Steve Talbott


In Flesh and Machines: How Robots Will Change Us (Random House, 2002), Rodney Brooks informs us that "we are nothing more than a highly ordered collection of biomolecules":

Molecular biology has made fantastic strides over the last fifty years, and its goal is to explain all the peculiarities and details of life in terms of molecular interactions. A central tenet of molecular biology is that that is all there is.

Apparently fearing that we will be insufficiently gripped by his message, Brooks (who is director of the Artificial Intelligence Laboratory at MIT) goes on to say in the space of three pages:

The body, this mass of biomolecules, is a machine that acts according to a set of specifiable rules....

The body is a machine....

We are machines, as are our spouses, our children, and our dogs....

We are nothing more than the sort of machine we saw in chapter 3, where I set out simple sets of rules that can be combined to provide the complex behavior of a walking robot .... we are much like the robot Genghis, although somewhat more complex in quantity, but not in quality....

I believe myself and my children all to be mere machines....

We, all of us, overanthropomorphize humans, who are after all mere machines (pp. 173-5).

Noting that some people may bristle at the word "machine" (not to mention the tedious repetition), Brooks acknowledges that he uses the word "to perhaps brutalize the reader a little". He feels the need to shake us free of any conviction that "we are special" -- meaning, in case you missed it, that we should accept our status as "mere machines".

Searching for the Bottom

As a reviewer facing an attempt at brutalization, perhaps I will be forgiven a touch of bluntness. When, using the characteristic language of the insecure reductionist ("we are nothing more than"), Brooks refers to molecular interactions and says emphatically, "that is all there is", he is engaged in a startlingly transparent argument from ignorance.

"That is all there is". What is all there is? Brooks makes the statement in order to intimidate, but is unwilling or unable to tell us anything at all about the brute, underlying, mechanistic reality he is pretending to beat us with. If there is one thing no respectable physicist since the mid-twentieth century would claim to do, it is to describe some sort of ultimate physical machine, let alone a well- understood physical machine. The entire movement of physics has been away from concrete machine models -- and, indeed, away from models altogether. Theories at the lowest level no longer even describe particular things or events that might be modeled. So when, with an air of settling matters for good, Brooks says "that is all there is", what hard reality is he shoving in front of our faces?

His general idea seems to be that there is some "mere" machine-stuff down at a lower level of explanation, and everything above it is "nothing more than" the compounded articulations of this mechanistic reality. Explanation proceeds from below upward. Our nature is fixed by the foundation on which it rests. "We are machines".

But if explanation proceeds from below, then you need to start with whatever is at the bottom. Your primary explanatory apparatus, in other words, is quantum weirdness. And the one place in science where you absolutely cannot find a machine, is amid the scarcely utterable perplexities of the quantum realm. Nevertheless, there is where Brooks would apparently nail down our determining and limiting nature -- the reality we are "nothing more than". Does he care to tell us a little about this reality?

Certainly not in Flesh and Machines. Moreover, even if he could extract something of our essential nature from the sphere of quantum effects, and even if he could demonstrate the machine-like character of this nature, there would remain the difficulty facing all bottom-up expositors of the human being: where is the bottom, since not one iota of evidence within physics suggests we have reached it, and no one even seems to know what it might mean for there to be a bottom or what would be its distinguishing features.

Physicist Steven Weinberg informs us that some of the (unknown) structures pointed to by currently accepted theory are smaller than the atomic nucleus by a factor of a million billion. In other words, it is at least as far from the atomic nucleus to these unknown structures as it is from the scale of ordinary life to the atomic nucleus. If the idea of explanatory levels makes any sense at all (it doesn't), and if our nature is determined by whatever constitutes the bottom, then so, too, the nature of atoms and molecules must be determined by the same inaccessible bottom.

But still, in his happy ignorance about whatever is "down there", Brooks can find satisfaction in the effort to whip us into submission with a phantom: "That is all there is". One wonders what he pictures in his own mind as he says "that". Here I have a nothing-more-than conjecture of my own: his idealized bit of ultimate matter is nothing more than a projection of the machines he builds; his imagined molecule is a kind of minuscule robot-homunculus.

Brooks might reply that, at some higher level of explanation, quantum weirdness (and the unknowns lying beneath it) manage to even themselves out, leaving us with the hope of identifying neat mechanisms at this higher level. But if he is not appealing to determination from a well- understood bottom -- if he claims to discern our mereness at some higher level -- what saves his choice of levels from arbitrariness? Why should we not, for example, take the living organism, or consciousness, as most fundamental and most revelatory for any explanation of the world? If we really want to understand the sustaining powers, the living energies, presented in the world, doesn't it make most sense to look for our understanding where those powers and energies are most fully developed and most explicitly manifest? In any case, whatever the level Brooks chooses, he needs to demonstrate, rather than simply assert, its essential mechanistic nature. He doesn't seem interested in the task.

If he had been interested, he might have begun by listening to the physicists. When Werner Heisenberg can say that atoms are not things, and when Steven Weinberg can say that "particles are just bundles of field energy", and when Richard Feynman confesses that "in physics today, we have no knowledge of what energy is " -- well, then, the natural question to put to Brooks is, Where's the beef? Show us your grand, microcosmic machine! And tell us what makes it merely a machine.

Order and Form

Brooks seems particularly inclined to find evidence of our machine nature at the level of biomolecules. Since he never tells us why these molecules should be regarded as machinery, there isn't much to quarrel with. But a couple of extremely brief comments may prove useful.

Two of the most longstanding and fundamental questions about living things are, first, How does the organism sustain its order and resist decay? That is, how is matter "caught up" in the organism's living processes in a way that suddenly and dramatically ceases when the organism dies and the inevitable processes of putrefaction and decay take over? Or again: what is the difference between the state of the same physical body before and after death? And second, How is the distinctive form of every organism brought into being and maintained?

Both these questions have tended to fade from view as biology has transformed itself from a discipline about organisms to a discipline about molecules. Researchers seem to believe (if they think about the matter at all) that if they can just get a handle on the molecular "machinery" of the cell and organism, the larger questions will somehow answer themselves.

What has actually been happening provides an ironic counterpoint to this expectation. It is less that the lower "mechanisms" are answering the larger questions than that the larger questions are simply reasserting themselves at the lower level. In particular, the problem of order and resistance to decay stares microbiologists in the face as soon as they cease averting their eyes. As Lenny Moss (who is both a cell biologist and philosopher) puts it, when we look at the molecular dynamics within the cytoplasm of the cell, what we see suggests that "biological resistance to thermodynamically driven entropic heat decay obtains all the way down to the most basic fabric of living matter" ( What Genes Can't Do, MIT Press, 2003, p. 91). Moss is one of many researchers looking at the complex chemical dynamics of the cell as a whole, and noting that there is no one-way chain of cause and effect determining the cell's order. This order (which is passed from one generation to the next) is irreducibly manifested in the cell as a whole, with each part (including the DNA) being effect as well as cause.

As to the other problem, that of organic form, one need only point to the decisive issue of protein folding. This is opening up into a complex and massive question of form that mirrors, and is organically inseparable from, the question of overall form in the organism. Harvard biologist Richard Lewontin, after citing some of the many factors affecting protein folding (and therefore protein functioning), continues:

These understandings, however, have not penetrated into the main structure of biological explanation .... What is needed is to move the issue of structure from the peripheral realm of a few special cases to a central concern of investigation at the molecular level. ( The Triple Helix, Harvard University Press, 2000, pp. 115-17)

Given that these two fundamental questions about life -- the question of order and the question of form -- simply reappear at the molecular level, what sense does it make to say, as Brooks does, that the living organism is now understood as merely a collection of biomolecules? Certainly the new questions will, to the mechanistically inclined, appear to demand standard mechanistic solutions. But by the same token, to the researcher who was willing to see the qualitative (and therefore non-mechanistic) unity of the organism as a whole, the new problems at the molecular level will appear as verifications of that original view rather than as refutations of it.

Incidentally, I recently saw a report that physicists have found the proton to be continually shifting in form. While it (whatever "it" is) can assume spherical form, apparently it also takes on other shapes, including that of a peanut and even a donut. (Remember, we're talking not so much about a "thing" as a pattern of forces or energies.) The world, from top to bottom, appears to consist of continually transforming form -- at the molecular level as well as at every other level -- and it's not at all clear how to map Brooks' picture of tiny machine-parts onto this essentially fluid and qualitative reality.

Losing Consciousness

The "we are nothing but" claim takes countless forms and comes at us from many sides. To cite just a few famous cases:

"You," your joys and your sorrows, your memories and your ambitions, your sense of personal identity and free will, are in fact no more than the behavior of a vast assembly of nerve cells and their associated molecules. As Lewis Carroll's Alice might have phrased it: "You're nothing but a pack of neurons." (Francis Crick)

There is no spirit-driven life force, no throbbing, heaving, pullulating, protoplasmic, mystic jelly. Life is just bytes and bytes and bytes or digital information. (Richard Dawkins)

Man has to understand that he is a mere accident. (Jacques Monod)

People, after all, are just extremely complicated machines. (E. O. Wilson)

I have a working hypothesis. Leaving Brooks (about whom I know almost nothing personal) aside and focusing instead on the wider rhetorical phenomenon: why do we so often encounter today the assaultive reductionist assertions, "the human being is only...", we are nothing more than...", "we are merely..."? The vacuity of the claims together with the characteristic aggressiveness of the authors certainly raises some interesting questions.

We get a hint of what is going on by considering the negative and belittling cast of the assertions. Clearly the authors, consciously or unconsciously, see themselves as bringing us down a peg or two. They could, after all, have reveled in the glories and depths and wisdom of the natural world revealed by science. But, no, they want to make sure we understand that we are only such-and-such. Apparently their own pessimistic sense of the matter is that we are discovering the human reality to be an impoverished one -- a reduced one, in fact -- this despite the usual explicit statements about what a stunningly rich and profound world science presents us with.

We may wonder, then, what is behind this sense of being reduced. Whence arises this feeling of inferiority requiring the use of pejorative and reductive language? My own suspicion is that one of the most fundamental insights of Carl Jung may throw light on the matter. The Swiss psychiatrist pointed out that when integral contents of the personality are lost to consciousness, an inferiority results. Moreover, the sense of inferiority gives rise to moral resentment and "always indicates that the missing element is something which, to judge by this feeling about it, really ought not to be missing, or which could be made conscious if only one took sufficient trouble" (Two Essays in Analytical Psychology, Princeton University, 1972, pp. 139ff.).

I feel free to mention this only because the way a personal lack leads to a sense of inferiority and then to moral resentment (so that we are happy to bring others down a peg or two, to our own level) is such a universal fact of life that none of us -- if we are possessed of the slightest self- awareness -- can fail to recognize the process in our own lives. So it is not a matter of pointing fingers at individuals or singling out any group as "morally challenged". But where we can recognize a broad social and intellectual pattern such as the one presented by reductionist rhetoric, we really ought to try to understand what is going on.

The lack Jung spoke of was a loss of contents properly belonging to consciousness. Where do we see such a loss more in evidence than in disciplines such as artificial intelligence, cognitive science, and reductionist science in general? Surely if there is any place where we should expect a loss of conscious contents, it is where such contents are commonly denigrated as a matter of principle. The main drift in these fields is to ignore as completely as possible the immediate contents of consciousness, reinterpreting them in terms of what is not conscious. It is hard to imagine a more direct route to the conditions of lack, inferiority, and moral resentment Jung describes. The entire project of these disciplines is to drop things from consciousness.

In other words, there is a truth of sorts in the reductionist claims. Or, you could say, there is truth in the plight of the reductionist. At the level of its symptoms the psyche never lies. A kind of practical reductionism really has been occurring as the inevitable result, over time, of theoretical reductionism, and we should not be surprised if it produces pathological results. In a society where the cry echoes from all sides, "You are nothing but a machine", we can rightly ask whether what we are really hearing is "I sense that I am becoming nothing but a machine and, dammit all, I won't tolerate anyone else being more than I am".

There is in this symptom, as Jung clarifies it, a paradoxical double aspect: on the one hand, a sense of inferiority, but, on the other hand, a compensating "psychic inflation" through which the ego assumes god-like pretensions. After all, am I, as one of the cognoscenti, not in the omniscient position of knowing the real truth about everyone else, while they remain benighted and self-deluded, ignorant of their "mereness"? We may wonder as well whether this tendency toward inflation helps to explain grandiose visions of our future as a race of cyborgs destined to become masters of the universe.

Another consequence of the loss of conscious contents, aside from a sense of inferiority and a compensating ego inflation, is that our conceptual resources for understanding the world are diminished. When we lose awareness of all but the machine-like in ourselves, we also lose the ability to conceive the world as anything but a machine. Those whose intellectual horizons are encompassed by digital machinery tend to see the world computationally, just as their predecessors saw the world in terms of clocks, cameras, steam engines, telegraph lines, and movie projectors. There is security in believing the world is like the things one knows best, and intellectual ease in draping one's well-practiced ideas like a veil over the Great Unknown.

Turnabout is Fair Play

I willingly own up to some mischief here. "Psychologizing the opposition" is a tactic more often abused than productively employed in worthwhile discourse. But I offer two justifications for the tactic, apart from what I take to be the demonstrable plausibility of the foregoing.

First, what I have said is no mere academic exercise. The future hangs in the balance of our self-knowledge. When we lose much of ourselves to the subconscious, we become blind to our own motivations. We may also seek external powers to compensate for our loss of internal mastery. And, in fact, what we see in much of science and technology today is precisely a blind drive toward power. The idea of inevitability has widely substituted for the submerged sphere of consciousness where we might have felt called upon to exercise personal responsibility for our own actions.

In the second place, Brooks repeatedly psychologizes his intellectual antagonists -- and in the most shameless manner. So I thought a little turnabout would be good medicine. To give one example of what I mean: in 1963 Joseph Weizenbaum wrote the program called ELIZA, which he designed to play the role of a psychologist who largely parrots back the responses of the patient. Weizenbaum was shocked at how dead-seriously many people took ELIZA while "conversing" with it. And he was even more shocked to hear psychiatrists speaking highly of the potentials for such therapy. So Weizenbaum wrote,

What can the psychiatrist's image of his patient be when he sees himself, as therapist, not as an engaged human being acting as a healer, but as an information processor, following rules, etc.?

This is the one and only statement by Weizenbaum that Brooks cites, and he proceeds, without any explanation whatever, to label it "nonsensical" -- unless you can call this gratuitous and unsupported bit of psychologizing an explanation:

Weizenbaum was fleeing from the notion that humans were no more than machines. He could not bear the thought. So he denied its possibility outright .... [He is] afraid to give up on the specialness of mankind .... It is intellectually too scary. (pp. 167-8)

Apart from the fact that Brooks offers no basis for his judgment about Weizenbaum's state of mind, he ignores the obvious possibility: maybe what Weizenbaum "could not bear" was what no one should be able to bear, namely, a completely nonsensical interpretation of machines and programs. The point, of course, needs to be argued, but Brooks casts his slur against Weizenbaum without making any reference to Weizenbaum's arguments -- arguments readily available in one of the classics of technology criticism, Computer Power and Human Reason.

*    *    *    *    *    *    *    *    *   

The New York Times Book Review blurb on its back cover proclaims Flesh and Machines

a stimulating book written by one of the major players in the field -- perhaps the major player .... [He] offers surprisingly deep glimpses into what it is to be human.

But so far as I can tell, Brooks makes no effort to discuss what it is to be human. He does, however, repeatedly express the faith that machines will turn out to be alive. I intend, in a later issue of NetFuture, to take up some of those points at which his faith comes closest to being argument. I will also have a great deal to say about Brooks' emphasis on the "rule-bound" aspects of people and robots.

© 2003 by The Nature Institute.

Steve Talbott (stevet@oreilly.com) is the editor of NetFuture an electronic newsletter where this article originally appeared.