Monday, January 18, 2010

Independent Discovery and Journalism; Consciousness in New Scientist

It is certainly not an article I expect from such a bastion of orthodoxy as New Scientist. Clearly, the commenters agree. As per my last post, yes, sometimes journalism will rise above its mandate and produce something approaching worth reading. (NS has also published Lee Smolin, whose writing could not be more different than the bulk of the magazine.)

The key is here:

"It is about the deep philosophical confusion embedded in the assumption that if you can correlate neural activity with consciousness, then you have demonstrated they are one and the same thing, and that a physical science such as neurophysiology is able to show what consciousness truly is."

Tallis fails to follow this through properly, I think. As most discussions of consciousness do, it begs the question. On one side, consciousness is physical, and the only thing left is to demonstrate it, while on the other, with Tallis, consciousness is clearly not physical, and it is just a matter of making the other side admit this. The whole point of the debate is whether the neural correlates, or some equivalent, are indeed consciousness itself or not; there's no use in just reasserting one's opinion.

I will now follow it through properly.

Assume we have the neural correlates exactly corresponding to consciousness, and we have completely described the brain. To use the eye, we can trace the trail of causation as photons enter the eye, as spikes travel through nerves and neurons, as the voluntary nervous system swings into action, and the person responds. ("That is one fine and dandy picture you got there.")

I take this model. I change one thing; I assume the correlates are unconscious. Does the model still work?

"The analogy fails as the level at which water can be seen as molecules, on the one hand, and as wet, shiny, cold stuff on the other, are intended to correspond to different "levels" at which we are conscious of it. But the existence of levels of experience or of description presupposes consciousness. Water does not intrinsically have these levels."

By definition, we have no direct access to another's consciousness, we can only correlate neuron spikes with consciousness reports, and liken those reports to our own experience. The reports, however, are themselves unverifiable, and if there's any systematic bias, they will be systematically wrong.

(As any Buddhist can tell you, there's lots of systematic bias. As long as you're not talking about consciousness, a good psychologist can also go on and on about systematic biases.)

I strip the unreliable reports out; the reports are not necessary for the model to work, and indeed should be predictions of the model, not components. I simply show, using our hypothetical exact model, how certain questions cause these answers. As per my one change, the reports are because of the neural correlates, not because of the contents of consciousness. If indeed the materialists are correct, a full causal description of the brain is simply a full description.

Now, hopefully, you can clearly see why Tallis is correct, that it is self-contradictory to look for consciousness in the brain. Materialism contradicts the very existence of consciousness; and indeed all materialists are in fact crypto-dualists. To reiterate, the neuroscientific persuasion of materialist measures a neural correlate, asks the subject what they're experiencing, and then correlates the two. Which is fine that far, but then they assume the subject's response was not caused by physical thing they just measured, but because of their conscious state, and then they assume the conscious state was the thing they just measured. This species works through the problem with dualism, and then assumes the answer is monistic.

To see this another way, to assume as above that the correlates are in fact unconscious is either to assume consciousness doesn't exist or to assume epiphenomenal dualism is true. The fact that the model still works if Ockham's razor is applied like this means that the model either assumes consciousness doesn't exist, or is dualistic.

(So, do you think my characterization of the camps is accurate?)

As a bonus, this kind of answer is not just philosophically wrong, but scientifically wrong as well.

I've seen complaints that dualism is just a myterious answer to a mysterious question. Neural correlates are actually a mysterious answer to a mysterious question. Ask the physicist question; how does this neural correlate know it is supposed to be happiness, and not surprise? As your neural cortex is processing red, why would even bother to know it is processing red, and not just do it? Consider further: it is a bunch of electrons, protons, and neutrons processing red and apparently also knowing that they are processing red. The photon itself is long gone at this point, there's no actual redness around. How do these particles know they're supposed to be correlating to red, and the particles in your hair know they're not supposed to be?

What is qualitatively different about being a neuronal neutron versus a hirsute neutron? How does it know? Because, if it can't know, then it must be the same, and either both unconscious...or both conscious.

Many materialists are aware of these kinds of issues, which tends to bring emergence into the discussion, but unfortunately that is even more absurd, because emergence hardly even pretends to be objective. Briefly, a flock is an emergent structure. Now, read Tallis' bit about water again. To say we became conscious because of emergence is to say we became conscious because of features of consciousness. Physics doesn't recognize flocks. Physics recognizes fundamental particles, and that's all.

To touch on journalism again, it seems obvious to me why Tallis didn't follow through properly; the whole article reeks of being crammed into a tiny space. (Long form Tallis.) While I can say Tallis should not have written the article, or NS should not try to stuff such a complex, difficult, frustrating topic into the do so consistently would violate Sturgeon's Second.

And now for some light fisking. I wonder if the article would be so glibly dismissesd if these errors were corrected.
"If it were identical, then we would be left with the insuperable problem of explaining how intracranial nerve impulses, which are material events, could "reach out" to extracranial objects in order to be "of" or "about" them."
Tallis fails to explain why the problem is insuperable. In fact, the perspective is new to me, so I have no idea if he's right, or even if he's on to something at all.

"Straightforward physical causation explains how light from an object brings about events in the occipital cortex. No such explanation is available as to how those neural events are "about" the physical object. Biophysical science explains how the light gets in but not how the gaze looks out."

I mentioned independent discovery, although that last sentence is more question-beggging.

Further down is,
"I believe there is a fundamental, but not obvious, reason why that explanation will always remain incomplete - or unrealisable."
"There are also problems with notions of the self, with the initiation of action, and with free will. Some neurophilosophers deal with these by denying their existence,"

Finally, Tallis trips up more seriously.

"It does not, as you and I do, reach temporally upstream from the effects of experience to the experience that brought about the effects. In other words, the sense of the past cannot exist in a physical system. This is consistent with the fact that the physics of time does not allow for tenses: Einstein called the distinction between past, present and future a "stubbornly persistent illusion"."

To LIFO... physics does not respect tense because there is only the present. Each fundamental particle only feels its immediate surroundings, both spatially and temporally, as per the principle of locality,[1] and the name for the temporally local is simply 'the present.' Particles don't care about the past or future because, almost by definition, the particles can touch neither.

However, your memories exist in the present. While we look backward by inducing a feeling of looking backward, it doesn't mean the process of remembering means moving backward in any way. The bar against correlation works both ways.

[1] I must note that while our current principle seems very local, as it basically means adjacent, it need not be, and locality can be back-defined; the local is simply the things a particle can touch, wherever they happen to be. For example, you could say entangled particles don't send signals faster than light, but rather bend space.

No comments: