The Indeterministic World Objection to Frankfurt Counterexamples

We got bogged down in an earlier thread, so let's try a different tack.  The following discussion draws upon Robert Kane, A Contemporary Introduction to Free Will, Oxford 2005, pp. 87-88.

In his seminal 1969 J. Phil. article, "Alternate Possibilities and Moral Responsibility," Harry Frankfurt enunciates what he calls "the principle of alternate possibilities," (PAP) namely, "a person is morally responsible for what he has done only if he could have done otherwise."  Frankfurt goes on to argue that PAP is false because there are conceivable scenarios in which an agent is morally responsible despite his inability to do otherwise.

Suppose agent Jones has a controller Black who makes sure that Jones does what Black wants.  So if Jones decides to do A, rob a bank, say, and Black wants Jones to do A, then Black stands aloof and allows Jones to implement his decision.  If Jones wavers in his resoluteness, or falters in his implementation of his decision, Black renders whatever assistance is necessary to keep Jones on course.  The point is that Jones cannot do other than A because Black won't let him.  Now suppose Jones decides on his own to do A:  he operates from his own motives and reasons without interference. On this supposition Jones is morally responsible for doing A if he does A.   He is morally responsible even though he could not have done otherwise.  For had he begun to change his plan or begun to falter in its execution Black would have intervened to keep him on track.

Now if scenarios such as the above are possible,  and it surely seems as if they are, then 'S is morally responsible for having done A' does not entail 'S could have done other than A.'

But if we think about it, we see that these Frankfurt examples give an incompatibilist who believes in free will no reason to abandon PAP.  Incompatibilists hold that (libertarian) free will and (causal) determinism are logically incompatible: they cannot both be true.  So if free will exists, then determinism is false.  And if determinism is false, then indeterminism is true.  If indeterminism is true, then free choices are not determined by earlier events and the laws of nature.  Jones choice is determined only at the instant at which Jones chooses, and is determined by Jones.  How then could Black control Jones' choice?  Suppose Black has all the powers of a Laplacean demon: in a deterministic universe he can predict any state from any temporally prior state.  These powers won't help him, however, in an indeterministic universe.  Before Jones chooses, Black cannot predict what he will choose.  He cannot foresee (by observing  electrical activity in Jones' brain , say, that Jones will choose A rather than B.  Black must wait for Jones to choose before he can know what he chooses.  But then it is too late for Black to interfere.  Jones will have made a choice, and indeed one that he might not have made. For Black to ensure that Jones will make the choice that Black desires him to make, Black must act prior to the time at which Jones chooses so as to bring it about that Jones chooses as Black desires.  But then Jones is not responsible for his choice.  Jones cannot be responsible for his choice if Black is part of the cause of the choice.

So if indeterminism is true, Frankfurt counterexamples  fail to refute PAP.    Either Black intervenes before Jones chooses or he does not.  If Black intervenes before Jones chooses, then Jones is not responsible for the choice.  But if Black waits for Jones to choose, then Jones has alternative possibilities.

In sum, whether Frankfurt-type counterexamples refute PAP seems to hinge on whether incompatibilism or compatibilism is true.  An incompatibilist who believes in free will can easily reject the Frankfurt counterexamples as explained above.   A compatibilist who is a determinist, however, could accept them.  But if you are a compatibilist and a determinist then you already have a reason to decouple moral responsibility and alternative possibilities.  So why would you need Frankfurt counterexamples?

Comments

[Philoponus]

Bill,

You have indeed broken new ground with this post and given us a chance to think about how we actually choose anything. Let’s take a very simple “soup or salad” choice. The waiter asks you “soup or salad, sir?” and usually you can give an immediate answer. Within a second or so, you decide and say “soup, please.”

So how do you so quickly choose soup? Well, you choose the one you like the best, and you can know this almost immediately. Your unconscious brain processes “soup” and then it processes “salad”, and it reacts aversively or pleasurably to the memories and associations it has stored for each. If “soup” lights up the pleasure centers and “salad” elicits a muted response, you will shortly (200-300 msec) decide for soup. If you are sitting in a fMRI while you are placing your order, the machine will read your brain’s differential reactions and predict with complete reliability which you WILL choose. Lots of actual experiments now corroborate this.

Now you say according to indeterminism “Jones’ choice is determined only at the instant Jones chooses”. But this is just plain wrong, because Jones’ brain needs to figure which alternative he wants the most before he can choose. This takes time, and we can watch in real-time as the brain is doing this necessary processing. We can tell well before you choose which option excites the brain the most, and this is one you will always subsequently choose.

Even in very simple, quick choices, the decider needs to figure out which option he wants most. This processing is done unconsciously and the result is signaled to the decider as pleasurable anticipation or aversion. The decider then chooses the one that he feels most positive about. If indeterminism claims my choices are undetermined until the moment I decide, then indeterminism is false.

Black or Evil Genie can ALWAYS watch Jones’ brain trying to figure out which option he really wants before he is able to conscious decide anything. Suppose Jones, aka Bad Bob, is at the point of completing the embezzlement he has been planning and the thought “I could get 10 years if I’m caught” pops into heads. I emphasize: Evil Genie has no control over Bob’s thinking about such things. Now Evil Genie watches how Bob’s emotional brain starts to react to this thought. If his brain treats like a joke and shrugs it off, Evil Genie does nothing. But if the amygdala starts to alert, Evil Genie stimulates the pleasure centers of Bob’s brain and these drown out the anxious response. Bob feels pleasure & confidence and reacts to this by thinking “I’m to smart for them to caught me. I’m going to be rich.”

Frankfurt-style counterexamples cannot be dismissed by an absurdly false "indeterministic" view of how we decide.

[Bill Vallicella]

Phil,

Your response indicates that you haven't understood the Kane-Widerker objection to PAP. A paper you should read is David Widerker, "Libertarianism and Frankfurt's Attack on the Principle of Alternate Possibilities," Phil. Rev. (April 1995, 247-261.

The basic idea is that Frankfurt's attack on PAP begs the question against the libertarian by assuming the truth of determinism.

Controller Black monitors Jones' mental goings-on and has the power to keep Jones on track. Suppose Jones at time t decides to kill Smith. This decision is a mental action and we agree that Jones is morally responsible for it. The question in dispute is whether the decision is such that both (i)Jones is morally responsible for it, and (ii) Jones could not have done otherwise than make that decision. If both conditions can be satisfied, then we have a case in which moral responsibility does not entail 'could have done otherwise.'

Now if libertarianism is true, then Jones' decision at t in not determined by prior events + the laws of nature. Black wants Jones to decide to commit murder. But if Black intervenes at t to ensure that Jones decides to murder, then, although there are in that case no alternative possibilities, there is also no moral responsibility on the part of Jones. Why? Because Black will be the cause or part-cause of Jones' decision. If, on the other hand, Black stands aloof, then there is moral responsibility on the part of Jones but also alternative possibilities. Why? Because his decision is not determined.

The upshot is that Frankfurt's attack on PAP as applied to decisions fails if libertarianism is true. The libertarian has been given no good reason to reject PAP as applied to decisions.

[Bill Vallicella]

Addendum: The neuroscience stuff, as interesting as it is, is a red herring because the Kane-Widerker objection focuses on CONSCIOUS mental actions such as deciding, choosing, and formining an intention. These actions constitute for the libertarian the basic loci of moral responsibility. See Widerker, p. 247.

[Philoponus]

Bill,

Correct me if I’m wrong, but I think I understand the logic of the manouevre you are describing. It just puzzles me that anyone would be tempted by it.

A libertarian, you say, can rescue a “could have chosen otherwise” version of the PAP from Frankfurt-style counters like Evil Genie by proposing that “decisions” are miraculous, neurally uncaused events that somehow pop into existence when an agent is confronted with a decision situation. And if “decisions” are instantaneous miracle-like events, then to be sure Evil Genie has no time or place to operate. Evil Genie can presumably still sense what is going on in your brain, and in fact has an uncanny ability to predict the miraculous “decision” that will pop out of your mouth, but there is no causal mechanism for her to influence. So the PAP is saved if human beings make “decisions”? Is that correct?

Would you agree that the first problem here is that libertarian “decisions” seem to have as much claim to reality as lunicorns? People who ride lunicorns can experience “decisions” all day, while the rest of us need healthy, active brains which do a phenomenal amount of fast neural processing to create the conscious events we know as decision making. Damage some key neural circuits and we cannot make decisions at all—another thing I’m sure the theory of uncaused “decisions” can easily explain.

You say people who reject miraculous “decisions” beg the question against the libertarians by assuming the truth of determinism? No, for two reasons. Neuroscientists don’t ASSUME that brain processes control conscious thinking & decision making. They spend lots of time in the lab discovering that if you witness such and such a brain process, you subsequently get such a conscious decision. And if you don’t have a certain type of brain activity, you don’t get any decisions at all. These are experimental facts and conclusions, not assumptions. Second, I don’t think we want to say that accepting the findings of neuroscience commits us to a fully (strongly?) deterministic model. Caused does not entail deterministically caused, yes? There is nothing necessarily fully deterministic about neuroscience.

Evil Genie does not need a fully deterministically model of brain processes because she has multiple neural pathways to help Bad Bob stay in control and do what he wants to do. She may not even know which way will ultimately work, but if she keeps trying, she must eventually succeed. This a complicated point we can explore further if you wish.

Sorry, but I don’t have access to the 1995 Phil Rev article you mention. Have I grasped the manouevre you were describing in the original post, or do I remain blinded by an infatuation with neuroscience?

[Clark Goble]

Many forms of Frankfurt examples depend upon free will being a kind of traditionally emergent phenomena of the brain. Thus they have the "mad scientist" merely modifying the brain. However it seems to me that these simply won't be persuasive to most LFW proponents who already tend to adopt some kind of agent essentialism. (i.e. the need for an irreducible agent who choses - either due to dualism or radical emergence)

Where I think Frankfurt examples tend to have an effect isn't as a formal argument against LFW proponents but more as a way of critiquing the language and context we use to think about freedom. That is if they change the intuitions of how we think about "could have done otherwise" in a non-rigorous and common sense fashion they probably will affect the intuitions people have in thought experiments. i.e. I think it interesting to look at as a rhetorical rather than formally argumentative tact. That said I do think it helps to highlight some of the issues (much like zombie arguments do even if those too aren't usually seen as terribly decisive)

[Clark Goble]

To add (and keep on the narrower topic) I think you are completely right in that Frankfurt examples do try and have it both ways. A particular outcome is determined, but what I think is important is that it is determined in a non-determined universe. However where Frankfurt examples appear to have more effect is to show that in this narrower kind of determinism we may not be free but we may be responsible. This is the implication that Fischer famously takes in his development of semi-compatibilism.

There are all sorts of interesting variations that Fischer develops. None of which ultimately convince any LFW proponent that I can see. The most interesting ones involve added "accidents" rather than mad scientists to make an outcome determined but still (purportedly) entail responsibility.

[Clark Goble]

Phil, I think that Mele has some excellent distinctions that I think Bill mentioned when Mele's paper was discussed. I know Bill has HTML links turned off, but you might wish to read the version in Mele's recent book.

http://books.google.com/books?id=MV0Ohi3z4LAC&pg=PA30

One should distinguish between deciding to initiate an act and the act of choosing to do. So if I want to walk to the door I first decide to initiate the act of walking to the door. I then choose to walk to the door. In the view of the Frankfurt example the mad scientist notes the former in order to eliminate the latter. However Mele would simply say that in this case our agent never gets to choose. He merely is able to choose to initiate. Therefore free will is absent.

What Frankfurt proponents need is a way to equate the two kinds of acts (decision and choice) into a single process. But as I noted LFW proponents need only suggest this is impossible. Certainly one can, as you do, suggest that LFW "decisions" are impossible. But I think even ignoring the metaphysics of LFW (which I admit I find problematic) there are problems. That's because even cognitively there is a distinction between deciding to initiate an action and the action itself. If we assign choice and free will to the action and not the decision to initiate the action then there's simply no place for an outside intervener to obtain the information in order to intervene. That's true even if one adopts a strict brain physicalism.

Now there are, as I alluded, more sophisticated versions of the Frankfurt examples that get around this. But then you run smack into the metaphysical issues which I suggested the LFW proponents already recognize. So ultimately what you end up having to argue against isn't free will but dualism or radical emergence.

[Bill Vallicella]

Phil,

Time just for a quick point. I believe I have shown that FCEs beg the question against libertarians and incompatibilists generally. If that is right, then the problems that libertarians face are not to the point.

I should think that a condition of adequacy for a genuine counterexample to PAP is that it be NEUTRAL on the sorts of issues that divide compatibilists and incompatibilists. This point is also made in other terms by Peter.