Skip to content

Freedom of the Will

February 6, 2012

 Blogging, as anything else, is a matter of tradeoffs. I’ve been working on an argument which I think is novel and interesting but which requires a significant amount of stage-setting. Because this blog is an attempt to share ideas in bite-size pieces (with the aid of a vast web of knowledge to which I freely link) I’ve had to abandon the hope of providing this argument—yet. Instead, I’m going to carry the argument out over a number of posts which will link to one another so that you might better orient yourself. In what follows, I’m going to offer a corrective to some sloppy—and very common—thinking about the idea of Freedom of the Will. This is important, because as ethicists have long understood, responsibility depends on freedom. Here goes.

Freedom of the Will: A Primer

For a long, long time, the common positions one can take on the question of the freedom of the will (I’ll be abbreviating free will and freedom of the will as FoW from here on) have been understood. The atomists of almost 2,500 years ago certainly conceived of the problem as we do and it isn’t unlikely that their precursors were in the same boat. With Socrates and Plato, the difficulty of maintaining both the infinity of the mental and the finitude of objecthood was already coming into focus as the philosophical problem par excellence.

What the atomists developed was a way of conceiving the myriad happenings at the molar level of chairs and tables as a perspectival illusion which results from a truly massive number of super-tiny indivisible units interacting according to a few, simple regularities. This view is familiar for the same reason that the name is familiar—scientists in recent history revived the term as they developed a similar model of causal reduction.

It is this recognition of the possibility of causal reduction, despite appearances, which sets up the FoW problem as we know it today. The argument hasn’t changed all that much.

Here it is:

  1. Assumption: I have FoW
  2. Assumption: Each and every molar event is really a collection of microscopic, atomic events which are governed by simple rules
  3. Assumption: All of my actions—moving my vocal cords around, tensing a muscle—are molar events
  4. By 2&3, each and every thing that I do is really a collection of microscopic, atomic events governed by simple rules
  5. Assumption: FoW consists in my being able to determine and govern my actions
  6. By 4&5, I do not have FoW
  7. By 1&6, I both have and do not have FoW. CONTRADICTION!

This is a standard reductio ad absurdam and it should be dealt with as such. If you are a FoW

Free Will

The options furnished by our reductio.

proponent, you can deny 2, 3, or 5 (or some combination, but what we’re interested in is what kind of content is possible given assumption 1 and by keeping the claim-pool as large as we can, we maximize information about the space of possibilities in which FoW is possible).

  • Indeterminism: Denying 2 can be done by abandoning the view that atoms are governed by simple rules (this has been tempting to philosophers and laypersons alike who think quantum mechanics will provide some means of escape from the problem). What you’re left with is an indeterminist position in which one can hold that the causal claim fails universally (the aforementioned proponents of the QM solution do this) or that, somehow, cause and effect steer clear of minded beings like us.
  • Libertarianism: Denying 3 is simply denying that my actions are truly molar events. More precisely, the position holds that my actions are something altogether different. By the beginning of the renaissance positions that denied 3 were becoming rather ugly: the self and the will, no longer claiming sovereignty over the whole domain of the body (because of the growth of anatomy, the bizarre physicalism of Hobbes, and the soulful Cartesian philosophy), curled inward and eventually was relegated to the job of res cogitans; literally, ‘thinking thing’. In the wake of Newton’s Principia and the growth of a mechanics which was not only expressively more powerful than Aristotle’s physics but which, in addition, multiplied the effectiveness of society as a whole by expanding its predictive powers, this position became unpalatable to most thinkers. This is still the case, though I suspect that those who worry about qualia are, at least, flirting with this impotent school of thought. Despite its lack of popularity in the academy, it might be the most common folk view on the matter.
  • Compatibilism: Denying 5 is tricky, but its rewards are great: completed successfully, our theoretical toolbox would include both the possibility of complete (insofar as physically possible) predictive power and an exalted, morally responsible view of the self. To be successful in this task, it isn’t so much that one must deny that FoW consists in the governance of one’s own actions (such a solution is possible only at the cost of all intuitive appeal) as that this governance must somehow be shown to be conceptually separable from the kind of governance that causal rules have over matter. Medieval philosophers, discussing this means of resolving a reductio said something to the effect of, “where there is a contradiction, introduce a distinction.”

