Review: John Searle's Freedom and Neurobiology.
John Searle, Freedom and Neurobiology: Reflections on Free Will, Language, and Political Power, New York: Columbia University Press, 2006
By Christopher Franklin
In this short book John R. Searle proves again why he is one of the most innovative and thoughtful philosophers of our time. Known best for his work on consciousness, Searle has recently turned his philosophical gaze towards agency (see also his book Rationality in Action). The book is composed of seemingly two unrelated chapters that both grew out lectures delivered in Europe. However, both chapters are attempts to move us closer to a solution to what Searle calls the single overriding question in contemporary philosophy: How do humans fit in? How do we square the fact that the world is wholly composed of “mindless, meaningless, unfree, nonrational, brute physical particles” with our ordinary conception of ourselves as “conscious, intentionalistic, rational, social, institutional, speech-act performing, ethical and free will possessing agents”? (p. 5)
Chapter 1 concerns the problem of free will and its attendant problems of rationality, consciousness, and the self. Chapter 2 focuses on social ontology and political power. Given the venue of this post, I found it appropriate to focus my discussion on Chapter 1.
Before diving into the free will problem Searle lays some ground work. It might seem anachronistic to say that the question of how we fit it is the most pressing question in contemporary philosophy. After all, haven’t philosophers been concerned all along with precisely this question? So what is so special about the current day? He answers this question as follows: it has only recently become possible to treat all these topics (e.g. consciousness, rationality, language, intentionality, free will, etc.) naturalistically; we can provide answers to these questions that are consistent with and a natural outgrowth of the basic facts. He takes the basic facts of the world to be described in the atomic theory of matter and evolutionary biology. Searle takes the world, as described by these theories, as a starting point.
Moreover, there have been recent developments in philosophy that allow us to tackle this question in a way that was previously impossible. All of the following have contributed to our privileged position: The move away from epistemology and its skeptical arguments about knowledge of the material world, the move from philosophy of language back to philosophy of mind, the destruction of a sharp line between philosophy and empirical disciplines, and also the great benefit we have from our philosophical predecessors, who over the last sixty years, have focused on answering questions in piecemeal fashion, providing distinctions and clarifications so that we can state theories in precise terms which admit of empirical verification.
Chapter 1 is entitled “Free Will as a Problem in Neurobiology”. The goal of this chapter is to take the first steps towards showing how the free will problem could achieve a scientific resolution. The task of the philosopher, as Searle sees it, is to provide a precise formulation of the problem of free will so that it admits of empirical and scientific testing. He thinks he has done this with the mind-body problem and he also wants to do this with the problem of free will. He wants to transform the traditional free will problem into a problem essentially about how the brain works. Let us now consider his attempt to move us towards this goal.
The logical form of the free will problem is that we appear equally committed to two positions which are inconsistent. We think, on the one hand, that natural phenomena occurs in such a way as to admit of deterministic explanations (contrastive explanations). When we explain why an earthquake occurred we do not merely explain why it just so happened to occur, we explain why, given the antecedent conditions, it must have occurred. There are sufficient causal conditions for the earthquake. However, there appears to be a small subset of human behavior that does not admit of deterministic explanations. We explain “free action” by citing the reasons we acted on. However, we ordinarily do not take ourselves to be giving deterministic explanations. Given the antecedent reasons, beliefs, and desires, we believe that we could have done otherwise. There are not sufficient causal conditions for free action.
The problem of free will arises from a feature of our experiences that Searle labels “volitional consciousness.” “The feature is that I do not sense the antecedent causes of my action in the form of reasons, such as beliefs and desires, as setting causally sufficient conditions for action….” (p. 41) This is strikingly different from perceptual consciousness; we do not experience our perceptions as being up to us in the way we experience some of our behavior as being up to us. Searle refers to this as the experience of the gap. This gap only occurs in a few places: (i) between reasons and the making of a decision, (ii) between decision and the onset of action, and (iii) for extended action, between the onset of action and its continuation. But on the other hand determinism appears just as convincing. We typically only accept explanations which cite sufficient causal conditions as true or adequate explanations for natural phenomena. Hence the problem of free will is the problem of a certain sort of consciousness, volitional consciousness.
Not only do we arrive at the postulation of a gap via experience, but this gap is also reflected in the logical structure of reason explanations. Such explanations are not deterministic in form, but instead specify a reason that an agent acted on. Ordinary causal explanations have the logical form: A caused B. However, rational explanations have the following logical form: “A self S performed action A, and in the performance of A, S acted on reason R” (p. 53). But a problem arises at this point: How in the world can such explanations be adequate since they do not cite casually sufficient conditions? They do not explain why, given A, B must have occurred and hence they do not furnish us with contrastive explanations. So what must be the case in order for rational explanations to be adequate? Searle believes that such explanations cannot be adequate unless we posit an irreducible, non-Humean self. He provides the following transcendental argument for the existence of such a self (pp. 53-55):
1) Explanations in terms of reasons do not typically cite causally sufficient conditions (assumption)
2) Such explanations can be adequate explanations of actions (assumption)
3) Adequate causal explanations cite conditions that, relative to the context, are causally sufficient (assumption)
4) Construed as ordinary causal explanations, reason explanations are inadequate (conclusion)
5) Reason explanations are not ordinary causal explanations (conclusion)
6) Reason explanations are adequate because they explain why a self acted in a certain way. They explain why a rational self acting in the gap acted in one way rather than another, by specifying the reason that the self acted on (conclusion)
Again, it is important to note that this is not a deductive argument but a transcendental argument; one that attempts to state what necessary conditions must obtain in order for there to be adequate reason explanations. Hence, the intelligibility of reason explanations requires the postulation of an irreducible, non-Humean self. Taking stock: the free will problem leads us to two other problems: consciousness and the self. Consequently, a solution to the free will problem requires a solution to these problems.
So what must the brain be like in order to realize free will? Searle’s picture of the brain is that at the micro level there are neurons, neurotransmitters and synapses. Consciousness is a higher-level or systemic feature entirely composed of the micro level elements. At the higher level we have decisions leading to intentions and at the micro level we have neuron firings causing more neuron firings. Moreover, consciousness at each instant t is completely determined by what is going on at the micro level at t. So the free will problem now becomes this: “If we suppose there is a gap at the top level in the case of rational decision making, how might that gap be reflected at the neurobiological level?” (59). We know there are no gaps in the brain: So what must the brain be like at the micro level in order to realize the gap at the higher-level of consciousness. Searle thinks that if the total state of the brain at any instant t1 is causally sufficient for the state of the brain at t2, then we do not have free will. Hence, there are two hypotheses: Hypothesis 1: for every brain state S1 at any instant t1, S1 is causally sufficient for S2 at t2. Hence, free will is an illusion. The experience of the gap is realized in a brain that is wholly deterministic. Hypothesis 2: the experience of the absence of causally sufficient conditions at the psychological level is matched by the absence of causally sufficient conditions at the neurobiological level. Hypothesis 2 envisages the brain as being such that the micro elements are sufficient to determine the conscious state at any given moment, but that the operation of the conscious self is what explains the movement from one conscious state to the next.
However, neither Hypothesis is very satisfying. Hypothesis 1, although consistent with what we currently know from brain science, leaves us with epiphenomenalism and this goes against everything we know about evolutionary biology. The problem with Hypothesis 2 is its extreme demandingness. It requires the brain to have the following three features: (1) consciousness functions causally in moving bodies, (2) the brain causes and sustains the existence of a conscious self that is able to make rational decisions and to carry them out in actions, and (3) the brain operates in such a way that the “conscious self is able to make and carry out decisions in the gap, where neither decision nor action is determined in advanced, yet both are rationally explained by reasons the agent is acting on” (p. 73). Searle thinks that (3) is the trickiest condition. Since quantum mechanics is the only indisputable form of indeterminism in the world, in order for Hypothesis 2 to be correct, consciousness must manifest quantum indeterminism. But this does not necessarily mean it is random. The whole does not necessarily have the properties of the parts. So randomness might be a property of the parts without being a property of the whole.
I would now like to turn to questions and comments about Searle’s intriguing project. This first point is somewhat nitpicky, but I think it is worth mentioning for clarity sake. I find Searle’s equating sufficient causal conditions with deterministic causal conditions objectionable. If A nondeterministically causes B, then wouldn’t A be a sufficient cause for B? Searle appears to simply move from sufficient causal conditions to deterministic causal conditions. I find this usage unfortunate and at times misleading. It would seem better to state the worry in terms of deterministic causal conditions and the experience of the gap as an experience of a lack of deterministic causal conditions.
Moving on to more substantive questions/comments, I worry that Searle’s program provides fertile ground for revisionism or skepticism about moral responsibility. Searle makes it clear that he does not know whether or not we actually possess free will. In discussing Hypothesis 2 he writes, “Notice that I do not ask, ‘How does the brain work so as to satisfy these conditions?’ because we do not know for a fact that it does satisfy these conditions…” (p. 71). Although Searle explicitly maintains that he is not concerned with moral responsibility, his view of our current epistemological status with respect to the existence of free will raises questions about whether or not we are justified in engaging in our ordinary practices of praise and blame (which might be thought to presuppose the existence of free will). Depending on one’s temperament, one might find an argument for skepticism concerning moral responsibility or an argument for revisionism about what justifies our ordinary practices lurking in this comment.
Lastly, although I applaud the spirit of Searle’s project, I worry to what extent the problem of free will is reducible to a problem in neurobiology. After all, “the free will problem” hardly denotes a single problem. So if even we were able to substantiate Hypothesis 1 or 2, would it really be correct to say that we have arrived at a solution the free will problem? I think this question deserves a negative answer. After all, one of the most pressing questions that is often denoted by the phrase, “the free will problem”, is whether or not free will is compatible with determinism. Searle himself simply asserts that he is using “free will” in such a way that it is incompatible with determinism. However, were we to discover that Hypothesis 1 is correct, could we, as fair minded philosophers, really announce that free will is illusion without first dealing with the question of whether or not it is actually compatible with determinism? And it is far from clear to me that this latter question is reducible to any scientific discipline.
It seems to me that the virtue of Searle’s program is an attempt to test certain hypotheses put forward as an answer to one part the problem of free will opposed to an attempt to reduce the whole problem to a neurobiological problem. Hence, it seems somewhat exaggerated to say that problem of free will just is whether the brain states at t1 are sufficient to cause the brain states at t2. Instead, such testing will help us in discerning the viability of certain libertarian programs. So again, I fine Searle’s approach quite advantageous when it comes to specific hypothesis offered as a solution to one aspect of the free will problem, but I am dubious about whether the whole problem can be reduced to science.
However, despite some of my disagreements with Searle, I am a huge admirer of this excellent piece of philosophical work. I am confident that it will generate interesting research programs and many responses from various fields of inquiry.
Christopher Franklin
Chris -- this is an excellent and very helpful review. Thanks for putting in the time to do it. A couple of quick points:
First, since I haven't read the book, a question of clarification. I take it that Searle thinks that neuroscience will help us figure out which of Hypotheses 1 and 2 is true. But just how is neuroscience supposed to be able to figure out whether "for every brain state S1 at any instant t1, S1 is causally sufficient for S2 at t2"? I guess I'm skeptical that any empirical inquiry can come to any definite conclusions about claims concerning causal sufficiency. Or, rather, if it does come to definite conclusions about that, it does so only by carrying some philosophical baggage along with it.
Second, I think on one plausible way of reading 'sufficient cause', it really is just the same as a deterministic cause. Suppose that C nondeterministically causes E. This means that there was an objective chance that C would not cause E. But if it's possible for C to occur without E occurring, then in what sense is C *sufficient* for E?
Posted by: Neal | February 22, 2007 at 11:12 PM
Neal,
I share your skepticism about the resources available to neuroscience to be able establish that S1 *deterministically* causes S2. But I am less skeptical about empirical disciplines being able to identify sufficient causes. But this of course takes me to your second point. The reason I do not like regimenting "C is a sufficient cause of E" as "C deterministically caused E" is that this appears to impliy that "C nondeterministically caused E" implies "C was not a sufficient cause of E." But, granting that nondeterministic causation is possible, this strikes me as being very odd. Do you want to say that there is no sufficient cause for E? It seems to me that all we need mean by sufficient is that the causal force that C exerted did not need to be supplimented by any other events occuring simulanteously with C.
All I really meant to point out is that it can be misleading to say there is no sufficient cause for E, at least to my hears that sounds a lot like saying there was no cause of E. It seems like if we admit the possibility of nondeterministic causation we ought to employ the notion sufficieny to indicate the minimum causal force required to bring about E.
Isn't this what we are doing when we identify certain laws of nature? However, I wonder if Searle can just say that he is attempting to formulate the free will problem
Posted by: Chris | February 24, 2007 at 05:38 PM
Good job, Chis!
Consider:
1) Explanations in terms of reasons do not typically cite causally sufficient conditions (assumption)
2) Such explanations can be adequate explanations of actions (assumption
3) Adequate causal explanations cite conditions that, relative to the context, are causally sufficient (assumption)
4) Construed as ordinary causal explanations, reason explanations are inadequate (conclusion)
5) Reason explanations are not ordinary causal explanations (conclusion 6) Reason explanations are adequate because they explain why a self acted in a certain way. They explain why a rational self acting in the gap acted in one way rather than another, by specifying the reason that the self acted on (conclusion)
I have various questions. First, consider the first premise. It is ambiguous between
1a) Explanations in terms of reasons do not typically cite conditions that are in fact causally sufficient.(assumption) and
1b) Explanations in terms of reasons do not typically cite causally sufficient conditions as such, i.e., as causally sufficient.(assumption)
I think that Searle is only entitled to 1b, and not 1a--but isn't 1a what is required for the argument?
Also, isn't 1b consistent with the idea that our explanations are often suited to our purposes, and our ordinary short-cuts can be filled in, if we are put under pressure (or in other contexts). When we say in a certain context that the heart pumps blood, this doesn't rule out a more complete and detailed biological (or even physical) explanation, which could be given in another context. What we "typically cite" isn't in any way decisive here.
Finally, it seems to me--and this raises some larger issues--that our reasons-explanations could in principle be irreducibly indeterministic without this imply anything about the metaphysics. That is, I would want to separate points about explanation from points about how the world works, as it were. I have been puzzled over the years by work in Action Theory that goes straighforwardly from points about explanation and description to points about the metaphysics (as it were). But haven't we learned to separate conceptual issues (roughly speaking) from metaphysics?
Posted by: John Fischer | February 25, 2007 at 10:39 AM
John,
I agree that there is an ambiguity in 1 between 1(a) and 1(b), but it is unclear to me why he is not entitled to the former. It seems like he could reply that he is simply assuming 1(a) in order to see where it takes us. He could argue that 1(b) is true and then claim he is simply curious what follows if our ordinary explanation practices are accurate, that is they really depict how the world is. Hence, I think Searle would, or at least could, grant you that he has not provided strong justification for 1(a), but that he does not need to justify it for his purposes. Would you agree?
I definitely agree with that you should not go straight from explanations to metaphysics. Randy Clarke has a great discussion of this in his book in regards to contrastive explanations and the luck objection. I think he is spot on. However, I think we can say that some of our explanations are inadequate on the basis of our metaphysics and this is all Searle needs. The transcendental argument only attempts to show that *if*
ordinary reasons explanations are adequate then our metaphysics must turn out a certain way.
Posted by: Chris | February 28, 2007 at 09:39 AM
John,
I agree that there is an ambiguity in 1 between 1(a) and 1(b), but it is unclear to me why he is not entitled to the former. It seems like he could reply that he is simply assuming 1(a) in order to see where it takes us. He could argue that 1(b) is true and then claim he is simply curious what follows if our ordinary explanation practices are accurate, that is they really depict how the world is. Hence, I think Searle would, or at least could, grant you that he has not provided strong justification for 1(a), but that he does not need to justify it for his purposes. Would you agree?
I definitely agree with that you should not go straight from explanations to metaphysics. Randy Clarke has a great discussion of this in his book in regards to contrastive explanations and the luck objection. I think he is spot on. However, I think we can say that some of our explanations are inadequate on the basis of our metaphysics and this is all Searle needs. The transcendental argument only attempts to show that *if*
ordinary reasons explanations are adequate then our metaphysics must turn out a certain way.
Posted by: Chris | February 28, 2007 at 09:39 AM
Chris!
Thank you for your excellent presentation and analysis of Searle´s book on Free Will.
I have a few qustions on the implications of your analysis.
1. Is hypothesis (2) compatible with a respectable scientific view on the brainfunctions?
2. Is the hyp(2) a necessary but not a sufficient condition for Free Will?
3. Is it reasonable to maintain that the quantum mechanic theory gives sense to the theory of Free Will?
4. Is the non-Humean soul something existing outside the brain?
Posted by: Per Ericson | March 21, 2007 at 10:20 AM