but who behaves exactly as if she had such properties—is a conceivable state of affairs. Thus, a way to attack (AUTOMATA), on the assumption that interactionism is false, is to deny that the following state of affairs makes sense: the antecedent of (AUTOMATA) is true, but the people in question do not behave any differently then they would if they had their "mental properties."

Verificationism argues that the state of affairs described would be unverifiable (if the people in question acted exactly as they would have if they had their mental properties, then it would be impossible in principle to verify that they lacked them) and therefore the supposed description of this "state of affairs" ("the antecedent of (AUTOMATA) is true, but the people in question do not behave any differently than they would if it were false") is cognitively meaningless. Unlike verificationism, logical behaviorism does not necessarily assume that all meaningful propositions are verifiable in principle, but it claims that all propositions about mental properties are logically equivalent to propositions about (physical) behavior. So if we were to suppose that we can imagine a world in which the antecedent of (AUTOMATA) is true and interactionism is false (a world in which the antecedent of (AUTOMATA) is true, but the people in question do not behave any differently than they would if it were false), we would be supposing one can imagine that the logically necessary conditions for the presence of mental properties are fulfilled but the properties are not present, and this is a contradiction. So if verificationism is correct, then the conjunction of "no interactionism" with the antecedent of (AUTOMATA) is cognitively meaningless, and if logical behaviorism is correct, the conjunction is self-contradictory. If either position is correct, we can refuse to count (AUTOMATA) as a true counterfactual without being forced to accept (NOT-AUTOMATA). Unfortunately, neither verificationism nor logical behaviorism appear to be tenable positions any longer.²²

Donald Davidson's position is that the raison d'être of talk about the mental is to enable us to "rationalize" the behavior of human agents. Moreover, he contends, this is its whole function; that is why there cannot be psychophysical laws. If this view is right, the conjunction of "no interactionism" with the antecedent of (AUTOMATA) describes a state of affairs in which the whole rationale for applying mental predicates applies, but those mental predicates do not "really" apply. Surely Davidson would regard this as an unintelligible suggestion!

Davidson is clearly not a verificationist, and he is certainly not a logical behaviorist, but his position has an interesting relation to both of these. Like the logical behaviorists, he believes that if you behave in all respects as if mental predicates applied to you ("all respects" here includes microphysical respects, if microphysical events can be come relevant to the interpretation of your speech and other behavior), then those mental predicates do apply to you; although unlike logical behaviorists, Davidson does not believe that one can write down conceptual truths linking particular mental predicates to particular behavioral predicates. Like the verificationists, he believes that if a mental predicate applies to you then an observer who was "omniscient" with respect to all physical facts about you and your environment could verify that it applies to you.²³ Indeed, we might call Davidson a psychoverificationista verificationist about the mental. My own rejection of the intelligibility of the various AUTOMATA scenarios we have considered will, however, depart from the strategy of defending Davidson at this point. I shall not make any of the metaphysical assumptions (about the one and only purpose of talk about the mental, etc.) that figure in Davidson's metaphysics of mind.

Turning now to physical reductionism (and it is a—highly sophisticated—version of physical reductionism that Kim is defending), those who take this approach have a number of possible strategies for dealing with (AUTOMATA) and the threat it poses to the causal efficacy of the mental. The simplest is just to argue that (AUTOMATA) is irrelevant. After all, if it is the case in the actual world that our so-called mental properties are just a subset of our physical properties, then any world in which the antecedent of (AUTOMATA) is true is a world in which mental properties are very differently "realized" than they are in the actual world. Even if mental properties would be "epiphenomenal" in such a "possible world," they are definitely not in the actual world, because they are physical and physical properties are by definition not epiphenomenal.²⁴

However, as I shall try to show in the succeeding lectures, reductionist physicalism is incoherent. If that is right, and we are not willing to be verificationists (or "psychoverificationists") and not willing to be logical behaviorists, on what other grounds could we possibly avoid accepting (AUTOMATA), and, with (AUTOMATA), the conclusion that since all physical events would go on in the same way even if our mental properties were stripped away, those mental properties are one and all epiphenomenal? How can we avoid Kim's disjunction: either our mental properties are physical properties in disguise, or they are epiphenomenal? (Kim tells me that he himself finds both disjuncts unattractive!)