Things Aren’t So Simple

Earlier, I said that denial of 2—that events are guided by regularities—was a means of keeping our assumption of FoW afloat. From the perspective of the reductio that we were looking at this is true. But reflection on the matter complicates things: there may be independent reasons to doubt that the lack of determinism in a physical system will yield FoW. If the deterministic control of an atomic physics can’t save us from the problem, neither can its opposite view: that the units which make us up are jumping around probabilistically or in an indeterminate manner. If being guided along a straight path by the arrow of time isn’t freedom, what more freedom is there in each moment’s event being the outcome of random, microscopic perturbations? The steady hand of a deterministic atomic theory is no more—or less—harmful than the palsied jitters of an indeterminate microstructure.

Something very peculiar is going on here. With the first argument, it was taken for granted that the determinacy of a system somehow precluded that system’s being free—that’s the motivation behind the sixth step in the argument—but later we saw that determinacy is no worse than its opposite as far as FoW goes. When one property, which is generally supposed to rely on another in some fashion (as the lack of FoW was thought to be a straightforward outcome of atomic determinacy), is invariant with respect to another, as FoW seems to be to the property of determinacy or lack thereof at the atomic level, you’re prima facie warranted in exploring the possibility that the distinctions that these properties cleave in the world-body are orthogonal.

Supposing that these distinctions carve at different joints, the next step should be to specify what FoW is if not the ability to guide one’s own actions in the sense that the laws of physics guide (better: articulate) the goings on at the atomic level. Now, there’s no such thing as ‘proper philosophical method’ (that’s an oversimplification—there’s no universally agreed upon method and, in fact, the search for method is a part of the philosophical enterprise), but as far as ‘best practices’ go, you can’t do much better than looking to actual cases in which a term is applied and attempting to say what it is that people are doing in so applying it. I want to be clear: philosophy is not a mere lexicographic or anthropological recording of regularities and it’s always possible that everyone is wrong about some question, but figuring out what people have meant by a term is one of the most powerful tools of the philosopher. For this method of analysis, the answer we seek is something that we have always already had.

In the parenthetical of the first sentence in the last paragraph, there’s just the kind of clue which will prove valuable for our discussion. What is the difference between the rules that I follow as a chess player—written down on a sheet of rules—and the rules that a collection of matter follows—written down in a physics text? What I’d like to suggest is that atoms simply act as they act and that we, at best, accurately articulate their doings while what humans freely do depends on some conception of appropriateness of doing. We can give reasons for an atom’s having done what it did, but there’s a richer sense of the term ‘reason’ which applies to the type of thing by which we can choose to do something.

My conjecture—and the moral of the story today—is that freedom of the will is a matter of choosing to act on the basis of reasons and that this is a feature that doesn’t depend on the features of the things of which we are constructed any more than the game of chess depends on the means by which pieces are moved around on the board. There is not some new, exciting scientific discovery which will show freedom to be a sham because freedom is a way of characterizing the relation between an actor and the reasons which that actor holds before her mind’s eye. Freedom is a notion that is internally related to the idea of the self. So, while one can abandon free will, one can’t hold that there exist selves that somehow lack it—freedom is a transcendental condition for the very possibility of having a self.

In a future post, I’m going to extend this argument and attempt to show you how freedom is something that can be created within a society and is not simply the type of thing that one always already has. The controversial thesis that will be approached at a later date is that metaphysical FoW is a regulative standard which should guide (and be tempered by, through dialog with) the political ideal of freedom. In preparation for that argument, feel free to skim my first post so that your mind isn’t polluted with the notion that there is some ahistorical perfected ideal of freedom which can be grasped without considering the institutions and networks of interaction in which people actually live and can be said to be free.


From → Philosophy, Value

  1. As a minor prelude to my question, I want to disagree with the implication that QM’s reliance on randomness means that it is not a causal theory. Though QM is not deterministic, in the sense that it is not possible to predict with certitude the future states of a system given a complete description of the present state, it is certainly causal, merely by virtue of its identity as a physical theory. What is physics, if not a theory which seeks to provide a universal accounting for the causes of events? (You may enjoy this paper concerning the interpretation of Bell’s theorem: However, in my view, the threat physics poses to FoW arises not from determinism per se, but from the (apparent) incompatibility of the notion of a “chooser” with that of a causal theory of physics. We can at least imagine accepting a theory in which future states of mind are dependent only and entirely on past states, without demanding of that theory that it predict precisely what those future states will be.

    This brings me to the bigger question I have for you. It seems to me that the crux of your argument – that “freedom of the will is a matter of choosing to act on the basis of reasons” – begs the question of the existence of FoW. You say that FoW is a feature that does not depend on the characteristics of the things of which we are constructed, but whether or not this is possible seems to depend on the definition of “choice.” Can we define “choice” (and by proxy FoW) as a sort of illusion, a feature that arises in any situation where a “chooser” appears to “choose” from multiple “alternatives”? This definition has the benefit of resolving the apparent conflict between FoW and rigid causality by casting FoW in terms of our descriptions and understandings of processes in the world, rather than in terms of the physical properties of their participants. But it leads to problems. By this definition, a chess computer exercises a choice when it makes its next move – the computer program is the “chooser”, the move is the “choice”, and the various moves the computer might have made, but didn’t, are the multiple “alternatives”.

    We probably cannot be satisfied with any definition of choice that results in our having to attribute FoW to a chess computer. But why? I think it has something to do with our intimate knowledge of the inner workings of a chess computer, knowledge that tells that there are three things, and only three things, that determine the program’s next move: 1. the programming of the computer, 2. the current and prior states of the board, and possibly, 3. a random number drawn from some arbitrary distribution. (If we substitute “laws” and “universe” for “programming” and “computer”, this also happens to be the nature of causality in QM.) A grandmaster from the 18th century, seeing that the computer appears to “choose”, far better than any human, between multiple “alternatives”, might well conclude that the program is exercising FoW during play – but we know that the computer’s play, though brilliant, is determined entirely by its programming, the state of play, and a roll of the dice, and this doesn’t leave much room for choice.

    So a naive observer might attribute FoW to a computer, but an observer with full knowledge of the computer’s functions would probably not. What does this mean? In my opinion, there are two major implications:

    1. that the question of the existence of FoW is not truly independent from the nature of objects in the world. If it were, then we should be satisfied that any configuration of objects and processes which appears to meet our definition of a “chooser” is exercising FoW, without having to care at all about the underlying physical realities of those objects and processes. However, when we learn that a chess computer operates in a strictly causal fashion, we are no longer satisfied that the computer has FoW, which shows us that the underlying physical realities do in fact enter into the equation. In fact, it shows us that if we can provide a rigorous causal account of a set of objects and processes, then that set is not a “chooser”.

    2. if you accept the first implication, then in order for us to be satisfied that something is a “chooser”, at least one of the following two conditions must obtain. Either
    a. we must be ignorant of the internal mechanisms of the “chooser”, or
    b. the “chooser” must be, or contain something that is, independent of causality.

    We would like to believe that the existence of FoW is not merely a consequence of our ignorance as to the causes of our action, so 2.a does not give us much confidence. But 2.b is even worse, because it is incompatible with a worldview where QM (or any other theory of physics) is true.

