The Philosophy of Mind, 1

by Stuart Kauffman

My purpose in the next several posts is to begin to explore major issues in the Philosophy of Mind. My hope is that it may be possible to bring to bear material from past posts to help resolve deep issues that have been with us at least since Descartes, in 1650. To do so, we need a brief outline of these issues.
Descartes famously postulated two kinds of "substance" in the universe, res extensa and res cogitans. Res extensa is, roughly the physical world. Res cogitans is, roughly, mind and consciousness. This view of two kinds of substances is called "dualism".
It was clear to Descartes that his dualism raised the deep issue of how mind/consciousness acts on the physical matter of the body, including the issue of how the mind can have a morally responsible free will. His own proposal was that the mind acted on the body via the pineal gland, a single gland in the brain.
With respect to res extensa, Descartes was an early mechanist: the world, and the bodies of animals and humans were conceived of as clock like machines, gears, escapements and so forth.
This early view was profoundly enriched by Newton's three laws of motion, universal gravitation, and the invention by Newton and Leibnitz of the calculus.
Recall that Aristotle had considered four causes, formal, material, final and efficient: the formal cause of a house is the blueprint, the material cause of the house are the bricks and mortar, the final cause of the house is your decision to build it, and the efficient cause is the actual process of building the house. As I have mentioned, Aristotle also offered a model for explanation in science: deduction: All men are mortal, Socrates is a man, therefore Socrates is mortal. As Robert Rosen pointed out in 'Life Itself', with Newton's laws in differential equation form, initial and boundary conditions, one has, say for a table of billiard balls, the initial positions and momenta of all the balls, the boundary conditions of the billiard table, then the physicist integrates the differential equations to get the trajectories the balls will follow. But as Rosen points out, integration is exactly Deduction. So Newton mathematized Aristotle's efficient case as integration of the differential equations for a system, and that became the "new" mechanical world view", for example Celestial Mechanics.
Note two things about Newton's triumph. First, from the perspective of an Alfred North Whitehead, or myself, who are both interested in exploring ontologically REAL Possibles giving rise to Ontologically real Actuals that give rise to Ontologically real Possibles, Newton's and Einstein's worlds are only Actuals. Thus, for Newton, what IS the Actual state of the billiard ball system in terms of positions and momenta, causally determines the next Actual state of the system, as revealed by integration of his equations of motion. Second, for Newton all events have sufficient Actual causes. There is no uncaused event. The forward and past (due to time reversibility of Newton's laws) trajectories of the billiard balls is entirely determined by the present state of the billiard balls's positions and momenta, and a sequence of Actuals that beget the next or preceding Actual.
With this in mind, let's return to a Cartesian dualism: We have a res extensa, described by Newton's deterministic equations. Now pass to a deterministic set of equations describing the dynamical neural behavior of the brain. Then we confront the profound problem: How does mind/consciousness - as we exprience it with our sense of free will - cause matter, res extensa to change? The standard philosophy of mind arguments are straightforward: 1) The brain, as a physical system is causally sufficient to generate the next Actual state of the brain, so there is nothing for mind to do. 2) Besides, there is no causal means by which Mind/Consciousness can act on matter.
We're stuck! One response was Idealism: this was Bishop Berkeley's theory that all is Res Cogitans, and what seem to be Res Extensa are so because they are held in the mind of God. There are a variety of Idealist positions spawned by Berkeley.
The modern view is called the "mind-brain" identity theory. It holds that causal events in the brain, say circuits of neurons firing creating a nonlinear dynamical system among about 10 to the 11th power neurons, constitute the causal behavior of the (classical) neurons, much like Newton, where Actuals give rise to the next Actuals in a state space of the firing activities of all 10 to the 11th power neurons at each moment, hence a flow along trajectories in this state space. But, since the mind and brain are identical, this very dynamical activity is mind and consciousness.
The mind-brain identity view is dominant today, and I agree with it, although not as formulated above.
  A subspecies of this mind-brain identity theory is called "connectionism". Here is the idea. A neuron firing or not can be thought of as the truth or falseness of a proposition, say, "An edge at angle X is here in the visual field". Then the firing of all the neurons is like a computer calculating, a Turing machine, and vastly many computable functions can be carried out by the brain, calculating the true or falseness of vastly many propositions.
The neurological correlate of this is easily seen by receptor fields. Hubel and Wiesel showed in the 1950s that a given neural ganglion at the back of your eye could be stimulated to fire by points of light directed at small spots on your retina, but inhibited from firing by directing the light at a small circle on your retina surrounding the central excitatory spot. This is an on center off surround receptor field. Later they showed that there are receptor fields that respond to short lines of light with specific spatial orientations, and these are called edge detectors, for it was soon realized that a set of these receptor fields firing together could detect a straight edge spanning a number of edge receptor fields oriented in the same direction.
This triumph led to the "grandmother" cell theory, in which a specific neuron would fire if and only if you looked at your grandmother. This theory has fallen into disfavor, as such cells have not been found. More, this Connectionist vision of the brain typically is associated with the claim that the mind is algorithmic, a computational machine like a classical computer. In a later blog I will suggest that this view is deeply wrong, the mind is not algorithmic. But that is for later...
So what are the philosophic problems with the mind-brain identity theory?
"Well", a philosopher of mind commonly says, "let's take the "meat-brain" part of the mind-brain identity. Once again, it seems as if the Actual state of the brain is causally sufficient for the next Actual state of the brain, so again, there is nothing for mind to do." (You may wonder why the philosopher gets to ask this question if mind and brain are identical, but an actual survey of a modest number of good philosophers shows that they do ask this question.)
Worse, there seems no way for mind to act on the brain - just as in dualism.
One response is to say the conscious mind is a mere "epiphenomenon", of no power on its own to cause anything in the world of matter. Here, consciousness is either an illusion, or in any case, ineffective in the real world. This may be the position of Tom Clark and Ursula, my co-blogger, in her recent post when she says of Ursula, "she did this and that", meaning her brain did it, not Ursula. (I may have misunderstood a very good friend, who will surely tell me if I've misunderstood.)
So what about morally responsible free will? Here is the standard dilemma. The mind-brain is a deterministic dynamical system, a la Newton, but different "neural" equations. So you were determined by that dynamics to kill the old man in the wheel chair. Not your fault. You didn't do it. No responsible free will.
Alternatively, we have a little or a lot of quantum chance, in the simplest case, like the random decay of a radioactive nucleus. So you are sauntering down the street, and by random chance, kill the same old man in the wheel chair. Not your fault, just a random event.
Again we are stuck. We can have no responsible free will.
I will offer an argument later about a coherent, decohereing-recohering mind-brain system that has consequences for the classical world via decoherence, but decoherence that is acausal, so mind can have consequences for matter without acting causally on matter. I think this is actually very hopeful. It answers cleanly one of the outstanding problems in the philosophy of mind - IF the mind-brain system is quantum coherent, decohering and recohering, which none of us knows as yet.
A responsible free will is harder.
But there are further problems:
Why did consciouisness evolve anyway? Suppose an algorithmic robot with sensors could calculate exactly what will happen in its world. Why bother to be "aware" of its world, just buzz around and plug yourself into power sockets and pop oil into your joints.
What use, if mind is a machine, such as a connectionist machine, is there in being conscious, having awareness, or "qualia"?
Now you might think that an answer is that the robot cannot compute its world exactly due to measurement error, or what is called "deterministic chaos", the famous butterfly effect where a butterfly in Rio can change the weather in Chicago because tiny changes in initial conditions lead to widely varying trajectories in state space.
But this answer won't do. The unconscious, but able computerized robot could sense the difference between its predictions about the world and the world, reset the initial conditions and recompute. Why be conscious? There seems no answer.
This blog is at the philosophy of mind 101 level, of course, but I hope it set in place the context of the problems we face. In my next blogs I hope to take up these issues.
Post Script: In a recent blog, 'Can a Changing Adjacent Possible Acausally Change History? The Open Universe, IV', I discussed 'Vast chemical reaction graphs' and the flow of a small amount of matter on such graphs. I claimed equilibrium would never be reached in the lifetime of the universe, that fluctuations would not damp out but be history dependent, due to those very fluctuations that do not damp out, and that, if we interpret Quantum Mechanics in terms of an ontologically 'real possible', the entire process is acausal. A new study, just published in the Proceedings of the National Academy of Science, USA, of the famous Murchison meteorite that fell in Australia about 40 years ago, used high resolution mass spectrometry and NMR, able to detect mass differences of 1 electron. The study detected 14,000 organic compounds from which, the authors say, millions of organic compounds can be made. The data suggest the meteorite may be older than the sun and carry compounds from early in the formation of the solar system. My addition is that space chemistry may indeed be a flow on a hypo-polulated "vast" reaction graph as we have discussed.
With respect to the philosophy of mind for later blogs, this flow, if interpreted as I have done above, is acausal, lawless but non-random in its historical contingency, and offers a potential way out of the bind about free will as either deterministic, hence not free, or merely quantum random like the decay of a single radioactive nucleus, hence just random, so not morally responsible as we just happen to kill the old man in the wheel chair.
'Scientists say that a meteorite that crashed into Earth 40 years ago contains millions of different carbon-containing, or organic, molecules.'
More information: High molecular diversity of extraterrestrial organic matter in Murchison meteorite revealed 40 years after its fall, Philippe Schmitt-Kopplin et al., PNAS February 16, 2010 vol. 107 no. 7 2763-2768, doi: 10.1073/pnas.0912157107

