The Mind as a Model-Maker

I just finished reading The Ego Tunnel by Thomas Metzinger, which is subtitled The Science of the Mind and the Myth of the Self. The book joins a steady stream of books from neuroscientists and philosophers (e.g. Descartes' Error and A Brief Tour of Human Consciousness ) all of which tell more or less the same story: The mind is nothing but the brain; the brain understands the world through model-making; the conscious self is but a part of that model and is therefore an illusion or a "myth." V.S. Ramachandran says it straightforwardly in A Brief Tour of Human Consciousness:

"Our brains are essentially model-making machines. We need to construct useful, virtual-reality simulations of the world that we can act on." [p. 105]

Thomas Metzinger's version is this:

"From a philosopher's point of view, there is a lot of nonsense in this popular notion. We don't create an individual world but only a world-model. Moreover, the whole idea of potentially being directly in touch with reality is a sort of romantic folklore; we know the world only by using representations, because (correctly) representing something is what knowing is." [p. 9]

What is so fascinating about these assertions is that the authors sincerely believe they are breakthroughs based on recent advances in neuroscience, but if they are, then the only thing to which neuroscience has broken through is a poorly understood version of the philosophy of Immanuel Kant. The Metzinger quote seems almost a direct lift from Kant; surely he is familiar with Kant, because he is a philosopher (and a German one at that). In fact, Kant is mentioned on page 26, but more on that passage later.

What makes a great philosopher great is his ability to plumb the full implications and meaning of a particular line of thinking, even if that line of thinking is ultimately mistaken in its foundation. Kant was a great philosopher, and if you want to understand the true meaning of the recent deluge of books on the philosophy of mind and neuroscience, read the Critique of Pure Reason and then the neuroscience books. Kant thought through this whole "mind is a model-maker" principle to its furthest implications, implications the more contemporary philosophers and scientists are still struggling to reach.

The key insight from Kant is this: If the mind is nothing more than a model-maker, if it doesn't have some direct contact with reality in its foundation, if all it knows is its representations of things, then whatever science or philosophy we construct will be a science or philosophy of those representations, not of the reality behind the representations. This is the basis of Kant's distinction between the phenomenal (things as they appear to us) and the noumenal (things as they are in themselves.) The recent philosophers/scientists of the mind, however, don't want a science merely of phenomena; they want a science of the things themselves. But they can't tell us how such a science is possible, so they just quickly brush the problem off and hope we don't notice. In the Metzinger quote above, he notes that knowledge is not merely representing something, but correctly representing it; there is a whole world of philosophy in that "correctly." How is that we can come to know that some representations are correct and others are not? The best answer Metzinger gives is this:

"Nor is it true that we can never get out of the tunnel or know anything about the outside world: Knowledge is possible, for instance, through the cooperation and communication of large groups of people - scientific communities that design and test theories, constantly criticize one another, and exchange empirical data and new hypotheses." [p. 9]

So knowledge is possible because we talk to each other and do science. This is straight question-begging, because science itself can get started only if we already are in possession of certain elements of knowledge (e.g. that mathematical knowledge reflects reality itself and not just our perception of reality. Kant held the latter position, which is why he insisted on the phenomenal/noumenal distinction.)

Metzinger discusses Plato's Cave on p. 22, but without understanding its full implications. The prisoners in Plato's Cave must be released from their chains in order to know the sunlit reality outside the cave. As long as they are chained in the cave and only watching shadows, it doesn't matter how clever their science is, how deep their philosophers think, or how much they talk to each other. Their science will necessarily be a science of the shadows, not a science of the sunlit reality outside the cave. Kant showed the limits of how far the philosophical prisoner could go: He could, through the a priori use of reason, come to know that his thought and his science was a science of shadows (the phenomenal) and not a science of real things themselves (the noumenal.) But he can't go beyond that; he can't know anything about the noumenal itself, only that it is something beyond his ability to comprehend through his science.

There is another implication of Plato's Cave that is sometimes overlooked. Even after the prisoners are released from the Cave and experience the sunlit reality outside the Cave, that experience will mean nothing to them unless their nature already has the capacity to benefit from that experience; they must be able to perceive and comprehend sunlit reality for it to mean anything to them. If their eyes are such that they can only perceive shadows (i.e. two-dimensional images of only two contrasting colors, in black and white without any gray), or their minds are such that they can only process the images of shadows, then the experience outside the cave will be worse than useless. It will be completely unintelligible and probably fatal. Plato, another great philosopher who understood the deep implications of his own thought, recognized this and so, in his story, the prisoners are initially baffled when released from the Cave. They initially refuse to believe that the world outside the Cave is the real world. It takes time for them to grow accustomed to the sunlit world, but that accommodation is only possible because the prisoner's nature already has the potential to understand it. Even while in the Cave, the prisoners have the faculties to perceive and comprehend three-dimensional, sunlit reality should they be exposed to it; those faculties are merely dormant in the Cave and need "waking up" outside the Cave. This is why Plato supposed that man was more a soul than a body, because the most significant aspect of his nature was that which enabled him to comprehend true reality when he encountered it. The soul is the part of man that is "woken up" in philosophy and begins to perceive reality as it is.

Now we may disagree with Plato's notion of human nature, but he at least understood the question and his answer is an attempt to address it. Kant disagreed with Plato about the connection of human nature with basic reality, and so concluded that man can never know more than the shadows (although he can come to know this fact itself, which was Kant's particular contribution.) The contemporary philosophers never reach the debate between Plato and Kant. They bake the Kantian cake by insisting that the mind of man is a model-maker through and through, but avoid eating the cake by insisting that their science of the mind is truly about the mind ("the brain") and not just the phenomena as they appear to scientists.

The real myth here is not the self, but the myth that a model-making mind can transcend its own models through science. Back to Metzinger's mention of Kant on p. 26 of his book:

"Philosophers like Immanuel Kant or Franz Brentano have theorized about this 'unity of consciousness': What exactly is it that, at every single point in time, blends all the different parts of your conscious experience into one single reality? Today it is interesting to note that the first essential insight - knowing that you know something - is mainly discussed in the philosophy of mind, whereas the neuroscience of consciousness focuses on the problem of integration: how the features of objects are bound together. The latter phenomenon - the One-World Problem of dynamic, global integration - is what we must examine if we want to understand the unity of consciousness."

Metzinger has gotten half of Kant but missed the other half. The blending of conscious experience into one single reality includes the empirical unity of consciousness, or the One-World Problem. But there is a deeper unity implicit, the transcendental unity of apperception, that is the unity of the thought of the scientist analyzing the empirical unity of consciousness. We might call this the One-Science Problem. Here is a concrete case: The neuroscientist examines the subjective experience of time of subjects in the lab. He notes how long they think different things have taken in different contexts, how long they think dreams are compared to how long they really are, that sort of thing. He stimulates various parts of the subject's brain and induces different time disorientations; the subject suddenly thinks it is tomorrow, or yesterday, or that lifting his hand took an hour instead of a second. The details don't matter. The point is that there is a distinction between the time experienced by the subject, and the time the scientist measures in his lab through his instruments; but both depend on structures of the human mind for their meaning. The time measured by the scientist in the laboratory, against which he judges the subjective time of his subject, has meaning only in terms of the concepts of before, now, after and duration that are basic cognitive elements of the human mind; science does not discover these cognitive elements but is constructed in their terms. What is the unity that brings these cognitive elements together into one science? It's not the empirical unity of consciousness, because the scientist is taking apart the empirical unity of consciousness in terms of his science. It is a unity that is not apparent in empirical experience, but is implied in our conduct of a science that permits us to decompose the empirical unity of consciousness. Kant called this deeper unity transcendental because it transcends empirics altogether; we know about it not through scientific analysis of empirical data but through pure reason, or thinking about what must be necessary for science to happen at all.

Either these basic cognitive elements of the mind are reflective of reality or they are not. If they are not (i.e. if our minds are model-makers through and through) then our science can't be a science of reality as such. Our construction of scientific theories to make sense of our experience may be models, but their basic cognitive elements (before, now, after, here, there, unity, multiplicity) had better not be, or we've got to eat the Kantian cake, frosting and all. If we don't want to end up with Kant, the only way is to accept that the human mind in its basic cognitive elements directly reflects reality. When we are confused about reality, it is not because our basic contact with reality is confused, or is merely a model of reality, but because we have constructed a false perception or understanding of reality through the elements of reality itself.

When reading a philosopher who starts with the principle that the mind is thoroughly a model-maker, but also wants to have a science that reflects reality itself, it is always possible to find a place where he surreptitiously slips out of Plato's Cave to find an anchor in reality, then slips back in the Cave and claims he never left. One of the favorite ways of doing this is by invoking evolutionary theory. We evolved as model-makers, you see, and we can have confidence that our models accurately reflect reality because organisms with poor models simply won't survive as well as organisms with better models. Over the millions of years of human evolution, this resulted in the very accurate models we human beings now possess.

That's a fine story, but it all hinges on this proposition: "Organisms with more accurate models of the world will survive better than organisms with less accurate models." Is this proposition a model of the world, or does it directly reflect reality? If it's merely a model of the world, then we can't use it to bootstrap all our other models, can we? It might be that, in true reality, it doesn't matter what sort of models organisms have, or they might not have models at all, or even that our model of the world as having "organisms" is itself mistaken. This tactic works only if it is granted that the evolutionary principle is an absolute anchor in the real world, not merely yet another model constructed by the model-making mind; which is what it must be if the mind is purely a model-maker. Typically, this assumption is never made explicit but is hidden in a bold assertion of evolutionary principle.