    So where does this leave our account of FoW? For one thing, there may well be ways out. For example, perhaps 2.a is not really so bad; perhaps, say, the problem of the subject specifically prevents us from gaining the same kind of understanding of our own workings as we have for the chess program. We may accept such an escape route so long as there exist no thinking machines aside from the ones in our skulls, but I think those days are numbered. Ultimately, I think we will have to accept that FoW arises from a kind of hazy thinking. It’s not hard to see why we wish to suspend our disbelief – hard determinism seems to be an enervating and profoundly demotivating worldview. Nonetheless, relying on the current ineffability of consciousness to support the existence of FoW amounts to the hope that a chain of causal processes, if only it is ever so long and complex, can have a non-causal outcome.

    • In philosophy, simple examples are highly important. They allow our little mammal heads to wrap themselves around some big ideas. There is, however, a danger in their application. The simplicity of a thought experiment might obscure or hide relevant details that are important for the connection to that larger set of ideas which are thought to be exemplified. As pragmatists have sometimes put it, when looking to the simplicity of an example one must always be careful that the difference doesn’t make a difference–that is, that the difference which constitutes the example as analog is different only in irrelevant features.
      The long and the short of it is that I think you’ve generalized too broadly from the simple case of the chess computer and that there are other possible ways that the distinction between freely willing, rule following beings and unfree rule-governed beings can be maintained. Our brains exist on a continuum which stretches from rocks and insects at one extreme to us and, ultimately even beyond, to beings far more capable than us but it need not follow that certain important features of dispositionally responsive compute-ers (I’ve separated the word out to emphasize the fact that this continuum runs along something like the capacity to deal with information in a given amount of time) come in degrees–or that those features which do come in degrees do not exist on continuums which are orthogonal to the computational continuum.

      I’ll get back to this issue in a second, as that part’s the meat of my response, but I first want to point out that my move to the denial of the fifth assumption–or rather, to its reinterpretation–is by way of suggesting that a non-deterministic microstructure is no better for freedom than a deterministic one. Whether we want to call a non-deterministic system causal or not isn’t a race in which I have a horse, so I’d be completely happy with the alternate means of constructing the second movement of this argument that you’ve suggested.

      Back to the interesting stuff. So, as I said, computational continuum, orthogonality, etc.
      The problem I have with your argument is that you begin by assuming that it is merely our ability to represent alternative moves among which–by way of some algorithm–a system makes a selection which warrants our saying that the system truly makes a choice. Let me be clear: it isn’t enough that we can predict a system’s behavior with the use of the language of intentionality (“it thinks, it wants, it believes”) for us to say that a system is intentional. If it were, then we need to give some serious thought to how we deal with our fellow thinkers–the thermostats–who live in a world of such great privation and servitude. Plainly, such a result would be disastrous if not practically, then at least ethically and theoretically.
      You’re in the neighborhood of truth when you say, “We probably cannot be satisfied with any definition of choice that results in our having to attribute FoW to a chess computer.” I’m afraid that you steer yourself into a different part of town altogether when you attempt to answer the question of why it should be so. I think it is true that the Grandmaster from a century before universal computation would conclude that the computer was exercising FoW, but I also think that he would think that there is a mind behind the screen exercising thought about position, strategy, etc. There’s no doubt that our knowledge of the construction of such a computer helps us to avoid mistakenly saying that it has FoW, but the increased material (rather than functional) understanding that we have can function as a negative condition on our judgment that what we’re dealing with is the functional equivalent of a thinking human (though with a thinner region in which to think). The dispositional responsiveness of a chess playing computer–the best chess playing computer–is barely even a parlor trick compared to the sophisticated computational powers of the dumbest rat.

      What separates us from chess-playing computers isn’t that we choose. Rats, computers, thermostats, etc. all make choices. What makes us different is that we represent ourselves as making choices. We have a substantive notion of the unity of a self in space and through time which allows us to construct beliefs and desires which are far more than merely dispositional responses to our environs. This is why I say that FoW is a central, transcendental feature of selfhood and that it consists not in being rule-governed but in holding rules before oneself and acting in accord with them.
      The point that you make about physical realities needing to be a part of the story is certainly one that i accept: each and every tokening of a belief or desire state will be instantiated in some token of a physical state, but there’s simply no reason to believe that there is any physical type which corresponds to a given mental type. Any token of a book is some token of matter, but the type of that book does not correspond to any type of matter–books can be on ereaders and on paper.

