The Mystery Of Consciousness by John Searle, PhD (1997, NYREV, Inc.)
My review (out of 5 stars):
One of the frustrations of reading the philosophical literature (true also of the neuroscience literature) is that the authors sometimes don’t address one another. When dealing with complex matters, in addition to reading the author’s point of view, it is sometimes even more helpful to know if the author has considered other points of view, and if he or she has rejected these, to discover why.
For this reason I was excited to pick up this old book by John Searle. The starting point for the book was a collection of book reviews he had written about the work of other authors interested in consciousness. I was hopeful the book then might shed a lot of light on what Searle believed and what other authors believed at the same time.
Instead, and much to my surprise, the book left me with a lower opinion of John Searle.
The book has the following structure: an Introduction, followed by 6 book reviews, followed by an epilogue. In the Introduction and Epilogue, Searle presents some of his own philosophical arguments. The Introduction revolves mostly around his Chinese Room thought experiment, which he is then able to return to in his commentaries of the books of several of the authors reviewed over the following 6 chapters. The Epilogue revolves mostly around what he calls “biological naturalism” – that brains cause consciousness. The reviews are of the work of Francis Crick (The Astonishing Hypothesis), Gerald Edelman (Bright Air, Brilliant Fire and The Remembered Present), Roger Penrose (The Emperor’s New Mind and Shadows of the Mind), Daniel Dennett (Consciousness Explained), David Chalmers (The Conscious Mind), and Israel Rosenfield (The Strange, Familiar, and Forgotten).
The authors of the books reviewed include 3 neuroscientists (Crick, Edelman, and Rosenfield – though Crick was a late-comer to the field of neuroscience), a physicist (Penrose), and 2 philosophers (Dennett and Chalmers). Thus it might be surprising that I, a neuroscientist, had read both of the reviewed books of philosophy but only one of the other books (Crick’s). But, in fact, this is what drew me to Searle’s book. I’m capable of critiquing the neuroscientific literature, but I was looking forward to a philosopher taking on other philosophers.
But that ended up being the worst part of Searle’s book. One gets the feeling from those 2 chapters that maybe Francis Crick is right in his prejudices. (Crick was unkindly and outspoken about the field of philosophy – to his detriment, I believe, as Searle rightly points out in his mostly-positive review.) These chapters are the only two in the book where we are treated to a rebuttal. After Searle’s review of each book was published in the New York Review of Books, Dennett and Chalmers wrote letters objecting to Searle’s analysis. It is interesting that, as far as I know, Crick, Edelman, Penrose, and Rosenfield did not do this.
At first I was impressed that Searle published these rebuttals in his book – it seemed rather generous for an author to permit objections to his work right there in the very volume he is selling under his own name. I was also looking especially forward to these chapters, as this would be an even more “live” exchange of ideas that gave the book so much promise in the first place. But then I realized that Searle published those objections only to have a space to also publish his objections to the objections. The exchanges were, for the most part, not illuminating. Dennett’s reply was vacuous, as was Searle’s reply to Dennett. (I felt like I was reading an elementary school cafeteria food fight.) The Chalmers reply was substantive – it pointed out the several errors of Searle’s logic – but Searle’s reply to Chalmers was merely a doubling down on the original logical fallacies. These were reinforced (tripled down?) in the book’s Epilogue (after an odd brief review of Rosenfield’s book which should have been placed earlier in the volume, or possibly combined with the Edelman chapter).
In the rest of this essay, I will focus on Searle’s final chapter. I have a lot to say about the Chinese Room (and therefore Searle’s introductory chapter as well), which might make for a post at a later time. (I admire this thought experiment and have used it in at least three classes to make separate points. I have objections to it as well, but I think it rightly deserves its fame as a thought-provoking gedanken.) In concentrating on the last chapter, I can capture some of the flavor of the Searle-Chalmers exchange and also explain why I’ve lost a little respect for this giant of philosophy from reading this book.
Much of Searle’s work in philosophy of mind has been to refute the notion of functionalism and to dispute the research program of what he calls “Strong AI”. His notion of this, strong artificial intelligence, is that “instituting the right program in any hardware at all is constitutive of mental states” (p. 13, italics in original). Thus, mind is a computer program, and brains don’t matter – they don’t cause minds (and thus consciousness), they merely provide the hardware for running the program. Searle believes that his Chinese Room thought experiment exposes Strong AI as incoherent. He is also deeply troubled by the notion that the causal nature of brains has been removed from the equation.
Before continuing, let me make a couple of things clear. I’m a neuroscientist. I’m sympathetic to the point of view that consciousness and the brain have a lot to do with one another. Like Searle, I believe consciousness has a natural (as opposed to supernatural) explanation. However, unlike Searle, I think the mind-body problem is what Chalmers would call a “hard problem”, and I’m grateful to philosophers for their analysis of why it’s so hard. I’m stunned by philosophers like Searle who ignore its hardness, and thus let neuroscientists like Crick get away with their own lapses in logic.
In the Epilogue, Searle says of computational theories of consciousness that they are “profoundly antibiological” (p. 190, italics in original). In making the case that “brains matter crucially” (p. 191, italics in original):
We know in fact that brain processes cause consciousness, and from this it follows that any other sort of system capable of causing consciousness would have to have causal powers of brains to do it.
This excerpt is just embarrassing. I’d expect something like this from one of my psychology students, not an eminent logician and professor of philosophy. Let’s take the two headdesk statements in turn:
We know that brains processes cause consciousness. No, we don’t, not by any acceptable evaluation of the word “cause”. The entire functionalist approach to consciousness was stimulated by a concern that maybe brains don’t cause consciousness, at least not in the direct sense of causing a motor movement. When a motor axon is stimulated by an electric current, the muscle contracts, and we can follow the entire chain of mechanism from electrical activity, to acetylcholine release, to nicotinic receptor binding, to muscle fiber depolarization, to calcium spikes in the cytosol of the muscle fiber, to myosin and actin interaction, to contraction. Brains cause muscle movements. But we have no chain of mechanism to follow to consciousness, and we are deeply skeptical of the ability to take this example of brain cause (brains cause movement) and apply it to consciousness. The mechanism of neural cause: the release of neurotransmitter on to a postsynaptic receptor – has no place in consciousness. There are postsynaptic receptors on muscles, but there aren’t any on the soul, or whatever you’d call the consciousness end organ. And when you start reaching for words like “soul” as a neuroscientist, you are much more apt to say, “hang on, back up, we must be thinking about this the wrong way around. Consciousness can’t be an end organ like a muscle. Brains can’t cause it in nearly the same way that they cause other things to occur, it must be more complicated.”
Now I’m not saying that functionalism is any better. And I’m not saying that substance dualism (i.e., we have souls) is any better. All I’m saying is it does no good to insist that we “know” that brains cause consciousness. We don’t.
What we do know is that our brain states are correlated with things in our conscious experience. We know that our conscious experience changes dramatically as our brain changes (through development, through injury). We know, or think we know, that consciousness can even go away during certain stages of sleep, during general anesthesia, or during coma. (There is a very important problem of how one verifies this – people woken from sleep almost always report at least some thought, and there are occasional troubling reports of patients we thought were anesthetized or in a coma who claim nonetheless to have had experiences. For a philosopher, Searle is far too willing to use the word “know”.) All well and good, and all useful data for trying to understand consciousness. But even if we take these data at face value, what does it say about functionalism? Does not the state of the computer change as the program is running? If we damage the hardware, does not the software stop working, or stop working as well? How do these things we know about the brain speak one way or another to functionalist notions of consciousness?
Any other sort of system capable of causing consciousness would have to have causal powers of brains to do it. How is it possible to read this sentence and find any interpretation other than the most trivial? Doesn’t this sentence essentially say “Any system capable of consciousness would have to be conscious”? The only way to avoid this sentence being trivial would be if we knew how the brain caused consciousness. Then we could re-interpret the sentence as: Any other sort of system capable of causing consciousness would have to… synchronize activity of its elements at 40 Hz (Crick)… integrate previous states with ongoing analysis in re-entrant fashion (Edelman)… form a representation of itself and integrate that with representations of current events (Rosenfield)… isolate quantum effects but then amplify those effects to the rest of the system (Penrose). (I don’t know if I’ve done justice to any of the proposed mechanisms of the other authors, but that’s not relevant to my point.) Searle uses the phrase “the causal powers of brains” repeatedly, but doesn’t seem to recognize that this phrase has as much force as “the causal powers of the soul” until there is meat on the bones.
Searle claims not to be a dualist and claims not to be in the grip of the Cartesian theater fallacy. In my more generous moods I’m willing to believe that. But without any more details, the consistent refrain that we know brains cause consciousness just invite us to imagine brains (or neurons) squirting out consciousness like they squirt out neurotransmitters. Brains squirt out consciousness but thermostats don’t. (Searle takes Chalmers to task for a chapter in The Conscious Mind called “What is it like to be a thermostat?”). Brains squirt out qualia (the quality of aspects of a conscious experience, like the redness of red or the painfulness of pain) but my Lenovo laptop does not.
Again, as a neuroscientist I’m temperamentally sympathetic to the intuition that brains play a role in these things but thermostats and run-of-the mill laptops don’t, but until I see how I am forced logically to entertain: 1) the possibility that thermostats are conscious and 2) the possibility that the relationship of brains to consciousness is not causal in the usual manner in which we think of causality (i.e., not like brains causing movement). Like many people I don’t like either of these choices, but I tend to be more attracted to the second choice (see my positive review of Alva Noe’s Out Of Our Heads), which is why (in that same Noe review) I accuse David Chalmers of being a panpsychist and therefore a bit kooky. Ironically, then, I agree with Searle that Chalmers’ conclusions are over the top, but unlike Searle I am forced to admit that there is no refutation of panpsychism available. I can suspect Chalmers is wrong, but I can’t fault him for raising a logical possibility. I can prefer Searle’s conclusions to Chalmers’, but at the same time I can see clearly that Chalmers ate Searle for lunch in the exchange printed in Searle’s own book.
Another place in the Epilogue where I think Searle fails to make his case well is in his discussion of Daniel Dennett’s work. On Searle’s p. 192, he quotes from Dennett and Hofstadter’s The Minds I p. 15:
[The mind] is an abstract sort of thing whose identity is independent of any particular physical embodiment.
Searle’s critique of this is that no one would ever think of making a similar claim about any other biological process such as photosynthesis or digestion. (These analogies appear consistently in Searle’s writing.) There are a couple of problems with this refutation. First, it’s question-begging. The refutation requires first an acceptance that consciousness, like photosynthesis, is a biological process, which is, of course, the notion being challenged. The tautology “Consciousness can’t be non-biological because, like digestion, it’s biological” isn’t very helpful.
The second reason the analogy doesn’t work is because “digestion” and “photosynthesis” are, to some extent, substrate-independent, abstract terms. Plants don’t “cause” photosynthesis, though they are a substrate in which photosynthesis occurs. Photosynthesis is just the process by which the energy of light is used to defy entropy locally: the building up of higher energy molecules from lower energy molecules. Our bodies use a different process for achieving entropy-defying effects: cellular metabolism, in which low energy molecules are used to synthesize adenosine triphosphate, a high energy molecule (whose stored energy can later be exploited for things like muscle contraction), generating carbon dioxide and water in addition. But cellular metabolism, like photosynthesis or digestion, is really an abstract term, because it might occur in many different ways (glucose, fatty acids, and amino acids can all be used in cellular metabolism; the word doesn’t define the hardware so much as it defines the process). For these reasons I see the analogy of digestion or photosynthesis to have much in common with “functionalism”.
This weakness is again exposed on p. 201 where Searle tries another analogy.
A hundred years ago it seemed a mystery that mere matter could be alive. And debates raged between mechanists who sought a mechanical, chemical explanation of life and vitalists who thought any such explanation was impossible, who thought that any explanation required us to postulate a “vital force,” an “elan vital” that stood outside of mere chemical processes and made life possible…. The mystery was resolved not just because the mechanists won and the vitalists lost the debate, but because we got a much richer conception of the mechanisms involved.
What makes this analogy curious is that it is a favorite analogy of both the functionalists (like Dennett) and the eliminative materialists (like Patricia and Paul Churchland) who are the very targets of Searle’s scorn. I applaud the sentiment of Searle (and Dennett and Churchland and Churchland) that it is time to roll up our sleeves and learn as much as we can about the brain, but I’d also advise Searle to look carefully at his analogy. One of his claims is that consciousness is a real property and that it is irreducible, and hence that Dennett and the Churchlands can neither eliminate it nor equate it with something else (like a computer program or conglomeration of memes). But these are not the features of “life”. Life is reducible: it is the sum of many independent processes of locally defying entropy (photosynthesis, metabolism), of consuming the environment (eating, breathing, absorbing), of reproducing (budding, fission, sex). In a sense, then, the mechanists won because they were eliminative materialists: life as a separate thing was no longer necessary as an explanatory concept. Life was redefined functionally.
Searle closes with a plea for us to do away with the old categories of dualism and monism; this historical baggage he thinks led modern philosophy to the twin incoherences of functionalism and panpsychism. If we jettison these things we can make progress with Searle’s biological naturalism, the only valid research program because it is the only one that starts from the a priori assumption that brains cause consciousness, an obvious fact. As Chalmers says in his short essay reprinted in Searle’s book, what is obvious to one is anything but obvious to another, which in and of itself should make one skeptical of Searle’s suggestion.
But the bigger problem is that it is not at all obvious that Searle isn’t just committing all of the same sins of those doctrines. In his Epilogue, he asks himself skeptical questions and answers them. One of them is “But in the end it seems to me your philosophy is just another version of materialism” (p. 210). His reply to this imaginary critic is that what distinguishes biological naturalism is its acceptance of consciousness as irreducible and subjective experience as something that can’t be explained away. This would lead me to say “But in the end it seems to me your philosophy is just another version of dualism.” He addresses this on the same page with: “I do indeed reject dualism”. But as noted above, this rejection seems to be in tension with the notion that brains cause consciousness if that cannot be reduced to other aspects of the physical world.
It looks like materialism, but, he tells us, it’s not. It looks like dualism, but, he tells us, it’s not. If I were to indulge in some Daniel Dennettesque turns of phrases, I would have to conclude that therefore, Searle’s philosophy sounds like nothing at all. And if you are struggling with the ambiguity of that sentence, remind yourself how many stars I gave the book.
I’ve been wanting to write about free will for some time, and I have decided this is the time. Or, as the old debates about free will go, deterministic causes have at last acted upon me to begin writing a post on free will.
Let’s make some issues plain at the outset. Metaphysically speaking, I am a determinist – I believe that there is no “mystery stuff” in the universe, and that everything has a cause that can be described by the unchanging laws of physics – though in principle, of course, we don’t know all of the causes in play at any one time and we don’t yet even understand all of the laws of physics. Human behavior is highly complex (complex = “consisting of many elements”) and therefore not fully predictable by us – but metaphysically, fully explainable without including nonphysical contributors such as a “soul”.
Besides the supernatural, a second way that people have suggested that human behavior might be unpredictable is to include a random factor. (This gets us no closer to free will, but it is a way of suggesting that human behavior is not determined.) So far I am unmoved by this suggestion. As far as I know, the strongest possibility for a non-supernatural random factor is quantum indeterminacy. There is no plausible mechanism by which such random effects at the quantum level could have any impact on molar events at the organismal level, however – put more simply, random events for electrons are not of the type that they can have any true effect on random events of a person, or a monkey, or a worm, or an amoeba. I stand behind the description of “no plausible mechanism” despite attempts by some doctors, physicists, and new age gurus to incorporate quantum indeterminacy in theories of free will and consciousness.
The purpose of this blog post, though, isn’t to defend a deterministic view of the universe. The purpose is rather to explore what the consequences might be, or should be, of the assumption that we are living in a deterministic universe. Some people are bothered by the implications. For example, one worry is that – if “we” don’t cause our actions, then we can’t be blamed for them either. One doesn’t “blame” the apple for falling from the tree. If murdering someone is the metaphysical equivalent of a falling apple, then how can we blame the murderer and execute him or put him away for life? Second, if we don’t cause our own decisions, why do we bother with anything at all? Why worry? Whatever happens is going to happen with no agency on our part. Shouldn’t we just adopt a passive resistance to the universe’s will, refusing to be pawns in its game?
The second example is particularly confusing, since if we have no choice about our decisions, that would include the “choice” not to participate in the game of life. By this reasoning, not making a decision is just as much an (impossible) decision as whichever decision we are rebelling against making.
In any event, I’m not bothered by either of these supposed implications of determinism, and the rest of this essay will attempt to explain why. I will take up the second example first.
Could a Robot Demonstrate Free Will?
The deterministic universe makes human beings sound like robots executing their programming, and thus turns beings that we think of as being free agents (us) into beings we think of as being automatons (robots). So one way out would be to show that the beings we think of as being automatons (robots) can in fact act as agents. Let’s see how that might work.
There is a mostly-effective episode of Star Trek: The Next Generation called The Quality Of Life in which an android robot, Lt. Commander Data, makes the decision to defy his commanding officer in solution to a moral dilemma.
In the episode, the Enterprise visits a space station experimenting with a new mining technology, and also with a new type of robot called the Exocomp. Exocomps are sophisticated in that they learn from their experiences, and have the ability to generate new protocols for fixing things on the station. The Exocomps, though, seem to be malfunctioning when they begin to defy orders. It turns out that the Exocomps only defy orders when there is an imminent danger to their own survival – they seem to be demonstrating self-preservation.
The human characters debate whether this new behavior is a feature or a bug, but when things get serious (Exocomp sacrifices are needed to save the human crew of the station), the commanding officer (Commander Riker who is senior officer on the Enterprise while Captain Picard is on the endangered station) decides that the time for philosophical debate is over and the time for action – sacrifice of the Exocomps – is nigh.
Rather than carry out the commander’s orders, Data – who himself is an android – instead sabotages the ship in such a way that the Exocomps cannot be used in this purpose. He does so understanding that he may be court martialed, and that his action will very likely cause Captain Picard and Geordi LaForge (Data’s best friend, or, not to beg any questions, Data’s “best friend”) to die. Picard will later tell Data that this was a very “human” decision. We the audience see a situation where Data seems to “defy his own programming” – he’s an android, remember, and one who generally follows orders and values the lives of his fellow humans very highly. He’s also not stupid – he understands that the Exocomp behavior is not definitive that they are equivalent in value to the humans on the station. He’s taking a risk.
The first thing we have to deal with is: is this possible, or is this just written for dramatic effect? Data is, after all, a character played by a human being (Brent Spiner), and so perhaps we are fooled into thinking a mere robot could behave in this way only because, after all, a human being is inhabiting the character. Indeed, I find Data’s behavior perfectly possible – not, perhaps, with what we knew about robots in 1992, when this episode aired, but surely now, in 2017, when robotics (and artificial intelligence) has continued to improve.
So let’s see if we can establish this a bit better. Imagine we are designing a robot to play Monopoly with.
In order for our robot to be an effective opponent, it would have to be able to value the spaces on the board appropriately. For example, Boardwalk is probably the most important space on a Monopoly board, because under the right circumstances, owning this property can crush your opponents. You can’t say that about too many other spaces on the board. Essentially, you’d have to own several of the other spaces to compete with one Boardwalk space.
As it turns out, the more valuable spaces also “cost” more. This suggests to the programmer one easy way to program the robot to value Boardwalk more. We might code a variable in the program that specifies Boardwalk’s value. For simplicity, we could set this value to the cost of the property, for example, Let BValue = 400. When managing its finances, the robot’s program will compare BValue to say its internal value of North Carolina Avenue (Let NCValue = 300). If the robot is on North Carolina Avenue, it might avoid purchasing this property given that Boardwalk – the higher-value property – is only a dice roll of 7 away, and we’ve also programmed the robot to calculate the probability of its next landing space (and probably the one after that). Of course, if the robot has more than $700 to spend, it might try to buy both.
However, as any good game player knows, the values of things change over the course of the game (even though the price does not – and this is how to win – to recognize when properties are worth more than they cost, or are worth less than they cost). For example, if the robot already owns Park Place, the value of Boardwalk is considerably higher. Owning both Boardwalk and Park Place is a Monopoly, and Monopolies allow you to bleed your opponents dry. The programmer might put in a subroutine that alters the value of BValue whenever the robot buys Park Place. It might be set to, say, 1200. (It might be made 3 times as desirable – and thus now the robot will forgo spending not only on North Carolina, but on other properties that aren’t 7 away from Boardwalk, but even less probable distances like 4, 5, 6, 8, 9, or 10.) What if the robot already owns Park Place and a Monopoly of the green or purple properties? This raises the value of Boardwalk even higher, because now there is an almost guarantee that opponents can’t make a dice throw that avoids enriching the robot on 1/4 of the board. This might make BValue = 2400.
What if, instead, the robot’s opponent owns Park Place? This might even make Boardwalk more valuable. The robot may want to buy Boardwalk to deny his opponent that Monopoly. Probably the highest value the robot should set is to avoid losing – it’s even more important that values to allow winning, because so long as you live to fight another turn, things may go in your favor. Once you are out, you’re out. So when an opponent has Park Place, we might set BValue at 1500. (And still higher if the opponent has green or purple properties.) But what if we aren’t playing a 2-player game? What if it’s a 4-player game? Then we might set BValue a little lower (1200?) because the important thing is that the owner of Park Place doesn’t get Boardwalk – not so much that the robot needs Boardwalk. With 2 other opponents at the table, there is a fair chance someone else will steal away Boardwalk, and so the robot can keep his $400 for more lucrative purchases (like maybe completing a Monopoly on the red properties).
Thus we get intelligent behavior through constantly adjusting the value of the properties. A robot’s “decision” whether to buy or not to buy would involve the weighing of all of these changing values against one another, along with other key details (how much money it has, where it is on the board, and so on).
But now the programmer has a new dilemma. What are the right settings? When we decided to triple Boardwalk’s value upon owning Park Place, was that right? Why not double or quadruple? How do we know how to make the robot maximally “intelligent”?
And so we might try a new strategy: let the robot “decide” – and in essentially the same way Monopoly enthusiasts decide: through experience.
Robots can get lots of experience very quickly. We simply have to program the rules of the game and then let the robot run a billion simulations of Monopoly internally with 2, 3, or 4 players. We need some sort of algorithm, probably involving neural nets, which allow actual game results to adjust the internal valuation. If the robot tended to lose games in which Boardwalk’s value was tripled, but win games in which Boardwalk’s value was quadrupled, the value will be set closer to quadrupled than tripled for the next game. The robot never stops tweaking these values, and neither do we humans.
Now there is nothing which occurs in this process which is anything but deterministic. And yet the programmer, sitting down to play with the “trained” robot, will not be able to anticipate the robot’s decisions. The programmer of course has complete knowledge of the program, but does not have complete knowledge of the simulations – the billions of games used to tweak the current values the program relies upon. The robot’s behavior may therefore “surprise” the programmer which is another way of saying that the robot seems to “defy” its program, or “rise above” it. The programmer might see the robot make a choice that he or she would not make and then end up losing to the robot as a result. The programmer can now even learn from the robot!
This was, in fact, what happened with the Exocomps. In the end of the episode, the Exocomps execute a strategy none of the humans had considered, including the human (or rather, alien!) who had programmed the Exocomps in the first place. The programmer did not have the Exocomps’ experience with fixing problems in the mining operation.
Likewise, in making his decision to defy his commander, Data appeared to rise above his programming. But Data had decades of experience in Star Fleet. He had observed his fellow officers wrestling with similarly weighty moral issues, and he had seem them come out well or poorly. He had seen his own Captain defend his (Data’s) very right to be treated as an equal (in what, in my opinion, is the finest Star Trek episode ever – all of the iterations of the show included – The Measure of A Man). Through all of this, the weights in various nodes of Data’s uber-complex neural network were changing, such that the value of the Exocomp, the value of the lives of his friends, the value of following orders, the value of taking a principled stand, and the probability given the evidence that the Exocomps really were “alive” all contributed to his final decision.
By any definition, Data was “deciding”, and his decision was based on “reasoning”, incorporating both his past experience (enormous experience) and the currently available information (his observations of the Exocomps and the situation on the station). Both the experiences and the information are unique to Data – no one else had his experiences, obviously, but also no one else had the particular information (coupled with the weighting of that information) that Data has. This is, of course, why we sometimes make bad decisions – either our experiences are incomplete, or our information is incomplete. It is also why the decisions of others are only partly predictable (only some of the experiences and information are shared between the person making the decision and the person making the prediction). Riker’s decision was to sacrifice the Exocomps; Data’s decision was to sacrifice the humans (and aliens). The programmer might expect the Monopoly robot to buy Illinois Avenue, but the robot might save his $240 instead.
Metaphysically, as stated at the beginning, the robot does not suddenly acquire a soul in this process, nor did the robot benefit (in some equally mysterious way!) from a random quantum event. But the robot was the agent of this decision, not the programmer, and not the current state of the board (this is the information required for the decision, not the decider itself). If we can come to view a mere robot as a deciding agent in a deterministic universe, we might also do the same with us.
Can We Punish People For Misdeeds In A Deterministic Universe?
What about the other question? I saved this one for last because I think it’s the easier issue. This answer does not rely on the previous defense of agency, though of course if you buy that argument there’s not really a second question to consider.
There are 2 ways to answer this question. One is to object to the notion that the exclusive goal of punishment is to chastise agents for making bad decisions. It is possible that one goal of punishment is to reduce the probability that other people will make the same bad decision. The second answer is to object to the idea that we should do nothing about a blameless act. We don’t blame an apple for falling from the tree, but does that mean we should just leave it there to rot, attract insects and other vermin, uglify our yard and make it smell bad too?
The first argument is this. If we live in a deterministic universe, then what causes people’s rational actions is: 1) their experiences and 2) the current information at hand. If their behavior is determined by these things, then the only hope we have to encourage good decisions is to promote certain experiences and certain information. (I see that there is a looming circularity problem here – can “we” really “promote” and “encourage” if we have no free will? It’s a problem, but it’s also a reality that we do promote and encourage things, so surely we can put that aside for the moment.)
Traffic accidents are more likely at high rates of speed. Therefore we make information available by posting speed limit signs. We also provide people with experiences that encourage obedience in the form of stiff penalties for speeding (which we may have experienced directly or indirectly through the lamentations of a friend or relative). People may also have had experience of losing control of their vehicle when speeding, or fearing for their lives when a speeder cuts them off. These provide the substrate for the deterministic decision making of “I choose not to speed today.” The reasoning that goes “Because the universe is deterministic, nothing we do matters, therefore we shouldn’t bother with speed limit signs and tickets” is incoherent! Because the universe is deterministic, if we don’t have these things, more people will die in traffic accidents. (And again, to address the circularity problem, the reason we do put up speed limit signs is because we have experience with this improving safety.)
The same then is true of punishments for violent crime, white collar crime, stealing from the cookie jar before dinner, and anything else you want to mention. Determinism requires punishment for acts which are harmful or dangerous, and if you don’t believe this, move to a country where the police are corrupt and the gangs rule the streets.
The second argument is this. For this argument I’ll use robots again, but now for the exactly opposite purpose. Let’s imagine we have a robot that’s unlike the Exocomps or Data in that it is pretty rigidly programmed, and thus its behavior is highly predictable. Its job is to clean apartments. Unfortunately there must have been some sort of mistake at the robot factory, or some computer virus has affected its program, because now in addition to cleaning apartments, it has the unfortunate additional behavior of murdering the people who own the apartments in their sleep.
What do we do with this robot? One might say “One does not blame the apple for falling from the tree. The robot didn’t mean to kill anybody, it was just part of its programming.” This is unlikely to be a winning argument. Probably instead we’ll have general agreement that this thing needs to be imprisoned, reprogrammed, or disassembled.
The point I’m making here is that I’m not sure where the notion started that we only correct mistakes if they were made voluntarily (in the metaphysical sense). That’s a non sequitur. In fact, imagine we were dealing with a robot that could be reasoned with. A Data, if you like. (Data did in fact go a bit crazy on a couple of occasions for dramatic effect.) Only in that case might you forgo punishment and instead attempt to reason. If a small child gets too close to the campfire, you yell “Stop!” and grab the child and move it to a safer location, then you watch it like a hawk. If an older child does so, you say “You better watch out, kiddo, you’re getting pretty close to the fire and that will really hurt.” The more potential there is for intentions and agency, the less you might have to resort to punishment.
Now of course punishment is a pretty complicated issue. We don’t hesitate disassembling the murdering cleaning robot, but capital punishment for people that commit heinous crimes is a more complicated issue. Likewise, a criminal who is found to have a brain tumor or to be developmentally delayed is likewise a difficult dilemma. The robot example does not trump ethical or political issues in these cases. It is merely meant to disabuse people of the incoherent reflexive belief that punishment is only legitimate if intent and agency can be established. That idea is an oft-used unexamined a priori assumption with very little to justify it.
According to the graphic above, the diseases phenylketonuria (PKU) and scurvy (vitamin C deficiency) couldn’t be more different. One, PKU, has a “highly genetic” etiology, whereas the other, scurvy, has an entirely behavioral/environmental cause. Both diseases nonetheless have the same mode of treatment: attention to one’s diet. Reasoning back from the “cure”, we might say both diseases are dietary diseases. However, another way of looking at it is that both diseases are caused by an enzymatic deficiency, and enzymes, as proteins, are specified by our genes. From that perspective, we might label both diseases “genetic”. In resolving these apparent paradoxes, we will also shed some light on why the nature/nurture debate is so thorny, and hopefully also dispel some errors in the way most people think about genetics, and errors in the way they think about diet.
The Cause of PKU
The reason PKU is placed on the far genetic end of the graphic is that its genetics are well-understood. PKU is an autosomal recessive disease – autosomal, meaning it is inherited via a non-sex chromosome, and so is equally likely to occur in males and females; recessive, meaning the disease-causing gene must be inherited from both mom and dad, making the disease relatively rare (for PKU, 1 in 12,000), and meaning it can occur in children whose parents show no signs of the disease themselves.
As I’ve written before, genes are segments of DNA that provide the recipe for making proteins from the chemical building blocks of proteins, amino acids. The gene that is disrupted in sufferers of PKU is called the PAH gene. This gene specifies the recipe for making phenylalanine hydroxylase. Babies born with PKU essentially have a bad recipe for making this enzyme, such that the enzyme just doesn’t work the way it’s supposed to – the way it does in the vast majority of people without PKU.
So what does phenylalanine hydroxylase do normally? It converts the amino acid phenylalanine into tyrosine:
In sufferers of PKU, who lack a functional phenylalanine hydroxylase, this reaction doesn’t happen. This is bad – very bad.
Why? Well let’s back up a bit. We’ve already established that genes are recipes for making proteins. We often hear things like our genes define who we are. Well – if all genes do is allow us to make proteins, then it must be equally true that our proteins define who we are. Yes, some gene specifies our eye color, some other gene our blood type – but the color of our eyes is determined by the proteins we make, and our blood type is named for a protein sticking through the membranes of our blood cells. Some proteins define differences between people, like eye color or blood type, but others determine whether we live or die. They determine the chemistry of our body, making some proteins absolutely essential for life.
When and how do we make proteins? When is all the time. We are constantly making proteins in every one of the 30 trillion cells in our body. We are massive protein factories. We never take a break from this activity until death. How we make them is that certain organelles in our cells interact with certain molecules in our cells to build our proteins one amino acid at a time. (A small protein like insulin has several dozen amino acids; larger proteins can have hundreds or thousands.) Again, genes specify the sequence of amino acids for a given protein. A mutation in a gene is a change in this code. The consequences of a mutation can be nil (if for example the change doesn’t alter the amino acid code), virtually nil (if a single amino acid is mis-specified but this amino acid doesn’t change the shape or electrical charge of the protein enough to alter its function), moderate (if the protein’s function is compromised slightly), or severe (if the protein becomes non-functional, as in the case of PKU).
In any event, having the right recipe to make a protein is only part of the problem. After all, if you have the right recipe to bake a cake, that’s not going to help you make a cake if you don’t have eggs and milk and flour in your kitchen. Likewise, having the recipe for making a protein is one thing – having the right amino acids is another.
We use about 20 amino acids to make all of our proteins. Of these 20, nine are considered essential amino acids – we must get them from our diet, or we die. One of these is phenylalanine.
The remaining 11 are not considered essential, because we can make them ourselves – if we have the right enzymes to do so, and if we have the right ingredients to do so.
One of the nonessential amino acids which is very important is tyrosine. The enzyme we need to make tyrosine is phenylalanine hydroxylase, as shown in the graphic above.
So we’ve identified the first problem that PKU causes: without a functioning phenylalanine hydroxylase enzyme, we can’t make tyrosine – and so a nonessential amino acid suddenly becomes an essential amino acid – we must get it in our diet. Without sufficient quantities of tyrosine, we can’t make dopamine, norepinephrine, or adrenaline, to say nothing of the hundreds of proteins requiring this amino acid in its recipe.
But the situation is even more dire than this. If this were just a matter of eating more tyrosine, PKU probably wouldn’t be so devastating. But the lack of the enzyme affects both sides of the chemical reaction shown in the graphic: not only does a PKU sufferer produce no tyrosine (the right side of the reaction), but the PKU sufferer will also build up high concentrations of phenylalanine (the left side of the reaction). This has several effects, including creating stress on the kidney to eliminate the excess.
More devastatingly, high phenylalanine levels disrupt the chemistry of the brain. The brain is protected by a blood-brain barrier that limits access to the brain by large molecules, presumably to keep toxins from affecting the nervous system. But this means that there has to be a way to allow needed large molecules access to the nervous system, and this is accomplished by what are called transport molecules. (These are proteins, by the way. Again proteins.) One such transport molecule is responsible for large, neutral charge amino acids. Think of this molecule like a single-file tunnel that works on a first-come, first-served basis. The problem is, when there’s excessively high levels of phenylalanine, almost every molecule that lines up for entry to the brain through this tunnel is phenylalanine – leading to low levels of valine, isoleucine, tyrosine, and other amino acids in the brain.
The result is devastating – small head size, severe intellectual delays, behavioral problems, depression, and reduced life expectancy.
How does one fix PKU? Gene therapy might be nice – that is, stick some cells in the body containing the right gene for phenylalanine hydroxylase, and let those cells crank out the enzyme. This is being tried, but so far with limited success.
What does work is careful control of the diet. All of the problems caused by this single gene are the result to too much phenylalanine and too little tyrosine. A diet low in the former and high in the latter can completely eliminate the symptoms of this disorder. This is, unfortunately, a pretty stringent diet, as many proteins in the food we eat contain phenylalanine. (When we eat the proteins of other species – beef, pork, chicken, rice, beans, corn – we break the proteins down into their amino acids prior to absorbing them. We then use these amino acids to make our own, human proteins. Think of amino acids like legos – we can destroy the spaceship our brother made of legos, rearrange those pieces, and build our own dune buggy.)
This stringent diet has to start right away – in fact, it’s especially crucial during development. For this reason, most babies have blood drawn within a few hours of birth to test for a small number of problems for which early diagnosis is crucial – and PKU is one of those problems.
With early diagnosis and strict adherence, an entirely genetic-caused disease (see the figure at the top of this post), is completely controllable using an entirely environmental/behavioral therapy.
The Cause of Scurvy
Scurvy, on the other hand, is listed on the extreme environment – nongenetic – portion of the figure. This is completely justifiable – and yet, just to show how thorny the nature/nurture debate is, I will also show how it would be possible to label scurvy as just as genetic as PKU.
Scurvy was virtually unknown until the age of exploration. In the early days of ocean voyages, the diet of the sailors (much more so than the officers) was relatively limited, and these voyages might last many weeks. Magellan’s years-long circumnavigation of the globe started with a crew of 237 and arrived with a crew of 18 – and the loss of men was probably mostly due to scurvy. (Magellan himself was impaled by a bamboo spear in the Philippines.) During the Seven Years War (in the mid 1700s) between the British and the French, a few hundred British seamen died from combat, and at least 60 times that number from scurvy.
Scurvy is a dietary deficiency of vitamin C. The word vitamin is a bit of a misnomer deriving from “vital amine”. We now recognize many vitamins that are not in the chemical class of an amine, and in fact, vitamin C is one such vitamin. Its chemical name is ascorbic acid, and it is a sugar acid. But the “vital” part of the word vitamin does retain its accuracy – vitamins are small molecules that are absolutely required for life, usually in very small amounts, which must be obtained from the diet. Of the vitamins, we require vitamin C in the greatest quantity, though still on the order of a few dozen milligrams per day.
Lack of vitamin C – scurvy – leads to fatigue and soreness, and then progresses to difficulty breathing (due to loss of red blood cells), bruising, bleeding, loss of teeth, and all sorts of other nasty symptoms as the connective tissues of the body slowly degenerate without repair, as vitamin C is necessary in the formation of collagen, a key component of connective tissue.
Because so little vitamin C is needed in the diet, scurvy can be rapidly corrected by eating foods containing vitamin C. Citrus fruits are, of course, excellent sources, and the British Naval habit of carrying limes on board for sailors to eat to combat scurvy led to the nickname “limeys” which was first applied to British seamen and later to British people generally. It is probably not an underestimate to attribute to limes the lion’s share of the credit for the formation and maintenance of the British empire, so devastating was scurvy to the maintenance of a strong Navy.
The rapid amelioration of scurvy by diet explains its position as a “completely environmental” disease on our initial figure. How then can I attempt to justify scurvy as a genetic disease?
Remember, what makes a vitamin a vitamin is that it must be obtained from the diet or death will inevitably result. Nutrition labels on our foods typically display the level of vitamin C per serving. But now check the nutrition label on your dog food or cat food. You probably won’t find vitamin C listed there, though you may find vitamin A, the B vitamins, and vitamin E. Why? Because dogs and cats don’t need vitamin C from their diet. Neither do rabbits, rats, mice, or lemurs.
Now, all of these species need to make collagen, and amides, and other things vitamin C is used for. And these species do use vitamin C to do the job. But unlike humans, dogs, cats, rats, and lemurs can make their own vitamin C from simple sugars. (Vitamin C is, after all, just a small sugar acid, requiring a simple chemical reaction to synthesize.) Again, molecular synthesis typically requires the right enzyme and the right building blocks. Humans have the right building blocks – we eat plenty of sugar – but we lack something that dogs, cats, rats, and lemurs have: the enzyme L-gulonolactone oxidase. If we had it, we’d make our own vitamin C.
The image to the right shows a partial evolutionary family tree. Species connected with the thick, black line, have a functional gene for L-gulonolactone oxidase. Species with connected with a thick, gray line have a nonfunctional copy of the gene. From this perspective, scurvy is actually a genetic disease – in the same way PKU is – it’s just a genetic disease that’s inherited by every human on the planet, rather than 1 in 12,000.
A Tale Of Two Diseases
Diseases that can be traced to a single, nonfunctional gene are pretty rare, which makes PKU something of a textbook example of the role of genetics in physiology. Diseases that can be traced to the lack of a single nutrient are also rare, which makes scurvy something of a textbook example of the role of diet in physiology. But if you dig a little deeper, these extremes start to disappear – PKU is fully treatable by diet (though it does take considerable effort), and scurvy, in principle, could be fully treatable by gene therapy by providing humans with the L-gulonolactone oxidase gene from a cat or a squirrel.
We are often bombarded with information about the role of genetics or diet in disease that concern the less extreme examples. We are told that scientists have found a gene associated with schizophrenia, or depression, or leukemia, or diabetes. We are advised that consuming probiotics, or antioxidants, or vitamin C, or plant protein could keep us healthy, or that too much salt, or cholesterol, or sugar, will make us unhealthy. Some people take extreme lessons from this deluge of information – maybe that certain problems are inevitable (“it’s in my genes”) or that other problems are easily solved (“just buy this supplement and avoid that food”).
But when the extreme cases are so plastic – when a genetic disorder is cured by diet and a dietary disorder is caused by lack of a gene common in the animal kingdom – how can we possibly take simple lessons that concern diseases we know to be a mixture of multiple genetic contributions and multiple environmental factors? The involvement of a gene in a disease does not imply inevitability, but it does represent an exciting tool to unlocking the interlocking effects of proteins to physiology – and once discovered, may lead to genetic, pharmacological, or environmental therapeutic approaches.
Psychologist Paul Rozin and his colleagues asked a fabulous question in a 1996 study investigating people’s attitudes about food:
“Assume you are alone on a desert island for one year. You can have water, oranges that grow on the trees, and one other food. Pick the food that you think would be best for your health (never mind what food you would like). Which of these foods would you pick:
Corn…Alfalfa sprouts…Hot dogs…Spinach…Peaches…Bananas…Milk chocolate”
Desert island questions are always fun, but they usually focus on recreational items. One CD. One movie. One book.
Here is a much more consequential question. One food – from a rather small list of choices – that you’d have to eat every day (along with water and oranges) in the hopes that it would keep you alive for 12 months.
What would you pick?
If you were like 39% of the 124 students surveyed, you picked spinach. Also popular were bananas (24%).
My preferred option, hot dogs, was selected by 17% of the respondents.
The authors of the study argued that only two of the foods stood a reasonable chance of keeping someone alive for 12 months – hot dogs and milk chocolate. These are more complete foods than the others on the list – containing reasonable quantities of all 3 macronutrients (fats, carbohydrates, and protein) as well as a healthy supply of minerals. In the case of hot dogs, in fact, there is very little nutritional requirements lacking other than vitamin A and vitamin C, which accounts for why the question specifies that oranges are also available. (In a previous version of the survey, oranges were lacking, making survival unlikely with any option. Still, you’d have lasted longest on the hot dogs or milk chocolate even then, although only 10% of respondents selected hot dogs or milk chocolate in the first version of the survey.)
Of course, hot dogs and milk chocolate are usually considered unhealthy foods, whereas peaches, spinach, sprouts, bananas, and corn are considered healthy foods. The idea of eating an unhealthy food for 12 months was apparently unthinkable for 79% of the respondents (hot dogs + milk chocolate), even though these were the most complete foods, nutritionally. When all you have is one food source, your best bet is to go with the more complete food source.
In general, of course, fruits and vegetables are considered to be healthy foods, and the perception is that meats are unhealthy. But we are animals – we need, for the most part, the same things that other animals need – and so by eating them, by eating meat, we have a better chance to consume all that we need in one sitting. Even the chocolate, which I hadn’t considered a viable option, contains fats and proteins associated with milk, an animal product, and thus is the second-best option on the list.
None of this is to imply that eating a year’s worth of hot dogs is a great idea when other options exist, and even in the original question there is the necessity of oranges. But the failure of most of the respondents to consider hot dogs or milk chocolate does speak to our black and white thinking about food.
This was made plain in another part of Rozin’s questionnaire. Here, he asked people to decide if a diet lacking salt is better than a diet containing a teaspoon of salt each day. Fully 51% of the respondents agreed with this statement (and another 18% considered these equally healthy options). For context (not provided to the respondents, of course), a teaspoon of salt is right around the recommended daily amount. By contrast, going completely salt free will kill you in about a month, and you’ll feel truly miserable after a week.
A second parallel question asked about a no fat diet vs. a diet in which you consumed the equivalent of 1 teaspoon of butter a day. Similarly to salt, 49% agreed that the fat free diet was healthier, and 18% said it was a wash. This is probably even more surprising a result than the salt question, because although you won’t die as quickly from a lack of fat, a teaspoon of butter contains only 6% of the recommended daily fat intake, and only 10% of the recommended limit for saturated fat. Furthermore, when asked to compare a no fat diet to a diet with the equivalent of 5 teaspoons of butter, which still doesn’t reach recommended levels, fully 79% said the fat free diet was healthier.
Again, salt and butter are seen as unhealthy foods. But this categorization omits a very important detail: foods are (for the most part) neither good nor bad; what matters is dose. As I’ve written about before, sodium is an essential micronutrient that your body cannot make or store. Since you lose sodium in sweat, urine, and feces, you simply have to replace it on a regular basis. Life does not go on without it. Furthermore, it is so critical to normal functioning that we have evolved mechanisms to regulate sodium levels (so-called hydromineral balance) so that we can tolerate excess sodium relatively easily. It is not until levels get well above physiological norms that problems arise.
Fat is less well understood. The primary reason to consume fats is for energy, but we can also get energy from other sources (carbohydrates and protein). But fats also make up key structures in our bodies, including all cell membranes, and thus fats are broken down and reformed into structures that we need to grow and to maintain our tissues. Fatty foods also contain fat-soluble vitamins, like vitamin E, that are necessary for life and which are difficult (impossible?) to obtain without consuming some fat. The icing on the cake is that some people believe that fat never deserved its negative reputation at all – though I always tend to favor the notion of moderation until controversies are settled.
Whereas some foods have a sinful reputation – like fats, salt, sugar, and carbs – others have a “health halo” – they are seen as good in all contexts. I was stunned to see that, in another part of the study, 21% of respondents agreed with the statement “A person cannot eat too many vitamins” and another 8% neither agreed nor disagreed. Likewise, 19% agreed (and another 18% did not disagree) that “A diet cannot have too much protein in it”. Some vitamins are quite dangerous in excess, and high protein consumption, though rare, can lead to kidney problems. As with “bad” foods that can be fine or even necessary at low doses, “good” foods can be harmful or even deadly at high doses.
We see the same kind of thinking with non-food items, such as medication. Recreational sports players will take Advil on a regular basis, before, during, and after pain, operating under the false belief that anything sold over the counter is always safe. Students hand out their ADHD medication as a substitute for a cup of coffee, on the grounds that their doctor wouldn’t prescribe (and their mother wouldn’t let them take) anything that could be dangerous. Activists exploit the success of marijuana in alleviating pain and nausea in cancer patients to drum up support for legalization of marijuana for recreational use, on the grounds that because it is helpful in one context, it is safe in all contexts.
It doesn’t help that we live in an environment saturated with a sensationalizing media. When you hear reports that this or that is linked to cancer, or heart disease, or diabetes, you never – really very close to never – hear anything about the important information: how much? How much of a certain food leads to how much of an increased risk? They don’t tell us, and we don’t ask. In a confusing and complicated environment in which we are bombarded with scary information at every turn, people fall back on black and white categories. It’s a shame, because as the political scientist Aaron Wildavsky once wrote: “The richest, longest lived, best protected, most resourceful civilization, with the highest degree of insight into its own technology, is on its way to becoming the most frightened.”
Frightened even of sodium and fat, two things you would die without. Frightened of hot dogs – even when they’re the only thing that can save you!
The Rozin article is:
Rozin, P; Ashmore, M, Markwith, M (1996). Lay American conceptions of nutrition: Dose insensitivity, categorical thinking, contagion, and the monotonic mind. Health Psychology, 15(6), 438-447.
The Wildavsky quote is in:
Wildavsky, A. (1979). No risk is the highest risk of all. American Scientist, 67, 32-37.
I got temporarily excited during Hillary Clinton’s nomination acceptance speech at the Democratic National Convention. After boisterous applause for a comment slamming Wall Street, buoyed by the enthusiasm of the arena, she shouted:
I believe in science!
This unleashed another round of applause from the crowd, and I have to admit, my heart swelled. (To borrow a Hillary phrase, prompting my wife to deadpan, “She should take something for that.”) A politician had just proclaimed her trust in the scientific method, and an arena full of people from all over America responded with approval. I had just enough time to raise my hopes for the next, oh I don’t know, 3 minutes of the speech? 2 minutes? 45 seconds? She continued:
I believe that climate change is real and that we can save our planet while creating millions of good-paying clean energy jobs.
And then… back to immigration. Science got one sentence, although note that even that one sentence had to share space with Joe Biden’s “three letter word: J-O-B-S jobs!”
This post isn’t about the scientific evidence for climate change or the merits of various public policy positions to combat it. What bothered me about that passing moment in Hillary’s speech is that for many politicians, climate change is the only scientific issue of our day. Worse, it has become a litmus test for politicians and for the general public. If you believe in human-caused global warming, you are pro-science. If you disbelieve, you are a knuckle-dragger. And so by boldly proclaiming her appraisal that science has proven that the climate is warming due to industrial activity, Hillary and her supporters can pat themselves on the back and move on in their sanctimony.
Here’s what I believe is great about science. Science is a system that forces you to weigh evidence and to accept that evidence even when it conflicts with your preconceived notions. That, in a sentence, is what science is – why it is good, why it is sorely needed. Understand, this is not to say that all scientists practice this ideal, and it is not to say there aren’t considerable problems in the day to day practice of science. But over the long haul, it is science – certainly not a particular brand of politics – that deserves the label “progressive”. Bad ideas are weeded out, and those with the best evidence survive.
So for me, you don’t demonstrate your scientific bona fides by taking one particular position. You do so if you favor evidence over your preconceived notions.
It takes no courage for a Democrat to stand before other Democrats and remind us that the scientific consensus is that human activity is warming the planet. That’s a softball in that environment. What would have demonstrated real courage would have been if Hillary Clinton then went on with my hoped-for 2 or 3 minutes:
“And by clean energy” – riotous applause – “I include nuclear power, the most efficient carbon-free energy source we already have the technology to use!” Silence. (Not to inject politics here, but wasn’t the “Iran Deal” all about keeping the Iranians from making nuclear weapons but allowing them to pursue “peaceful” nuclear technology for power generation? Why do Democrats think it’s okay for the Iranians to develop nuclear power but favor inefficient wind farms and solar fields to nuclear power here at home?)
Although you can now hear a pin drop, I imagine Hillary continuing. “I believe also in the science that demonstrates that transgenic crop technology is not only safe, but actually increases yields, decreases the need for new farmland, lowers carbon emissions, and is safer for the environment!” The camera now zooms in on Bernie Sanders, squirming in his seat. She seems to be boring a hole through his chest as she continues, “We will oppose unnecessary GMO labeling laws, recognizing that such regulations would decrease consumer choice, favor large corporations, increase the price of food, and demonize a promising technology!”
Personalize it, Hillary. “When I was Secretary of State, I traveled to some of the poorest countries on Earth. I saw the faces of young children, blinded by Vitamin A deficiency, and met mothers who had buried their children far too young. As President, I will stand up to anti-science crusaders like Greenpeace to ensure that technologies like golden rice become available to all those in need!” Several spectators walk to the exits. A gentleman in the front row with a No GMO hat faints.
Keep harping on your record, Hillary. “I have made a career fighting for access to health care, especially for young children. My administration will continue to do so, placing special emphasis on ensuring that all children have access to life-saving vaccines. I strongly rebuke Robert Kennedy, Jr.’s nefarious demonization of vaccines, and I part with our current President and my opponent Donald Trump in that I unequivocally deny the fraudulent vaccine-autism link!” The Massachusetts and California delegations suddenly become dizzy.
Show your personal growth, Hillary. “And speaking of health care issues, let me also clarify my position on so-called complementary and alternative medicine. Although I was previously sympathetic to this quackery, having learned more, I now recognize that this is one of the major ways in which our nation squanders precious health care resources. I no longer consider Dr. Mark Hyman an advisor on these issues.”
“Instead I will support evidence-based biomedical research. My administration will pursue bipartisan increases to biomedical research funding, which we recognize requires the use of animal models.” Delegates that give time and money to PETA and the Humane Society of the United States break out in a cold sweat.
“My administration will also fully support NASA and exploration of the universe through telescopic observation. I challenge the delegates of the great state of Hawaii to overcome unscientific superstition and support bringing cutting edge research to the Big Island.” The Hawaiian delegation heads for the exit.
“And speaking of unscientific superstition, let me make amends for my earlier embarrassing comments about aliens having visited Earth. When I made those comments I was uneducated on not only the enormous distances between stars and the impossibility of traveling at speeds approaching the speed of light, I was also ignorant on the psychological sciences on how false beliefs are easily formed. When I said there couldn’t be so many stories of UFOs unless they were real was a too-credulous comment on my part, and one that I regret.”
Suddenly the entire roomful of delegates – ones that had lustily applauded science belief when the topic was first broached in a fit of self-congratulation – are themselves experiencing regret. Possibly at their own scientific ignorance, but more likely at having nominated a woman who believes not only in the science consistent with their preconceived notions, but even in the science that does not. How audacious!
There’s a funny Seinfeld episode in which George Costanza has been experiencing some good fortune, but then becomes worried about a spot on his lip that might be cancer. He is discussing this fear with a therapist. He says: “God would never let me be successful. He’d kill me first. He’d never let me be happy.” His therapist replies: “I thought you didn’t believe in God?” He answers: “I do for the bad things.” Take a look at some common left leaning views on climate change, nuclear power, transgenic crops, alien visitations, vivisection. One wonders if many at the Democratic National Convention do believe in science – but only for the bad things.
Don’t get me wrong, I’m not making any political points in this reaction. Donald Trump didn’t do any better in defending science or in promising to make decisions as President with science in mind. But preaching to the choir is easy. Leading – which sometimes means taking your followers where they don’t really want to go – is hard. I suppose it is nice that Hillary Clinton wants to be known as a pro-science leader, but if she really wants to be one, she has to adopt a scientific worldview that favors carefully collected evidence over preconceived notions. To conclude she is pro-science, well – I need to see more evidence.
Every once in awhile, it is worth repeating the Isaac Asimov quote that inspired the title of this blog:
If knowledge can create problems, it is not through ignorance that we can solve them.
Yet ignorance breeds fear, and fear, apparently, breeds petitions over at change.org. Facebook served up this little gem for me today: a petition entitled Say No To Genetically Modified Mosquito Release In The Florida Keys, posted by Mila de Mier. Her goal of 200,000 signatures is nearly met, despite the fact that there are only 25,000 residents of Key West, and there are less that 100,000 residents in all of Monroe County.
It turns out this petition is quite old (4 years or so) and that the project she was attempting to block has apparently received approval and may be currently underway. But having experienced the pain of reading the petition, I can’t let it pass without comment. And even if it is old news, apparently it’s still out there, and so it should still be countered.
The rambling text of the petition certainly qualifies under the heading of ignorance, unless it qualifies as willful lying. Here it is, with commentary.
Right now, a British company named Oxitec is planning to release genetically modified mosquitoes into the fragile enviroment [sic] of the Florida Keys.
The environment of the Keys certainly qualifies as fragile, but it is hard to understand how mosquitoes could damage that fragile environment. Hurricanes, certainly. Another catastrophe with an offshore oil well, perhaps. Multiplication of the lionfish, maybe. But mosquitoes? Indeed, killing mosquitoes might normally require the widespread spraying of insecticide, which, depending on the insecticide in question, might indeed be a challenge to a fragile ecosystem. So shouldn’t someone worried about a fragile ecosystem be standing up and applauding Oxitec’s environmentally friendly mosquito solution?
The company wants to use the Florida Keys as a testing ground for these mutant bugs.
“Mutant bugs” certainly sound scary, especially if you glance up at that mutant bug image (which I borrowed from the petition itself) – it looks like a meth-crazed (or maybe tomacco-crazed), bloodthirsty killer, outfitted by science with superhuman (supermosquito) powers of destruction. But “mutants” are rarely more powerful than naturally-selected forms with hundreds of millions of years of evolutionary fine-tuning, and in this case, the mutant and its offspring are designed to harmlessly die.
Even though the local community in the Florida Keys has spoken — we even passed an ordinance demanding more testing — Oxitec is trying to use a loophole by applying to the FDA for an “animal bug” patent. This could mean these mutant mosquitoes could be released at any point against the wishes of locals and the scientific community. We need to make sure the FDA does not approve Oxitec’s patent.
Now I’m lost. The petition is supposed to be directed at Adam Putnam, Florida’s Commissioner of Agriculture. He doesn’t work for the FDA. In any event, the FDA has already issued a preliminary opinion that the project will have no significant impact on human or animal health, to say nothing of the fragile environment of Key West.
Nearly all experiments with genetically-modified crops have eventually resulted in unintended consequences: superweeds more resistant to herbicides, mutated and resistant insects also collateral damage to ecosystems.
This is an outright lie. Certainly, the so-called Roundup-Ready crops are more resistant to the herbicide Roundup, but crops aren’t weeds, and their resistance to Roundup does not produce resistance to any other herbicide. Yes, application of herbicide to crops will slowly select for weeds resistant to that herbicide, but that has nothing to do with genetic modification of the crops – it has to do with the schedule of herbicide application, which is equally true for conventional crops. More importantly, this sentence is a complete red herring – the genetically modified mosquito can’t produce superweeds. It’s a complete non sequitur. What exactly is Ms. de Mier (and 170,000 signatories) worried about?
A recent news story reported that the monarch butterfly population is down by half in areas where Roundup Ready GM crops are doused with ultra-high levels of herbicides that wipe out the monarch’s favorite milkweed plant.
I’d like to see this “news story” indeed. No one “douses” their crops with Roundup, be it conventional crop or Roundup Ready crop. And research shows that monarch butterfly populations are not limited by the availability of milkweed. Now, if monarch butterfly populations are declining, that is worthy of investigation. But nothing Oxitec is proposing to do with mosquitoes has anything to do with butterflies. Again, if you are worried about butterflies, you should be standing on the rooftops cheering Oxitec and the Florida Keys Mosquito Control District for attempting to eliminate mosquitoes without spraying insecticides that might affect beneficial insect populations.
What about our native species of Florida Keys Bats. Are there any studies being conducted to see if these mosquitoes will harm the native bat population? Why would we not expect GM (genetically modified) insects, especially those that bite humans, to have similar unintended negative consequences?
I’m trying to understand this, but it’s really hard. What similar negative consequences is she talking about? Because there are crops that are Roundup Ready and Roundup reduces milkweed and milkweed is necessary for monarch butterflies and mosquitoes are genetically modified and they bite humans… is the concern that we won’t be able to feed human babies milkweed any more? I’m lost!
But let’s go at this a different angle. Oxitec wants to kill mosquitoes, and clearly Ms. de Mier is worried that if an Oxitec mosquito, doomed to die, bites a human, then the human might die as well – or at least suffer in some way. But this concern conveys complete ignorance about how the mosquitoes are modified. Admittedly, the technology is complex, but the biologists at Oxitec, and the FDA, and thousands of scientists worldwide do understand the technology, and do have justifiable confidence that Ms. de Mier’s fear is completely unfounded.
First, it is female mosquitoes that bite humans, but Oxitec only genetically modifies and releases male mosquitoes. Thus, even if the GM mosquitoes could pass on some toxin by biting (and they can’t – but even if they could), only non-biting mosquitoes are modified. But second, and far more importantly, the modification causes the mosquitoes to manufacture a protein that inhibits gene transcription. This protein quickly causes the cells of the mosquito to cease functioning. In fact, Oxitec had to build in the ability to turn off this gene prior to releasing these mosquitoes, or they wouldn’t even mature in the lab. Because the agent is a protein, it is completely harmless to any organism (such as a bat) that might eat the insect, as proteins are normally and thoroughly broken down into amino acids before being absorbed into the body. And even if Oxitec made a mistake and made a few modified female mosquitoes, and they happened to bite you, and the protein managed to accumulate in the fluids that enter the blood of the person bitten by the mosquito, the injected protein would be in such minuscule amounts compared to the size of the human, or dog, or cat bitten that the effect on any cellular machinery would be too small to measure. If, indeed, it could have any effect, given that the protein is optimized to work on the gene transcription process of insects.
And by the way, if you are worried that a protein made by a GM mosquito might hurt a bat that eats it or a human bitten by it, shouldn’t you be just as worried for a bat eating or a human bitten by a mosquito doomed to die from an insecticide spray? After all, insecticides which are sprayed aren’t proteins and may have a much longer active lifespan. I’m not saying you should be worried about this either – I’m just pointing out that the fear of the technology is entirely due to ignorance, not knowledge.
Will the more virulent Asian tiger mosquito that also carries dengue fill the void left by reductions in A. aegypti?
As far as I can tell, the Asian tiger mosquito is active during more of the day, and therefore might be more likely to bite when people are out and about. The more bites, the more virulent (capable of causing harm). This is, so far, the one bit of cleverness in Ms. de Mier’s plea. Of course, if the tiger mosquito moves in when the aegypti mosquito dies out thanks to Oxitec, then we would expect to see rises in disease rates. Will Ms. de Mier then greet, with relief and celebration, the dramatic reduction in dengue fever cases in areas where the Oxitec mosquitoes have been released?
Will the dengue virus mutate (think antibiotic resistant MRSA) and become even more dangerous?
Well, after making one halfway decent suggestion, we return to kooky town. If you are worried about a virus mutating, then you should reduce the virus’s ability to get into hosts where it can replicate and – you know, mutate. Genetically modifying a mosquito won’t increase the rate of dengue virus mutation. Indeed, killing off dengue’s host in great numbers is the surest way to reduce the rate of viral replication and therefore mutation.
There are more questions than answers and we need more testing to be done.
If there’s one reliable, laughable bit of hypocrisy you always hear from the Luddites of the world it is this – “We’ve got to stop testing this technology because we haven’t tested it enough!” Whether it’s Greenpeace destroying a test field of potentially life-saving golden rice, or Ms. de Mier and her 170,000 petition signers trying to stop a field test of the Oxitec mosquito, you can be sure they don’t really want more testing. Ms. de Mier, this is how testing works. As the FDA has already ruled and as any professional biologist can tell you, the testing thus far has been more than ample to determine that field tests can proceed. We know these things are safe to humans and animals, and now we need to find out if it’s effective in lowering transmission of disease.
Having exposed Ms. de Mier as either an ignoramus or a charlatan, let’s check out the comments and see who is signing this petition.
I don’t live in Key West, but I am sick and tired of Monsanto and other biotech companies using the general population as their laboratory! I can control what food I put in my mouth, but I cannot control their poisons blowing onto the crops that I eat, nor can I control getting bitten by mosquitoes! PLEASE do not let this insanity continue! No more genetically-mutated crap on this Earth!Sandi White, Lowell, MI
Oxitec is not Monsanto. If you don’t want poisons blowing onto your crops, you should favor development of GM crops which reduce pesticide use, particularly insecticide use. Crap cannot be genetically modified, only the organisms (like Sandi White of Lowell, Michigan) which produce it (like the comment above).
I am certain that, though Oxitec claims that these mosquitoes will be harmless and/or beneficial, sooner or later it will be discovered that something is horribly wrong with these mosquitoes. Genetic engineering is in its infancy. Common sense dictates that the release of an experimental organism – one that breeds uncontrollably and will undoubtedly transmit antigens to humans and other hosts – into the natural environment is both moronic and irreversible.Seth Casson, Kihei, HI
We should make all public health decisions on the basis of the certitude of Seth Casson of Kihei, Hawaii, right? Genetic engineering is hardly in its infancy; it’s been used for several decades and is responsible for major medical advances such as the ready supply of insulin for diabetics and the creation of mouse models for neuroscience research. The Oxitec mosquito can breed, this is true (that’s the point of releasing them) but the larva will die before maturity. How is this breeding uncontrollably? And can we all agree Mr. Casson has used the word antigens while having no idea what it means?
I am also sick of Monsanto and other biotech companies using us as guinea pigs. We really DO NOT need to let loose GM mosquitos into the environment. Whatever happened to the USA being a country “for the people, by the people”? We were never asked if we wanted GMOs released into our environment and polls show that 90% or more of citizens don’t want them. It makes me incredibly sad and angry that the US has become a falsely “democratic” nation. There is very little democracy left if we have no voice.
Mairin Elmer, Fallbrook, CA
Oxitec is not Monsanto. Would Mr. or Ms. Elmer vote tomorrow to rid the world of insulin, or cheese, or other products largely available due to genetic modification? I think if you announced to the world they’d have to give up inexpensive cheese, that 90% figure would drop right quick. Heck, 80% of Americans oppose food with DNA (at least unlabeled). If we get rid of food with DNA, try surviving on salt. It’s the only food eaten in quantity that has no DNA. Of course, the food idiots tell you salt is bad for you too.
I’m signing because I want these atrocities to stop. You can’t mess with Mother Nature & not have something bad happen, they don’t know what they’re doing!!!
Karen Whissen, Newark, OH
Ah yes, the Frankenstein gambit. You can’t mess with Mother Nature, says Ms. Whissen, pounding angrily on an iPhone constructed of rare metals mined from the earth’s crust.
There’s not much point in going on, I suppose – by definition, if someone signed the petition, the comment is unlikely to be scientifically grounded. Perhaps, instead, we should take some comfort in the fact that only 170,000 people signed the petition, and not a single one of them could justify that signature with a coherent rationale.
Out Of Our Heads: Why you are not your brain, and other lessons from the biology of consciousness by Alva Noe, PhD (2009, Hill and Wang, New York)
I routinely teach a course formerly called The Psychobiology of Consciousness and currently called The Mind-Body Problem. Although I am not a consciousness researcher per se, I was drawn into the field of physiological psychology because of my fascination with this topic. Like many introspective people, I “discovered” John Locke’s inverted spectrum problem long before I’d ever heard of John Locke: if you and I are both looking at a red apple, how do I know that your experience of red is the same as mine? You might see it the way I see the blue sky, or a yellow dandelion; yet having learned the term “red” for that experience – the experience of looking at such an apple – you call it red and beyond that verbal agreement, neither of us have direct access to one another’s subjective phenomenology. Later, as a graduate student, I learned that there was such a thing as blindsight – a neuropsychological syndrome usually caused by damage to primary visual cortex in which a person becomes blind – yet can paradoxically can recognize objects by sight if forced to guess at their identity.
These examples convinced me that the best way to understand the mind-body problem – the question of how a physical brain can create ineffable subjective experiences (“red”, “cold”, “sourness”) – would be to become a sensory neurobiologist. Furthermore I began to study the taste system – because of all the sensory systems, that was the one that seemed to have the most circumscribed phenomenological experiences. Tastes were sweet, or sour, or bitter, or salty, and that was about it. (Yes strong and weak, and yes umami or oleogustic, but nonetheless, a more manageable range than millions of colors or thousands of auditory pitches.) Furthermore I styled myself as a researcher in “taste quality coding”, which is to say, I was interested in understanding the patterns of neural activity correlated with those particular experiences. In that respect my work was in the tradition of Francis Crick and Christof Koch’s suggestion that people interested in consciousness should begin to search for the neurobiological correlates of consciousness – brain activity associated with a particular feature of conscious experience.
Even at the beginning, though, I think I knew there was something wrong with this approach. There’s a danger in taking the word “coding” too seriously. When we taste something, some of our taste buds detect the molecules of our food, and cause electrical signals to stream towards our brains. Eating a sweet apple versus a salty pretzel both cause this electrical activity, but presumably the activity is different in some way for the apple than it is for the pretzel – hence we can tell the difference, and hence we experience sweetness in one case and saltiness in another. Whatever that difference is we might call the code for taste quality. Like a code, the meaning (“sweetness”) is in a different “language” (a barrage of electrical impulses). However, a code implies decoding – someone or something will translate the message and experience the sweetness as a result. But is this really what happens? There’s no little guy inside of our brains that decodes the message. Our brains operate on the language of electrical impulses: there’s no need for a decoding at all. This was a thought illusion one of my scientific heroes, Robert Erickson, tried (mostly in vain) to disabuse his colleagues of. One colleague who was sympathetic was Bruce Halpern, whose article “Sensory coding, decoding, and representations: Unnecessary and troublesome constructs?” must have pleased Erickson when it was presented at a festschrift in his honor.
Regardless of concerns about decoding, there is still the question of where our subjective experiences come from. The working assumption of Crick and Koch, obviously, is that they come from brain activity. Most people believe that only organisms with brains are conscious – I am conscious, the rock is not. My dog is conscious, my tomato plant is not. But if this is right (and when I get around to talking about Alva Noe, I will point out that he does not think this is right – or rather, that this is not the whole story) – then there is an interesting problem. Our brains are made up of 80 billion neurons (and hundreds of billions of glial cells) which are not in physical contact with one another, yet we seem to have only one unified consciousness. How is such a thing possible? (And Noe would chime in here: and why is the skull a magical barrier?)
Imagine we were to remove one of these 80 billion neurons. Or a million. Or a billion. Such things happen all the time of course, as a result of aging, neurodegenerative disorders, strokes, head injury. These events may change someone’s behavior, but they do not eliminate consciousness. But how far could we go? How many could we eliminate? (One could ask the reverse question: when does consciousness emerge in embryological development?) There’s really no principled way to give an answer to this question. I think, in fact, that it was because of this problem that the renowned philosopher David Chalmers proposed a radical solution. Unable to draw a line, Chalmers proposes that no, we’re wrong, the tomato plant is conscious too. And so is the rock. Chalmers proposes that consciousness is a fundamental property of the universe, like mass, and that (somehow) the magnitude of the consciousness is proportional to the amount of information involved. If this sounds loopy, I think it does too. If I get around to reviewing one of his books, I’ll say more.
Alva Noe (remember him? This essay is about him!) has a very different answer to this conundrum. Noe believes the mistake is to start with the premise that consciousness occurs inside of us, inside of our brains. He doesn’t believe the neurobiological correlates of consciousness will reveal anything about the mind-body problem. Instead of going inward, more and more restrictively (as Penrose and Hameroff do, with their idea of consciousness as a product of the quantum states of microtubules – an idea even loopier than Chalmers’ panpsychism), Noe goes more expansive. Noe suggests that consciousness is not something in us but something we do – and that it encompasses (is encompassed by?) our interactions with the world (including all that we are perceiving at the moment and all that we are acting upon). We should be looking not for consciousness in our brains, or even worse, in some small part of our brains (the microtubules of Penrose or the dynamic core of Gerald Edelman and Giulio Tononi), but rather in the dynamic interactions of a situated agent in its locally-accessible environment.
This may also sound like a loopy idea, but I don’t think so. Consider the following exercise I have my students try in the first week of class. Take a pencil and close your eyes. Now draw a tree on a piece of paper. As you move the pencil, ask yourself the following question: as you are guiding the pencil, do you in some sense “feel” the paper through the tip of your pencil? Most people do. (And the golfer “feels” the ball hitting the club, the blind man “feels” the grass with his cane, the gardener “feels” the roots of the bush with her rake.) Of course what’s really happening is the pencil, or the golf club, or the cane, or the rake, is vibrating against our hand and fingers in a way that we’ve learned to ascribe that to that other feeling. Except that’s not quite right either, since if we are our brains, what’s really really happening is that the vibrations against our hands and fingers are causing neural activity in the hand region of primary somatosensory cortex (or somewhere “beyond” that in the neural circuitry). Or maybe the first description is right after all. Noe would argue for that more expansive view of our bodies as extended. The voice from across the room is experienced as being across the room, not in our auditory cortices and not in our eardrums.
These kinds of examples are discussed in Noe’s Chapter 4 (Wide Minds) where he also reviews some of my favorite studies from my class. There is the rubber hand illusion, in which an experimenter touches a fake hand which is visible to the subject while simultaneously touching (in the same relative location) the subject’s actual hand (hidden from view). Over time, the subject experiences that rubber hand as part of his or her own body, and have the experience that the touch is being felt from the rubber hand itself. (If you watch the video linked here, be warned that the explanation provided for the effect falls into the usual trap that Noe objects to in his book).
Noe addresses related experiments in his Chapter 3 (The Dynamics of Consciousness) which is the chapter where his book really begins to gather momentum. Here, he addresses the rewired ferret experiments of Mriganka Sur. These technically arduous and brilliant experiments (with one outstanding flaw, in my opinion, which maybe I will write about another time) essentially produced ferrets in which information from the eyes was redirected to primary auditory cortex. These ferrets behaved like they still experienced vision despite this redirection, and features of the auditory cortex developed a visual cortex like character. In the battle, in other words, between the brain (I’m auditory cortex, therefore you shall hear) and the dynamic interactions of a situated agent in its locally-accessible environment (to coin a phrase), the latter wins. The sensory-motor contingencies were visual, so the experience was visual, despite the identity of the brain region.
Related, Noe also describes another favorite of my Mind-Body class: sensory substitution, especially the work of Paul Bach y Rita. Rita was interested in developing a technology that might help the visually impaired. In the original incarnation, blind subjects were seated in front of a large TV camera, which they could direct at an object. The camera’s view would then be translated as little electrical tingles on the subject’s back, isomorphic to the scene. So if the camera was pointing at the letter X drawn on a chalk board, the subject would feel a X-shaped set of tingles on his or her back. The technology improved over time, so that now the camera can be placed in a pair of sunglasses, and the electrode array is placed on a small pad worn on the tongue. Although Noe oversells the phenomenon a bit in his description, Rita describes the experience as visual or quasi-visual – at least, it is unlike touch. This phenomenology emerges once the subjects have some experience with the system, and is much more powerful when the subjects are in control of the camera. That is, pointing the camera at a stationary X is much less useful than panning the camera (now, by moving the head back and forth) – a behavior that is also very visual in nature. Even more exciting, users can duck to avoid objects or, alternatively, catch them. When visual objects approach us, they “loom” – they grow bigger. This does not occur (in nearly the same way) with somatosensory stimuli – so experienced users of this system immediately equate a spreading of the electrical tingles with an approaching object. They also quickly learn how to move their heads to get more information about an object, again, not a natural somatosensory behavior. Again, he have a case where the sensory-motor contingencies seem to specify the conscious experience rather than the brain area activated (here, the tongue region of somatosensory cortex). The dynamic interactions of a situated agent in its locally-accessible environment, once again, is explanatory. (See also a recent exciting paper by Julia Ward and Peter Meijer.)
There are problems with Noe’s ideas too. Phantom limb pain is a difficult condition faced by many amputees in which they continue to feel their non-existent limb – and often it feels excruciating. The usual explanation is that the lack of neural inputs from the hand to the somatosensory cortex produces a change in the brain so that this area is now dominated by inputs from other places – such as the face. Touch to the face is then felt in the hand (a case of the brain region winning over sensorimotor contingencies). There are also dreams and hallucinations – where sensorimotor contingencies would seem to have no explanatory power for a phenomenological experience – where the only thing that seems to be happening (correlated with the experience) is neural activity. To his credit, Noe takes on these situations. In some cases I found his explanations compelling (as with dreams) but in other cases less-so (as with phantom limbs).
Noe’s Chapter 6 is titled The Grand Illusion, which is how I first came to know of Noe’s thinking (he authored a paper called “Is The Visual World A Grand Illusion?” which I have used for many years in class). His answer to this question, by the way, is essentially “no”. Since this is probably most people’s answer to the question, one would wonder why such a paper would need to be written, which means I must do some explaining. Consider the examples I gave at the start of this essay. When we hear the sound of a distant voice, we experience the voice as coming from far away. In a sense, this is an illusion: the only reason we can detect the voice is that air molecules (set in motion by the speaker’s vocal cords) cause our ear drums to vibrate. Nothing about the way they vibrate indicates the origin of the voice that set them in motion. Likewise, we see the world in 3 dimensions: my coffee cup I see as being at arm’s length, my door is several feet away. But this too is something of an illusion. The only way I see these objects at all is that the reflection of light from them falls on my 2-dimensional retina. The brain, it would seem, creates the illusion of 3-dimensions. (Obviously a useful illusion, as it proves to be accurate when I reach for the coffee cup.)
But furthermore – so the story goes – we experience our visual worlds as being all in focus, and we experience ourselves as being able to easily detect changes in our environment. But a moment’s experimentation should prove that very little of the world is in focus: concentrate on any word on the screen of this essay, attempt to keep your eyes still, and notice that only that word is in focus. Also consider that magicians can easily fool audiences with sleight of hand tricks in which we fail miserably at detecting changes in our environment when we are distracted. (This is related to the psychological phenomenon of change blindness.) Noe describes the fact that many philosophers and psychologists have made much hay of these phenomena: that we have a false belief about the completeness of our perceptual worlds – and that this is the grand illusion. Noe argues that we do not in fact have false beliefs – or at least, that our behavior belies this. We are constantly shifting our eyes, tilting our heads. The artist does not look once at his or her portrait subject and draw from memory; the artist is constantly studying and restudying the subject throughout the sitting. We do not act as though we build up a representation of the world in our heads for constant consultation – we do not have to. The detail is not in our heads, it is in the world. Our feeling of the completeness of our perceptual experience is not, Noe would say, an illusion of the completeness of an internal representation of the world but rather an awareness – based on a lifetime of experience – that we have access to all that rich detail by employing the right, basic skills – eye movements, head movements, body movements. Again, Noe is reinforcing the point that consciousness is not in us but rather consists of what we do – the skills that we use to interact with the world.
For the neuroscientist – and for the taste quality coding theorists of the world – this hits home. Much of the program of sensory neuroscience has been based on understanding how stimulus features are represented in neural activity. In Chapter 7 (“Voyages of Discovery”), Noe takes on the giants of my field – David Hubel and Torsten Wiesel – Nobel Prize winning neurophysiologists. (Theoretical critiques aside, Hubel & Wiesel’s contributions to neuroscience are unassailable.) Noe notes that their discovery of the responses of visual cortex neurons – in anesthetized animals – was responsible for decades of research and thinking in neuroscience focused on understanding feature representation (which reached its most Baroque form with the probably misguided work of the genius David Marr.) The kind of reification of the duties of neurons or brain areas, and the eventual (also misguided) “modular brain” theories of cognitive science, are a long way from the warnings of Erickson and Halpern, cited earlier, that representations and internal models may not be necessary to explain behavior. (Mental representations or fuzzy modularity may still have some utility – but Noe would probably disagree.) Noe’s critique of Hubel and Wiesel was certainly the boldest part of the book, and for that reason, one of the most important.
In the end, then, I found Alva Noe’s book full of important ideas. He reviewed a number of key phenomena in psychology and neuroscience. He called out the hidden dualism of active programs in neuroscience. As effective and as thought-provoking as the book was, though, it still didn’t help me understand why that apple was red, and why it tastes sweet. The how of the mind-body problem still nags, but in part thanks to Noe’s writings, I am excited that we may have a better idea of the where.