How Can Mind Act On Matter?

 

by Stuart Kauffman

Rene' Descartes in about 1650, when Descartes posted two substances, Res Extentsa and Res Cogitans. Res Extensa is to be the material, and for Descartes the mechanical world of gears and direct physical causation in which the human body which was thought as a machine. With Newton's laws in differential equation form, initial and boundary conditions, Descartes' machine became the deterministic trajectories in the prestated "state spaces" of classical physics. Res Cogitans is the world of the "mind", brieflly, our subjective conscious experiences, or what philosophers call "qualia".

dierk schaefer/via flickr
Can we use quantum mechanics to understand how our minds can act on matter?

Since Descartes, three central problems have plagued our thinking about mind and matter: 1) How can mind act on matter? It seems that the current state of the brain is sufficient for the next state of the brain, so there is nothing for mind to do. And there seems no way for mind to act on matter anyway. 2) How can we have a morally responsible free will? 3) The "hard problem", what is consciousness, after all? Three and a half centuries of trying have produced no satisfactory answers to any of these questions. Indeed, John Searle, a famous philosopher of mind, has famously quipped, "Not only do we have no idea what consciousness IS, we have no idea what it would even be like to have an idea what consciousness is!".
Notwithstanding three and a half centuries of difficulty, in the next three blogs, I will propose what I take to be the start of answers to all three questions. I follow Roger Penrose who first suggested in his "The Emporer's New Mind" that consciousness has something to do with quantum mechanics. However Penrose ties his attempt to quantum gravity.
I will take a radically different approach: To issue 1) I will propose in this blog that mind and brain are "identical", but that the total system is both quantum and classical, more that it is a system poised in a realm between persistent acausal decoherence from quant;um choherence partially or completely to "classicity", and recoherence partially or completely to quantum coherence. Thus, quantum mind can have consequences for classical matter acausally, without acting on causally on matter.
In principle, this answers "how mind can act on matter",. It does doesnot act causally on matter, but can do so acausally by acausally losing phase information from the mind-brain system to the environment, hence decohering from the quantum "mere possibilities" to classical matter. I will present below the evidence, theoretical and now experimental that supports this possible poised state of the mind-brain system, reversibly decohering and recohering from quantum to classical and back.
In the next blog, I will confront the familliar dilemna: If the mind-brain system is determinsitic, we have no free will at all. If the mind-brain system is sometimes quantum, hence purely random on any interpretation of quantum mechanics, then we are not morally responsible for our random actions. I will show that a partially decoherent-recoherent mind-brain system is almost certainly lawless in its detailed behavior, that that behavior can nonetheless yield historical behavior that is highly non-random in its historical "becoming" and that there is an interpretation of my hypothesis for the poised mind-brain system which allows the system to "decide" and acausally yield specific classical behaviors. This will be my basis for a morally responsible free will. I emphasize that at this stage it is more important to conceive of a possible answer than to be right in scientific detail.
In the third blog I will make the concrete suggestion that conscious, the hard problem, can be solved if we identify this poised quantum decohering-recohering mind-brain system with consciousness itself. I will propose speculative, but possible and ultimately testable identifications of this hypothesis with specific neural correlates comprised of the neural transmitter molecules in synaptic vesicles, and/or the post-synaptic receptors for those neurotransmitters and/or transmembrane ion and other channels on dendrites and nerve axons. Again any possible hypothesis, particularly testable ones, can hope to be more helpful than Searle's quip above.
As a scientist, I think my hypothesis, at present, is "just possible", but certainly improbable, ultimately testable, and may be the most hopeful and investigatable set of hypotheses we now have .
The last paragraphs are brave language. Here is why I do not like the reigning best hypothesis, which derives from the view of the mind as algorithmic: I do not think the mind is algorithmic! I discussed this in my last blog, "Is The Human Mind Algorithmic?"
  A cogent discussion of the algorithmic view is put forward by Daniel Dennett in "Freedom Evolves", a fine book. Dennett rightly notes that John Conway's famous Game of Life, played on a large square lattice that can grow indefinitely in size as specified in a moment has the following properties: 1) The cellular automata rules of the game of life define for each square on the lattice its next state as a definite, defined logical, or Boolean function of the eight neighbors to that square and the square itself. All squares have the same logical rule. Each square can take the values 1 or 0, "white" or "black". We therefore know completely the "physics" of the completely deterministic system in which all squares update their values, 1 or 0, simultaneoujsly each time a discrete imaginary clock ticks. 2) Conway showed that the game of life could give rise to patterns of black and white squares, called gliders, which propagate across the lattice. The speed of the gliders moving across the lattice constitutes a "speed of light" in this physics and for our purposes, the size of the ever finite lattice need merely increase as fast or faster than this speed of light for the remaining results. Next, glider "guns" can emit gliders. Finally, an wonderfully, Conway showed that these gliders and guns could constitute a universal Turing machine!
Now, recall the famous "Halting problem" in computer science. Here the issue is that we want an algorithm which can examine any other algorithm which is to compute an answer, print it and halt. Our hoped for examining algorithm is asked to say whether the examined algorithm will halt in finite time as it works on its infinite Turing machine tape. Turing famously proved that the Halting problem is formally undecidable! That is, we cannot have a formal procedure, that is, an examining algorithm, which can say ahead of time what the examined algorithm will do, halt or not halt. The essential result is that, although we know completely the underlying Boolean function,or "physics" of the Game of Life on the growing two dimensional lattice, we cannot say ahead of time what the behavior of the every growing lattice will be.
The above remarkable result, firmly proven, is the posterchild for "emergent" behavior. And indeed it is emergent behavior: Once we have the above, the behavior on the growing lattice cannot be deduced from the underlying deterministic physics. Wonderful result.
Nevertheless, reducible to the underlying physics of the Game of Life or not, Dennett's Turing machine glider and glider gun system remains perfectly algorithmic!
What is wrong with this beautiful emergence as a model for mind?
I see an enormous pair of problems, despite Dennett's attempt:
1) As I argued in my last blog,"Is The Mind Algorithmic"?, the mind is not algorithmic. I refer the reader to that last blog, but briefly: Can you finitely name all the uses of a screw driver?
2) And is there a mathematics in which all those uses are theorems! I surely do not see such a mathematics. And if not, it is not true that everything that unfolds in the universe is mathematizable. If you object that here we assume a responsible free will for the possible uses of a screw driver, well yes I do. But the same is true for the emergence of Darwinian exaptations in the biosphere 2 billion years ago where we need no appeal to conscious thought at all. That unfolding becoming of the biosphere, I claim cannot be finitely, or even denumerably infinitely, prestated, nor can it be mathematized. The becoming by Darwinian exaptations cannot be denumerably infinitely prestated because to do so we would need an effective procedure to create an ordering relation, like the integers, listing the first, second, third...to infinity possible preadaptation, but in "Breaking the Galilean Spell" I think we agreed that we cannot prestate all the possible Darwinian exaptations into the Adjacent Possible of the bisophere. It seems we can have no ordering listing of those unnamable possibilities.
With respect to 1) above, can you finitely, or denumerably infinitely, name all the possible uses of a screw driver, I think you cannot. Watch: The screw driver can be used to screw in screws, pop open paint can lids, wedge doors open, wedge doors shut, peel putty, used to stab an assailant, tied to the end of a bamboo pole to spear a fish, used with a rock to chop down a (small) tree, leaned agains a wall with the flat side of the tip of the screw driver at right angles to the wall, and used to prop up a square piece of plywood leaning against the wall and supported by the screw driver from falling on a valuable pot,.....
Notice two features of the above. First, relational features matter in most cases above, for example the angle of the screw driver to the wall, and its being turned properly to support the square plywood board. To prestate both all the possible relational features of the screw driver with all entities in the universe from, say atoms to molecules to plywood boards and brick, to fish....from two to many "things" at a time, and simultaneously to prestate all the uses or purposes to which the screw driver can be put, seems impossible. Certainly if spacetime is taken as continuous, it is second order infinite, and no finite list of relational features is possible. no finite or even denumerably infinite list is possible, for, again, no ordering relation among the above, a first, second, third, use of a screw driver, can be devised. How would we create such an ordering?
As I argued in the previous blog, what computer scientists do is list a finite set of "affordances" of screw drivers, "is a", "does a", has a", "needs a", and do neat things. But if the weird use we or James Bond or McGiver wishes to make use of the screw driver to accomplish is not deducible from those affordances, then there is no algorithmic way to get to such uses. My claim is that not all possible relation features and uses of screw drivers can be deduced from any finite set of affordances, nor is there any mathematics that can entail all such relational uses, but we humans, in particular Bond or McGiver in a pinch, find the inventive uses all the time. The mind is not algorithmic.
If not, we have, as yet, no theory of mind.
I think a deep issue arises here: I am claiming that something is impossible, but have not proved it. Is it possible to prove such a claim about no ordering being possible? I don't know how. And I will take two lines with respect to this.
First, it may be what philosophers call a "category mistake" to seek a mathematical proof from that which is not open to mathematical proof. Maybe we just need to think.
Second, I have and here again suggest that the co-evolution of the quantum-classical system in its environment is beyond law, hence cannot be algorithmic. In particular, for a quantum system in a quantum environment, when the former's quantum degrees of freedom are roughly uniformly partially decoherent, the Schrodinger wave equation cannot be propagated unitarily to preserve probability for all the quantum possibilities. While the Schrodinger equation for a fully coherent quantum system is lawful, we simply do not know if a quantum system losing phase information to its environment is describable by a law. If such a system is not describale by law, that constitutes a proof that such processes are real, and non-agorithmic, so not all that unfolds is algorithmic.
Having attempted again to counter the best present view of the mind as algorithmic, I now proceed to my own, radically different approach based on a quantum cohereing, decohering, recohering mind-brains system.
Let's look again at Cartesian dualism, Res Extensa and Res Cogitans. But Res Extensa according to DesCartes, Newton and classical physics is a deterministic dynamical system. Then the state of the brain - Res Extensa, is sufficient for the next state of the brain and there is nothing for mind to do. Worse, there is no way for Res Cogitans do do that something to the brain.
The situation is not helped by the mind-brain identity theory, for again philosophers do ask, and we can ask: If the brain-meat of the brain is a sufficient determinsitic condition for the next state of the brain, there is nothing for mind to do. Worse, the "mind" conscious part of the mind-brain system has no causal way to act on the brain!
Claims to stochastic equations for the dynamics of the brain will not help us, for these are merely "epistemological" stochastic events such as chaos when we cannot measure initial conditions accurately enough. The system is still deterministic ontologically.
But, critically, the above dilemma is purely stated in classical physics.
The world is not limited to classical physics, quantum mechanics also applies.
Now let me state again my hypothesis then defend it: The mind-brain system is quantum coherent, persistently decoheres to classicity and recoheres again to quantum coherent. The mind-brain system, since acausal decoherence takes an interval of time at least a femtosecond, can exist in a poised realm, see previous post, in which most or all its degrees of freedom are partially decoherent as it acausally loses phase information to the environment.
Then the mind does not act causally on the meat of the brain. Rather, the mind decoheres acausally to a classical "meat" brain state that has consequences for the classical aspects of the mind-brain system. Then mind can act on the brain, but does so acausally via decoherence.