      The rest of the argument doesn’t go through assuming that the first part is stopped in its tracks. I planned on saying something about the Chinese room and about how Turing’s paper, which Searle’s paper is a response to, can be seen as an argument for the very method–broadly hermeneutic–which I’m suggesting lies at the heart of all judgments of mindedness or lack thereof. Obviously I didn’t do this. Partially because I don’t think I need to, partially because it will be fodder for a future post, and partially because my girlfriend just got online and I’d rather talk to her. (See, everything is a matter of tradeoffs. The post has come full circle, do I win a prize or something?)

      Final, related thought: We don’t just act according to that which happens outside of us and stimulates us, we also look back at our history to develop a sense of what our acts have been and what they mean so that we can act appropriately in the future. We don’t simply act according to rules in our brains, we construct ways to cope with those rules–thereby moving beyond them, ‘sublating’ them in Hegelese.

  2. The Unwelcome Pundit permalink

    You know that I stand right where you do on this topic. Yet I do have to make a slight quibble. The initial argument that you outline in the beginning doesn’t actually appear to be a reductio. If it were, then one of the lines between the first assumption and the contradiction would use the assumption as a premise. But as you can see, none of them actually do. Therefore, it’s just a conditional proof with a set of assumptions that you wish to challenge. This of course has no bearing on the content of your overall argument.

  3. Which assumption isn’t used as a premise? My understanding of a reductio is just that a given set of assumptions gives rise to a contradiction and, as a result, you’ve got to pick at least one of them to toss out. Is there a stricter notion that I fail to recognize?

  4. The Unwelcome Pundit permalink

    Line 1 is not used anytime before line 7. In a strict reductio, you would actually have to use line 1 in the course of your reasoning between 1 & 7. For instance, would you call this a reductio:

    1. Not B
    2. If A, then B.
    3. A.
    4. B (from 2 & 3).
    5. B and not B.

    and then try to argue against the truth of A? It seems that the cleaner version would just be to say that 2-4 is the argument you want to attack, and then to attack it by undermining the credibility of 3.
    Contrast that argument with this:

    1. The square root of 2 is rational.
    2. sqrt(2) can be expressed as a simplified fraction sqrt(2)=a/b. (by def. of rational)
    3. Hence, 2 = (a/b)^2.
    4. Rearranging, 2 x b^2 = a^2.
    5. Therefore, a^2 is even.
    6. That means a is even, since its factorization will contain a 2.
    7. Since the ratio is simplified, b must be odd.
    8. However, since a is even, a^2 will be a multiple of four.
    9. Recalling b^2 x 2 = a^2, that means b^2 must contain a factor of 2 also, hence being even as well.
    10. If b^2 is even, so is b.
    11. Contradiction between 7 and 10.

    Notice that here we actually used the original false premise (1) in the rest of syllogism to develop the contradiction. Had we proved directly that sqrt(2) was irrational, the statement of (1) would have been superfluous. I think that’s the difference.

  5. I will admit that expressing the problem as a reductio is unclean, but I think it better brings out a tension in the actual positions that people are holding and gives a good ‘lay of the land’ for determining where to go next. There seem to be at least a few sources who treat the weaker form of reductio as a first class member of the class, but I’m willing to accept that on some formulations of the reductio, what I’ve offered is either not a reductio or only trivially one.

    Here’s a link to a site that makes the same error (assumption 3):

    My intent is just to show that you can’t hold all of the assumptions to be true at the same time without falling into contradiction and that, therefore, there’s a decision to be made.

  6. The Unwelcome Pundit permalink

    That’s true. Like I said, it was only a quibble.

Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out /  Change )

Google+ photo

You are commenting using your Google+ account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )


Connecting to %s

%d bloggers like this: