« Nietzsche, Guilt, and Incoherence | Main | Samuel Alexander on Emergence »

Friday, May 15, 2009


Feed You can follow this conversation by subscribing to the comment feed for this post.

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?


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.


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.


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.


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?


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?


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.


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?


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.


Perhaps you should define 'global property.'

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.


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.


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.

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

The comments to this entry are closed.

My Photo
Blog powered by Typepad
Member since 10/2008



January 2023

Sun Mon Tue Wed Thu Fri Sat
1 2 3 4 5 6 7
8 9 10 11 12 13 14
15 16 17 18 19 20 21
22 23 24 25 26 27 28
29 30 31        
Blog powered by Typepad