I can't think of a philosophical way to disprove this, and creating an experimental conditions whereby we could even tell if mental states change without brain states relevantly changing seems mind-buggeringly hard.

Mind you, as supervenience is compatible with all theories for the mind except perhaps the most extreme of substance dualisms, then this is not a problem that worries me.

Is there a way to disprove this though? — LibnizMakesMeThrowMyBookAway

Yes. Supervenience assumes (or logically relies) on two assertions, (1) Reductionism and (2) determinism. Disprove either and the concept of supervenience falls.

I can't think of a philosophical way to disprove this, and creating an experimental conditions whereby we could even tell if mental states change without brain states relevantly changing seems mind-buggeringly hard. — bert1

Make a distinction between the philosophically way (abstraction) and scientific way (empirical or practical).

OP asks the former.

It's not just philosophy any more, supervenience of mindstates upon brainstates and neural correlates are essentially the same concepts. Whether this implies that consciousness is just something the brain does or whether it's enabled by the nervous system and sensorimotor constraints too is a different question; I think it's pretty incontestable that there are neural correlates of consciousness in a loose sense at this point.

This might be a bit left-field, but there have been some cases of patients who have had massive brain abnormalities due to some congenital condition or disease, who have nevertheless been able to function normally. One was reported yesterday Doctors find air pocket where part of man's brain should be. There was also a much stranger case reported a couple of years ago about a subject who was apparently missing more than 80% of brain mass. Such anomalies certainly challenge the conventional understanding of the brain: mind relationship.

A related point is the phenomenon of neuroplasticity. This is something that has only been discovered in the last 20 years or so; previously there was a firm belief that the brain had as many cells as it was going to have at the time of birth, and that thereafter brain mass was always lost. However I believe it’s now been found that the brain can re-wire itself so as to activate totally new kinds of abilities. And also it can re-wire itself so as to compensate for loss or damage. All of which seems to suggest a ‘top-down’ causal chain whereby the mind, thought, or intentional action actually produces physical re-configuration of the organ itself; ‘mind over matter’, to use the hackneyed old phrase.

Finally, the idea that a brain state ‘means’ or ‘repreents’ anything at all, is highly questionable. What ‘means’ or ‘represents’ are signs, images and words. And the way we understand what they mean is through interpreting meaning via language and rational inference. Whatever else can be said of brains. I don’t feel that neural activity operates according to the rules of syntax. Instead language and abstract thought are used to infer how the brain (or anything else) does its work. So I think understanding neural functions to ‘represent or mean’ anything is a kind of deep category error.

How would you go about disproving reductionism then, in relation to mind-brain supervenience.?

Trigger Warning:
This doesnt seem like a good enough argument to disprove mind brain supervenience...because the fact is, if say i clone the man with pockets in his brain, it would seem unlikely that his clone wouldnt possess the same mental state or qualia as the original.

To disprove supervenience we would need to observe a change in mind state over a time interval in which the brain state did not change. Since brain states are always changing - think of all the subconscious processing necessary to keep our heart pumping and physiology regulated - there is no time interval in which brain states do not change. So it looks like the theory cannot be tested.

Even if we could somehow determine with certainty which parts of the brain were related to consciousness (neural correlates) and focus on only changes in those parts (and I doubt that a perfect separation of that type will ever be possible in practice), I don't think a test would be possible. Neural activity occurs at levels so tiny that any attempt to observe it in complete detail will change it. So it would not be possible to ascertain whether the neural correlates had changed.

Sorry, I am a bit late, forgot to reply.

Proving that the conceptual understanding of the whole can't be sensibly reduced to its (underlying) parts without fundamentally shifting our conceptual understanding (or approach). Reductionism in other words: the whole is simply the sum of its parts -- nothing more. If you prove that the whole is more than just the sum of its parts, then you basically proved, emergentism (a form of non-reductionism).

The Mereological fallacy surrounding neuroscience is a great example.

Now-- this is by no means a rebuttal of supervenience. It is an ongoing philosophical debate, with a high level of abstractions that goes beyond the purpose of this thread.

Note that I am merely providing some insights and giving some examples -- by no means do I intend to argue either side.


edit: also see mereological nihilism.

It's not just philosophy any more, supervenience of mindstates upon brainstates and neural correlates are essentially the same concepts. — fdrake

I don't think that's right. Supervenience is an asymmetrical relationship: a dependence relation. To say that the mind supervenes on the brain is to say the mind depends on the brain. Conversely it could be said that the brain supervenes on the mind. Correlation would obtain in either case.

Do you think correlation is sufficient for both directions?

Yes. Supervenience assumes (or logically relies) on two assertions, (1) Reductionism and (2) determinism. — Kitty

No, it does not.

Well, the reductionism part may or may not be true, depending on how reductionism is defined. The loosest definition simply equates reductionism with supervenience, although most often something like Nigelian reduction is meant (i.e. deduction of the laws of the reduced system from the laws of the reducing system). If that is the sense of reductionism, then supervenience certainly does not logically depend on it.

And determinism is not relevant either. You can have a deterministic system supervene on a non-deterministic one (e.g. thermodynamics and statistical mechanics), and the other way around (statistical mechanics and molecular dynamics).

Rationally speaking, if it looked at from an analytical perspective instead of an objective perspective, Hume's theory comes into mind when speaking about the validity of mind-brain supervenience.

To disprove supervenience we would need to observe a change in mind state over a time interval in which the brain state did not change. Since brain states are always changing - think of all the subconscious processing necessary to keep our heart pumping and physiology regulated - there is no time interval in which brain states do not change. So it looks like the theory cannot be tested. — andrewk

I think your criterion is too strict - it would apply to any dynamical systems. We don't necessarily need to test supervenience in such direct, literal way. First of all, in order to even talk about supervenience, we need to have two commensurate theories, i.e. we need to have a theory of brain and a theory of mind, and we need to be able to relate these theories to each other, e.g. by relating mental states to neurophysiological states. If we could do that, then it might be possible, through theoretical analysis, to find an instance where the same neurophysiological state can correspond to two different mental states. How do we test that theoretical analysis? Not necessarily directly, as you suggest, but by validating each of the two theories against observations.

Yes, it all depends on how you define supervenience -- no one holds dominion over its meaning. But if you define supervenience without the concepts of reductionism and determinism, one may ask whether the purpose of the term is not lost, and how it contrasts with emergentism.

I'm not sure what you're asking here. Are you asking whether correlation is sufficient to constitute a a dependence relation? If that's the question, then I would say 'no', regardless of the "direction". It might be necessary, though.

On second thoughts numerical correlation doesn't mean anything like A supervening on B.

Two sequences A,B.

A = (1,3)
B=(-1,2)
Cov(A,B)=(-1,1)(0.5,0.5)=-0.5+0.5=0

No A changes without B changes, no B changes without A changes, correlation 0. Odd.

"Nigelian reduction" you mean Nagelian reduction? I kept on thinking about Nigella Lawson cooking :rofl:

Sorry, man, you lost me there. My brain is soft when it comes to that kinda stuff. :cry:

Yeah, sorry :blush: E. Nagel.

It's a constructed example of when you have two sequences of measurements, A and B. Whenever A changes, B changes. Whenever B changes, A changes. The sequences of measurements have no (0) correlation (the Cov calculation). So I don't think nonzero correlation (being correlated) is sufficient or necessary for any supervenience. So when I said earlier 'neural correlates' (like the kind you get in fMRI studies) are going to be observed instances of supervenience because they're observed instances of correlation between mind and brainstates... I was wrong.

But I think I was wrong for an interesting reason. Supervenience doesn't imply the presence of statistical correlation, and statistical correlation doesn't imply the presence of supervenience (this would need another demonstration, but empirically easy to find).

I'm still not sure what this means, fdrake. But...

Whenever A changes, B changes. Whenever B changes, A changes. — fdrake

would not the fact that the changes are parallel constitute correlation?

In a loose sense yes, in a mathematical sense no.

K, then entailment in a logical sense?

Correlation between two variables neither implies nor implies the negation of supervenience.

In other words, if I clone Ben 2 from Ben 1, Ben 2 should have the same psychological qualia/mindset/thinking as Ben 1, given that their brain states are similar. — LibnizMakesMeThrowMyBookAway

The problem here is that things change as time passes. Ben 1 and Ben 2 might be "the same" at the precise moment in which you make them to be 'the same" (although that is really impossible so "similar" is a better word), but with each moment of passing time their difference will increase due to the separation between them. Therefore Ben 1 and Ben 2 will never have "the same" thinking it will just be similar, and this is what we already observe between two distinct people anyway.

(1) Reductionism and (2) determinism. — Kitty

Neither of these are assumptions, but facts of existence upon which all is predicated.

There is something that I have often found puzzling that might reveal a problem for mind-brain supervenience. But it probably depends on mind-brain supervenience entailing epiphenomenalism. It seems to me that epiphenomenalism would have to be the case with mind-brain supervenience, but I'll have to give it more thought. Anyway, here's the gist of the idea:

If my behavior is entirely explained by a series of microphysical causes, without appeal to any extra mental causes, then how does my behavior come to refer to my mental states? We know about our minds and we talk about them. Can information about something get into a system without that something somehow influencing that system?

If mental states depend on physical states, but physical states depend only on other physical states, the mental has no influence on the physical and minds cannot declare their existence in behavior.

Consider that if the behavior of a brain is entirely explained by pure physical causes, whether or not there are any accompanying mental states, those mental states would make no difference to that behavior. The brain would behave the same whether or not mental states exist. The mental states would put no evidence of themselves into the physical system. There would be no reference to them in behavior. And so behavior that seems to refer to mental states doesn't really refer to them. When we talk about mental states, we are talking nonsense. In other words, if I say that I am conscious and I describe what it is like to experience seeing the color red, my behavior would be the same whether or not there really was any such experience. The behavior has nothing to do with the presence of mental states, since they have no causal efficacy.

If epiphenomenalism were true, and mind-brain supervenience seems to entail epiphemenalism, we'd never know anything about our minds or ever say anything about them. Whatever it is we are talking about, it surely isn't our causally inefficacious mental states.

If the physical brain state is what determines the form of the mental state, the mental state won't contain any information about itself or other mental states. Mental states would have no effect on mental states. So the mind, while there might well be an experience, would never refer to itself or its experience. It would have to have some sort of causal influence on its own form in order to come to contain some reference to itself. At least partly, its form would have to depend on the mental and not the physical. And the physical, since physical behavior is involved in our talking about mental states, would have to partially depend on mental states and not fully on physical states.

We could diagram the causal chain as being like the following in the case of epiphenomenalism:


M1 M2 M3
^ ^ ^
P1 > P2 > P3


The mental states are strictly effects, having no causal outflow. And we could remove the mental states without changing the physical.

If, however, mental states were to have some sort of causal influence on the physical state, there would have to be a gap in the physical causal chain. If the behavior is partly explained by the mental states and not fully explained by the physical, then the physical isn't causally closed. But if the physical is causally closed and all physical happenings are fully explained by physical causes, any causal influence of a mental state would have to involve overdetermination.

Consider that for a mental state to come to refer to experiential qualities, to contain any information about mental states, its form must be partly determined by the mental and not fully determined by the physical. In other words, the mental state would have to be partly underdetermined by and independent of the physical and so perfect supervenience wouldn't hold. It would be somewhat independent of its physical base. The underlying or previous physical states wouldn't fully explain the form of the mental state.

So if the mind supervenes on the brain perfectly, how do you explain the existence of discussions of the puzzles involving the mind?

This seems a serious problem for mind-brain supervenience, no?

OK, still not getting it, but I won't press further. :smile: