« Sense and Sublimity | Main | Trends in Labor Productivity »

February 22, 2007

Comments

Nato

Physicalists don't believe in *radical* freedom of the will. It's like the leap between an analytical linguist that recognizes that lexicons and grammars are highly flexible and heterogeneous in use and a deconstructionist who takes seriously Derrida's now-it's-facetious-now-it's-not claim that there is infinite play in words and nothing has any stable meaning. Likewise, will is a phenomenon so flexible, durable and extensible (i.e. free) that it's pointless to reduce its causality to that of the atoms instantiating it - to physicalists. To others, it would seem will must be utterly irreduceable and not be causally lawful to be "free" in the important respect.

Nathan Smith

"Physicalists don't believe in *radical* freedom of the will... To others, it would seem will must be utterly irreduceable and not be causally lawful to be 'free' in the important respect."

I would say, not "in the important respect" but "in any intelligible respect." The phrase "*radical* freedom of the will" is an attempt to isolate the view that there is a separate mode of causation called choice, and treat it as somehow odd; at the same time, it suggests that there is some possible middle ground between choice being real and not real. I don't understand what the nature of that middle ground could be, and I suspect it is a mock-concession to the powerful intuition that there is choice.

One sometimes hears the "compatibilist" view that what is really meant by free will is a lack of threats or constraints. Of course, "I did it of my own free will" does tend to mean somewhat more than that the action was strictly voluntary; it is somewhat more emphatic, and implies an absence of particular constraints that some in the audience might assume. But this is an irrelevant semantic anomaly. In practice we know the difference between a voluntary and an involuntary action quite well. We know that the man whom the Nazis force to shoot his comrade on pain of death nonetheless *chose* to pull the trigger.

One understands the motivation for denying free will if the physics one believes in is deterministic. If the universe is only atoms bouncing around like billiard balls, exchanging momenta with mathematical exactitude, then the universe is perfectly deterministic and there is no room for free will. When one introduces Heisenberg's uncertainty principle and all that, when physics ceases to be deterministic, it becomes clear that there is *room* for freedom of the will. Bertrand Russell argued that even in a world of sub-atomic randomness physicalist determinacy rules out freedom of the will, because non-determinacy is only present in extremely small things, but he was clearly mistaken. A huge boulder could be balanced at the top of a hill so as to be able to be tipped either way by the slightest breath of wind. Or, as the saying goes, "A butterfly flaps its wings in China and changes the weather in New York." Systems can be rigged up so that very small changes have very large effects. And it is a principle of living organisms, and human brains in particular, that small things can have big results. It is almost as if the whole miraculous contrivance of life were oriented towards that end.

If the combination of sub-atomic randomness with the nature of the brain makes human beings' behavior non-deterministic, then physics is not exactly incompatible with freedom of the will, but the existence of such an entity as a free will is nonetheless rather at odds with the ontology implicit in the methodology of the physical sciences.

I regard the intuition of free will as foundational. No evidence could cause me to doubt it because there is no evidence which is more foundational than that intuition. Perhaps I can be mistaken even in that, but if so, there is nothing for it but to be a total skeptic, for there is certainly nothing else that I can trust more. That is perhaps a separate reason to mistrust the physicalist ontology.

Nato

Well, it certainly is an attempt to isolate that position, since otherwise it would be a very confusing conversation.

I'm glad you bring up intelligibility, because incompatibilist positions are, to me, quite in danger of unintelligibility. If someone decides to do something, what causes that decision? If the answer is "nothing", then how do we square that with the idea of "will" in which some stable part of a person's composition drives the decision-making? This goes just as much for ideas that mark free will as being possible due to quantum indeterminacy. If the answer is "the soul", then the next question becomes: "what is that?" Now, we don't have to answer that fully for the position to remain intelligible, but it would seem that one would have to accept souls as having some properties and the properties themselves either being the result of other things or being themselves causeless. If the soul is the result of its trajectory through existence, then it would stand in the same relation to free will that physicalist minds do. If causeless, then it would seem one's will is ultimately arbitrary and contingent on the whim of God, the Universe or whatever. I suppose this is a fine view, except that using the incompatibilist analysis of responsibility, no one would be responsible for anything and there is no will, free or otherwise.

Everything is either caused by something or caused by nothing. There just isn't another position. Thus, so far as I can tell the self-made self (in one form or another) is the only way out of the nihilist pit, whether one is monist, dualist, pluralist or anything else. We can talk about intuitions of radical freedom all day long, but if the intuition leads to a ridiculous result, perhaps there's something wrong with it.

The intuition that we have free will is not the problem. The insistence that this free will is constituted of radical causality seems very much a philosophical add-on, born of less rigid commonsense ideas that can be supported if slightly deflated. One must ask, "what do we want from 'free will' and why?" to discover what exactly we want to discuss when we refer to free will. I don't think one is constitutionally incapable of clarifying intuition.

Nathan Smith

The question "If someone decides to do something, what causes that decision?" is certainly an interesting one. To answer, "X caused the decision" seems to point to determinism. To answer, "the person's decision was its own cause," which might be offered as an attempt to defend free will, seems to point to circularity. And if the question turns out to be somehow meaningless or misconceived, this would be surprising, since it seems simple and clear enough.

(Although, on the other hand, causation was one of the targets of Humean skepticism: how can we ever know that two events have the relation of *causation* to one another? I think one could, in principle, accept Humean skepticism about our ever being able to know about causal relationships, without accepting the skepticism about all induction.)

The question is an invitation to reflect on the relations between the intuition of free will and the fact that people have "reasons" for their decisions. Yes, I made that left turn because it's the shortest way to work; but I could have gone straight. For any action one can offer many reasons, indeed almost an infinite number: I took the left turn to get to work; I took the left turn because I didn't want to detour through the traffic on Sullivan Street; I took the left turn because I don't want to be late; I took the left turn because the light had just turned green; I took the left turn because I don't want to lose my job, etc. All the reasons are true; some would be regarded as evasions in some contexts but are true nonetheless. I will account for my action differently depending on my audience, citing different reasons, sometimes because it's convenient to conceal some reasons, but sometimes quite honestly, because different interlocutors are aware of different information about the circumstances leading up to my action, and need to be informed of different things to make my action explicable. We typically assume, though, that even given all the "reasons," a person could have chosen differently.

To the extent that a merely *logical* objection to free will is being made here--"everything is caused either by something or nothing; there just isn't any other position," and free will is "nothing"-- we can answer that with the following formulation. Consider the possible relationships between Situation A and Event B, subsequent in time to Situation A. (Let Situation A be understood to be a comprehensive account of all relevant facts about the world at a point in time.) Event B, or some aspect of Event B, might be *deterministically caused* by Situation A, in that, given Situation A, it was inevitable that Event B would occur. Or, Event B, or some aspect of Event B, may be one of many possible events/aspects-of-events subsequent to Situation A, subject to *randomness,* in the spirit of the Heisenberg Uncertainty Principle perhaps, though of course ordinary people are also perfectly accustomed to the idea that things sometimes happen "by accident." Or, Event B may be one of many possible events/aspects-of-events subsequent to Situation A, subject to *choice.*

The answer, then, to the question: "If someone decides to do something, what causes that decision?" should run something like this:

"When the question you have asked is asked about a *particular* action, it is conventionally answered by the provision of 'reasons' for the person's action, but even in the case of a particular action, if you're looking for *complete* causes, *deterministic* causes, you will not hear what you are looking for. Common-sense accepts at least three accounts of the genesis of events: more-or-less-deterministic physical causation; choice; and randomness or accident. Of course the three interact with each other, and any given event is likely to be the result of all three, and because of this ordinary language is not especially scrupulous and the word 'cause' will be casually applied to physical causes and the 'reasons' that a person makes a choice, but in the latter case it is taken for granted that there is an irreducible element of choice, so that if you were to insist on being given a complete set of 'causes' for a person's decision which were entirely determinative of the person's decision, common-sense would find the line of inquiry impossible to satisfy not only in practice but in theory. When the question about the causes for people's decisions is posed in a general way, then the answer either has to be an abstraction from the conventional answer, a vague 'Well, people have reasons they do things, no doubt,' or 'Depends on the situation,' or else it will be a question: 'What do you mean by cause?' Which would only lead back to the open question, 'Why do you think that causation, in *that* sense, is applicable to people's decisions at all?'. To say 'people's decisions are caused by nothing' is not appropriate, first because it seems to be an attack on the common-sense view that people have reasons for their actions *at all,* and second because 'nothing' can't be a cause, thus the statement is artificially paradoxical. More apt would be statements like 'People's decisions are not fully determined by prior causes,' or 'While there are some parallels, there is no reason to believe, and perhaps, if certain intuitions are accepted, compelling reasons to deny, that the model of deterministic physical causation which it is often convenient to assume in analyzing physical and chemical phenomena, can be extended in a comprehensive way to cases where human will is involved.'"

Having dispensed with the logical objection, we are left with the challenge of trying to understand the nature of free will through introspection. What was the moment when we made that disastrous decision? Why is it that so-and-so doesn't even seem to be *tempted* when opportunities for underhanded or dishonest gain appear? Is there any way that I can make myself stop thinking mean thoughts about that person? What is the nature of my control within my own thoughts?...

Nato

First off, my logical objection is not to "free will" per se, but to the incompatibilist conception of free will, which I regard as either mistaken or incoherent. The idea of "choice" as a third route between deterministic and stochastic causation is perfectly opaque. What sort of sense can one make of the phrase "decisions are not fully determined by prior causes"? So, that is to say, some part of your action is not determined by your personality, your soul, or any other preexisting condition, yet is not stochastic. So what is it? How does this mysterious concept provide anything we want from free will?

Chance gives us nothing we want from free will. "Why did you do that?" someone asks, and we reply "Because a beta particle struck a neuron just so and caused a discharge that cascaded into the decision to help the needy." That's not deterministic, but it sure doesn't allow for ownership of our decisions. Determinism, broadly speaking, allows for our personality and composition to determine our future actions, but since the chain of causality leading to our personality and composition can be traced farther back than our own existence, incompatibilists feel they have to rule it out. So we have the incompatibilist "choice," which appears to translate to something like "the uncaused cause that nevertheless is the product of will, which is also uncaused."

What I find most painfully ironic about this is that if it weren't for what seems to be a purely theoretical hangup, much of Nathan's account would mesh perfectly with Dennett's account of free will. We self-control, trying to manage our temptations, better ourselves, judge our dispositions and consider how to change them if we find them wanting. *That* is the essence of free will, not special causation.

PS. "Nothing" can certainly be a cause of a truly stochastic process. What causes this atom versus that to fission at this time? Nothing, it just happened that way, say the equations. Tom thinks this is an artifact of theoretical primitivity, but I don't really have a personal position on it except to say that it seems that I cannot rule out "nothing" as a cause. Further, in a fully deterministic system, everything is caused by everything else in some sense, requiring anyone wishing to make use of causal discussion to chunk systems and processes in a conceptually tractable manner that allows for proximal causes, medial causes and so on. This fits Nathan's third paragraph exactly. In certain contexts, however, there's no way to chunk a phenomenon so that there's grounds for selecting any proximal or medial cause, and one could meaningfully claim that nothing caused them.

Nato

I should mention that some physicalists uncritically accept the incompatibilist notion of free will, and thus pronounce it an illusion.

Nathan Smith

re: "The idea of 'choice' as a third route between deterministic and stochastic causation is perfectly opaque."

I don't see how choice is any less opaque than causation. If you say, "Why did the ball drop?" and I answer "Because I let go of it and gravity pulled it down," and you go on to ask, "But why did the release of the ball, combined with the force of gravity, cause the ball to drop?", I don't see what answer I can give. To say, "It always does that" is, of course, neither here nor there.

In a way, causation is more problematic than choice. If Situation A precedes Event B, how can we ever know it *caused* it. Are we relying on some kind of induction-- Event B always follows Situation A-- in which case we are, at best, relying on the validity of induction? But actually, even if induction is valid, how can we ever know that Situation A and Event B were not co-caused by unobservable Condition C? How do we refute the Humean causation-skeptic?

The question may seem like a perversity of which only a philosopher would be capable. But I think the effort to answer it would be helpful in dispelling a certain naive confidence in our knowledge about causation which makes people think an insistence that choice be reduced to causation is somehow cogent.

Nathan Smith

re: "I should mention that some physicalists uncritically accept the incompatibilist notion of free will, and thus pronounce it an illusion."

John Searle accepts the incomptabilist notion of free will, by no means uncritically but because it's clear to him that that is the only interpretation of the free will intuition.

Nato

"I don't see how choice is any less opaque than causation"

I believe the discussion involved dividing causation into three categories: deterministic, stochastic and "choice". If choice is not a kind of causation, then how does one describe it?

On the other hand, Nathan may be comparing choice-based causation and deterministic causation. There are a number related senses of deterministic causation as well as different analyses (eg actualism) thereof. There's a lot of different ways folks phrase the issues, but I think there's a generally agreed idea of deterministic causation describing a link between the present, what came before and what will come later. How contingent one views the whole tapestry to be and with what words one may wish to apply to any given circumstance are open issues, of course, but I don't think there's any denial of the idea that if the past was different the present would be different, and for the present to be different, the past would have to be different. Why does the ball drop? Because the deterministic rules say that's how everything works. Why do things work that way? We cannot answer the infinite regress any more than we can answer why there is something rather than nothing - they may ultimately be the same question.

What does "choice" do in the same position? If we make reference to a person's personality, we would seem to be referring to a preexisting state of the world plus rules of behavior. This would seem to be a deterministic mode of explanation. If we attempt to escape this mode by making reference to uncertain boundary cases, that would be a fairly ordinary softening of a deterministic explanation due to uncertainty of initial conditions. If we eschew both of these modes of explanation, what have we left to say? Do we make a separate unexplainable phenomenon of each choice, where every decision is no more answerable than why there is something rather than nothing?

As for Searle the asseverating essentialist - I find his stuff perfectly hopeless since it always seems to end in admitted mystery (subjectivity) and paradox (free will). Maybe if I could bear to read more I would come to some better appreciation of his virtues but for now I place him with Block in being one of the most obfuscatory thinkers in analytical philosophy today.

Nato

On another tack, I would want to point out that from a certain perspective one might view it as illegitimate to say that the ball fell because of gravity, because that's just the gross effect of many small interactions between innumerable subatomic particles. This is not a useful or meaningful position, however. Consider the more frequent commentary that when we slam our hands down on a desk, we never *really* touch the desk, because what's *really* happening is that electrons of the matter constituting the desk are repelling electrons of the matter constituting our hand and none of the matter really *touches*. It's a fine thought-picture, but its counterintuitiveness comes from an insistence on a certain idea of what "touching" means that doesn't translate well outside its proper domain. Electrons repelling just *is* what is going on when we think about touching in ordinary life, and the minutia of the interactions doesn't invalidate any part of the idea of touching that we really care about.

Nathan Smith

"I believe the discussion involved dividing causation into three categories: deterministic, stochastic and 'choice'. If choice is not a kind of causation, then how does one describe it?"

Choice is a cause of other things; and in a different sense, choice itself has causes. But the causes of a choice are not determinative; this is perfectly well-understood by common sense. However strong my reasons for telling a lie, I can defy them and tell the truth. If the reasons *are* determinative-- if I am hypnotized or tricked into telling a lie-- then it is understood that choice is not involved; people will say that I did not *choose* to tell a lie in that case. If you want to think of the world as a tapestry of deterministic threads of causation, then choice is an interruption of those threads of causation.

Randomness seems to me another interruption in the threads of causation. The uranium atom may disintegrate during time-interval (t, t+1), or not; if it does, can it be said that anything "caused" this? We do use the word to apply to random cases. I run a red light and am said to have "caused" the accident, even though the odds were 99% that my running the red light would *not* have had that result. Similarly, we might say that my anger *caused* me to kill my wife's lover, even though I could have chosen differently. But I'm not sure that the semantic practice of using one word, "cause," in relation to the three situations-- determination, choice, and randomness-- is proof that the three belong in a category together.

One can imagine a universe which is painted, in three dimensions of space and one of time, by some playful god, and creatures within that world whose perceptions and consciousnesses move that world in one direction, observing it in a particular way, so that they can see the "present" and, through the foggy lenses of memory, the "past," but not the "future." The god has arranged that people's perceptions are thus related to the time-spectrum; but he did not paint in in chronological order, and with respect to some time-outside-of-time in which the creator-god lives, distant futures were painted, perhaps, before the present. In this imaginary world, there is no causation. That is, the god caused this whole world to be, but between the created things and events in this world there are no causal relationships.

How would we know the difference between that world and our own? The strange thing is: to me, it seems almost that in a deterministic world, causation, though supreme, loses its meaning. All future ages are determined *now*; in a sense, it is as if they already exist, and that I, one creature, have the peculiar characteristic of being able to look backwards but not forwards in time, seems of little relevance. It "A causes B" means "if A, then B; if not A, then perhaps not B", then this statement seems meaningless, for if A is the case, then not-A cannot be, it is an impossibility.

It seems to me that the fact of choice motivates the notion of causation and makes it meaningful. Why am I interested in causation? Because I make choices, and I wanted to know what the results of my choices will be. Also because I make choices, A and not-A are both possibilities, so "if A, then B; if not A, then perhaps not B" is meaningful. Does it follow that choice is more fundamental than causation?

Nato

"It seems to me that the fact of choice motivates the notion of causation and makes it meaningful. Why am I interested in causation? Because I make choices, and I wanted to know what the results of my choices will be. Also because I make choices, A and not-A are both possibilities, so "if A, then B; if not A, then perhaps not B" is meaningful. Does it follow that choice is more fundamental than causation?"

An excellent analysis, except that there's nothing in there that needs choice to be fundamentally outside the chain of ordinary causation. More on this later.

Nathan Smith

What is meant by "ordinary causation?" Deterministic causation, or randomness?

Nato

"What is meant by "ordinary causation?" Deterministic causation, or randomness?"

Here is about half the heart of the matter. It would appear that our uses of "causation" are not quite in registry, which would confuse the issue. I want causation to refer to everything that we must account for in projecting from the present. "Things are in state Y, which indicates that X occurred" if projecting backwards, with "and Y will probably result in Z tomorrow" being the forward version. The projection does not, of course, have to fit the broad determinism of the first example, which can be extended as "but X is a stochastic event, so there's no ordered connection back to W." X is nondeterministic (unordered) causation, X leading to Y is deterministic (ordered) causation, and Y probably leading to Z is either the result of lopsided probabilistic (nondeterministic, partially ordered) causation or represents limitations on the information available with which to project. This account of causation is not meant to presuppose that there exists truly random events(X), nor do I presuppose there are any events that are fully ordered(X->Y), but of course, causation as a concept is worthless unless there is at least some order, so Y->Z is the minimum.

As an important side issue, we perceive varying shades of randomness based mostly on how predictable the phenomenon is. Earthquakes, a candidate for determinsitic causation if ever there was one, strikes us all as entirely random except in a regional sense. There's considerable equivocation in the ordinary conception of causation because we are familiar with the possibility of making a "random" event predictable and thus nonrandom, if circumstances furnish us the relevant information. Talk of what caused what else sometimes assumes that with enough information we'd be able to put our finger on the very event, sometimes speaks only of what can reasonably known, and sometimes addresses - usually in a vague sense - the question of whether a chain of events had any nonrandom origin.

Of course, if one assumes that all events have fully sufficient (deterministic) causes, then one may assume the actualist stance, in which only the actual is possible. There's a sense in which this would be a true stance, but it's a worthless one, since, as Nathan notes, the whole motivation for causative ascription assumes that one can affect one's future. Since actualist causation leaves no room for agency, then there's no reason for that kind of causation to interest us. The general actualist can also be extended to encompass stochastic causation by simply noting that whatever random event eventuates cannot be affected without reference to other such uncontrollable events, so there remains no room for discussion of differing paths of events. The simplicity that recommends both positions is exactly what disqualifies them from usefulness.

I'm not sure exactly what "ordinary causation" is, either, since I threw it out there as a general reference to "however Nathan is using the word." Whichever way that is, though, there's nothing about the choice-to-causation motivational chain that requires choice to be different from causation. Unless one is an actualist, but I'm assuming that we can all dispense with that self-negating perspective.

All of which is just some elucidation of causation-related issues, without attempting any sort of synthesis of the overal issue. So, more later.

Nathan Smith

Let me try my hand at "elucidation of causation-related issues." Certainly this is a challenging and interesting topic...

It seems to me that the word "cause" has meaning in (at least) three types of situations: deterministic situations, random/accidental/probabilistic situations, and choice situations.

In deterministic situations, "A causes B" means: "if A is the case, then B will occur; if A is not the case, B either will not occur, or may or may not occur depending on other factors." By "deterministic situations," I am not talking about a comprehensively deterministic world, but only about a particularly causal link: A *determines* B, but the rest of the world may be nondeterministic.

In random/accidental/probabilistic situations, "A causes B" is more complicated, but it means something like: (a) A increased the probability of B occurring, (b) B did, in fact, occur, and perhaps (c) A is somehow special or irregular or noteworthy such that it deserves to be set apart as a cause from many other circumstances or conditions which also increased the probability of B occurring, but which a typical hearer of an account of B will take for granted. This definition of causation is dependent on a notion of probability, but probability seems, if anything, more philosophically problematic than causation. Probability-conditional-on seems to be more fundamental than probability *per se*. That is, if you are asked about the probability of event X, you will ask about conditions A, B, C, etc., which are inputs to the likelihood of X; if you don't know them, you will, explicitly or implicitly, estimate or assume the probabilities of the conditions first; and if, having estimated the probability of X in this way, you get new information about the truth or not of A, B, and C, you will amend your estimates of the probability of X. This makes it seem that probability is an observer-dependent property, a property not so much of X as of a particular observer's relation to X. If our account of probabilistic causation depends on that kind of probability, it becomes non-objective in principle. So "probability" must be turned into probability-conditional-on-the-entire-state-of-the-world. I think, in fact-- to justify this belief would require an extensive discussion-- that the gap between this probability-conditional-on-the-entire-state-of-the-world and subjective probability is not as wide as it might seem to be. Indeed, a more serious problem with probability is that it is inherently unobservable. We can observe event X; we cannot observe all the possible worlds in which X0, X1, X2, X3, and all the other events that, in the situation of randomness, could have happened instead of X, occur; still less can we observe all the possible worlds and measures their respective *ex ante* probabilities-conditional-on-the-entire-state-of-the-world. We can, perhaps, observe "situations like A" (A being the antecedent of X) many times, and see how many times they result in events-like-X; but how do we define the relevant "like." All this is deeply bewildering when you try to analyze it philosophically, and it strikes me as profoundly mysterious and even a sort of miracle that in practical life, common-sense is able to understand probabilistic causation. An argument about the philosophical bewildering-ness of probability to a certain kind of rationalistic analysis feels like a mere prelude to a breezy conclusion: "But of course we all understand these things by common-sense," a sort of secular echo of what some religious would say, "But of course we know by faith that..."

Finally, in choice situations, "A caused B," or rather, "A caused B to do X," means: (a) A was, and was recognized by B as, a "reason" (in a sense that is specific to describing the motivations of conscious beings) for B to do X, (b) B did X. But in this case, unlike the case of deterministic causation, B could have acted differently. Indeed, B might not even have chosen to recognize A as a reason to do X at all. Or B could have done X for some other reason.

One cannot put this "A caused B to do X" into the present tense without changing its meaning. "A causes B to do X" is certainly intelligible, but it would typically describe an action X that is involuntary or habitual. "Talking about his secret crush makes Tony blush." "Swearing makes Mother get angry." The statement suggests that Mother gets angry involuntarily, or out of habit. Or perhaps she gets angry because she is following a rule based on a principled belief; she still has free will in principle-- she could start cursing like a sailor herself if she chose!-- but she has decided to relinquish it in favor of *obedience* to a code of conduct which requires her to discipline her children.

"A causes B to do X" refers to situations in which choice is either absent or is being downplayed; the situation is deterministic or quasi-deterministic, as is the applicable definition of causation. The absence of choice makes possible the prediction implicit in the present indefinite tense of the verb. "A caused B to do X," by contrast, refers, and must refer, to a completed action, because, since choice is involved, the causal link is not predictable or generalizable.

Nato

"A causes B to do X" refers to situations in which choice is either absent or is being downplayed; the situation is deterministic or quasi-deterministic, as is the applicable definition of causation. The absence of choice makes possible the prediction implicit in the present indefinite tense of the verb. "A caused B to do X," by contrast, refers, and must refer, to a completed action, because, since choice is involved, the causal link is not predictable or generalizable."

"Reaching a door just ahead of another causes (ceteris paribus) Sarah to hold the door for the other person." This would seem to be a valid generalization, since Sarah demonstrably holds the door open at least 99% of the time that nothing physically impedes her. It would seem she chooses to do so every time. Or, perhaps, we could say that she has chosen to inculcate in herself the habit of behaving this way so that she doesn't have to reason her way into the choice every time. Whatever the case, we conclude Sarah is conscientious and polite, not that she's a deterministic automaton.

One might respond that we take her free will on faith and so dismiss evidence of regularity in her behavior, but this seems an unnecessary theoretical complication. I just don't believe that we pretheoretically regard regularities in action to be evidence for or against free will. Wouldn't it be more parsimonious to conclude that what we want from free will is "acting in a way responsive to our own reasons (ceteris paribus)" not "acting with radical unpredictability"?

Aside from motivations to *want* free will, there is, of course, the vivid impression that at any moment we could decide to do something differently than we have always done, and, well, none of the points of view in question dispute that! A hard determinist might say that you are fated to do something different in instance #207,342 because of some exact neural coincidence, but it would still be different, and the crucial events would all happen inside *you* and would be available to *your* introspection. Of course, this doesn't answer if we could have done something differently *that one time* or, more immediately, "Could I do either this or that right... NOW!" It would seem that there are many ways of looking at this last scenario - perhaps we just don't have enough information about ourselves to be sure what we're going to do. This is a possible corner case, but more of the time it would seem we know what we're going to do but represent to ourselves many options. We imagine or carry with us the awareness of differing causal narratives indicating that *if* we behave in some way, something else could happen. This leads to the idea that we have radical causation, but does it do so pretheoretically or because as we grow older we learn to dismiss radical causation from external causative forks but theoretical concerns persuade us to refrain from extending the same metaphysical stance to ourselves*?

A critically important side note: because people are very, very complex systems, even the hardest determinist would be badly mistaken to conclude a person's behavior to be easily predictable even with a great deal of microphysical evidence, because as Nathanael (sort of) notes, we're chaotic systems that amplify very small differences at arbitrarily large rates. Thus, the super neuroscientist with near perfect knowledge of the initial conditions of anyone's mind at the minutest level would quickly find the subject's best friend giving as good or better predictions than the model. For all intents and purposes, even deterministic minds are unpredictable except at the level of "intent" and "will." Thus "choice" is a real pattern in the world even if it has no radical causal power.

*In *practice* we still try to examine our own behavior and look for causes behind our decisions, establishing a chain of causation just like we learn that many things that initially seemed random or unpredictable have an underlying pattern if you know where to look. The theoretical denial only precludes total causation to obviate the (imagined) possibility that we are just meaningless automatons and/or cannot exercise self-control. Of course, those who wish to avoid responsibility are happy to abdicate claims to self-control in any manner available. Such people would not be pleased by the conclusions of Dennett's "Control and Self-Control" or "Could Have Done Otherwise." My gf was actually a little indignant when she thought Dennett's position didn't allow any room for shading moral evaluation based on extenuating circumstances.

Nathan Smith

re: "One might respond that we take her free will on faith and so dismiss evidence of regularity in her behavior..."

In a sense, I'd agree that we take the free will of *others* on faith. We are aware our own free will through introspection, but in others we only assume free will on the basis of a belief that they are "like us" in some essential but difficult to articulate way. That belief, or "meta-belief" as I sometimes call it, cannot be proven or even adequately articulated, etc., etc.-- I've made this argument before...

re: "I just don't believe that we pretheoretically regard regularities in action to be evidence for or against free will."

Well, I agree that regularities in action don't obviously point one way or the other, but the belief in free will certainly *is* pretheoretical. Nato's mention of "reasons to *want* free will" seems to miss the point: I don't see how free will would even be conceivable, let alone desirable, if we didn't have it, and have immediate and perpetual experience of it which theory is an effort to understand.

re: "Wouldn't it be more parsimonious to conclude that what we want from free will is 'acting in a way responsive to our own reasons (ceteris paribus)' not 'acting with radical unpredictability'?"

I don't know what is meant by "radical" unpredictability. Non-certain predictability, to be sure, but not "more" unpredictable, as far as I can tell, than in the case of randomness, though the nature of the limited predictability in the two cases is somewhat different. And by all means, we are "acting in a way responsive to our own reasons," but those reasons are not *determinative* of our actions. They are the lawyers and the witnesses, but the will holds the gavel and gives the verdict.

Nato

When I use "reasons" I mean to use the somewhat traditional sense of the word, meaning motivations; in aggregate they *are* the will. The will can be inconstant and we sometimes do things that surprise ourselves, but in the main, what we will is, if we have *free* will, a result of our own values, moods, dispositions, goals, etc. That's my position, anyway.

Also, concerning radical unpredictability - quantum randomness is also radically unpredictable, as (according to the equations as they now stand) it is not possible to predict certain processes even in principle. Free will as put by incompatibilists is, I understand, usually considered to have some *other* kind of radical unpredictability than that in quantum physics, but I've never seen an intelligible explication thereof.

Finally, I wish to say that, yes, certain parts of the fleshed out incompatibilist idea of free will are entirely pretheoretical as indisputably true in the important respects. However, I resist the claim that the pretheoretical portions are isomorphic with the incompatibilist formulation and counter that the additional portions are precisely what make the concept(s) either unintelligible or mistaken.

Nathan Smith

re: "When I use 'reasons' I mean to use the somewhat traditional sense of the word, meaning motivations; in aggregate they *are* the will."

I experience reasons and the will as different things, so I would first insist that this equation is counter-intuitive. Reasons are things my will encounters/experiences, and can respond to, or override.

Now, it might be said that even if the will overrides Reason A, it always has Reason B for doing so. And no doubt a "Reason B" could always be formulated by a person articulate enough to do so. (Sometimes people say they did something "for no reason" or "I can't explain it," but we can take that to mean that they don't remember, or can't explain, the reason, or perhaps are concealing it.) But it still doesn't follow that the will is reducible to its reasons, because reasons do not have a quantifiable strength so that they can be balanced against each other and the will's choice depends only on the reasons.

As for Nato's objection that he's "never seen an intelligible explanation of" free will (what he calls "radically unpredictable" free will), a certain inexplicability may be a property of all things that are experienced directly through introspection. Could we offer an intelligible explanation of what *red* feels like to someone who did not know it? (Not the physical processes that cause the sensation of red, of course, but the internal experience of red.) I believe in the reality of the experience of red because I have experienced it. Likewise, I believe in free will from direct experience of making choices, of confronting multiple possibilities and deciding, of being the source of actions that then occur in the world.

We're making a valiant effort to argue here, I think on some sort of desire to believe that argument can achieve things which really, perhaps, it cannot. I can't prove that I had free will by turning back time and taking the roads not taken, just as Nato can't prove that I lack ("radical") free will by creating a comprehensive compilation of my "reasons," quantifying their strength and elucidating their actions, and showing that my choice was determined by them. We're ultimately falling back on intuitions, and as a result we're at something of an impasse.

Nato

Aside from discussions of what is pretheoretical and what is not, I don't think that there's much that's empirical here. The only reason I make my empirical claim regarding intuitions is that at any time my interlocutor can claim a conclusion as a foundational, axiomatic intuition that further reasoning could not contravene, and if things are to proceed, I need to try to cast doubt on the claim of a foundational intuition. It's a bass-ackwards way of arguing but sometimes it works.

The fundamental nature of this argument is definitional, not empirical. It comes in two parts: 1)If choice is to be defined as *requiring* a causative role that is neither determined nor random, then it would seem that the definition would have to include some intelligible account of what that role would be. If one wishes to define choice as something in some ways determined and in others random, that still seems to be within the overall compatibilist rubric, since compatibilism is not determinism. 2)laying aside the claim that we feel ourselves to have causation that is neither determined nor random, what distinctions do we want to make between free will and otherwise, and why?

Some, like Searle, leave #1 as a "mystery" while still insisting we need radical causation or something similar (e.g. Penrose) to get a good answer out of #2. Thus there's a strong prima facie case (to them) for allowing #1 to remain undefined because the alternative is the loss of all meaning in #2. This makes a sort of sense, but of course, I think they're wrong on #2 as well.

Regarding the intelligible explanation of red, I would refer back to the corpus of responses to Frank Jackson's thought experiment of Mary the color scientist.

Nathan Smith

re: "The fundamental nature of this argument is definitional, not empirical. It comes in two parts: 1)If choice is to be defined as *requiring* a causative role that is neither determined nor random, then it would seem that the definition would have to include some intelligible account of what that role would be."

I'm trying to figure out by what kind of "account" is being asked for here.

"An account" could mean some description of what it is like to choose, of the experience of choice. I have offered some of that, Nato has actually offered a bit, and of course, I could offer more. Or I could refer him to novels, or ethical theory, which deal with choice in different ways. It seems to me, though, that none of that would convince Nato of anything.

It seems to me that what Nato really wants is some account of how free will could function in a causal way at the level of atoms and molecules. How the soul intervenes within the probabilistic motions of protons and electrons and, beginning from a level below that which can ever be observed by physics, sets in motion changes that work their way up to the macro-level, making the physical brain dependent upon the motions of a supernatural soul.

It might work something like that; we'll probably never know. Meanwhile our method for studying free will needs to rely on introspective evidence, in which the free will intuition is an inherent part. What's getting in the way is a mental habit, rooted in the ideology of our own peculiar times, of assuming that atoms and molecules are somehow more fundamental than anything else. An "account" of free will that reduces it to atoms and molecules will probably never be available, but that's not really a problem; the problem is in the physicalist assumptions that give such an account undue importance.

Nato

"It seems to me that what Nato really wants is some account of how free will could function in a causal way at the level of atoms and molecules."

I think this is the source of the confusion here. For the purposes of this discussion I'm perfectly willing to bracket mechanisms by which will affects the world. Do atoms suddenly swerve in response to choices? Does our will affect the probability of certain quantum interactions? Whatever - help yourself to whatever mechanism you want, since there's no need for specificity at definition time.

So if I'm not asking for that, for what *am* I asking? I want an account of will that makes sense of it having no antecedent causation yet is not random. If our decisions are not (fully) dependent on our composition, then what is the missing part? Saying "our will" is what decides (and nothing more), then it would seem that we're into a question-begging definition of "our will is the part of decision that is neither predisposition nor randomness." One might say we can help ourselves to this undefined third category because we have an intuition that it exists, but that's bizarre. We have an intuition that *what* exists? Something defined as being not two other things. It amounts to nothing more than the asseveration that predisposition and randomness are inadequate to explain our intuition of free will. To then reference the ability of predisposition plus randomness to address the free will problem is illegitimate.

All that said, there's nothing about predisposition and randomness that entails or even suggests physicalism. The compatibilist formulation of free will generally suggests problems of will are entirely orthogonal to monism/dualism/pluralism.

Nathan Smith

"One might say we can help ourselves to this undefined third category because we have an intuition that it exists, but that's bizarre."

Why is it bizarre? Isn't it normal to recognize that things exist from experience, and then to develop theoretical frameworks such that they are able to accommodate these things? And choice is not "defined as not being two other things" any more than red is defined as being not green and not blue. Yes, our experience of choice is such that it rules out choice being either deterministic causation ("predisposition?") or randomness, but that is a sort of side-effect of what choice is, and of what randomness and deterministic causation are. Choice is an experience in its own right. Randomness and deterministic causation are experiences too, not unproblematically perceived to be sure, but experiences of a sort, and they are different from the experience of choice. One can't deduce the idea of deterministic causation, still less of randomness, *a priori*, either; experience plays a role in motivating in us these ideas.

What I can't understand is why Nato should regard causation and randomness as somehow less odd and/or more fundamental than choice.

Nato

"Isn't it normal to recognize that things exist from experience, and then to develop theoretical frameworks such that they are able to accommodate these things? And choice is not "defined as not being two other things" any more than red is defined as being not green and not blue. Yes, our experience of choice is such that it rules out choice being either deterministic causation ("predisposition?") or randomness, but that is a sort of side-effect of what choice is, and of what randomness and deterministic causation are."

We know from experience that we are at times represent to ourselves (imagine) possible courses of action and there's sensation we go through on the way to pursuing a course that we lable "choice" or something similar. The sensation is of ourselves settling on (or lunging at) the course of action rather than, usually, something external propelling us at the action or suddenly finding ourselves pursuing a course of action with no idea how we got to be doing so. This inevitably leads to the intuition that whatever drives our choices is identified with us and that our choices are (largely) nonrandom.

The further conclusion that the causation inside us is utterly unlike what we see around us (as opposed to different because it's *not us*) seems to be a theoretically-motivated step. It's not a foolish step - as I've mentioned before, agents are uniquely well predicted using the intentional stance in which reasons (in the broad sense) and epistemic obstacles play roles unlike they do in the inanimate world. It also doesn't generally seem clear to us how But it's still a theoretical response to non-introspective findings.

Everything is still either caused by something or by nothing, and it would seem the only intelligible answer is that will is something*. Our pretheoretical intuitions seem to require that whatever the will is, we constitute it and it is not (entirely) random. Our experience also provides us a difference between what predicts a teddy-bear's actions and what predicts our sibling's. We could conclude that the kind of causation that leads the bear to lay in a seat is not the kind that leads Emily to sit in the same seat. Are these two of several qualitatively different types of causation narratives or are they separated by a metaphysical divide? One could bring reasoning to bear on this question to conclude one way or another, but the basic experience doesn't establish the metaphysical divide as an irrefutable explanandum or undeduced axiom.

It is at that point that the emptiness (or inscrutability) of the incompatibilist's definition of "will" becomes a problem. Of course, if such a definition could made to avoid unintelligibility, tautology, self-reference and contentlessness, then the incompatibilist would certainly be in the running for #1 above. Maybe it would also show a way through to taking #2 by storm, though I posit this without being able to remotely conceive of either.


*I suppose it one could just define "will" to mean "randomness", but this trades unintelligibility for contentlessness.

The comments to this entry are closed.

My Photo

Only use a payday cash advance as a last resort.

Categories

Blog powered by Typepad