The Pairing Objection to Substance Dualism

As I understand the  Pairing Objection to substance dualism it goes like this.  Let m1 and m2 be mental tokens of type M and b1 and b2 brain tokens of type B, and suppose that M-type events cause B-type events. Suppose m1 and m2 both occur at time t, and b1 and b2 both occur at a slightly later time t*. Suppose further that m1 is in Tim's mind, m2 in Tom's mind, b1 in Tim's brain and b2 in Tom's brain. What makes it the case that m1 causes b1 rather than b2, and that m2 causes b2 rather than b1? What insures that m1 is paired with b1 and m2 with b2? How, on dualist interactionist assumptions, can we insure that the picture looks like this:

m1 --> b1
m2 --> b2

and not like this:

m1 --> b2
m2 --> b1?

For example, what insures that Tim's desire for sex causes changes in his brain and body that eventuate in certain familiar actions as opposed to causing changes in Tom's brain and body?  Note that one cannot appeal to spatial contiguity to establish the right pairings for the simple reason that m1 and m2 are not in space on substance dualism. Thus one cannot say that m1 pairs with b1 because it is nearer in space to b1 than to b2. It is also clear that neither temporal contiguity nor temporal precedence can establish the right pairings. For both mental events occur at t, and both brain events at the later time t*.

One will be tempted to say that a mind controls the body it is embodied in, not some other body. This is true, but doesn't solve the pairing problem since embodiment is a causal notion: Tim is embodied in that body over which he has (some) causal control, but what insures that this is Tim's body? It is circular to say that Tim's mind causes changes in Tim's body rather than Tom's because Tim's mind is embodied in Tim's body — given that Tim's mind is embodied in that body over which he has causal influence.

The Pairing Objection can be met if we reject an assumption about causation   on which it is based. The assumption is that an event-sequence is a causal sequence in virtue of its falling under a law of nature. The assumption, in other words, is that what makes m1 the cause of b1 is the fact that the m1-b1 sequence instantiates the law that every M-event causes a B-event. On this assumption, there is nothing to distinguish the m1-b1 sequence from the m1-b2 sequence. For they are alike in being instances of the law, and as noted, nothing else distinguishes them.

Borrowing from Michael Tooley (Sosa and Tooley, eds. Causation, Oxford 1993, p. 173), the assumption can also be put as follows:

Weak Reductionism with Respect to Causal Relations. Any two worlds that agree with respect to all of the non-causal properties of, and relations between, particular events or states of affairs, and with respect to all causal laws, must also agree with respect to all of the causal relations between states of affairs.

Spatial contiguity, temporal contiguity, and temporal precedence are examples of non-causal relations. Clearly, on the above assumption there is no way to distinguish the two different pairings displayed above. This implies that the Pairing Problem may be construed as an objection to Weak Reductionism with respect to Causal Relations rather than as an objection to interactionist dualism. In other words, an interactionist dualist who understands his position will not subscribe to Weak Reductionism as just stated; he will adopt a realist theory of causation according to which certain event-sequences are intrinsically causal.  To say that an event-sequence is intrinsically causal is to say that its causality does not derive from instantiation of a law.

What we need, however, is an independent reason to reject Weak Reductionism, a reason independent of the mind-body problem. If the Pairing Problem arises in a purely physical situation, then we have an independent, non-question-begging reason to abandon Weak Reductionism with respect to Causal Relations. So consider the following set-up. (What follows was suggested to me by John Foster's "A Defense of Dualism" in The Case for Dualism, pp. 14-15.)

There are two light bulbs, B1 and B2, in an enclosed space. They are indistinguishable: they are of the same wattage and manufacture, each is connected to a 120V AC power source, etc. When either of the bulbs is on, it causes the temperature in the enclosure to increase by X degrees in an interval of time I. Suppose that this is because of a law of nature according to which running a current through a filament causes it to emit heat. Suppose B1 and B2 are both switched on at the same time, t, and kept on for interval I. During I, the temperature in the enclosure increases by 2X degrees. Clearly, each light's being on causes its own temperature increase in the enclosed space. There are two distinct physical causes c1 and c2 and two distinct effects e1 and e2.

But what makes c1 the cause of e1 rather than the cause of e2, and c2 the cause of e2 rather than the cause of e1? The causal pairings are not fixed by the causal law and the non-causal conditions.  The precedence and contiguity conditions are the same.  This shows that there must be more to causation than instantiation of a law. It shows that Weak Reductionism with respect to Causal Relations is false, or at least not obviously true.

If we abandon the assumption that causal relations can be completely explained in terms of non-causal properties/relations and causal laws, then the problem of psychophysical causal pairings no longer arises. If, on the other hand, we hold fast to Weak Reductionism, then, since the Pairing Problem arises both in the mental-physical and the physical-physical case, this problem cannot be taken to be an objection to interactionist dualism.

Since the Pairing Problem cannot possibly arise if Cartesian egos are spatial, the dualist could also take the view that they are spatial even if immaterial:  they are located where their corresponding brains are located.  See William G. Lycan, Giving Dualism Its Due, p. 12:  "It may be wondered wherein minds are immaterial, if they are spatially located. In at least two ways: They do not have other physical properties such as mass or charge; and unlike brain matter, they are not made of atoms or subatomic particles."

Comments

[doctor logic]

Your physical(ist) example of a pairing problem doesn't seem to be analogous.

The light bulb problem is analogous to the original pairing problem in the case that our instrumentation cannot distinguish b1 from b2. Imagine that the fMRI machines in which Tim and Tom are laying are wired together so that they show just a single numerical sum of the activities of their brains instead of showing two pictures as they normally would. In that scenario, no one would say that any theory of mind was at fault for making the brain effects indistinguishable.

I tried to think of another physical scenario that would do what you want. The double-slit experiment came to mind, but that physical duality says that multiple causes effectively become a single cause under certain microphysical conditions. IOW, it won't really suit your purpose because we're then considering a situation analogous to a mind meld.

Thank you for the clear description of intrinsic causality. I assume that in your picture, the will causes a person to choose A instead of ~A without there being a law specifying that A would be chosen over ~A. That is, the will is the intrinsic cause of the choice.

The problem seems to me that, to the extent the selection was not lawful, it was fundamentally random. In other words, if A causing B without the event being an instance of a general law, is just like having a one-time law that says A causes B in this unique situation at this time. That instance of causation becomes a brute fact of the universe, and brute facts are precisely random facts. That would imply that choice-making is fundamentally random, like a quantum decay. Indeed, quantum decays are isomorphic to this situation. You have some global constraints from conservation of energy/momentum/charge/etc, but beyond those constraints, the outcome is described by a one-time, unique law for each decay. That's the recipe for fundamental randomness. Is there an escape?

[David Brightly]

Bill,

When this question arose earlier I gave a physicist's objection (http://maverickphilosopher.powerblogs.com/posts/1157414389.shtml ) that your account of the light bulb experiment was inadequate because E1 and E2 were indiscernible. You counter-objected that I was confusing what we know with what is, for surely the events E1 and E2 occur even if we can't discern them. This time I will object that your account is simply wrong. For E1 = 'temperature rises by X degrees' and E2 = 'temperature rises by X degrees', so E1=E2. Hence you are saying that only one effect event occurs, viz, 'temperature rises by X degrees'. But this event does *not* happen. The temperature rises by 2X degrees.

[Bill Vallicella]

David,

Are you sure you are not confusing type-identity with token-identity when you write, "For E1 = 'temperature rises by X degrees' and E2 = 'temperature rises by X degrees', so E1=E2"? There is no doubt that these two events are type-identical. But I would insist that they are token-distinct. That just means that there are two of them, that they are not numerically one and the same.

[Bill Vallicella]

Doc,

I am having a hard time following you. The structure of my argument is like this:

1. The Pairing Objection presupposes a certain purely nomological theory of causation according to which an event-sequence is constituted a causal sequence solely by its instantiation of a law of nature.

2. If the dualist has a reason to reject this theory of causation, then he has a reason to reject the Pairing Objection.

3. The light bulb example gives the dualist a reason to reject the theory of causation presupposed by the Paining Objection.

[Matt Hart]

Bill,

Ok, I agree that you've sent the pairing at-a-time problem packing, but what about pairing over-time? Why is it that my mind is forever causing this body to move, and not some other hunk of matter? Bit of a stretch to call it coincidence. I feel tempted to reach for a theistic explanation at this juncture (I find it better than any alternative), but I was wondering if you had anything up your sleeve? Maybe something involving embodiment?

[David Brightly]

Bill,

If we were to do the experiment with one light bulb for a second time I'd agree that two events of the type 'temperature rises by X degrees' had taken place because we would see two distinct rises of X degrees well separated in time. What is your criterion for deciding that in the two bulb experiment *two* events have taken place in the box over the *same* period of time?

[Bill Vallicella]

David,

If each bulb causes the temp. to increase by X degrees, and you agree that when both bulbs are on the temp. increases by 2X degrees, then it seems there would have to be two separate effects that add up.

I suppose you find it arbitrary of me to say that there are two effects of two causes rather than one effect of two causes.

Suppose 10 gals of water come out of spigot A in a bath tub and 10 gals of oil from spigot B. Seems to me there are two effects of two causes and that the two effects add up to 20 gals of fluid.

[Bill Vallicella]

Matt,

God can be put to work, but only when all else fails, on pain of the dreaded deus ex machina. Isn't pairing-over-time solvable in the same way, by holding that the causal relation cannot be analyzed in terms of constant conjunction?

[David Brightly]

Bill,

Do you have in mind a law of the form 'a bulb carrying a current for interval I causes an effect of an X degree rise in temperature in the box'? Two bulbs would then seem to produce two such effects. But isn't this to ignore an implicit ceteris paribus which would rule out a second source of heat in the box? In other words, the two bulb outcome cannot be deduced from the law.

Yes, in this thought experiment involving temperature, I do find it very strange that you find two effects. I have an array of arguments against this analysis. Here is one: Temperature is a global property of the box and its contents (we may assume the air inside is well-stirred). An effect of this type is a rise in temperature of X degrees over a given time interval, ie, a well-specified change in a global property. I find it incoherent to suggest that there could be two synchronic yet distinct changes to a global property. Can you give me a clear example of this? The oily bath doesn't work. I see it as either two effects, viz changes in two distinct properties, 'quantity of water in bath', and 'quantity of oil in bath', or as one effect, viz, a change in the property 'quantity of fluid in bath'. In the former the two running taps are each the cause of the corresponding change in quantity effect. In the latter, the two running taps are contributary causes of a single effect.

But perhaps it doesn't matter. What we have here might be called 'divisibility of causes and effects'. For even in the one bulb experiment we can think of there being two causes, viz the current flowing in the left hand half of the bulb's filament, and the current flowing in the right. And we can generalise to n causes by dividing the filament into n segments. If you insist on divisible effects I will insist on divisible causes and the upshot will be that we will have to agree that all the causes contribute to all the effects. But then the putative pairings evaporate and we no longer have reason to abandon WRRCR.

[Bill Vallicella]

David,

Perhaps you should define 'global property.'

[David Brightly]

Hi Bill,

By 'global' I mean 'applying to all spatial parts without local variation'. Colour, for instance, need not be global, and a change to red at one end of an object could be accompanied by a change to green at the other. But under reasonable conditions (which I think are implicit in the thought experiment), such as that the air is well stirred and the time interval I is long enough in relation to the rate of heating for all parts of the box to reach thermal equilibrium, the temperature will be uniform throughout the box.

[Bill Vallicella]

David,

Let me try a different example. There are two matches alike in all respects. They are side by side but not touching. They are struck simultaneously in such a way that each match head 'experiences' the same degree of friction. Both ignite at the same time and in the same way. we can say the following:

1. Striking 1 causes ignition 1 and striking 2 causes ignition 2.
2. The noncausal factors are the same: with respect to both matches the striking occurs before the ignition, and the striking is spatiotemporally contiguous with the ignition.
3. On a nomological theory of causation, what makes an event sequence a causal sequence is the event sequence's instantion of a law.
4. Suppose there is a law that governs the situation, something like: Every match-striking (with sufficient force, in an oxygen-rich environment, etc) is followed by a match-ignition.
5. If the nomological theory of causation were true, then there would be no difference between S1 causing I1 and S1 causing I2, and no difference between S2 causing I2 and S2 causing I1. But there are these difference. Therefore, the nomological theory is false.

[David Brightly]

Bill,

I agree that in the account you give in your last comment there is no requirement that S1 cause I1 rather than I2. But why is the 'law' in (4) formulated purely in terms of event types? Surely nobody would ever think this strong enough? One needs to say something like: Every match-striking event S is followed by a match-ignition event I and if match M participates in S then it also participates in I.

[Bill Vallicella]

Laws have to be formulated in terms of event-types.