Fine, I have suggested a solution to the problem of how mind can "act on" brain acausally. But the mind keeps acting on brain, so I am forced to assume that the classical, or classical (for all practical purposes - as some physicists say) brain can recohere partially or completely to the coherent quantum state.
In the remainder of this blog I will discuss evidence supporting decoherence and more tentative evidence supporting recoherence. In my next blog on a possible view of a responsible free will I will very tentatively identify the sites of such coherence, decoherence and recoherence as neurotransmitter molecules in synaptic vessicles, their post-synaptic receptors, and transmembrane channels in dendrites and possibly axons.
Decoherence is well extabished and is the bane of people trying to construct quantum computers. Different physicists differ on whether decoherence can proceed completely to classicity, or classicity "for all practical purposes", hence whether classical behavior always maintains some residue of quantum aspects.
Four lines of evidence support recoherence:
Shor's quantum error correction theorem establishes mathematically that injection of "information" about quantum phases (and amplitudes), can correct a quantum computer such that its quantum degrees of freedom return to coherence. This, of course, requires that the quantum system be distinguished from the quantum environment that supplies the phase and amplitude information, hence that the quantum system be "open".
Hans Briegel in the Physics Department at the University of Innsbruck, Austria, has published two theoretical papers claiming that a molecule can pass repeatedly from a quantum coherent "entangled" state to a classical state and back.
Recently, as noted by a commenter on my previous blog about a Poised Realm between quantum coherence and classical behavior, it has been proposed that a poised resonant state that is partially decoherent may account for high temperature superconducting.
Most striking, there is direct experimental evidence based on the familiar chlorophyll molecule, which carries out photosynthesis, and the "antenna" protein which is wrapped around the chlorophyll molecule. The normal time scale for decoherence is 10 to the- 15 seconds, a femtosecond. These workers have shown experimentally that coherence in chlorophyll at 77K, where K is Kelvin, absolute temperature and room temperature is 300K, can last at least as long as 7000 femtoseconds, or almost a nanosecond. This discovery is leading to a new field of quantum biology. Now it is thought that the antenna protein somehow "suppresses" decoherence. This is currently directly experimentally testable by using mutants of the antenna protein. But there is this further thought: chlorophyll, as a quantum system, is losing phase information to its environment. It is hard to imagine how the antenna protein can prevent this loss of phase information. Instead, it seems equally or more plausible that the antenna protein is acting, like Shor's quantum error correction algorithm above, to inject phase and amplitude information into chlorophyll. No one knows, so I will assume that the antenna protein injects phase and amplitude information into the chlorophyll molecule. Then it is hard to see how that injected information can exactly match the phase information chlorophyll is losing to the environment. If not, the temporal behavior of chlorophyll as a molecule in its envronment, in which many or all quantum degrees of freedom of chlorophyll are partly decoherent, cannot be described by the Schrodinger equation which is only able to unitarily propagates all the quantum possibilities together with all their phase and amplitude information. In short, chlorophyll is likely to be an open system gaining and losing phase and amplitude information all the time from and to its environment, and therefore it is not clear that a law describes its behavior. In particular, we do not know how that phase information is being lost to the environment, either in general, or due to Popper's argument in a Special Relativity setting as described in a previous blog.
In summary, it is plausible that both quantum decoherence and recoherence partially or totally to a quantum coherent state can happen, perhaps lawlessly from a quantum system to its environment - the universe.
Then my proposal is that the mind-brain system is such a quantum coherent, decohering to classicity, perhaps for all practical purposes, and recohering system, in which mind has acausal consequences for classical brain matter without acting causally on brain matter.
How mind acts on matter has a proposed answer.

The Habitual Brain: How Routine Action And Thought Are The Structure Of Life

by Alva Noë

Habits are a blessing, and a curse.
What is a habit?
It is a skillfully routinized manner of action or thought.
Some philosophers have unwisely supposed that a habit-free existence would be preferable to our own. In this they are mistaken. But the deeper error is the supposition that such an existence might even be possible.
A habit-free existence would be a robotic existence; it would be one in which nothing could be taken for granted. But if nothing can be taken for granted, you can’t get started on anything. How could you talk, if you couldn’t take your own fluency, that is to say, your own habitual mastery of words, meanings, and ways of talking, for granted? How could you read the newspaper? Imagine that you had to think about and decide where to put your feet in the morning! Without habits nothing recognizable as a human or even animal form of life would be possible. To have a mind like ours, you need habits like ours. The fields of robotics and AI should note this well: instead of making clever robots, they should try to design machines with habits! Habits of thought and action form the skeletal structure of life as we know it.
I have been told that Goethe said that architecture is frozen music. Actually, architecture is frozen habit.
  The park near my house has two kinds of paths. There are paved paths laid out by the park’s original designers. And there are unpaved paths created by pedestrians who were unwilling to be constrained by those latter. Both are bona fide pathways. One set were planned and executed. The other arose out of the habitual activity of the park’s users.
An environment, natural, or human-made, reflects a compromise between design and habit. Exactly the same is true of our bodies. Indeed, habit is, or ought to be considered as, a fundamental category for biological thought.
Consider this question: What makes a region of cortex visual rather than auditory? What explains the assignment of visual function to a cortical field? Answer: experience. The assignment of function is the result of species-wide habits of dynamic exchange with the environment. This is why when an individual mammal is born blind, its “visual” structures get conscripted to perform non-visual functions; function answers to the animal’s real-world transactions. Although genetic factors — which themselves reflect ancestral environmental pressures and adaptations — largely govern the growth of anatomical structures in the brain, non-genetic, environmental (epigenetic) factors determine their function. The blind mole-rat, for example, has eyes, but they are tiny and disused; despite this, the retina of the developing mole-rat sprouts cellular projections into what would be, in related mammalian species, “visual” areas of the brain. In the mole-rat, however, these morphologically visual areas perform mainly non-visual sensory functions. (On the evolution of cortex and the competition between genetic and epigenetic factors, I recommend Leah Krubitzers’s “The magificent compromise: cortical field evolution in mammals.” Neuron 56, 2007: 201-208. I rely on this paper here.)
A lot of human somatosensory cortical real estate is devoted to the hands; it wouldn’t be if the hands weren’t so well-suited to human needs and predicaments. And of course the hands wouldn’t be so well-suited if not for the correspondingly adequate cortical infrastructure. This is a virtuous circle. We see this sort of circle everywhere. We can digest milk because we drink milk; we drink milk because its nourishing and we can digest it. Mutations in genes governing cell growth and death can affect the anatomy of a region of cortex. But likewise, cortex gets affected by mutations in genes governing the growth of sensory appendages such has hands, limbs and bills. This is why platypus touch-cortex is hugely devoted to supporting the use of the bill as a perceptual organ. (Another example of this sort of circular “cortical magnification” is the enlargement of auditory cortex in echolocating bats.)
All this shows that how an animal habitually lives is a crucial element in understanding the biological processes whereby the animal — taken as an individual or as a species — becomes the sort of animal that it is.
I believe that habit plays an even deeper, more constitutive role in shaping mind and brain. Let’s go back to the question what makes a region of cortex visual as opposed to auditory. I don’t now mean, what causes it to acquire this function. I mean, what is it for it to have acquired this function. Or better even: what is the function it acquires? I have argued — first in a series of papers with the late philosopher Susan Hurley, and later in my book Out of Our Heads: Why You Are Not Your Brain and Other Lessons From the Biology of Consciousness (Hill and Wang, 2009) — that you can’t explain the distinctively visual character of neural activity in a cortical region without looking at the embedding of that neural activity in a behavioral context. Movements of the body produce sensory change. As Paris-based psychologist Kevin O’Regan and I had argued, to each of the distinct sensory modalities (seeing, touching, etc) there corresponds a distinct way in which movement produces characteristic sensory events. Hence the hypothesis: what makes neural activity visual (for example) is not the intrinsic character of the neural activity, but the way the neural activity depends on and varies as a function of the animal’s movement. Or to sum up: the brain is visual when its supports habits that are vision-like, it is auditory when it supports touch-like habits, and so on.
Mrganka Sur, the MIT neuroscientist, surgically altered newborn ferrets, splicing retinal cells so that they sprouted connections into what would normally be auditory fields. What happens as result of this rewiring of the animal’s sensory periphery to the cortical heart of the brain? Does the rewired ferret come to hear with its eyes? No, it comes to see with its auditory brain. And this is exactly what our proposal predicts. “Auditory” cortex does visual work when it is harnessed by and integrated within a vision-like sensorimotor dynamic of exchange with the environment. True, as critics have observed, such rewired cortex actually takes on the sorts of lower-level organizational properties (receptive fields, etc) that one finds in “normal” visual cortex. But this further supports our hypothesis. For after all, what drives this plasticity is the fact that the neural region is entrained in vision-like sensorimotor dynamics, and moreover, the brute fact of the existence of these low-level properties explains nothing on its own.
We like to think that we can explain what we are, and what we can do, in terms of the brain. But it turns out that the brain itself, and its role in consciousness and cognition, is best understand in relation to what we are and what we can do. Either way, habit plays a basic role in making sense of what we are.
Coming soon: Why habits are also a curse.

Does Thinking Happen In The Brain?

mollycakes/via Flickr

by Alva Noë

  

Inside each of us there is a thing that thinks and feels and wants and decides. Each of us is that thing.
This is the traditional view of mind, the view that has dominated establishment research into cognition and consciousness for the last 500 years.
Contemporary scientists — neuroscientists as well as other cognitive scientists — by and large take this basic schema for granted no less than Descartes did. Of course, today’s thinkers believe that thing inside us, which is the self we are, is a bit of our flesh (the brain). Descartes, for his part, could not conceive of how mere meat could produce mind, so he supposed that mind was an immaterial something. But this difference, it turns out, and as I argue in Out of Our Heads, is merely technical. Despite having learned so much about the anatomy and physiology of the human brain in the last century, we don’t actually have a better account of how consciousness and cognition arise in the brain than it arises out of immaterial soul-stuff.
This last claim is not controversial, not really. But then why are we so certain, as a scientific and as a popular culture, that the secrets to our nature lie inside us, in the brain?
Answer: We can’t imagine an alternative to this “you are your brain” idea that does not end up giving up on science. Either you are your brain, or you are a mystery.
But this is mistaken.
  It’s like thinking that the only alternative to finding the value of money in the chemical composition of the bank notes would be supposing that monetary value is magical! And so here, is the case that interests us: there are good, sound, naturalistic approaches to the self and consciousness that give up the shopworn prejudice that we are identical to pieces of ourself inside our heads. The brain is necessary for our life, but it is hardly sufficient. A human being, like every living being, is a locus of densely interwoven coupling with the world around us. We make consciousness dynamically, in our exchange with the world around us. Ultimately, if we want to understand consciousness, we need to go out of our heads and look at the way we are embodied and also bound to and embedded in the world around us.
My topic for today, and next week, is recent work by John Dylan Haynes and his colleagues at the Humboldt University in Berlin on what they call Brain Reading. Let me say right at the outset: This is exciting work. It’s smart work. What interests me, though, is how easy it is to misunderstand the research itself and jump to the conclusion, entirely unwarranted, that it lends support to the Cartesian (= derived from Descartes) neuroscience from which we so urgently need to free our scientific and cultural imagination.
Let’s begin at the beginning: Mental acts and processes of all sorts — for example, feeling excited, thinking about someone, hoping for something to take place, savoring the taste chocolate — happen only thanks to the activation of systems in the brain and nervous system.
Some people confuse this truism — that mental episodes have neural correlates — with the further, entirely different claim that thoughts, feelings, experiences of taste, and the like, are, simply, events in the brain. This latter claim is not a truism. It is not even true.
How my car drives depends on what goes on inside its engine. Modifying the engine affects the driving behavior. Speeding up, slowing down, and such like, in turn, affect the engine. But it would be bizarre to say that the driving really happens in the engine. If my car has no wheels on it, or is up on the lift, it won’t drive, whatever happens inside the engine.
Brain stands to mind the way engine stands to driving. Or, to shift comparisons: to insist that thinking and feeling happen in the brain is rather like insisting that speech — talking — happens in the brain. We could not speak without the brain, to be sure. But speech also depends on many other physical processes — such as articulatory movements in the mouth and throat, and also respiratory activity. And of course it depends on social circumstances, and needs. People speak, and they do so thanks to their brains, and mouths, and throats, and much else besides (e.g. the existence of socially shared linguistic practices!)
Many mental processes or capacities depend on patterns of activation in specific areas of the brain. For example, visual recognition of faces is associated with neural activity in a region that has come to be known as the fusiform face area (FFA), whereas episodes of seeing houses (and other place-related items) are associated with activity in a different region (the so-called parahipocampal place area, PPA). Given correlations such as these, it is should not be surprising that facts about experience (e.g. that one is looking at a face) carry information about a person’s neural states. And vice versa. It ought to be possible, at least in principle, to find out that someone is experiencing a face by direct observation of neural activity in FFA.
Enter John Dylan Haynes and his colleagues at the Humboldt: prospects for “brain reading” have been greatly improved thanks to recent work by Haynes and his colleagues at the Humboldt in Berlin. They have developed techniques that allow one to determine a person’s cognitive state on the basis of neural data, even when, as happens to be the case, there is no one-to-one mapping between the occurrence of that cognitive state and neural activity in a specific location. Making use of computer-based multivariate pattern recognition software, they are able to use information about the brain’s total state gathered from brain imaging tools (fMRI) to make reliable predictions about a person’s mental state, in real time.
Is this brain reading, really? And what does it tell us about the neural substrates of experience? Does the possibility of this sort of brain reading lend support to the stronger claim that mental events are neural events? I turn to this question next Friday.