PETER VAN INWAGEN: THE MYSTERY OF METAPHYSICAL FREEDOM

-- The Determinism and Freedom Philosophy Website --

This is an account of his present thinking by an excellent philosopher who has been among the two or three foremost defenders of the doctrine that determinism and freedom are incompatible -- that logically we cannot have both. In his 1983 book, An Essay on Free Will, he laid out with unique clarity and force a fundamental argument for this conclusion. What the argument comes to is that if determinism is true, we are not free, since our actions are effects of causal circumstances in the remote past, and those circumstances are certainly not up to us. To that line of thought, in the article below, by way of the supposition of a world of angels, he adds something new. This is a fundamental difficulty with the freedom that we cannot have if determinism is true. The difficulty, indeed a mystery, is one having to do with the opposite of determinism -- indeterminism. 

---------------------------------------------------------------

There are many kinds of freedom—or, as I prefer to say, the word ‘freedom’ has many senses. In one sense of the word, an agent is "free" to the extent that his actions are not subject to control by the state. It is, however, obvious that an agent may be free in this sense but unfree in other senses. However little the state may interfere with my actions, I may be unfree because I am paralyzed from the waist down or because I am subject to a neurotic fear of open spaces that makes it impossible for me to venture out of doors or because I am so poor that I am unable to afford the necessary means to what I want to do. These examples suggest that freedom is a merely negative concept—that freedom is freedom from constraint, that freedom consists in the mere absence of constraint. If freedom is in this sense a negative concept, this explains why there are many kinds of freedom: there are many kinds of freedom because there are many kinds of constraint. Because there are political constraints, there is political freedom, which exists in their absence; because there are internal psychological constraints (such as neurosis), there is psychological freedom, which exists in their absence; because there are economic constraints, there is economic freedom, which exists in their absence—and so on.

When we turn from politics and psychology and economics to metaphysics, however, we encounter discussions of freedom—discussions involving words like ‘freedom’, ‘free’, and ‘freely’—that it is hard to account for if freedom is no more than a negative concept. Consider, for example, the following words of Holbach:

Man’s life is a line that nature commands him to describe upon the surface of the earth, without his ever being able to swerve from it, even for an instant . . . Nevertheless, in spite of the shackles by which he is bound, it is pretended he is a free agent . . .

Or consider the ancient problem of future contingents, which would seem to depend on considerations different from those adduced by Holbach, for it has only to do with whether statements about future events must be either true or false, and has nothing to do with causation and physical law. Consider, again, the problem of divine knowledge of future human action. Consider, finally, the problem of evil and the attempts to solve that problem that appeal to the freedom of creatures and the alleged impossibility of a free creature that is certain to do no evil.

I think it is fairly evident that the concept of freedom that figures in the discussions raised by these metaphysical problems is the same concept. I think it is not easy to see how this concept could be understood as a merely negative concept, as a concept that applies to any agent just in the case that that agent’s acts are not subject to some sort of constraint.

Consider, for example, the problem of free will and determinism, the problem that is raised by the above quotation from Holbach. Although my present actions may be determined by the laws of nature and the state of the world before my birth (indeed, millions of years ago), it does not follow that this state of affairs places me under any sort of constraint. A constraint on one’s behavior is an impediment to the exercise of one’s will. If the state places me in chains, then my will to be elsewhere, if I attempt to exercise it, will soon come into conflict with the length and solidity of my chain. If I am an extreme agoraphobe, then my will to go about the ordinary business of life will come into conflict with sensations of panic and dislocation the moment I step out of doors. If I am very poor, my will to own a warm overcoat will come into conflict with my lack of the price of the coat. It is things of these sorts that are meant by ‘constraint’. And it is evident that determinism places me under no constraints. It is true that in a deterministic world, what my will is on a given occasion will be a consequence of the way the world was millions of years ago and the laws of nature. It is true that in a deterministic world, whether my will happens to encounter an obstacle on a given occasion will be a consequence of the way the world was millions of years ago and the laws of nature. But it is certainly not inevitable that my will encounter an obstacle on any given occasion in a deterministic world, and even in an indeterministic world, my will must encounter obstacles on many occasions. Indeed, there is no reason to suppose that my will will encounter obstacles more frequently in a deterministic world than in an indeterministic world. Anyone who believes that freedom is a negative concept will therefore conclude that the so-called problem of free will and determinism is founded on confusion. (So Hobbes, Hume, Mill, and many other philosophers have concluded.)

The situation is similar with the problem of divine knowledge of future human actions. We are often told that there really is no problem about this, since the fact that God knows that one is going to tell a lie (for example) in no way forces one to lie. Since God’s knowledge does interfere with the exercise of one’s will, since the false words that issue from one’s mouth are the words that it was one’s will to speak, God’s knowledge that one was going to lie is consistent with the lie’s being a free act.

All this can sound very sensible. And yet one is left with the feeling that the freedom this leaves us with is, in Kant’s words, a "wretched subterfuge." This feeling can be embodied in an argument. The argument is, to my mind, a rather powerful one. If the argument is correct, then freedom is not a merely negative concept. Or, at any rate, there is a concept of freedom that is not a merely negative concept, and this concept is a very important one. It is this concept, I believe, that figures in the metaphysical problems I have cited. I will call it metaphysical freedom. In calling it metaphysical freedom, however, I do not mean to imply that it is of interest only to the metaphysician. I believe that this concept is also of importance in everyday life, and that the concept that metaphysicians employ is just this everyday concept, or perhaps a refinement of it. (I should be willing to argue that all concepts that we employ in philosophy or science or any other area of inquiry are either everyday concepts or explicable in terms of everyday concepts.)

In ordinary English, the concept of metaphysical freedom finds its primary expression in simple, common words and phrases, and not in the grand, abstract terms of philosophical art that one is apt to associate with metaphysics. (The situation is similar in French, German, and Latin. I should be surprised to learn of a language in which the concept I am calling "metaphysical freedom" could not be expressed in simple, common words and phrases.) It is true that philosophical analysis is needed to distinguish those uses of these simple words and phrases on which they express this concept from other uses on which they express other concepts. Nevertheless, in particular, concrete contexts, these simple words express that very concept of freedom (not, as we shall see, a negative concept), that figures in metaphysical problems like the problem of freedom and determinism. But perhaps the meaning of these abstract remarks will not be clear without an example.

One of the simple words that expresses the concept of metaphysical freedom in English is ‘can’. What are we asking when we ask whether I am free to tell the truth to morrow if it has been determined by events in the remote past and the laws of nature that when, tomorrow, I confront a choice between lying and telling the truth, I shall lie? Only this: ‘I am free to tell the truth’ means ‘I can tell the truth’, and ‘I am not free to tell the truth’ means ‘I cannot tell the truth’. Metaphysical freedom, therefore, is simply what is expressed by ‘can’. If we accept this thesis, however, we must take care to understand it properly. We must take care to avoid two possible sources of confusion: the ambiguity of the word ‘can’ and false philosophical theories about what is expressed by certain sentences in which it occurs.

As to the first point, the word ‘can’ is extremely versatile, and can be used to express many ideas other than the idea of metaphysical freedom (a fact illustrated by this sentence). One example must suffice. In negative constructions, ‘can’ sometimes expresses an idea that might be called ‘moral impossibility’. One might say to a hard-hearted son, "You can’t refuse to take your own mother into your house"—even though one knows perfectly well that in the sense of ‘can’ we have been discussing he certainly can refuse to take his own mother into his house because he has already done so. We must take care that if we propose to use the simple word ‘can’ as our means to an understanding of metaphysical freedom, we do not allow our understanding of metaphysical freedom to be influenced by any of the many other concepts this simple word can be used to express. The best way to avoid such influence is not to rely on the word ‘can’ alone in our attempt to understand metaphysical freedom, but to examine also as many as possible of the other simple, ordinary words and phrases that can be used to express the concept of metaphysical freedom (or unfreedom). To illustrate what I mean, here are three sentences in which idioms of ordinary speech other that do not involve ‘can’ are used to express the concepts of metaphysical freedom and unfreedom:
 
 

—He will be able to be there in time for the meeting
 
 

—You must not blame her for missing the meeting; she had no choice about that
 
 

—It was simply not within my power to attend the meeting.
 
 

(Oddly enough, the phrase ‘of his own free will’ does not express the concept of metaphysical freedom, despite the fact that ‘free will’, as a philosophical term of art, means just exactly what I mean by ‘metaphysical freedom’. To say that someone attended a meeting of his own free will is simply to say that no one forced him to attend the meeting. The phase ‘of his own free will’ thus expresses a merely negative concept, the concept of the absence of coercion.)

False theories about the meanings of philosophically important words and phrases abound, and the philosophically important word ‘can’ is no exception to this generalization. There are those who, recognizing the importance of idioms like ‘I can do X’ for the metaphysical problems of freedom, have simply insisted that this word means something that supports their favorite philosophical theories. An example of such a theory would be: ‘I can do X’ means ‘There exists no impediment, obstacle, or barrier to my doing X; nothing prevents my doing X’. I will not argue specifically for the conclusion that that this theory is false; the argument I will later present for the incompatibility of metaphysical freedom and determinism, however, will have the consequence that this theory about the meaning of ‘I can’ is false—since, if the theory were true, metaphysical freedom would be compatible with determinism. At this point, I wish merely to call attention to the fact that there do exist tendentious theories about the meaning of ‘I can do X’.

If we consider carefully the meaning of ‘I can do X’ (‘I am able to do X’; ‘It is within my power to do X’) do we find that the idea expressed by this form or words is a merely negative one, the idea of the absence of some constraint or barrier or obstacle to action? It would seem not. It is true that the presence of an obstacle to the performance of an action can be sufficient for one’s being unable to perform that action. But it does not follow that the absence of all obstacles to the performance of an action is sufficient for one’s being able to perform that action. And the idea that ability could consist in the absence of obstacles does seem, on consideration, to be a very puzzling idea indeed. To see this, let us examine carefully the relation between the concept of ability and the concept of an obstacle. We should note that not just any obstacle to one’s performance of an action is such that its presence renders one unable to perform that action—for some obstacles can be surmounted or eliminated or bypassed (for short: some obstacles can be overcome). Let us ask a simple question: which obstacles to the performance of an action are such that their presence renders one able to perform that action? Why, just those obstacles that one is unable to overcome, of course. And it seems fairly obvious that the concept of an obstacle that one is unable to overcome cannot be analyzed or explained in terms of the concept of an obstacle simpliciter. (Is the concept of an obstacle that one cannot overcome the concept of an obstacle such that there is some "decisive" obstacle to one’s overcoming it?—No, not unless a "decisive" obstacle is understood as an obstacle that one is unable to overcome . . . .) These reflections suggest very strongly that the concept expressed by the words ‘I can do X’ or ‘I am able to do X’ cannot be a merely negative concept, the concept of the absence of some sort of obstacle or barrier or impediment to action. But let us turn now to the question of the compatibility of determinism and metaphysical freedom. I shall present an argument for the conclusion that determinism is incompatible with metaphysical freedom. Since, as we have seen, determinism and metaphysical freedom are compatible if metaphysical freedom (the concept expressed by ‘I can do X’) is a merely negative concept, this argument will be in effect an argument for the conclusion that metaphysical freedom is not a merely negative concept.

As Carl Ginet has said, our freedom can only be the freedom to add to the actual past—for the past is unalterable; it is what we find ourselves with in any situation in which we are contemplating some course of action. (Or to put this point in the terms I have been recommending, all we can do, all we are able to do, is add to the actual past.) And, unless we are bona fide miracle workers, we can make only such additions to the actual past as conform to the laws of nature. But the only additions to the actual past that conform to a deterministic set of laws are the additions that are actually made, the additions that collectively make up the actual present and the actual future. This is simply a statement of what is meant by determinism, which is the thesis that the law of nature and the past together determine a unique future. Therefore, if the laws of nature are deterministic, we are free to do only what we in fact do—that is, we are unable to act otherwise than we do and are ipso facto not free in the sense in which the term ‘free’ is properly used in metaphysics.

This little argument has great persuasive power, and it is probably no more than an articulation of the reasons that lead, almost without exception, the undergraduates to whom I lecture to join Kant in regarding the merely negative freedom of Hobbes and Hume as a wretched subterfuge. If the argument is correct, as I have said, it refutes the idea that metaphysical freedom is a merely negative concept, for the past and the laws of nature are not impediments to the exercise of one’s will. But, more generally, we may well ask what we are to say of this argument and its consequences, for these consequences go far beyond establishing that metaphysical freedom is not a negative concept. One possible reaction to the argument would be to say, with Holbach, that, because determinism is true, we therefore do not possess metaphysical freedom. (An epistemologically more modest reaction would be to say that, because we do not know whether determinism is true, we do not know whether we possess metaphysical freedom.) I shall return to the possibility that we lack freedom (or that we do not know whether we have freedom). For the moment, let us see where the argument leaves those of us who would like to say that we are free and that we know this. Many philosophers have regarded it as evident that we are free, and have accepted something like our argument for the incompatibility of determinism and metaphysical freedom. These philosophers, therefore, have denied that the world is deterministic, have denied that the laws of nature and the past together determine a unique future.

These philosophers (among whom I count myself) face a difficult problem. They assert or postulate that the laws of nature are indeterministic. One might ask how they know this, or what gives them the right to this postulate. These are good questions, but I will not consider them. I want to consider instead another question that these philosophers must answer: does postulating or asserting that the laws of nature are indeterministic provide any comfort to those who would like to believe in metaphysical freedom? If the laws are indeterministic, then more than one future is indeed consistent with those laws and the actual past and present—but how can anyone have any choice about which of these futures becomes actual? Isn’t it just a matter of chance which becomes actual? If God were to "return" an indeterministic world to precisely its state at some time in the past, and then let the world go forward again, things might indeed happen differently the "second" time. But then, if the world is indeterministic, isn’t it just a matter of chance how things did happen in the one, actual course of events? And if what we do is just a matter of chance—well, who would want to call that freedom?

It seems, therefore, that, in addition to our argument for the incompatibility of metaphysical freedom and determinism, we have an argument for the incompatibility of metaphysical freedom and indeteminism. But the world must be either deterministic or indeterministic. It follows that, unless one of the two arguments contains some logical error or proceeds from a false premise, metaphysical freedom must be a contradiction in terms, as much an impossibility as a round square or a liquid wine bottle. We may in fact define the problem of metaphysical freedom as the problem of discovering whether either of the two arguments is defective, and (if so) of locating the defect or defects.

The problem of metaphysical freedom, so conceived, is a very abstract problem. Although, for historical reasons, it is natural to think of the problem as essentially involving reference to the physical world and its supposedly intransigent laws ("man’s life is a line that nature commands him to describe on the surface of the earth . . ."), it does not. For suppose that man’s life is in fact not a line that nature commands him to describe on the surface of the earth. Suppose that nature presents us with two or seventeen or ten thousand lines inscribed on the surface of the earth, and says to us (in effect), "Choose whichever one of them you like." How could it be that we really had any choice about which "line" we followed, when any deliberations we might undertake would themselves have to be segments of the lines that nature has offered us? Imagine that two of the lines that nature offers me diverge at some point—that is, imagine that the lines present the aspect of a fork in a road or a river. The common part of the two lines, the segment that immediately precedes their divergence, represents the course of my deliberations; their divergence from a common origin represents diagrammatically the fact that either of two futures is a possible outcome of my deliberations. My deliberations, therefore, do not determine which future I shall choose. But then what does determine which future I shall choose? Only chance, it would seem, and if only chance determines which of two paths into the future I follow, then how can it be that I have a choice about which of them I follow?

The problem of metaphysical freedom is so abstract, so very nearly independent of the features of the world in which agents happen to find themselves, that it could—it would; it must—arise in essentially the same form in a world inhabited only by immaterial intelligences, a world whose only inhabitants were, let us say, angels.

Let us consider such a world. It is true that if there were only angels, there would be no physical laws—or at any rate there would be nothing for the laws to apply to, so we might as well say there would be none. But if we assume the angels make choices, we have to assume that time (somehow) exists in this non-physical world, and that the agents are in different "states" at different times. And what is responsible for the way an angel changes its states with the passage of time? One possibility is that it is something structurally analogous to the laws of physics—something that stands to angels as our laws of physics stand to electrons and quarks. (I’m assuming, by the way, that these angels are metaphysical simples, that they are not composed of smaller immaterial things. If they were, we could conduct the argument in terms of the smallest immaterial things, the "elementary particles" of this imaginary immaterial world.) This "something" takes the properties of the angels at any time (and the relations they bear to one another at that time: the analogue, whatever it may be, of spatial relations in a material world) as "input," and delivers as output a sheaf of possible futures and histories of the world. In other words, given the "state of the world" at any time, it tells you what temporal sequences of states could have preceded the world’s being in that state at that time, and it tells you what temporal sequences of states could follow the world’s being in that state at that time. Maybe it couldn’t be written as a set of differential equations (since nothing I have said implies that the properties of and relations among angels are quantifiable) as the laws of our physical world presumably can, but I don’t think that affects the point. And the point is: either "the sheaf of possible futures" relative to each moment has only one member or it has more than one. If it has only one, the world of angels is deterministic. And then where is their free will? (Their freedom is the freedom to add to the actual past. And they can only add to the actual past in accordance with the laws that govern the way angels change their properties and their relations to one another with time.) If it has more than one, then the fact that one possible future rather than another, equally possible, future becomes actual seems to be simply a matter of chance. And then where is their free will?

I said above that this way of looking at a postulated "world of angels" was one possibility. But are there really any others? We have to think of the angels as being temporal and as changing their properties with the passage of time if we are to think of them as making choices. And we have to think of them as bearing various relations to one another if we are to think of them as belonging to the same world. And we have to think of them as having natures if we are to think of them as being real things. Every real thing that is in time must have a nature that puts some kinds of constraints on how it can change its states with the passage of time. Or so, at any rate, it seems to me. But if we grant this much, it seems that, insofar as we can imagine a world of non-physical things (angels or any others) we must imagine the inhabitants of this world as being subject to something analogous to the laws of physics. If this "something" is deterministic, then (it seems) we can’t think of the inhabitants of our imaginary world as having free will. And if this "something" is indeterministic, then (it seems) we can’t think of the inhabitants of our imaginary world as having free will. Thus, the "problem of metaphysical freedom" is a problem so abstract and general that it arises in any imaginable world in which there are beings who make choices. The problem, in fact, arises in exactly the same way in relation to God. God, the theologians tell us, although He did in fact create a world, was free not to. (That is, He was able not to create a world.) But God has His own nature, which even He cannot violate and cannot change. (He cannot, for example, make Himself less than omnipotent; He cannot break a promise He has made; He cannot command immoral behavior.) And either this nature determines that He shall create a world or it does not. If he does, He was not free not to create. If it does not, then, it would seem, the fact that He did create a world was merely a matter of chance. For what, other than chance, could be responsible for the fact that He created a world? His choice or His will? But what determined that he should make that choice when the choice not to make a world was also consistent with His nature? What determined that His will should be set on making a world, when a will set on not making a world was also consistent with His nature? We should not be surprised that our dilemma concerning metaphysical freedom applies even to God, for the dilemma does not depend on the nature of the agent to whom the concept of metaphysical freedom is applied. The dilemma arises from the concept of metaphysical freedom itself, and its conclusion is that metaphysical freedom is a contradictory concept. And a contradictory concept can no more apply to God than it can apply to anything else.

The concept of metaphysical freedom seems, then, to be contradictory. One way to react to the seeming contradiction in this concept would be to conclude that it was real: metaphysical freedom seems contradictory because it is contradictory. (This was the conclusion reached by C. D. Broad.)

But none of us really believes this. A philosopher may argue that consciousness does not exist or that knowledge is impossible or that there is no right or wrong. But no one really believes that he himself is not conscious or that no one knows whether there is such a city as Warsaw; and only interested parties believe that there is nothing morally objectionable about child brothels or slavery or the employment of poison gas against civilians. And everyone really believes in metaphysical freedom, whether or not he would call it by that name. Dr Johnson famously said, "Sir, we know our will’s free, and there’s an end on’t." Perhaps he was wrong, but he was saying something we all believe. Whether or not we are all, as the existentialists said, condemned to freedom, we are certainly all condemned to believe in freedom—and, in fact, condemned to believe that we know that we are free. (I am not disputing the sincerity of those philosophers who, like Holbach, have denied in their writings the reality of metaphysical freedom. I am saying rather that their beliefs are contradictory. Perhaps, as they say, they believe that there is no freedom—but, being human beings, they also believe that there is. In my book on freedom, I compared them to the Japanese astronomer who was said to have believed, in the 1930s, that the sun was an astronomically distant ball of hot gas vastly larger than the earth, and also to have believed that the sun was the ancestress of the Japanese imperial dynasty.)

I would ask you to try a simple experiment. Consider some important choice that confronts you. You must, perhaps, decide whether to marry a certain person, or whether to undergo a dangerous but promising course of medical treatment, or whether to report to a superior a colleague you suspect of embezzling money. (Tailor the example to your own life.) Consider the two courses of action that confront you; since I don’t know what you have chosen, I’ll call them simply A and B. Do you really not believe that you are able to do A and able to do B? If you do not, then how can it be that you are trying to decide which of them to do? It seems clear to me that when I am trying to decide which of two things to do, I commit myself, by the very act of attempting to decide between the two, to the thesis that I am able to do each of them. If I am trying to decide whether to report my colleague, then, by the very act of trying to reach a decision about this matter, I commit myself both to the thesis that I am able to report him and to the thesis that I am able to refrain from reporting him: although I obviously cannot do both these things, I can (I believe) do either. In sum: whether we are free or not, we believe that we are—and I think we believe, too, that we know this. We believe that we know this even if, like Holbach, we also believe that we are not free, and, therefore, that we do not know that we are free.

But if we know that we are free—indeed, if we are free and do not know it—, there is some defect in one or both of our two arguments. Either there is something wrong with our argument for the conclusion that metaphysical freedom is incompatible with determinism or there is something wrong with our argument for the conclusion that metaphysical freedom is incompatible with indeterminism—or there is something wrong with both arguments. But which argument is wrong, and why? (Or are they both wrong?) I do not know. I think no one knows. That is why my title is, "The Mystery of Metaphysical Freedom." I believe I know, as surely as I know anything, that at least one of the two arguments contains a mistake. And yet, having thought very hard about the two arguments for almost thirty years, I confess myself unable to identify even a possible candidate for such a mistake. My opinion is that the first argument (the argument for the incompatibility of freedom and determinism) is essentially sound, and that there is, therefore, something wrong with the second argument (the argument for the incompatibility of freedom and indeterminism). But if you ask me what it is, I have to say that I am, as current American slang has it, absolutely clueless. Indeed the problem seems to me to be so evidently impossible of solution that I find very attractive a suggestion that has been made by Noam Chomsky (and which was developed by Colin McGinn in his recent book The Problems of Philosophy) that there is something about our biology, something about the ways of thinking that are "hardwired" into our brains, that renders it impossible for us human beings to dispel the mystery of metaphysical freedom. However this may be, I am certain that I cannot dispel the mystery, and I am certain that no one else has in fact done so.

-----------------------------------------------------

For another paper by Professor van Inwagen, go to Van Inwagen on Free Will.

Home: Determinism & Freedom Philosophy

Home: T.H. front page