Up to David L. Thompson's Homepage
Outline by Section:    


In our previous colloquium, Brian Penrose gave us a triadic analysis of freedom (taken from MacCallum): Freedom always involves an agent, an obstacle and an action. What kind of freedom we are discussing is defined primarily by what obstacles count as impediments. The obstacle for negative, political freedom, for example, is coercion.

The freedom with which I will be concerned in this paper is defined by its purported obstacle: causal determinism. I will not be discussing civil, social or political freedom, but Free Will.

Causal or physical determinism ("determinism" from here on) is the claim that "all things in nature behave according to inviolable and unchanging laws of nature" (Enc. of Philosophy, Vol 2, p. 364).  Laplace, for instance, claimed that if we knew all the laws of nature, and the initial conditions, we could in principle predict the occurrence of any material event at any time in the future. Such a belief in determinism is at the very core of the modern scientific world-view; if we didn't think the universe were governed by laws, there would be little point in pursuing science. (Note that appeal to quantum indeterminism in no way undermines causal determinism in this sense - the sun is still determined to rise in the east! Even if micro-particles obey probabilistic laws, the laws of chemistry, of neurophysiology and so on remain as "deterministic" as ever. Laplace would be quite satisfied to use these laws to predict all macroscopic events.)

Free will, ("freedom" from here on) has been held by many to be incompatible with determinism. Such incompatibilism leads to a dilemma: either we are material, hence determined, beings and cannot be free; or we are free and hence cannot be material beings. B.F. Skinner opted for the first horn of the dilemma and denied human freedom. Descartes chose the second horn and declared that since we are free, we, that is, our minds, must be immaterial, beyond physical nature; that is, he is a non-naturalist.

By "Naturalism" I refer to the claim that everything in the universe, including every human being, is material and so is governed by scientific laws. Evolution by natural selection has led gradually to the existence of human beings with minds. Minds are somehow the product of the functioning of brains and so can eventually be understood without appeal to spirits, ghosts or other principles which are super-natural. I have great respect for science and am tempted to accept naturalism, but I hesitate out of the fear that it leaves no room for human freedom, morality and responsibility. And indeed some naturalists such as Baron d'Holbach have simply declared them to be illusions.

Daniel Dennett offers me a way out of this dilemma so I don't have to choose between giving up freedom or giving up science: He proposes that the two are compatible. Compatibilism has been around in one version or another at least since Hobbes. The recent popularity of the position that the mind can be understood as the information processing function of the brain, however, has led Dennett to update, in this contemporary context, the claim that while we are material beings and so are determined, we need not be afraid, for we can still have freedom, or at least some kinds of freedom.

What kinds of freedom? He claims we can have all the freedoms worth having. When people want freedom they don't want Absolute, Metaphysical freedom (with the exception of some confused and obstinate philosophers); they want freedom in the sense of being able to control their own lives by means of the information they have; that is, they want to be able to deliberate and make choices about their future on the basis of reasons. These kinds of freedom, Dennett maintains, are compatible with determinism. There may be some concept of Metaphysical Freedom which is not compatible, but once we sweep the confusions aside we will find that that wasn't worth having anyway.

The point of Dennett's arguments is not to prove that we are free. He holds that we could not be discussing the matter unless we were. His aim is to show that our initial fear that freedom and determinism are incompatible is a mistake based on our confusion about what kinds of freedom are worth having. When we analyze in concrete detail the actual things we want when we say we want "freedom," we will discover that not only are they compatible with determinism, but that determinism is a necessary condition for us having them.


So what do we want when we value "freedom"? We want a number of things, says Dennett, and we have various fears that determinism means we can't have them. But when we look at these fears in the face they vanish like bugbears. Dennett claims that we want self-control; we want to be rational, that is, to be able to govern our conduct by reasons; we want our deliberations to influence our actions; and we want to be able to avoid evil, to be able to "make a difference" and not be forever faced with the inevitable. Each of these is freedom, in a way. I'll look at each in turn, and we'll see that we can have them all, and determinism too.

Otto, the Incompatibilist, in his usual, self-disciplined manner, argues: This argument is based on a confusion, says Dennett. He claims it is a mistake to assume that if an action is under our control, then it cannot be subject to causal determinism. To show this, we need to analyze the notions of control and self-control.

Self-control needs to be contrasted with two alternatives: being controlled by another, and being out of control. To say that I am controlled by another implies that the other is an agent, that the other has some purpose, and that he or she has the means to impose this purpose on me.

Consider a typical way we speak of "control." NASA can use a radio unit to remotely control a Viking spacecraft and so can force it, like a puppet, to carry out the controllers' purposes. If the spacecraft, however, is likely to travel out of effective radio range, NASA would be wise to install a computer so that the spacecraft controls itself. It then acts not as a puppet, but as a robot. NASA must, of course, design it so that it seeks some goal and must give it the information it needs to do so, or the means to get that information. The space craft is then a self-controller.

It is possible that either the remote control system or the self-control system might break down. In that case the spacecraft will be out of control. It will then be incapable of carrying out either its own internally programmed goals, or those of a remote controller. Any information it has will be irrelevant.

The crucial point is to notice that, in all three cases, the spacecraft is determined. Control and self-control are possible only because of determinism, and even in the case of being out of control, the spacecraft is still causally determined.

Consider another example. In a game of billiards we try to control the ball, but in roulette if someone effectively controls the ball (with a magnet?) we feel cheated. In roulette, the ball is supposed to be uncontrolled. But no one proposes that the roulette ball thereby escapes the laws of nature. Knowing that the roulette table is fully subject to deterministic laws still does not tell us whether it is controlled or not.
Laplace's superior intelligence would presumably fail to see the difference [between billiards and roulette], and if offered a job in a casino, would be just a capable of controlling a game of roulette as any billiards player would be capable of controlling a game of billiards. (Daniel Dennett, Elbow Room, p 61. All future references will be to this text.)

The point of this analysis is that "determinism does not in itself 'erode control.'" (p. 72) Far from being incompatible with causation, control, self-control and even being out of control are all possible only because of causal determinism; they are modes of being determined.

Human beings want to be self-controlled. We find ourselves with goals and we attempt to achieve these by searching out the information we need to accomplish them. Many things are incompatible with such self-control. We rightly fear control by others such as military dictators, torturers, brain washers, TV commercials, neurologists with brain probes, subliminal hypnotists controlling our actions, or evil demons. (Maybe only the philosophers among us worry about the latter things.) Many of us worry about getting out of control, of losing control of ourselves; Lou Gehrig's disease, MS, or Alzheimer's come to mind as extreme cases. There are real dangers to our self-control, but determinism is not one of them.

Dennett considers this fear to be based on false personification. "The environment, not being an agent, does not control us (p. 61)." Nature is not an agent; it has no desires for us; it doesn't make us do what it wants, for it has no wants. We may cease to be self-controlled and go out of control, but that doesn't mean we fall into the clutches of some Laplacean demon and become other-controlled. If I am out of control, then there is no coherent purpose, my own or anyone else's, to my actions. The most we can say here is that when we are in full control of ourselves we are responding rationally to the world because Mother Nature has "designed" us to act in accordance with reasons; but that, presumably, is what any self-controlling being would want. We are, if you like, designed (by evolution) to be self-controllers. (Cf. Sartre's "destined to be free.")
  Accepting defeat by the rationality of Dennett's argument, Otto the Incompatibilist offers another reason for his position: If a naturalist like Dennett is to explain freedom, then he must give an account of how acting for a reason, that is, teleological action, fits into the natural, causal scheme. "How could reason ever find a foothold in a material, mechanical universe? (p. 21)"

Dennett offers the following naturalistic story. Once there were no reasons, only causes. Then simple replicators, such as amoebas, appeared which, in so far as they were replicators, had an "interest" in their environment being such as to be conducive to their survival. The most successful organisms were those which developed the capacity to respond differently to those things in the environment which were objectively "good for it" and those which were "bad for it." In other words, these organisms came to respond to things in their environment not just by responding to their objective physical properties, but by categorizing them from the viewpoint of the organism's interests and acting accordingly. In a very primitive sense, then, these organisms acted for "reasons;" they took in some substances because they were nutritious and rejected others because they were noxious. Organisms came to respond to the informational features of the world, the "meaning" the things had for the organism, not just to their brute physical properties.

A mouse, for instance, responds with similar evasive action to the spring of a cat or the swoop of an eagle not because they have anything in common from the viewpoint of physics, but because both are categorized as "predators" from the mouse's point of view, they both carry the same information for the mouse: time you were out of here! Given the mouse's interest in survival there are reasons to act in one way rather than another. That is, the mouse's action is purposeful and guided by information.

Acting for a reason need not have anything to do with knowledge or consciousness, as Aristotle long since pointed out. The amoeba knows nothing. But the distinction between knowing and not knowing should not be confused with another distinction. An organism can do the right thing either by pure chance or on purpose. An amoeba rejects poisons on purpose, that is, because it is evolutionarily designed to do so. A species not so designed might survive briefly by chance, but not for long!

The crucial thing here is to see that, for Dennett, acting for reasons is not opposed to, or incompatible with, being moved by causes. To act for a reason is to allow a cause to influence the organism only "with respect;" it involves a kind of delayed reaction. Gravity acts on a planet without regard for any purpose or interest the planet might have. How a stimulus influences an organism depends no longer just on the physical nature of the stimulus, but on the goals of the organism and on its history. Whether a cell absorbs a molecule of sugar depends not just on the sugar, but on the goal of the cell, whether it "needs" more sugar right now or not, which in turn depends on when it last "ate." This is not a violation of causality, but a way of structuring causation so as to subordinate it to organic teleology. This structuring is itself causally effected by the internal organization of the organism. Far from reason and causation being in contrast, acting for a reason is only possible in a causally deterministic world. This is bootstrapping at its best!

Well, at least Dennett doesn't include rocks and streams! I am very suspicious of all-or-nothing theories in the philosophy of mind. The Platonic-Cartesian Axis substantialises the mind's powers into objects which are either there or not there. One cannot have only a bit of a soul, or 35% of a cogito, or be only half free, they say. Yet we all know that consciousness and freedom grow slowly in a child and a naturalist at least could be expected to claim that their appearance in the course of evolution is gradual. The alternative would be to have God intervene supernaturally at some point: "ZAP! You're free." For Dennett, amoebas act to some extent on the basis of information, and so they are more "free" than, say, planets. That Dennett's theory sees human freedom as growing gradually out of a more general development of the capacity to act for a reason seems to me to make his position more credible, not less.

Nevertheless, Dennett holds that human being are unique in that we are capable of knowing the reasons for which we act. We not only form representations of things in the world, we also represent to ourselves our goals, our reasons, by means of language or other symbols. Such self-representation is, for Dennett, the essence of consciousness. Just as representation of the physical world gives the amoeba a certain distance from brute causation, so consciousness gives us some leverage over our own goals. We may even take on projects beyond the biological ones of survival or reproduction.

Yet the basic analysis is the same. To act for a reason, whether the reason is represented to oneself or not, is not to go against causality, but to structure causal processes in a way which contributes to our interests, which allows us to act for the (our?) good. If, then, freedom is to be found in acting rationally on the basis of information, it is not only compatible with causation, it is one particular structure or mode of causation.

Stymied by these reasons, Otto deliberates about his next move. Ah! He has it! Dennett responds by analyzing the difference between Laplacean calculation and human deliberation. From the viewpoint of Laplace's demon, any future state of the universe can be predicted by calculation. All that is needed is to calculate the future state of every microscopic particle in the universe at that future time. Note that there are no short-cuts; it won't do to calculate only particles on the earth, since we need to ensure that no comet will collide with the earth during the time we're considering. Only a practically infinite, probably divine, calculator could possibly handle the computations needed to say what will happen in this room five minutes from now. What is a poor finite human creature with time restraints to do?

Deliberate! Unlike calculation, deliberation is a heuristic process by which an organism uses quick and dirty methods to get a rough estimation of what is likely to happen in its vicinity on the basis of the limited information available to it. Only a few factors are considered, those which are likely to be relevant. Only macroscopic agglomerations enter into consideration. Rough approximations rather than exact calculations speed up the process. Within a fraction of a second I can estimate the state of this room five minutes from now. Of course, my estimate won't be certain, nor exact, but how else could us finite beings be expected to get along in the world? Better to act now after a brief heuristic deliberation and perhaps get it wrong than to sit calculating for centuries after the moment has passed!

No, says Dennett, this is to confuse determinism with fatalism.

Consider a genuine instance of fatalism. If I am sucked out the faulty window of a plane, I have about three minutes while I am falling to the ground to deliberate philosophically about what I should do. But this is futile, as Otto correctly sees. What is about to happen cannot be the result of, will not be influenced by, my deliberations. On the other hand, when I first reserved my seat on Wing-and-a-Prayer Airlines, I did deliberate about their safety and checked my Consumer Reports, the best heuristic method I could think of. Well, I made a mistake. But this was not fatalistic; if my deliberations had led me to read the most recent Consumer's Update, I might well have decided otherwise. My buying the ticket was a real (causal) result of my deliberations; deliberation made a difference.

In a deterministic world, we can still distinguish between those events which, though determined, are the result of deliberation and those which are determined in "a simple, informationally insensitive way." (p. 105) What we want is that as many events in our lives as possible not be subject to fate, but be events about which we can deliberate in a manner that is effective. And Dennett is an optimist: While he grants that "there are indeed conditions under which deliberation is futile ... it is our good fortune that these conditions are abnormal in our world." (p. 106)

Determinism is not fatalism. The argument that we should reject determinism because it renders deliberation futile and so undermines the possibility of freedom is based on confusion. What Otto really fears is fatalism, not determinism. What we want is that our deliberations be effective, and in a deterministic world they often are.

But Otto is determined: Again, Dennett claims we must analyze the notion of "avoidable" more carefully before acquiescing in this argument.

Suppose astronomers discover a comet which is about to crash into the earth and cause a major disaster. At the last moment another comet, previously unnoticed by astronomers, appears in just the right place to prevent the first from hitting the earth and so the catastrophe is avoided. "The" catastrophe? There was never going to be one! The second comet has been on track from all eternity, so the collision was never in the cards.

From the Laplacean point of view nothing could ever be prevented, so from this perspective it would be meaningless to distinguish the avoidable from the inevitable. A finite knower, however, has to make predictions and develop expectations on the basis of limited information. It is when these expected events do not occur due to the actions of a deliberator that we speak of avoidance.

Against the background of that anticipated future (that misanticipated future) we see the events that happen as an instance of prevention. ... All the verbs of 'making a difference' ["prevent," "bring about," "thwart," "disturb the universe"] involve a tacit comparison between the way the world was apparently going to go and the way it turned out to go." (p. 125-126)

Unlike Laplace's demon, finite deliberators live in a world structured by their own categories. We don't live in a world of micro-particles; we live in a world of trees and birds, of exams which are passed and wars which break out. In particular, we classify events in terms of whether our deliberate action can make a difference. The sun will rise tomorrow no matter what I do; that is inevitable. Failing my exam is avoidable for I can decide to study for it and thereby prevent the "normal," expected outcome from occurring. Such distinctions make sense only from our macro point of view, from the pragmatic perspective of a human being with limited information. They differentiate between events which happen as a result of deliberation and those whose causal chains do not pass through deliberation. If we as finite deliberators are to be efficient, we must clearly distinguish between those situations whose expected outcomes are avoidable, and so are worth deliberating about, and those situations which are "inevitable."

Imagine someone arguing, "Since determinism is true, thermostats don't control temperature," or "Since determinism is true, planets would move in ellipses even if there were no gravity." Determinism "works through" the thermostat; it works via gravity. In a similar way, Dennett is claiming, it can work through deliberation. A causal sequence which does not pass through the deliberative process we label "inevitable;" if it passes through deliberation we call it "preventable," or "avoidable." Hence "inevitable" doesn't mean the same as "determined," and "avoidable" doesn't mean "non-determined."

So Otto is incorrect in saying that if determinism is true then all events are inevitable. Some events (though determined) involve deliberation, and it is with respect to these events that we can "make a difference."


Otto is now bursting with frustration and can no longer prevent himself from blurting out, on impulse, that:

In the face of this outburst, Dennett coolly explains that Otto's wanting such freedom or feeling that he has it guarantees nothing. It is quite possible that such feelings are illusions (especially among those who believe themselves to be gods.)

Otto (cooling down a bit):

"Really?" asks Dennett. Are you sure that's what you want? Even Descartes, who claimed to have the freedom of God, said, "In order to be free, there is no need for me to be capable of moving both ways." For instance, if I am faced with a difficult philosophical objection, do I really want to have the arbitrary choice of killing or not killing the objector? If I thought for one moment that I could do such a thing, I for one (Dennett for another) would be devastated. I know there are thugs on the streets of New York who could kill you if you objected to the colour of their jacket; they are governed by their impulses and their decisions are close to arbitrary. When my parents brought me up, they worked hard to make me into an adult who could not act like that, and since I took over my own rearing, I've been trying to achieve the same thing. I want my actions to be guided by reasons, to depend on my history, to head in the direction that goodness points. If arbitrariness is what Otto ultimately means by freedom, then I don't want it (and I hope he doesn't get it!) That kind of freedom is not worth having.

A tortured Otto responds:

If Otto was trying to change the past, yes, it was totally useless, probably neurotic; he should have tried valium.

But a more charitable interpretation is to see Otto not as trying to change the past, but as trying to change the future. According to Dennett, this is what most of us who ask ourselves, "Could I have done otherwise?" have in mind. I want to learn from my experience, if possible. Are events like this beyond my control? Are there some factors that I left out of my deliberations in the past that I should remind myself to include in the future? Lying awake asking whether I could have done otherwise is usually an attempt to rationally reconsider my policies, to re-programme myself so that in similar circumstances in the future I will act differently.

So Otto's sense that "I could have done otherwise" is either (A) an illusion, (B) a reference to a kind of freedom not worth having, or (C) a rational attempt to reconsider his deliberative policies. He is mistaken in thinking that it is an indication that he has some ultimate, godlike, arbitrary metaphysical power beyond the deterministic world. And if it were that, he wouldn't want it.


Dennett holds that the information processing model of the mind allows us to have our cake and eat it to. If he's right, we can accept naturalism and brain-based minds without having to give up freedom and dignity. The hardware of the brain may be governed by deterministic laws, yet on the level of mental functioning we could still respond to reasons, and so we could still be self-controllers who deliberate effectively, act purposefully, pursue the good, avoid the avoidable, and even reprogramme ourselves on the basis of experience. That is real freedom, the only kind worth having. And it would be possible for us not in spite of our being determined, but precisely because we are determined.