Built on Facts

An exploration of physics, and the search to understand our universe

Built on Facts header image 2
← Something Completely Different That Point is in the Mandelbrot Set →

What’s in your head?

June 6th, 2008 · 3 Comments

Tom Levenson has a guest post at Cosmic Variance that I think illustrates a fundamental flaw in some of the modern ideas about neuroscience. Like some physicists who embrace Many Worlds or other “interpretation” ideas without empirical evidence, many neurosciencists have done essentially the same thing in their own discipline.

The first part of Levenson’s post describes some of the triumphs of neuroscience. He’s completely right. Despite being a field in its infancy, the amount of knowledge gained about the brain in the last decade or two is nothing short of staggering. We’re making tremendous strides in understanding the structural reasons within the brain for mental illness, PTSD, and just about everything else that our minds do. That our thoughts and feelings are generated by the brain is almost too trivially true to bother discussing. But Levinson extends this:

Instead, the trend of current neuroscience seems to argue that the enormously powerful sense each of us has of a self as distinct from the matter of which we are made is false. Our minds, our selves may be real—but they are the outcome of a purely material process taking place in the liter or so of grey stuff between our ears.

And this is where I bail out. Our thought and our mental states are certainly the outcome of a purely material process in our skulls. But to blithely equate thoughts and mental states with “our minds, our selves” seems like quite a leap. I left some brief thoughts in the comments of the post, but decided that such an interesting subject also warranted a bit fuller discussion here.

Consciousness in the sense of “what is it like” is something you and I clearly possess. We can trivially observe it in ourselves. It is possibly an emergent property of the hundred billion or so neurons in our heads (Levenson would say so), and so a mechanical or electronic replica ought to be similarly conscious. So let’s say you built an electronic replica of a human brain and hooked it up to sensors which provided the equivalent of senses so the brain could interact with the outside. Is it conscious? “Yes, because it does everything the wet human brain does.”, many would say. As a physicist, I worry about this. I can say that gravity and static electric fields are both inverse square laws and thus do the same things because both are inverse square laws, and I’d be totally right. But first I have to prove experimentally that both are in fact inverse square laws. This step has been entirely skipped in the assertion that the brain generates consciousness. Is the computer’s consciousness falsifiable even in principle? If not, it’s not science.

I brought up the far-fetched thought experiment of considering our galaxy. It has roughly the same number of stars as a human brain has neurons, and like the neurons they mutually interact. Is the galaxy conscious? Clearly not. But really, what’s so clear about it? What exactly is different qualitatively about the information processing of gravity and the information processing of lots of cells that gives one the ability to generate consciousness but not the other? It probably can’t be what they’re made out of - both could be equally well simulated on the same Turing machine, and I don’t think there’s any reason to believe the “wetness” of biological brains is somehow privileged. So it’s actually quite difficult to say that a galaxy is not conscious if in fact the bare fact of information processing is necessary and sufficient to produce consciousness. This seems implausible, and so I remain unconvinced about that premise. Some people (including one commenter on the post) are in fact totally willing to believe that a galaxy is conscious, but I’m not. Not without evidence.

Now don’t worry. I’m not advocating some strange dualist homunculus. Conscious could well be an emergent property of the brain. There’s just right now exactly zero prospect of testing that experimentally, and I would not be shocked if it turns out to be impossible in principle. Though I would also not be shocked if it doesn’t.

All that is just a very informal heuristic for describing the problem. Let me briefly state a more concise version. It’s possible to conceive of things which do not seem conscious, but are (galaxies, in my silly example). It’s also possible to conceive of things which seem conscious, but aren’t. All the philosophical arguments are so much empty speculation - give me facts. Until an experiment can differentiate between the various combinations of “seems conscious” and “is conscious”, such speculation is simply not science.

Tags: Looking Beyond

3 responses so far ↓

  • 1 Will TS // Jun 6, 2008 at 7:58 pm

    It’s a tricky experimental question. How do I know you are conscious and don’t just seem conscious? If a zombie seems conscious, is that consciousness any less real than the consciousness you display? How do I know if I’m conscious and not just appearing to be conscious? It seems like consciousness is such an imprecise term that it can’t be empirically recognized.

  • 2 Blake Stacey // Jun 10, 2008 at 12:46 am

    So it’s actually quite difficult to say that a galaxy is not conscious if in fact the bare fact of information processing is necessary and sufficient to produce consciousness.

    Straw man argument.

    I would consider it painfully obvious that a mere measure of “information-processing power” is not sufficient to tell whether an entity is “conscious”, and I’d expect neuroscientists to agree. (I’m vain like that.) After all, merely measuring the processing power of the computer on my desk won’t tell you whether it’s running Windows XP or Ubuntu Linux. If the brain is damaged, consciousness changes; perhaps new memories no longer form (patient H.M.), or impressions from different senses can no longer be combined to form a single perception of an object (patient K.E.). We already know that disrupting the connections among neurons alters how people act, think and feel. An equivalent number of processing units could be rewired into a configuration of equivalent complexity while completely destroying the peculiarities which made the original configuration worthwhile.


    I believe it was D. Knuth who said that almost all errors in published mathematics were preceded by the word “obviously”. Chinese Room and philosophical zombie type arguments tend to support the idea that the way people act, think, and feel need not necessarily have anything to do with the experience of consciousness at all. Personally I find both those arguments and their counterarguments unconvincing. I don’t think philosophy is useful in answering questions of fact in the first place, only for finding a starting place for thinking. The problem is that as obvious as it is to you (and me - this is mostly Devil’s advocate really) that galaxies aren’t conscious by virtue of the fact that the information being processed is different, there’s exactly no experimental way to test that seemingly obvious fact.

    Dennett and others believe that this isn’t a problem at all, but I think the Leibniz objection is too robust to be defined away:

    “And supposing there were a machine, so constructed as to think, feel, and have perception, it might be conceived as increased in size, while keeping the same proportions, so that one might go into it as into a mill. That being so, we should, on examining its interior, find only parts which work one upon another, and never anything by which to explain a perception.”

  • 3 Paul Murray // Jun 10, 2008 at 2:12 am

    The big problem is defining the terms.

    The point at which you “bail out” is so fine that I can’t make it out. That yes, thoughts are the *outcome* of material processes but they are not *the same as* those processes.

    For myself, I can’t help noticing that me when I am drunk, or angry, is a very different me from the me that writes this. “Self” is a chimera.

    Matt replies: It looks to me that “Self is a chimera” is functionally the same as “This sentence is not in English” - problematic by the very act of making the statement. You have a conscious experience of writing your comment and I have a conscious experience of writing this reply. If conscious experience is a chimera I’m not exactly sure what wouldn’t be a chimera. But not everything has experiences, and we have no way of knowing what has experiences and what doesn’t. And therein lies the major difficulty in trying to find objective data one way or another about what produces conscious experience and what doesn’t.

    You’re probably right about definitions being unclear here. This is a notoriously difficult thing to discuss without talking past each other. And I can’t shake the nagging suspicion that the “hard problem” (if as per Dennett it exists at all) is in the same class of problem as trying to prove or disprove solipsism. Science can’t do it by virtue of the impossibility of experimental test, and philosophy can’t prove anything at all.

    And finally I want to emphasize the fact that I’m just thinking out loud here. I tell people all the time not to put too much stock in any statement not backed by data. Data is abundant in the empirical questions of how our brains work and what they do, but nonexistent in the questions of just what conscious experience is. Since my speculations are about the latter, and should not be taken as anything other than just that… idle speculation for fun about a very interesting topic.

Leave a Comment