Removing the Head, or Destroying the Brain: Qualia, Part IV

The final twig on this branch of the discussion.

Several years ago, I read an account of a researcher who was trying to create a device that is to olfaction what cameras are to vision. Chemical sensors of various types would be wired to a learning neural net.

At first, this researcher’s team worked hard to produce the right kinds of sensors to produce the right kinds of signals so that the neural net could identify the important properties of chemicals. Trying to determine what those properties were, how their presence should be signaled, and setting up the wiring so that the net could learn to respond to the important signals and so efficiently identify different substances was quite difficult, and the project floundered.

Eventually, they had a revelation. Why bother trying to predetermine what was ‘important’ about chemicals when the neural net could do so itself? Just plunk down a diverse set of sensors that respond to as many different properties as possible, create a complex signal from the result, and let the neural net determine which features were important to distinguishing between substances.


Introspection is not a particularly useful way to gain insight about psychology. But as the ‘qualia’ aspects of cognition are claimed to be part of our experience, it’s reasonable to look at our experiences in order to study qualia.

I tried that.

I quickly found that every attempt I made to describe my experiences upon detecting a stimulus was only a set of references to other experiences. I could describe them in terms of similitudes with motion and speed, temperature, pressure, color, odor and taste. But when I tried to describe those things, all I could do was refer to other, seemingly elemental experiences.

Where was the ‘sensation’ when I looked at something red? Where were the special properties that weaker concepts of qualia insisted were there? I’d already rejected the hypotheses that involved qualia being some magical, logically-incoherent spirit properties. But even the weaker and potentially true ones didn’t match what I felt.

I have no conscious insight into how my liver or pancreas are working. There’s no reason to believe that I would have conscious insight into how a module of my mind was operating if my mind wasn’t set up to monitor its workings. But qualia are supposed to be something we’re conscious of. Why, then, could I not identify any aspect of my conscious experience that matched any of the definitions of ‘qualia’ I’d come across?

‘Faces’ is a concept that’s built into our perceptual nervous system at a basic level. We have found and identified cells and structures specifically tuned to react to visual patterns that even vaguely resemble human faces, and we can analyze the properties necessary to trigger those reactions. But why not other properties?

Babies don’t come preprogrammed with relationships between intentions and movements. Natural selection instead provided modules that generate intentions, and modules that affect movements, and let feedback between modules generate the proper connections. Each individual result is probably unique in its details, although similar at a high level of analysis, in the same way that all human brains are both similar and unique. There’s no need to postulate complex, built-in signals for all the important parts of experience. That’s only necessary for complex features, like faces, that we need to identify very quickly. For the basic things? Why not simply provide sensoria capable of responding to certain properties (like the wavelengths of light) and let the brain develop categorical associations on its own?

Parts of the brain could then enter into arbitrarily-different states, depending on the nature of the signals being sent, and those states can then be referenced as something like pointer variables in computer programming. They’d be arbitrary ‘names’ that exist only to quickly reference and distinguish between recognitions of difference. No further properties would really be necessary.

That was when I really realized that I don’t have qualia. Maybe some of you out there DO have them. But the only thing I sense when I look at a color is that it’s a particular color, and associations between it and other properties.

I have no ineffable experiences not reducible to logical interaction. I don’t even have special states beyond association and recognition. Just senses and namestates.

I think I feel good about that.


10 Responses to “Removing the Head, or Destroying the Brain: Qualia, Part IV”

  1. Maybe the best hypothesis here is that you in particular don’t have qualia. When I see something red, I definitely don’t just see that it’s red; I perceive its redness. Anything propositional about it is created after the fact of experience.

  2. I don’t think anyone can not experience qualia; even knowing a color’s properties and associations is a type of subjective experience. It’s just that qualia really isn’t non-transferable in the theoretical sense. It’s only non-transferable because our brains are not evolved with a direct qualia-transferal channel. We can transfer limited emotions and concepts through speech via associations. But if someone never saw the color red in their life, they would have nothing to associate it with.

  3. “When I see something red, I definitely don’t just see that it’s red; I perceive its redness.”

    First, I will note that your claim is incompatible with the ‘ineffable’ qualia definition.

    Second, your claim is compatible with the lesser, weaker use of the term. It is conceivable that you really are experiencing something fundamentally different than I am. However, your claim (and the many similar to it that I have encountered) is not sufficient to convince me of its accuracy, as my postulating a simple misunderstanding on your part is sufficient to explain why you make the claim if it is not actually true.

    “I don’t think anyone can not experience qualia.”

    An argument of this nature, but with different specific content, is often made by Christian fundamentalists. They assert that everyone knows that God exists, and atheists / agnostics are denying the obvious truth because they’re, in essence, depraved and evil.

    I have just as much respect for your argument as I have for theirs.

  4. After rereading your comment, Jake G., I think I was too harsh in the above response. Please accept my apologies for including you, even temporarily, in a category with Christian fundamentalists.

    You seem to be using ‘qualia’ to refer to a very minimal concept that is much, much more defensible than even the weaker version I refer to.

    Your final claim, though, is clearly wrong. I will make a brief post discussing why.

  5. Hey, no harm done melendwyr; I did not express my position as clearly as I would have liked anyway.

    Just to clarify, I do not really disagree with your position in this series. But I do think that qualia, the subjective part of its definition, is not in contradiction with transferability or a reductionist explanation. According to some definitions, this still qualifies as qualia, and others not. So probably just a disagreement over semantics.

  6. I say that, rigorously speaking, there is no such thing as ‘subjective’. Just objective things that we can’t conveniently perceive.

  7. Any time one thinks a concept is meaningless (in the sense of ‘no such thing’) a choice must be made whether to be eliminitavist about it. If you can simply define ‘subjective’ things as a particular kind of objective thing that we can’t conveniently perceive, then it doesn’t seem obvious that the term serves no good purpose.

  8. As long as you don’t start making exceptions for ‘subjective’ things that don’t apply to ‘objective’ ones, I have no problems.

    Not all rectangles are squares, etc.

  9. Aren’t some substances more “important” to our olfactory sense than others? That rotten egg smell, for instance, can be detected with very few parts per million. Don’t they have to tell the neural net about that sort of thing?

  10. But what molecular properties are most useful for detecting such toxic substances?

    And what if wanted to seek out one particular kind of odor — say, to specialize the circuit for drug- or explosive-sniffing. Trying to explicitly evolve sensors to detect that signal and only that signal is really tough.

Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out /  Change )

Google+ photo

You are commenting using your Google+ account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )


Connecting to %s

%d bloggers like this: