Maverick Philosopher

Footnotes to Plato from the foothills of the Superstition Mountains

An Argument for Libertarian Freedom of the Will

First the argument in nuce, then a detailed explanation.

P1. I am morally responsible for at least some of my actions and omissions.
P2. I cannot be morally responsible for an action or omission unless I am libertarianly free with respect to that action or omission.
Therefore
C. I am libertarianly free with respect to at least some of my actions and omissions.


Posted

in

by

Bill Vallicella

Tags:

Comments

48 responses to “An Argument for Libertarian Freedom of the Will”

  1. Mike Avatar
    Mike

    “Compatibilists believe that moral responsibility can be had ‘on the cheap’ if , roughly speaking, a person is free in the sense of not being constrained to act or refrain from acting by anything external to the agent.”
    Certainly some compatibilists hold that S is free only if (i) S could have done otherwise and (ii) in cases where S’s actions are causally determined, S could have done otherwise.
    That seems no cheaper than the incompatibilists who hold that pap governs freedom (fwiw, some hold of course that pap governs responsibility but not freedom). Compatibilists who hold (i) and (ii)–Lewis, for instance–argue that the backtracking counterfactual ‘if you had done other than what you did, the past would have been different’ is true in these contexts, and that is compatible with the additional fact that you cannot change the past. There are, of course challenges to this position (Hasker’s interminible challenges, for instance), but any deeper challenge must go to deflationary (anti-essentialist) views of modality. That’s where these interesting forms of compatibilism come from.

    Reply
  2. Bill Vallicella Avatar
    Bill Vallicella

    By ‘on the cheap’ I meant that the compatibilist gets freedom without having to pay the price of PAP (Princ of Alternative Possibilities) or the price of commitment to the notion that one can be the ultimate source of an action. What the compatibilist does is redefine ‘free’ is such a way that FW and determinism become logically consistent. But to me that appears as nothing but an evasion of the issue. The issue is: am I free either in the UNCONDITIONAL ‘could have doen otherwise’ sense or in the sense of being the ultimate source (unmoved mover if you will) of my action, or both?

    Reply
  3. Philoponus Avatar
    Philoponus

    This is a very well framed argument, Bill. There is nothing in the logic of your presentation that I can take exception to. What I’d like to do is offer a few examples that may suggest that we do hold ourselves and others responsible for our actions without assuming a “libertarian free will (LFW).” I am NOT going to argue in any way that the LFW is a fiction or an illusion or just plain false.
    Let’s prime ourselves by recalling a Frankfurt-style example. These examples show that we are willing to hold someone responsible for his actions even if he could not have done otherwise. Bob is a greedy and venial young man who was wormed his way up to a position of authority in the Trust Department of a big bank. Bob decides to embezzle $10 million from his clients. Unbeknownst to Bob, he is also the victim of a secret CIA programme, code-named Project Zombie, that has implanted electrode in his brain (during his recent hemorrhoid surgery). Bob now has a “controller” who continuously monitors key areas of his brain and can alter his brain states and behaviour at a moment’s notice. The controller learns of Bob’s plan to embezzle and is delighted. He decides to insure Bob carries out the embezzlement by standing ready to alter Bob’s brain as needed. If Bob hesitates in the least, the controller will create in his brain an irresistible urge to steal the money. In the event, the controller needs to do nothing. Bob embezzles the money without the least hesitation. The controller does nothing except insure that Bob could not have done otherwise. Frankfurt says: Bob embezzled the money of his free and greedy will. Granted he could not have done otherwise because of his controller, yet he is still responsible for what he did. I think we agree.
    Now consider the case of John, a happily married and very social surburban husband, a respected contractor, “Man of the Year” according to the Chamber of Commerce. John has a small problem. When his wife is out of town, John brings teenage boys to his home, forces them to have sex, and then murders them, disposing of their dismembered bodies under his house. John has done this 33 times in the last 15 years. Parts of John’s brain don’t work right. His amygdala, which should incapacitate him with anxiety and revulsion as he even contemplates these deeds, is silent. His Nucleus Accumbens instead lights up with pleasure at the thought of torturing and killing. Etc. The cause of Bob’s brain pathologies was almost certainly the years of physical abuse and neglect he suffered as a youth. They literally rewired his brain in a way that creates an unfeeling psychopath. John is a rare example of what Aristotle calls the bestial man. His nature transcends mere vice. John’s full name, by the way, is John Wayne Gacy.
    John is finally caught and brought to trial for capital murder. He pleads, credibly, that given his damaged brain he could not help do the horrible things he did. Child abuse and the damage it inflicted upon his young brain were the ultimate causes of his inhuman behaviour. The prosecution barely bothers to resond to these neurological/psychological claims, arguing only that he did what he did knowingly and with deliberate purpose. The jury quickly convicts and Gacy is sentenced to die.
    The jury is later polled about what they thought about his child abuse and damaged brain defense. The upshot of their answer is it didn’t matter. He did what he did and he knew what he was doing. Guilty, and let’s fry the bastard. I agree. I also agree with Gacy’s claim that he probably could not have done otherwise given the abuse that had damaged his brain. The jury and I both accept these “ultimate” and compulsive causes of his behaviour and still convict him. Don’t say this is illogical, because it isn’t. There is a choice here, a choice that reflects how we as individuals and as a society hold others responsible for they have done. We don’t care about what causes ultimately formed Gacy’s brain and character. Maybe he could not have done otherwise. It does not matter. We hold him responsible and we punish him. This is the choice we have made in that matter of moral & criminal responsibility, and we doesn’t care about “ultimate sources” of his behaviour and whether someone like Gacy could have done otherwise. It just doesn’t matter whether someone like Gacy is an agent with LFW. He is responsible for his deeds.

    Reply
  4. Mike Avatar
    Mike

    “But to me that appears as nothing but an evasion of the issue. The issue is: am I free either in the UNCONDITIONAL ‘could have doen otherwise’ sense or in the sense of being the ultimate source (unmoved mover if you will) of my action, or both?”
    Yes, I took you to be saying something like this. But I noted that not every compatibilist takes this route. Some agree that pap is relevant to freedom. So, they assert all of the following, where L and H are the causal laws and factual history until sometime t, A at t states that some action A is performed at t, [] is logical necessity.
    1. []((L & H) -> A at t)
    2. <>-A at t
    (2) gives the unconditional possibility that the agent performs -A at t. The unconditional possibility in (2) is true because there is a world in which the agent does -A at t and -(L & H) is true. So, the backtracking counterfactual in (3) is true, where the agent can (in a weak sense) render false the conjunction (L & H).
    3. -A at t []-> -(L & H)
    That gets pap satisfied and causal determinism true. The argument is of course more subtle and complicated, but that is the general idea. See Lewis’s, ‘Are we free to break laws?’, Theoria (1981).

    Reply
  5. peterlupu Avatar
    peterlupu

    1) Phil has offered several arguments against Bill’s LFW. One of these arguments is a Frankfurt-style counterexample to the link forged by LFW between moral responsibility and PAP: i.e., that the presence of alternative possibilities (commonly phrased as ‘could have done otherwise’) is a necessary condition for moral responsibility. The general structure of the argument is that there are cases where we hold A morally responsible even when A *could not have done otherwise*.
    2) Frankfurt-Style Examples: (I shall use Phil’s example in this case.)
    PAP: If A is morally responsible for doing X, then A could have *done* other than X.
    (i) Bob decided to do X (e.g., embezzle $X from a bank), where X is an immoral act;
    (ii) There is an Evil-manipulator: i.e., someone arranges things so that if Bob wavers in any way from his intentions to X (embezzle the bank), then Bob will be caused by certain instruments (in this case an electrode implanted in his brain) to X.
    (iii) Bob decides to go through with X on his own;
    (iv) It is not the case that Bob *could have done otherwise* (i.e., refrain from Xing);
    Why? Because if Bob would have decided to refrain from Xing, then the Evil Manipulator would have caused him to X anyway.
    (v) Bob is morally responsible for doing X.
    Therefore,
    (vi) PAP is false.
    3) Wonderful example! There is just one problem with it: it is wrong! The blame for it being wrong falls evenly upon the proponents of PAP as well as those who think Frankfurt-style examples are counterexamples to the intuitions behind PAP.
    4) The proponents of PAP have regularly formulated the principle in terms of the phrase ‘could have *done* otherwise’. This phrase is systematically ambiguous because the word ‘done’ is ambiguous. The ambiguity involves the following two interpretations:
    (I) could have *behaved* otherwise: i.e., engaged in a behavior (non-intentional) that is different from the action actually undertaken;
    (II) could have *acted* otherwise; i.e., intentionally refrained from or acted in a manner different from the action actually undertaken;
    5) Frankfurt-style examples insure that the person could not have *behaved* otherwise (because of the Evil Manipulator). But they do not show that the person could not have *acted* otherwise: namely, in Phil’s example, Bob could have intentionally refrained from embezzling the bank. Of course, if Bob would have intentionally refrained from doing so, then the Evil Manipulator would have insured that Bob would embezzle the bank anyway; but, and this is the crux of the matter, under such circumstances Bob’s behavior would not have been an *action* of his. Bob’s embezzling behavior would have been induced by the Evil Manipulator. Bob’s intentions, volitions, etc., would have had nothing to do with his behavior. Therefore, Bob could have acted otherwise, even if he could not have behaved otherwise. All we need in order to block Frankfurt-style examples is to rephrase PAP as follows:
    PAP*: If A is morally responsible for doing X (in the sense of acting), then A could have *acted* otherwise.
    6) The application of the above argument to Phil’s other cases should be fairly obvious. What is less obvious is whether we *ought* to take the line articulated above or Phil’s recommendations with regard to our legal system, social policy, and even personal conduct. But these matters require a whole new line of thinking.
    peter

    Reply
  6. Bill Vallicella Avatar
    Bill Vallicella

    Thanks, Phil. I am not sure I have a good response to your challenging comment.
    First, the case of Bob. The purpose of your Frankfurt example is to show that moral responsibility does not require that the agent could have done otherwise. Bob could not have done otherwise than embezzle, and yet we feel that he is morally responsible. But we should distinguish between freedom of action and freedom of will. Might it not be that moral responsibility for a physical action like embezzling does not require ‘could have done otherwise’ while moral responsibility for the will to embezzle, or the decision to embezzle, does require ‘could have done otherwise’? In your example, the controller stands ready to intervene after the decision has been made, and only if Bob falters in the implementation of his decision. Your example does not show that ‘could have done otherwise’ is not required for moral responsibility for the decision. Or does it?
    I’ll bet that in the literature this response and been made and someone has come up with a more fiendish Frankfurt counterexample to meet it.

    Reply
  7. Bill Vallicella Avatar
    Bill Vallicella

    Phil,
    Your discussion of the case of John Wayne Gacy is not as clear. We will agree that people like Gacy have to be locked up to protect the rest of us. But I should think that punishment for the sake of retributive justice could not be justified unless Gacy was responsible for his deeds, and for him to be responsible for his deeds, I should think that at least one of the following two conditions would have to be met: (i) he could have done otherwise, or (ii) he (not some diseased potrtion of his brain) is the ultimate source of his actions.
    The question concerns the rational for punishment. If you are saying that punishment in the service of retributive justice is justified even if Gacy’s bad behavior was driven by brain malfunction, then I disagree with you. And if you are saying that Gacy is responsible whether or not his bad behavior is due to brain malfunction, then I don’t agree.

    Reply
  8. Bill Vallicella Avatar
    Bill Vallicella

    Peter,
    Brilliant response. I just now read it after writing my responses to Phil. What you are saying may be related to what I was saying in my first response.
    You point out that ‘done’ is ambiguous as between ‘behaved’ and ‘acted.’ Phil’s example shows that Bob could not have behaved otherwise, but not that he could have acted otherwise. I would add that ‘action’ is ambiguous as between ‘behavior’ and ‘intentional action.’ I distinguished between freedomo f action (behavior) and freedom of will. That seems to run parallel with your distinction between freedom of behavior and freedom of action.

    Reply
  9. Bill Vallicella Avatar
    Bill Vallicella

    Mike writes,
    >>1. []((L & H) -> A at t)
    2. <>-A at t
    (2) gives the unconditional possibility that the agent performs -A at t. The unconditional possibility in (2) is true because there is a world in which the agent does -A at t and -(L & H) is true. So, the backtracking counterfactual in (3) is true, where the agent can (in a weak sense) render false the conjunction (L & H).
    3. -A at t []-> -(L & H)
    That gets pap satisfied and causal determinism true. The argument is of course more subtle and complicated, but that is the general idea. See Lewis’s, ‘Are we free to break laws?’, Theoria (1981). A at t)
    2. <>-A at t.
    O am I missing something?

    Reply
  10. peterlupu Avatar
    peterlupu

    Bill,
    Thanks. I must believe that someone already made these points after Frankfurt’s paper was published. Philosophers from all people should be sensitive to the distinction between action and mere behavior. Unlike mere behavior, the individuation of actions needs to include intentionality or one of its cognates. We know this, among many other reasons, because one and the same behavior could be used to perform several actions simultaneously: e.g., imagine a party attended by both Jane and her ex-husband John. Suppose I really like Jane but despise John and so in an opportune moment I tell Jane knowing full well that John can hear me: “Jane, you look marvelous since you divorced John.” By uttering this one sentence (behavior) I have committed two different actions: (i) I flattered Jane; and (ii) I insulted John.
    Consider the example of someone who during sleepwalking kills someone. We surely do not hold them responsible for the unfortunate outcome. The principal reason for not holding them responsible is because while their behavior indeed caused the unfortunate outcome, their behavior is not identical to any action they have undertaken; the unfortunate outcome therefore was not done by the sleepwalker intentionally.
    There is a deeper question lurking here. What are the conceptual grounds for linking moral blameworthiness/responsibility to *intentions* in the manner you, I, and others think appropriate? One could imagine a concept of blame that is purely causal; e.g., someone might be blamed or in some form held morally “contaminated” just because they happen to be in the proximity of a morally blameworthy action. Similarly, Phil’s concept of responsibility is characterized only in terms of the *causal* outcome of behavior regardless of any considerations pertaining to the antecedent intentions, etc.
    So we certainly can distinguish between a causal conception of moral blame vs. an intentional conception of responsibility.
    I think the Old Testament may be systematically ambiguous between these two notions.
    And this brings me to the connection between your response above to Phil’s Frankfurt-style examples and mine. There is indeed an intimate link, because your distinction suggests that while someone might have freedom of behavior; e.g., they are not physically paralyzed, but they might not have the freedom of will; and vice versa. Therefore, someone might be free to behave in certain ways without their will being free to guide the behavior they undertake. A behavior unguided by the will of the agent is not an action.
    So in Phil’s example we hold Bob responsible because he had the freedom of the will to refrain from embezzling the Bank, even though he did not have the free will to prevent his behavior from embezzling the Bank because his behavior would have been under the control of someone else’s will. That I think is your intuition and I agree with it.
    peter

    Reply
  11. Philoponus Avatar
    Philoponus

    What I hoped to do in giving you the JWG story—all the details are factually correct to my knowledge—was emphasize that holding someone responsible for his actions is not an armchair or theoretical exercise. We make judgment of responsibility everyday, sometimes in venues where the consequences of these judgments are very serious. Twelve jurors heard all the evidence about Gacy’s childhood and brain dysfunction, including some credible claims that he could not have done otherwise when tempted sexually by young boys. The prosecution did not take the bait and begin to argue about “could not have done otherwise.” Their experts concentrated on the issues of whether he knew what he was doing—yes, he planned meticulously—and did he know what he was doing was wrong. Yes, again. Those two points settled the matter in the minds of 12 jurors. Notice that the defense tried to float something like your libertarian concept of responsibility, and the jury would have none of it. In fact the concept of responsibility they brought to judgment was a compatibilist one. I agree with them. I and almost everyone I know judge responsibility without concern for the (mostly unknowable) ultimate sources of behaviour. You may personally want to embrace a more restrictive notion of responsibility. Well, try it and see whether it is practicable. I think you will find you can rarely if ever settle matters about ultimate sources, and so you will end holding almost no one, including psychopaths like JWC, responsible for their horrific acts. In this sense I predict you will find your libertarian notion of responsibility utterly dysfunction. In jury selection, by the way, prosecutors are careful to flag people who hold what the manuals call “deviant” notions of responsibility (and proof) and exclude them for cause. Someone who was worried about whether JWC could ultimately have done otherwise would never have made it on the jury.

    Reply
  12. Philoponus Avatar
    Philoponus

    Peter,
    Your distinction, stated with exemplary clarity, is absolutely correct as a response to classic Frankfurt-style cases. I know your objection and I tried to anticipate it with some nuanced details in the story of Bad Bob. I did not dwell on it, however, because it is, for a few years more at least, a sci-fi scenario, and I think the case of JWC is much more important in showing why we should reject a libertarian notion of responsibility.
    Obviously, if we turn bad Bob into a Zombie-on-command, then in your terms he behaves but does act. But I imagine the controller acting more subtly, so that Bad Bob never even knows someone is “prompting” him to go forward with his plan. The controller watches some key systems in Bob’s brain that create aversive reactions (amygdala, insula, etc). Within a half second or so of seeing any beginning activation in these areas, he strongly stimulates Bob’s nucleus accumbens and Bob’s brain goes awash in dopamine making him feeling happy and confidence about his embezzlement. Bob joyously re-affirms his intention to steal and never feels any reluctantly to proceed. He acts, I suggest, in your terms, though he could not have acted otherwise. What do you think?

    Reply
  13. Malcolm Pollack Avatar
    Malcolm Pollack

    It seems that the question coming into focus is: what are the boundaries of “Bob”? As Dennett remarks in this context: “If you make yourself small enough, you can externalize everything.”

    Reply
  14. peterlupu Avatar
    peterlupu

    Libertarian Free Will, Morality, and Social Order
    In several posts Bill and Phil challenged us to think hard about the relationship between LFW and morality, law, social order, and even our personal judgment and conduct. Bill presented a conception of LFW which requires that an agent could have *acted* otherwise in order for them to be free (PAP) and, therefore, to be held morally responsible. While PAP is only a necessary condition for LFW, it introduces extremely complicated considerations involving the nature of action, their relationship to the mental state of the actor, intentionality, the will, and so forth. Phil, on the other hand, views PAP as irrelevant to moral (and legal) responsibility. In light of Phil’s view, several questions arise: Does Phil think that there is an alternative notion of FW other than LFW which is required by moral responsibility? What is it? Is it an adequate conception of FW?
    1) I wish first to dispense with some arguments from practice raised by Phil in his recent post (May 24, 07:03 AM) as well as in some others.
    (i) Phil argues that the jury in the JWG’s case found him guilty despite evidence that JWG could not have acted otherwise. This supposed to show that PAP played no role in the jury’s verdict. Let us suppose we grant this fact. What does it prove? It certainly does not prove that it should not have played a role. Jurors regularly issue verdicts that are puzzling to the rest of us: we are familiar with this fact in our legal system. It is not perfect. Nothing follows from these considerations regarding LFW.
    (ii) Phil argues that judgments of responsibility frequently demand urgent practical decisions and therefore cannot be hindered by considerations involving PAP and ultimate-agency which are *theoretically* in nature and frequently cannot be discerned in a timely fashion:
    “holding someone responsible for his actions is not an armchair or theoretical exercise. We make judgment of responsibility everyday, sometimes in venues where the consequences of these judgments are very serious.”
    (a) The fact that some circumstances demand urgent judgment about matters of responsibility does not mean that in making such judgments under these circumstances we ignore completely assessing matters pertaining to PAP and ultimate-agency. What it does mean is that if we make such assessments, we do it on the basis of partial information and, thus, risk getting it wrong. So in such cases we sacrifice accuracy in favor of quick response because on balance a quick reaction is more important.
    (b) Even if urgency demands a quick response which compels us to ignore considerations regarding PAP and ultimate-agency, it does not mean that we *endorse* the view that such considerations are irrelevant to the situation. Suppose a doctor in an emergency room must react quickly in order to save the life of a person and cannot get timely access to certain medical information (e.g., whether the person is allergic to certain medication); the doctor administers a treatment without consulting the appropriate medical information thereby risking an allergic reaction. The doctor’s decision in this situation does not reflect his expert opinion that such medical information is altogether irrelevant to administering this particular treatment. On the contrary, the doctor would have preferred to have timely access to such information. But when such access is not practicable and time is of the essence, the doctor must make a judgment under conditions of uncertainty and does makes the most reasonable judgment that fits the circumstances. Similarly, even if practical circumstances demand quick reaction that must ignore considerations of PAP and ultimate-agency, we do not thereby think that they are irrelevant altogether. Rather we think that in certain practical circumstances we must react without full information, even information that is essential to the case. “Thou shall not legislate based upon uncommon cases”.
    (iii) Of course, there is a much bigger point lurking in Phil’s position. The bigger point is that questions regarding PAP and ultimate-agency and their role in moral responsibility belong to that class of philosophical or theoretical questions that are irrelevant to everyday affairs and, therefore, are irrelevant period. This philosophical temperament is really the driving force behind Phil’s other two arguments. Are philosophical questions of a certain kind irrelevant to everyday affairs? And if so, are they futile exercises? I shall leave these questions for a later discussion.
    2) What is the conception, or regulative ideal, behind viewing LFW as necessary for moral responsibility? The regulative ideal is that of an autonomous moral and epistemic agent that
    (i) when faced with the need to act does so based upon considering a variety of available options;
    (ii) takes into account a variety of factors that are relevant to their potential conduct;
    (iii) based upon a sincere and reflective decision making, the agent opts for one of the alternative courses of action;
    (iv) the agent proceeds to intentionally produce the action so chosen.
    I am sure I have missed several component of this picture. But the idea should be fairly clear. Here we have PAP, ultimate-agency, decision making and autonomy. It is something along these lines that we find the libertarian conception of responsibility.
    3) We all fall short of this regulative ideal and in many ways. In light of this harsh fact we might ask: Should a society embrace this regulative ideal in the manner they organize their social institutions, legal system, and norms? Should we as individual be guided by this regulative ideal in our everyday life? I could not conceive of a human society that is worth its name if it were to altogether reject this regulative ideal. But, then, how do we resolve the discrepancy between the ideal and the fact that we all fall short of this ideal? How should society deal with these facts? How should we as individuals deal with these shortcomings?
    4) My answer to the first and second questions is that we resolve the discrepancy in a way we do in any other area where reality falls short of an ideal. We make an effort to acquire as much knowledge as possible about the human condition and how the diversity and individual differences and histories impact human action in ways that depart from our regulative ideal. On the basis of such knowledge we introduce operational mechanisms that attempt to deal with different classes of incongruity from the ideal, recognizing all along that these are instruments of our own making and that they are far from perfect. Therefore, we are in an unending process of reflective equilibrium where new empirical knowledge about human beings and proven deficiencies about our instruments eventually combine to revise or altogether abandon them. At the same time we attempt as a society to create enabling conditions so that individuals are able on their own initiative to inch closer to the ideal of autonomous agency (a project our society seems to be further and further away from implementing).
    5) How shall we view matters as individuals? Simple: The privilege to be viewed as approximating an autonomous agency has a price: the price is that we must treat others as similarly approximating to one degree or another autonomous agency. The “some degree or another” is the tricky part because it may (and in fact does) represent a very wide spectrum from close proximity to the regulative ideal of autonomous agency to a very remote distance. It is those who are situated in this remote distance that we may find hard to tolerate. But we cannot treat them in accordance with a different standard on pain of eventually turning our back upon the regulative ideal we embraced and, hence, change our conception of ourselves in the process. It is not easy to actually live with imperfection when we have a glimpse at perfection. And it is tempting to seek a perfect solution that will wipe out all imperfections. There is no such a thing. All allegedly perfect solutions ended up being some of the worst nightmares in human history. We cannot create good in seven days, years, seven hundred or even seven hundred thousand years. We cannot only take ourselves closer to the good inch by inch, day by day, one human suffering after another.
    6) The above is only a small part of the way I see the questions raised by Bill and Phil here. I let others take the stage from here.
    peter

    Reply
  15. Malcolm Pollack Avatar
    Malcolm Pollack

    Peter, you wrote:
    “2) What is the conception, or regulative ideal, behind viewing LFW as necessary for moral responsibility? The regulative ideal is that of an autonomous moral and epistemic agent that
    (i) when faced with the need to act does so based upon considering a variety of available options;
    (ii) takes into account a variety of factors that are relevant to their potential conduct;
    (iii) based upon a sincere and reflective decision making, the agent opts for one of the alternative courses of action;
    (iv) the agent proceeds to intentionally produce the action so chosen.”
    If we were sending a space probe to Pluto, where it would be beyond the range of timely control from Earth, we would wish to design its onboard computer system so as to:
    (a) identify circumstances that require it to take some sort of action in the service of its mission or its self-preservation;
    (b) generate a collection of possible actions to be considered;
    (c) model, as well as it can, the likely outcomes of each candidate action, based on its current knowledge and past experience;
    (d) choose, from amongst the projected outcomes, the one that seems optimal according to its evaluative criteria;
    (e) produce the action so chosen;
    (f) update its knowledge base so that the result of the action actually taken in the particular context is available to inform future deliberations.
    Nobody here would argue that this space probe instantiates LFW in the strongest sense, but it seems to meet the criteria outlined in Peter’s comment.
    The objection will likely be raised that the probe is a designed object, and thereby, in (iv), piggy-backs on the choices made by its designers. The Darwinian compatibilist/materialist will respond that human brains are designed objects also.
    The objection may also be made that the space probe is not *consciously* reflecting on its options (we assume that the stipulation in (iii) that decision-making be “sincere” is satisfied). The response will be that all this shows is that consciousness, therefore, appears to quite beside the point, and is apparently not required for deliberative choice. (He may also point out that even we humans make unconscious choices all the time, and that even in cases of conscious reflection the actual moment of decision-making seems hidden: we reflect, and then the decision “appears” in our consciousness.)

    Reply
  16. Bill Vallicella Avatar
    Bill Vallicella

    Peter,
    Much depends in a discussion like this on what exactly we mean by an action. You suggest that actions are individuated by the intentions behind them. In the case of ‘Jane, you look marvelous since you divorced John’ there is one piece of lingusitic behavior which expresses two actions since the speaker had two intentions: to flatter Jane and to insult John. How many of the following are true:
    1. Every action is intentional; there are no unintentional actions.
    The typical muscle spasm or burp would then not be an action.
    2. There are both mental and physical actions. Pulling a trigger intentionally is a physical action; deciding to pull a tigger is a mental action.
    3. Intentions are mental actions.
    4. There are basic actions, actions we perform directily, not by doing something else.
    5. Intentions are basic mental actions.
    6. There is no action without an agent.
    7. No action can have two agents. If you help me lift a chair, there are two actions, yours and mine. The lifting of the chair is the product of two actions but is not itself an action. If this idea could be worked out, then Bob’s embezzling is not an action.
    I am not affirming these points, just indicating some of the questions that arise. Action theory appears to be a prerequisite for getting clear about FW.

    Reply
  17. Bill Vallicella Avatar
    Bill Vallicella

    Peter,
    Your “Libertarian Free Will, Morality, and Social Order” makes some excellent points. What you say about regulative ideals is fruitful. We might approach it like this.
    Libertarian moral responsibility (LMR) presupposes libertarian free will (LFW). Ascription of LMR to ourselves and others seems to be necessary to have a society worth living in. I for one do not want to live in a society in which people are punished regardless of their responsibility for the crimes alleged, to take just one example. Or a society in which all agents are adjudged morally on a par because all are just products of their environment. Decent human interaction seems to require the ascription of LMR and with it LFW. Now it may be that only some of us some of the time approximate to LMR. Still, we should take it as a regulative ideal: we should hold ourselves and others LM-responsible. We should not excuse ourselves and others on the basis of environment or upbringing or genetic endowment. That seems to be Peter’s suggestion.
    But what if hard determinism is true? Then the ideal of LMR is unattainable and so (arguably) cannot be an ideal. As I have said before, an ideal to be an ideal must be realizable. (This is a variant of the ‘ought’ implies ‘can’ principle.) What if LFW is simply impossible? Galen Strawson, I seem to recall, has an argument that LFW is impossible because a LF agent would have to be causa sui, which is impossible. If LFW is impossible, then so is LMR. The upshot would seem to be that LMR and LFW cannot be regulative ideals.
    But there is a way out. Saul Slimansky (U of Haifa) argues that hard determinism is true and the LFW is an illusion, but a necessary life-enhancing illusion that we cannot afford to let go by the boards. More on this in a separate post.

    Reply
  18. peterlupu Avatar
    peterlupu

    Malcolm,
    Thanks for your comment and for reminding me (nay, insisting) how important it is to view all of this agency paraphernalia in the context of and presence of consciousness. Without it, none of the clauses I have stated above as characterizing autonomous agency make sense at all.
    Incidentally, I did not consider the clauses I presented as exhaustive or jointly sufficient; nor did I view them as the most eloquent statement of what is necessary. My focus was upon the overall picture. A theory of autonomous agency of the sort I have tried to describe will have to include a comprehensive theory of action and much more besides (see the latest comments by Bill). My own embrace of emergence involves an additional layer of puzzling questions that require an account that would legitimize the notion in the first place. All theories on the table carry IOUs that are pretty hefty.
    This last point takes me to a new and improved version of materialism. Its founder is most likely Dennett and it deserves the title of “anthropomorphic-materialism” (actually it deserves a much crisper title, but I cannot come up with such presently; so this one will have to do). The new materialist strategy is not to identify the mental with the physical (identity theory) or even call for the elimination of mental terminology (eliminative materialism). The new strategy is to appropriate mental terminology–use distinctively mental terms to apply to clearly non-mental items. The strategy is ingenious: “If you can’t beat them, steal them.” By beginning to apply mental terms to non-mental items prior boundaries between these categories begin to erode; new linguistic practices are introduced; and before we know it we will find as completely natural the practice of calling our thermostat an “intentional system” that has an “intentional stand” towards the temperature in our room and being so “intentionally generous” and “intentionally caring for our well being” that it intentionally monitors the temperature to our full benefit.
    (Lately I lost one of these wonderful intentional friends; he became schizophrenic or something, they tell me. They replaced him with a new and improved “intentional system”. The new chap’s “intentional stance” appears to be more efficient [electrical costs are down], but I find the new one less charming and less authentic.)
    However, I am somewhat concerned about the consequences of this anthropomorphic materialism regarding our moral, legal, and social order. It may inadvertently shake our social order and bring about massive unrest and require subsequent changes to our moral and legal system. Undoubtedly, we are facing a new turbulent era of emancipation and “civil rights” struggles. Our current practices of viewing these thermostats, cars, computers, probes, etc., as property, mere slaves, subject to purchase and trade will have to radically change. A new movement will emerge and demand that exploiting them as slaves must change and that we must liberate them, give them civil rights as well as perhaps voting rights.
    It is interesting how the history of ideas took us full circle back to an old tradition, one which materialists and naturalists were then so eager to oppose: namely, anthropomorphic explanations. And here we are now facing the new and improved materialism fully embracing anthropomorphic practices unashamed and in full public view. Interesting!
    A probe, a thermostat, a car, a computer, and my chair are not autonomous agents; they have no intentions, they do not deliberate, think, have desires, feel a wire-ache, feel sad or angry, and they do not act. They are not moral agents and should not be afforded the usual protections of moral and legal laws. Some of their behavior is more complicated than others and chairs do not behave at all. Evolution does not “design” anything; evolution just happens. But, of course, all of these matters are up for grabs in the new and improved era of “anthropomorphic materialism.” These are my views, anyway. If you think otherwise, you are going to have to explain to me why we should not take the kind of moral and legal stand towards them that we take towards us.
    peter
    P.S. Despite the light tone of the present response, I am presenting a serious argument. It is simply that if you want to conceive of these objects as agents, then you must bite the moral and legal bullet (and much more besides) that flows from such a stand.

    Reply
  19. Malcolm Pollack Avatar
    Malcolm Pollack

    Peter,
    First, I must quibble regarding your assertion that “Evolution does not ‘design’ anything; evolution just happens.”
    Surely you would agree that such features of living creatures as a bird’s wing, a snake’s fang, etc, etc. are exquisitely well-formed for the uses to which they are put (flying, injecting venom, etc): uses that further the “interests” of the creatures that wield them. One might say “they just happened”, but this seems absurd; clearly wings and fangs are the result of a process that shaped them in accordance with their utility for the jobs they do. In any meaningful sense of the word “for”, a bird’s wing is “for” flying. Would you disagree with that?
    The question, then, is: what do we mean by the transitive verb “design”? You can insist that to say “X is the product of a process of design” means that “X was created by a mind that consciously intended to create X for explicitly understood and consciously represented purposes of its own”, but this is simply to insist on a particular definition of “design”, and then to leave us in need of a new word for what evolution does – and in need of some other conceptual framework for the origin of all the fantastic machinery of living creatures.
    It seems far simpler to follow William Paley’s intuition at least partway, and acknowledge that something “designed” the bird’s marvelous wing, so perfectly sculpted for flight. It is simply that we need to let go of our (perfectly understandable) pre-Darwinian intuition that all design must be the work of a conscious, teleological agent. My own feeling is that both what we do and what evolution does is nicely captured by the word “design”. But this is really just a quibble about language; surely it is clear to all that *some* generative process gave the bird its superb wings and the snake its lethal fangs. Call it what you like.
    But you are quite right that once we assert that all living things, by virtue of their being products of evolutionary “design” (or whatever term we settle upon), are intentional systems – and that therefore their artifacts, such as a map of Paris, a bee’s food dance, or a thermostat, are too – we must then find a way to draw a line between those that deserve moral consideration (and ascription of responsibility) and those that don’t.
    Is consciousness the right criterion? Perhaps for moral consideration it is; certainly I think most people’s notion of whether it is morally OK to drop a lobster into boiling water, for example, turns directly on whether the lobster can consciously experience its horrible death. On that view we need not worry about a Thermostat’s Rights movement anytime soon.
    But responsibility? We make unconscious choices all the time, and certainly we hold people responsible for them if they lead to misfortune. We also hold people responsible for unconscious omissions, such as the father who forgets to take his infant child out of the back seat of the car on a hot day.
    We also acknowledge many states of “diminished responsibility”, in which the usual deliberative apparatus cannot function normally, despite there being no lack of conscious awareness. The Gacy case, and the “Twinkie Defense” in the Dan White trial, are examples of this, and of course it is also the reason we do not hold children or the severely retarded responsible as we do adults.
    Finally, and most importantly of all: the ascription of responsibility itself can affect the intentional system to which such responsibility is ascribed – *but only if that system is subtle and flexible enough to be so affected*. We know empirically that a normal adult human is such a system (and a thermostat, being far too simple, clearly is not!), so we ascribe responsibility to such people (along with concomitant rights and privileges). Indeed, the ability of such a system to be self-reflectively altered by the presumption of responsibility is the primary criterion for ascribing such responsibility in the first place! The inability to be effectively deterred by such moral considerations, and ascriptions of responsibilty, as would suffice for a normal mind, is the hallmark of the criminally insane.
    So: there are two issues here, moral consideration and legal responsibility. The first, I think, may well require consciousness; the ability to experience subjective suffering – though we have already seen there are difficulties regarding questions of potentiality here, as in the case of a comatose (or even sleeping) person.
    The second requires that the system in question be sufficiently flexible that the very ascription of responsibility itself can affect the system’s actions. Most humans clearly are such systems (though it is hard to draw a bright line between those that are and those that aren’t). Toasters are not.
    As things get more complicated we may find it harder to define the boundaries and “edge cases” (Isaac Simov’s prescient novel “I, Robot” dealt with exactly that difficulty) – but I think this limns, in broad strokes, the answer to your objection.

    Reply
  20. Malcolm Pollack Avatar
    Malcolm Pollack

    Peter,
    One further comment: all that the “new and improved materialism” seeks to do is to decouple consciousness from intentionality. It is not at all anthropomorphic to do so, I think; far more anthropomorphic would be to impute *consciousness* to all the other systems you mention.
    What it does, rather, is to assert two premises:
    1) All intentional systems and artifacts are the product of some process of design.
    2) It is *not* the case, however, that only intentional systems themselves are capable of doing the design work. We have recently learned that there is another process capable of creating design in the world (and therefore intentional systems), namely the mindless Darwinian process of evolution and selection.
    To deny 2) is to impose an arbitrary limitation on the natural world on the basis of a persistent anthropocentric intuition. To adopt such a view seems, I think, empirically false – indeed quite plainly so, given the obvious “aboutness” of all living systems.
    *This*, it seems to me, is the real anthropomorphic error, not that committed by the Dennetian materialist.

    Reply
  21. peterlupu Avatar
    peterlupu

    Malcolm,
    1) Consciousness is a necessary condition for anything to have intentionality. Decoupling the two is changing the very concept of intentionality.
    2) Design requires purpose. Evolution as a natural process cannot involve the concept of a purpose. It is nothing more than a more complicated process not unlike rain and rain is not falling for the purpose of irrigating the trees. It is just falling and thus the trees are irrigated. To think otherwise is to change the meaning of ‘design’ and commit the fallacy of anthropomorphizing.
    3) “We have recently learned that there is another process capable of creating design in the world (and therefore intentional systems), namely the mindless Darwinian process of evolution and selection.”
    Tell me which part of current evolution theory involves creating something for a purpose? No part of evolutionary theory is ineliminably teleological in character. The original Darwinian theory did involve evolution theory in teleological processes, but those elements were now purged from the more modern version of the theory.
    4) “indeed quite plainly so, given the obvious “aboutness” of all living systems.”
    What “aboutness” are you talking about? Are trees about something? Flowers? Insects? except perhaps survival. But, to assert the proposition “all living beings are *about* survival” does not use the term ‘about’ in its normal sense associated with intentionality; namely, object-directedness. Rather the use of ‘about’ here is metaphorical and means that all living beings have a survival instinct. Just like when we say that John is all about fame.
    5) We can introduce by stipulation a new meaning to old words in the service of whatever philosophical position we fancy. And we can even justify such stipulations by exploiting fancy scientific theories for the purpose. But doing so will not change the facts that thermostats do not have intentions; certainly, not in the sense in which human beings do. Thermostats do not intend to maintain temperature in a room like I intend to do the same by opening the windows. Thermostats do not have an idea of a temperature and intend to manipulate the environment for the purpose of maintaining a comfortable temperature in the room because thermostats do not have purposes; they are designed to operate in a manner that causes certain state to take place which might be the goal of a human being. But they do not themselves set these ends and then operate in order to bring about such ends.
    peter

    Reply
  22. peterlupu Avatar
    peterlupu

    Oops,
    Malcolm, Sorry only read the second post and not your first one. Will read carefully both tomorrow morning and respond to both. While we disagree on these matters, I hope we can remain cordial like always.
    peter

    Reply
  23. Malcolm Pollack Avatar
    Malcolm Pollack

    Peter,
    Thanks, and I look forward to your response.
    I will say this, though, about your point 1): if you make this assertion it appears that you must deny that we have an “unconscious mind” in the traditional psychological sense. Surely if we have unconscious cognitive processes (and do you really assert that we do not?) they are still “about” various things. To assert that they are “potentially conscious”, as is sometimes suggested, seems rather desperate; they are not conscious at the time they happen, and are still “about” something *then*. (To say that they are knowably “potentially conscious” *in virtue* of their being about something is simply to beg the question, of course.)
    If the definition of “intentionality” is such as to make it simply a tightly shrink-wrapped synonym of “human consciousness and its artifacts”, then it seems to me the term loses most of its discriminatory utility, and we need a new word.
    Also, it seems that your 1) commits you to the view that all of the signalling done by various living creatures (I am fond of citing, as a particularly vivid example that Bill is probably by now tired of hearing about, the food-dance of the bee.) It seems that 1) forces you to defend one of two propositions:
    (a) The bee’s dance isn’t “about” the food at all (or about anything else, either);
    (b) Bees are conscious.
    Regarding (a): entomologists have learned to decipher the bee’s dance. Surely it is indeed “about” the food. Indeed, why would the bee bother at all, if not?
    This leaves (b). But are bees really “conscious”? Are ants? And there are far simpler creatures that engage in this sort of communication.
    It seems the simplest way out of these difficulties is to pry apart the “forness” and “aboutness” of intentionality from the “subjective awareness” of consciousness. Why should they be so inextricably wedded in the first place? They are very different notions.

    Reply
  24. Malcolm Pollack Avatar
    Malcolm Pollack

    Peter,
    One last remark. You wrote:
    “What “aboutness” are you talking about? Are trees about something? Flowers? Insects? except perhaps survival. But, to assert the proposition “all living beings are *about* survival” does not use the term ‘about’ in its normal sense associated with intentionality; namely, object-directedness. Rather the use of ‘about’ here is metaphorical and means that all living beings have a survival instinct. Just like when we say that John is all about fame.”
    I am referring to the way a pit viper’s heat-sensing organ is “about” the prey it detects, for example. If the snake were not hunting warm-blooded creatures, such infrared-tuned sensors would be useless.
    Even more to the point is “for-ness”. I quite agree with you that the rain is not about, or “for”, anything at all. But the living roots of a tree are very much “for” absorbing water. They were designed – if you don’t like the word “designed”, then we can say “optimized”, or “tuned” – to do so (but “designed or “engineered” seems most accurate to me).
    This is a critical distinction between living systems – which can be seen as having “interests”, even if they have no conscious awareness of them – and the inanimate world. The essential point is that the living systems are unique, and qualitatively distinct from the rest of the world, in that as varying self-replicators, they were brought into existence by the reiterative process of evolution. Rain and rocks weren’t. And all this “forness” and “aboutness” is, I think it is safe to say, quite independent of consciousness.

    Reply
  25. peterlupu Avatar
    peterlupu

    Language, Communication, and Aboutness:
    Malcolm, I shall divide my responses into several posts each focusing on one or two issues. In the present post I shall focus upon language and communication.
    When I speak of a language I mean a natural language (designated ‘NL’; a formal language, ‘FL’; and a notation ‘NT’).
    A. Preliminaries.
    1) While every NL can be (potentially) used for the purpose of communication, not all forms of communication constitute a NL. A NL is characterized by a specific syntactic, semantic, and pragmatic structure. I shall now give as brief a description of standard NL-theory as I can along these lines. My purpose is to distinguish Language and communication and link ‘aboutness’ to language but not communication.
    1.1) The most promising Syntactic Theory of NL currently available is some form of generative grammar in Chomsky’s sense that features certain universal principles and a generative structure.
    1.2) A Semantic theory of NL consists (possibly) of two sub-components: (i) a meaning theory; and (ii) a truth/reference theory. (Note: Some think that (i) is dispensable in favor of (ii); i.e., a theory of truth-conditions (e.g., Davidson), while others (notably Katz) think it is a separate theory.) A meaning theory assigns meaning to complexes on the basis of the meaning of the parts and the manner they are combined. Such a theory will also determine how sentences express propositions. If one adopts a truth-theoretical account of meaning, then the burden of the later will fall upon the truth-theory alone; or else one will reject propositions altogether (Quine, Davidson). A truth-theory will determine the truth conditions of sentences on the basis of the referential properties of the parts and the manner in which they are combined to forge the logical form of wholes.
    1.3) A Pragmatic theory is a theory of the use of language for the purpose of communication. Such a theory presupposes some syntactic and semantic theory and attempts to discern general principles that pertain to contextual features of language in use. Are there general pragmatic principles? I don’t know. There may be! Since pragmatic theory deals with language in use, it will appeal to the communicative intentions of speakers.
    1.4) A FL is a language that is constructed for a certain purpose where the syntactic and semantic rules are stated explicitly and introduced by stipulation. Naturally, the construction of a FL requires at some point the use of or the presence of a NL. A Hybrid language is a language that combines a NL with elements of a FL: e.g., mathematics, logic, physics, statistics, etc. A notation, NT, is a set of marks or symbols introduced by stipulation for an antecedently well-defined purpose (and no other purpose): e.g., a music notation.
    2) A *correlation* is a mapping (any-thing to or onto any-thing) of a symbol, mark, event, object, action, proposition, state-of-affairs, time, space to or onto any symbol, mark, event, object, action, proposition, state-of-affairs, time, space (a special case is when a mark may be correlated with itself).
    2.1) Correlation is not identical to causation: we cannot infer cause and effect from mere correlation. Some correlations indicate an underlying causal connection, others indicate a common cause, and still others do not indicate any causal connection whatsoever. Hence, I shall use the following convention: when I use correlation without any associated additional content I shall use the word ‘correlation-‘; on the other hand, when I use correlation to indicate the presence of potentially additional content, I shall use the symbol ‘correlation+’.
    2.2) A correlation-, or mapping-, should not be confused with any of the following: (i) a meaning relationship; (ii) a referential relationship; (iii) an *aboutness* relationship; (iv) a pointing relationship.
    Examples:
    (a) “Smoke *means* fire” is a common phrase to indicate a correlation+cause between the presence of smoke and the presence of fire together with the potential presence of fire as the *causal agent* of smoke. But smoke is not a linguistic item and therefore it cannot be taken to have the linguistic meaning: i.e., smoke means that-fire. However, one could introduce by stipulation a correlation+ (mapping+) between the presence of smoke and something else (the presence of enemy, or the presence of the one who makes the smoke, etc.,) for some communicative purposes. A correlation+ so introduced *points to* the presence of the correlate, whatever that correlate may be. However, establishing such a correlation+, where the correlatum (say fire) *points to* a correlate (say the presence of an enemy), requires some prior semantic notion such as pointing, reference, means-that etc. Therefore, some rudimentary form of NL must already be present antecedently to introducing such a convention.
    (b) “The rings of a tree refer to (mean, signify, etc.,) its age” is another common phrase. While the rings of a tree are correlated with the years (another artificial construction) of a tree, they do not mean, refer to, signify, or are about the number of years the tree lived. We have discovered that such a correlation- (or mapping-) exists in some cases and we use this correlation- to infer the age of the tree. There may be a complicated causal link between the correlatum (the rings) and the correlate (the age of a tree), in which case we have here a case of correlation+cause. I do not know. But even if such a causal relationship exists, it cannot be taken to be equivalent to a relation of meaning, referential, aboutness, etc.
    2.3) We can construct a FL (or a fragment thereof) that includes a notation system. For certain purposes, we may introduce by an arbitrary stipulation a *correlation-scheme* such that certain symbols are correlated with certain objects (or sets of such) or sounds (in the case of the music notation). Once we create such a correlation-scheme, we can use it (again by stipulation) to serve as a *referential-scheme*. By so doing we convert by a stipulation the correlation- into a correlation+reference. However, it is imperative to understand clearly that (a) the initial correlation- we have introduced by stipulation is NOT identical to the relation of reference; and (b) That the correlation+reference relationship we have so established PRESUPPOSES the relation of reference and, therefore, cannot be taken in and of itself as an account of reference or replace the relation of reference. This confusion is the source (in my opinion) of much of the issues surrounding Quine’s so called “inscrutability of reference” thesis and some of the subsequent controversies regarding reference. But that belongs to a whole other set of issues that I hope I will have the opportunity to explore some day.
    2.4) Correlations can be used for the purpose of non-linguistic communication. Certain facial expressions are correlated with certain moods or emotions. Thus, one can use the correlatum (facial expression, for instance) in order to invoke an association with a particular correlate (say a mood); thus, indicating the presence of a certain specific mood; provided the communicator has reason to believe that the communicatee knows the relevant association. Most of the miss-named enterprise called “body-language” relies upon some such correlation+association, where the ‘knows’ is typically taken to be some form of tacit-knowledge.
    B. Principles:
    (P1) A correlation is nothing but a mapping from one or more set of things into or onto another set of things. By itself a correlation cannot be taken to indicate the presence of any other substantive relationship between the correlated things or sets of such.
    (P2) While all cases of causation, linguistic meaning-relations, referential-relations, signification, aboutness, and pointing involve a correlation (or a purported correlation), not all cases of correlation are cases of causation, meaning-relations, referential-relations, aboutness (purported or actual), or pointing.
    (P3) The ability to use, convey and grasp a correlation- is not identical to the ability to use, convey and grasp a correlation+.
    (P4) The ability to use, convey and grasp a correlation+causation or correlation+association is not the same as the ability to use, convey, and grasp pointing, reference, meaning, aboutness, etc.
    C. Conclusions:
    1) Let me now turn to your example of the “bee-dance”.
    You say:
    (a) “Also, it seems that your (1) commits you to the view that all of the signalling done by various living creatures (I am fond of citing, as a particularly vivid example that Bill is probably by now tired of hearing about, the food-dance of the bee.)” is not about anything at all.
    Note: I had to add the last 6 words to the quotation in order to complete the sentence Malcolm intended to express.
    (b) “…entomologists have learned to decipher the bee’s dance. Surely it is indeed “about” the food. Indeed, why would the bee bother at all, if not?”
    1.1) I maintain that the bees-dance is not a case of pointing, referring, meaning, signaling, or being about anything. To think otherwise is to commit a category mistake. The bees-dance is a correlation- between a correlatum; a certain type of movement, and a correlate; i.e., the presence of food. Perhaps, research might indicate that it is even a form of correlation+ (e.g., correlation+cause or correlation+association). What the etymologists did in “deciphering” the bee’s dance is to identify that a certain motion of the bee is in fact a correlatum for a certain correlate: i.e., that the bee’s movement indicates the presence of a correlation- (or potentially a correlation+cause or correlation+association).
    1.2) Is the bee’s movement a case of a relation of intentionally-signaling-toward-food or a case of signaling-aboutness-toward-food to other bees? Absolutely not: by principle (P4) the use and grasp of signaling (in the semantic sense) or aboutness is not the same as the use and grasp of a correlation- or even a correlation+ (cause or association). There are many cases where correlations- or correlation+causation are used in the animal kingdom. Some animals use smell (or sound) as a correlatum of prey. We can train dogs to use smell as a correlatum for the presence of illegal drugs. We can train them to bark when such a correlation is present. What we do under such circumstances is train the dog to forge a 3-way correlation. The first part of the 3-way correlation: correlatum=smell + correlate=presence of substance; the second part of the 3-way correlation: correlatum = first-correlation + correlate = barking. It will be a confusion to interpret the barking (or other behavior) of a dog upon smelling the presence of an illegal substance as stating, meaning to say, pointing, refereeing, or intending to do any of these things towards the illegal substance.
    1.3) Guess what: I maintain that dogs (and other animals) are conscious animals. My point was that consciousness is a necessary, but not a sufficient condition in order to get reference, aboutness, etc. Even very intelligent dogs cannot comprehend the concept of the most basic form of reference or aboutness; namely, pointing. They sniff your finger when you try to point.
    2) It is an understandable mistake, but nonetheless a mistake, to conflate the concept of correlation with any of the semantic concepts such as pointing, referring, meaning, aboutness, etc. It is understandable because of the following reason: these semantic notions might have *emerged* (Bill, note!) from certain types of correlations in some way yet to be understood. But once they so emerge, they formed a completely different category that is irreducible to the correlations from which they emerged. I have tried to articulate (very briefly I might add) the relationship and the differences between mere correlations and the other semantic notions.
    3) The above represents a first shot at developing this tangled subject of correlations, intentionality (as aboutness), and semantics. The concept of correlation and its role in semantics is an extremely complex and interesting topic. Correlations (known by the name of ‘mapping’) play an important role in formal logic, semantics, model and set theory as well as in all empirical branches. It is a fundamental concept and it is one that is part of the repertoire of many animals. Without it they could not have survived. So it is quite understandable that materialistically inclined people attempt to reduce intentionality and other semantic notions to this fundamental notion. A good case in point is Malcolm’s bee-dance example. But for the sake of science and of intentionality, such a reduction ought to be resisted.
    peter

    Reply
  26. peterlupu Avatar
    peterlupu

    Note:
    Due to time pressure the above post is not as well organized and as thorough as I would like it to be. Many issues and connections that need to be developed were either not done at all or done in a very rudimentary fashion. It can be taken as a blue print.
    For instance, directional correlations are an important class of correlations. Pointing, for instance, might have emerged from a directional correlation between a finger and an object plus something else. Also I have not developed to my satisfaction the full relationship between communication and correlations plus various other properties and relations. Another inadequately developed subject is the connection between correlation and aboutness. Aboutness is an aspect of the intentional which is a directional correlation plus something else. What is the something else? My purpose in the above post was to bring attention to the tendency to confuse semantic concepts with correlations. The rest will have to wait for another time.
    peter

    Reply
  27. Malcolm Pollack Avatar
    Malcolm Pollack

    Thanks, Peter for that substantial response. I’ll be on the road all day; it will certainly take some time for me to read and digest it all, and to make a worthwhile reply.
    Thanks also for the six extra words you edited into my comment – they were there originally, but I somehow must have dropped them as I was pasting the comment together.
    This is (obviously) a difficult and important topic, and there are many areas of contention. I greatly appreciate your patience and thoroughness in addressing my objections, and in sharing your knowledge of the work that has already been done; my only aim is understanding.

    Reply
  28. peterlupu Avatar
    peterlupu

    Malcolm,
    You have a great trip. I my self am going on the road tomorrow morning for a couple of days: hence, the urgency of posting the above today. I appreciate your cordial attitude and I think we both seek understanding even if we end up disagreeing.
    Have Fun
    peter

    Reply
  29. peterlupu Avatar
    peterlupu

    Phil,
    The more nuanced case:
    “But I imagine the controller acting more subtly, so that Bad Bob never even knows someone is “prompting” him to go forward with his plan. The controller watches some key systems in Bob’s brain that create aversive reactions (amygdala, insula, etc). Within a half second or so of seeing any beginning activation in these areas, he strongly stimulates Bob’s nucleus accumbens and Bob’s brain goes awash in dopamine making him feeling happy and confidence about his embezzlement. Bob joyously re-affirms his intention to steal and never feels any reluctantly to proceed. He acts, I suggest, in your terms, though he could not have acted otherwise. What do you think?”
    Let me note that whether or not Bob knows about the Evil Manipulator (EM) does not matter. The only thing that matters is the *fact* that there actually is such an EM who causes Bob to behave in such and such a fashion. That fact already determines that Bob’s subsequent behavior no longer counts as an action. Why? Suppose that John would never could even entertain the thought of embezzling a bank, let alone form a plan and an intention to do so: it goes against his well-entrenched character. Nonetheless, EM can gain access to John’s brain and do the very same thing that he did to Bob. Now, John is going to proceed to behave in a manner that embezzles a bank even though such behavior would never flow from his own character. Once the causal mechanism of manipulation is envisioned, it can be introduced with regards to anyone, anytime, and with respect to any form of behavior regardless of who they are otherwise. Hence, behavior produced by EM is not an action.
    What about a case where the EM causes Bob to form the intention/desire to embezzle which in turn causes Bob to behave in a manner that embezzles. This is I suspect your scenario above. I maintain that under such circumstances what we have here is a case where the EM creates in Bob a desire which disguises as Bob’s own intention (acts like an intention) to embezzle (i.e., phenomenologically it appears to Bob as if he himself formed the desire and hence an intention). Still, since as a matter of actual fact the desire was not formed by Bob and fake-intentions are not intentions, the subsequent behavior is not an act.
    The above raises the following question: What are the proper substitutes for EM? Abused upbringing, social factors, mental or physical diseases can all be taken as naturally occurring EMs. How many substitutes for EM am I willing to accept that legitimately exempts a person from acting on their own free will?
    I do not have a general answer to this question. We have empirically identified selected cases. Some are willing to push the envelop so far that virtually anything could function as an EM. I do not. But where the boundaries lie is a complicated question for which I do not have a principled answer.
    peter

    Reply
  30. Malcolm Pollack Avatar
    Malcolm Pollack

    Again:
    “If you make yourself small enough, you can externalize everything.”

    Reply
  31. Philoponus Avatar
    Philoponus

    Peter,
    Thank you for responding to the case of Bad Bob.
    I agree with you that when we have an evil manipulator who creates intentions in someone and forces him to act out of character, we have zombie-like behaviour, and not what should be properly called action. Those, however, are not the facts in the Bad Bob case. Bad Bob is rotten to the core, and the plan to embezzle is his and his alone. Even if the controller assists him in caring out his plans, I do not agree that he CAUSES Bad Bob to do anything. Let me explain.
    Suppose you hatch a plan to rob a bank to supplement your summer income. Malicious person that I am, I decide to support you and encourage you at every stage of the preparations to make sure you carry out your criminal scheme. If you have any doubts, I am there to persuade you to persist. If you need something—a nice ski mask, for example—I get for you. I loan you my car to drive to bank and toast you out the door with champagne. You go and rob the bank.
    Did I CAUSE you to rob the bank? Well, I’m certainly an accomplice, and perhaps without me being there to constantly encourage you, you would have abandoned your plan. But I don’t think I CAUSED you to rob the bank. The controller in the Bad Bob case acts in a very similar way, but instead of giving Bad Bob a speech of encouragement and some champagne, he gives him a little dopamine to calm his nerves. His control is more reliable because he has access to Bob brain’s. He doesn’t mess around Bob’s plans or intentions at all— he is not wired into the pre-frontal areas where he would need to be to do that. What I say the controller does is assist Bad Bob. Bob is no more a zombie than I am after 5 Fat Tires! Nothing the controller does compromises Bad Bob’s ability to plan and execute his own plan. Bad Bob is acting, not behaving like a zombie. If I give you alcohol and encourage to rob a bank you’ve been planning to rob, YOU ROB THE BANK. You are not CAUSED to rob the bank by me.
    So, Bad Bob is responsible for the embezzlement which he performed with no help from the controller, though he could not have ACTED otherwise because the controller stood by ready to help.

    Reply
  32. Bill Vallicella Avatar
    Bill Vallicella

    Phil,
    You should avoid the word ‘zombie’ in this context because it muddies the waters. ‘Zombie’ is a technical philosophical term and I used it that way some posts ago. It has nothing directly to do with the FW issue.
    Suppose you have an indiscernible twin just like you molecule-for-molecule, atom-for-atom. He is behaviorally, materially and functionally indiscernible from you, except that he is not conscious. This twin is Zombie-Phil. Phil is conscious while Zombie-Phil is not. If determinism is true, then both are deterministic systems. Phil’s being conscious does not entail that he has LFW.
    You write, “I agree with you that when we have an evil manipulator who creates intentions in someone and forces him to act out of character, we have zombie-like behaviour, and not what should be properly called action.” That’s not right if you are using ‘zombie’ as we have been using it and in the manner I just explained. If an EM forces Bob to act out of character, this does not make of Bob a zombie for the simple reason that Bob remains conscious.
    So injecting the word ‘zombie’ just confuses matters.
    I take Peter to be saying something like the following. There is behavior and there is action. Action by its very nature is intentional in the sense that ‘behind’ each action is an intention. Behavior, on the other hand needn’t be intentional. The heliotropism of a plant is a characteristic behavior of some plants, but there is no intention behind it. We can speak of the behavior of chemical agents and reagents without presupposing intentionality.
    Peter’s point is that from ‘Bob could not have behaved differently’ it does not follow that he could not have acted differently.
    But Peter must have the last word on exactly what he means.

    Reply
  33. Philoponus Avatar
    Philoponus

    Hi Bill,
    I am amused but not surprised that “zombies” is now a technical term in philosophy. I am tempted to some snide comments, but I shall restraint myself. Be that as it may, the point I miss to state again is that you cannot dismiss Frankfurt-style examples with a distinction.
    Of course we can distinguish action and mere behaviour that fails to be intentional. What does that have to do with Bad Bob, who acts as he has intended and planned to act for a long time? If the “controller” needs to act, he merely rewards and helps Bob to persevere in the face of some nascent doubts & worries. Alcohol would have the same effect. Peter seems to think that the controller needs to CAUSE Bad Bob to embezzle in this scenario. This seems completely wrong to me.
    There are interesting and complicated issues afoot here regarding action and influence, but arguing that Bob could only have behaved differently just doesn’t work. Bad Bob in the end does exactly what he intends and plans to do, with or without help from his controller.

    Reply
  34. peterlupu Avatar
    peterlupu

    Phil says:
    “…, the point I miss to state again is that you cannot dismiss Frankfurt-style examples with a distinction.
    Of course we can distinguish action and mere behaviour that fails to be intentional. What does that have to do with Bad Bob, who acts as he has intended and planned to act for a long time?”
    1) Phil, I am puzzled. What is the basis of the claim that one “cannot” disarm Frankfurt-style examples “with a distinction”? Is this claim based upon some unique property of Frankfurt-style examples that renders them a-priori immune from refutation based on distinctions? If so, then I for one would like to know the nature of this special and mysterious property that Frankfurt-style examples enjoy. Alternatively, you might be deriving this claim from a more general philosophical position that distinctions cannot play a substantive role in disarming alleged counterexamples to philosophical thesis. If so, then I think the general principle from which you derive this claim is provably false. However, it is possible that what you meant to say is rather something along the following lines:
    “…, the point I miss to state is that you cannot dismiss Frankfurt style examples with” the distinction between actions and behavior.
    The last 6 words I inserted replace Phil’s original words so as to render the claim specific to the distinction I proposed between actions and behavior. This alteration is consistent with and it is inspired by the rest of the quotation from Phil’s post. So now we can restate Phil’s claim as follows:
    Phil’s Claim: It is not possible to dismiss/refute Frankfurt-style examples based upon the distinction between actions vs. behavior because this distinction, while legitimate on its own grounds, is irrelevant to the question of whether Frankfurt-style examples are genuine counterexamples to PAP.
    We shall assume that Phil’s intention was to assert the above claim and so we shall proceed to evaluate this particular objection to my argument against Frankfurt-style examples.
    2) The structure of the debate:
    2.1) PAP: A is morally responsible only if A could have acted other than he actually did.
    There are two clauses to the necessary condition on moral responsibility:
    PAP(a) As a matter of fact A committed act x at t;
    PAP(b) It is possible that A committed act y instead of x at t.
    [where y is not identical to x.]
    2.2) Frankfurt-style examples challenge PAP(b). The examples have the following form:
    (Fi) We stipulate that PAP(a) holds: i.e., an agent performed some act x at t;
    (Fii) We construct a case where PAP(b) fails: i.e., it is not possible that for any act y different from x, the agent performed y at t;
    (Fiii) We judge that the agent is morally responsible: i.e., the agent is morally responsible for performing x at t under circumstances (i) and (ii).
    Claim 1: In order for a Frankfurt-style example to be a genuine counterexample to PAP it must feature F(i)-F(iii).
    Do we agree on that? Suppose we do.
    2.3) Claim 2: It is not possible that F(i)-F(iii) are simultaneously satisfied by any situation conforming to Frankfurt-style examples: in particular I maintain that Frankfurt’s or Phil’s examples fail to do so.
    3) I shall now prove Claim 2. The proof involves three stages:
    3.1) Stage I: The Good Bob’s Case.
    (I shall use Phi’s example as a blue-print for the argument).
    Imagine a good Bob who simply could not embezzle a bank: such an act goes against every fabric of his character. Suppose that unbeknownst to Bob, an Evil Manipulator (EM) gets a hold of Bob’s mind/brain either by hypnosis or by inserting a chip in his head. Now, EM can control Bob’s behavior. Suppose Bob, acting within his character, never formed any intentions or plans to embezzle a bank. Suppose EM manipulates Bob’s mind/brain so that Bob embezzles a bank. Question: Knowing the details of the situation as I just described it, do we hold Bob responsible for embezzling the bank? No, we don’t. Why? Because given the circumstances of the case, we do not consider embezzling the bank by Bob an *act* performed by Bob, even though it certainly was a *behavior* in which Bob engaged. Why don’t we consider the behavior performed by Bob an act of his under these circumstances? Because we know that the behavior in question was not conceived by Bob; he did not form any intentions to proceed with such behavior; and that the behavior in question was caused by the manipulation and control of EM.
    3.1.1) The good Bob’s case is neither a Frankfurt-style example nor is it a case that violates PAP. Why? Because the good Bob case violates F(i): i.e., the good Bob’s embezzling behavior is not a case of action. PAP requires that the behavior considered for moral responsibility must be an action the agent performed. Correspondingly, Frankfurt-style examples require that the behavior under consideration must be an act that the agent intended and carried out this intention F(i). Since in the good Bob case embezzling the bank is not an action performed by Bob, it violates clause PAP(a). Hence, it fails to conform to F(i). We correctly judge Bob not morally responsible for embezzling the bank. So in the good Bob’s case the distinction between mere behavior and action is handy in order to determine that good Bob is not morally responsible. PAP is confirmed.
    3.1.2) Thus, the proponents of PAP as well as Frankfurt-style examples agree on the good Bob’s case.
    3.2) Stage II: The flip-flop Bob case.
    Imagine a case that begins just like Phil’s Bob’s case. Bob on his own forms the intention to embezzle a bank. He plans the details. Unbeknownst to Bob, EM gains control of his mind/brain either by hypnosis or via a chip. EM’s plan is that if Bob changes his mind, then EM will cause Bob to go through with embezzling the bank. Flip-flop Bob changes his mind. He decides against embezzling the bank at t. While this flip-flop is going on with Bob, EM is busy manipulating a couple of other minds. 24 hours after flip-flop Bob changed his mind and failed to proceed with embezzling the bank, EM realizes what is going on. EM activates his control of Bob’s mind and Bob proceeds to embezzle the bank.
    3.2.1) What are we to say about flip-flop Bob’s case? Well, we can certainly say that flip-flop Bob is not as good a person as good Bob. Flip-flop Bob formed the intention to embezzle the bank. So he is certainly responsible for that intention. Flip-flop Bob had the opportunity to act otherwise: he had the opportunity to change his mind and refrain from acting on his original intention to embezzle the bank. Flip-flop Bob took advantage of this opportunity and did act otherwise: he refrained from proceeding according to his original intention to embezzle the bank. Flip-flop subsequent behavior of embezzling the bank was not in accordance with his latest intention. This behavior was fully induced by EM. Therefore, flip-flop Bob’s behavior of embezzling the bank was clearly not an action. Therefore, by PAP he is not morally responsible for this behavior: clause PAP(a) is violated. Likewise, flip-flop Bob’s case is not a Frankfurt-style example because it violates F(i). And we arrive at these conclusions by invoking once again the distinction between mere behavior and action.
    3.3) Stage III: The Bad Bob’s case.
    Now, imagine bad Bob’s case exactly as specified by Phil’s original example. Bad Bob forms the intention to embezzle the bank. EM gains control of bad Bob’s mind/brain either by hypnosis or by way of a chip implanted in his brain. EM’s purpose is to insure that bad Bob is not going to flip-flop like flip-flop Bob did. Learning from the case of flip-flop Bob, EM is fully focused upon bad Bob. EM wants to make sure that if bad Bob shows any signs of flip-flop, EM can immediately activate the controls he implanted in bad Bob’s mind/brain without wasting any time.
    3.3.1) Let us focus on EM a bit. Why did EM implant a control mechanism in bad Bob’s mind/brain? Well, I suppose EM wants very badly to insure that bad Bob goes through with his intention to embezzle the bank? But, now, here is a pivotal question:
    Pivotal Question: Why would EM be so concerned about whether or not bad Bob goes through with the intention to embezzle the bank that EM goes through the trouble of gaining control of bad Bob’s mind/brain so as to insure that bad Bob does go through with embezzling the bank?
    3.3.2) The reason EM is so concerned and the reason he goes through the trouble of gaining control over bad Bob’s mind/brain is precisely *because EM knows that the possibility is open for bad Bob to change his mind and decide to refrain from acting on his intention to embezzle the bank*. The very existence of EM in Frankfurt-style examples is predicated upon the very assumption that the agent could have changed his mind; that he could have acted otherwise; that he had this option.
    3.3.3) Suppose that this assumption is not in force. Suppose that there is some sort of a guarantee that once bad Bob forms the intention to embezzle the bank, he will never change his mind: his personality will not allow him to do so. Given this bit of knowledge, there is simply no reason for the existence of EM and there is no reason for EM to take steps to insure that bad Bob will not flip-flop. Such a possibility is excluded. Thus, the very existence of EM in Frankfurt-style examples represents the fact that and it is an acknowledgement of the very possibility that the agent could change his mind: that the possibility exists that he could act otherwise. And that is all that PAP requires. But, now, if the very existence of EM (or any other equivalent device) represents this acknowledgement, then the very structure of Frankfurt-style examples violates F(ii): i.e., one of the clauses required in order to view Frankfurt-style examples as genuine counterexamples to PAP.
    4) Here is an alternative perspective on the situation. I have argued that the very existence of any controlling device in Frankfurt-style examples is predicated upon the assumption that the agent could have *acted* otherwise, for in the absence of such an assumption, a controlling device would be unnecessary. Therefore, we are justified to assert the following propositions:
    (P1) Any Frankfurt-style example requires the presence of a controlling device.
    (P2) The presence of a controlling device entails that the possibility that the agent will act otherwise exists.
    4.1) However, every genuine Frankfurt-style example also has the property that the controlling device insures that the agent will *behave in conformity with his original intention* even if the agent does change his original intentions. The pivotal issue now is this: how should we interpret the phrase “behave in conformity with”? We have two alternatives: we can interpret this phrase as meaning that the agent could *act* in conformity with his prior intentions or we can interpret the phrase as meaning that the agent could *behave* in conformity with his prior intentions. Of course, these two alternatives correspond to the distinction I have proposed previously. Suppose we proceed to interpret the pivotal phrase as meaning that the agent could *act* in conformity with his prior intentions. Then the following line of reasoning leads to a straightforward contradiction:
    (P3) The presence of a controlling device entails that the possibility that the agent will act otherwise does not exist.
    But, now, (P2) and (P3) entail a contradiction. For according to any Frankfurt-style example we must assume that a controlling device exists. Therefore, from (P2) and (P3) we can derive:
    (P4) the possibility that the agent will act otherwise exists & the possibility that the agent will act otherwise does not exist.
    4.2) Therefore, in order to avoid the conclusion that Frankfurt-style examples cannot be counterexamples to PAP because they are inherently inconsistent we must resist interpreting the pivotal phrase as meaning an act. So we must interpret the phrase as follows:
    (P5) The presence of a controlling device entails that the possibility that the agent will behave otherwise does not exist;
    With (P5) in place, Frankfurt-style examples are perfectly coherent. However, they no longer threaten PAP because with (P5) in place as the only possible interpretation, a Frankfurt-style example is in violation of F(ii): i.e., it is not an example such that it is not possible for an agent to perform any act y different from x at t. We have been forced to admit that no Frankfurt-style example can entail such a conclusion on pain of incoherence. The most a Frankfurt-style example can entail is that an agent cannot behave otherwise at t, which does not violate clause PAP(b) and fails to conform to condition F(ii); therefore, no Frankfurt-style example can be a counterexample to PAP.
    5) The above argument is a proof of Claim 2 above: no coherent form of Frankfurt-style examples can fulfill all the conditions required from an example that is a genuine counterexample to PAP. Any appearances to the contrary either fail to appreciate (P2) or avoid the incoherence by conflating behavior with action. Once the consequences of (P2) are recognized and the distinction between behavior and action is explicitly employed so as to avoid an internal incoherence in Frankfurt-style examples, it become perfectly clear that such examples *cannot* constitute a genuine counterexample to PAP.
    6) There is nothing in Phil’s last response quoted above that challenges any of the elements of the above argument. First, as we have seen, the distinction between behavior and action does help in solidifying our judgments about moral responsibility. Second, Phil’s contention that this distinction has nothing to do with the fact that “Bad Bob, … acts as he has intended and planned to act for a long time” is completely besides the point. No one in this dispute denies that bad Bob acted based upon his prior intentions. The pertinent question was whether Frankfurt-style examples show that he *could not have acted otherwise* and yet he is morally responsible. I have shown that no Frankfurt-style example can be interpreted as simultaneously showing these last two things on pain of internal incoherence. Therefore, Phil has not addressed my original argument, expounded here in more detail. I do not see how the current version of the argument can be resisted. Which premise would Phil, or anyone else, deny? Phil is not denying the distinction between mere behavior and action. He cannot deny the validity of the argument from (P1)-(P3) to the contradictory (P4). He cannot deny the truth of any of these premises. Therefore, the only option left to rescue Frankfurt-style examples from incoherence is to interpret such examples along the lines of accepting (P1), (P2), rejecting (P3), and accepting (P5). But this combination cannot constitute a counterexample to PAP.
    peter
    P.S. The length of the present post is not my responsibility because it was fully cause by some unknown EM.

    Reply
  35. peterlupu Avatar
    peterlupu

    The light version of the argument and another proposal:
    1) As Bill noted in a recent post, my previous argument against Frankfurt-style examples was this. Once we appreciate the distinction between behavior and action, we see that from the fact that the agent could not behave otherwise, it does not follow that he could not have acted otherwise. Therefore, Frankfurt-style examples *are not* counterexamples to PAP.
    2) The argument presented in my recent lengthy post is that the form of Frankfurt-style examples entails both that the agent could not have behaved otherwise and also that he could have acted otherwise. Therefore, the standard Frankfurt-style examples *cannot* be on pain of inconsistency counterexamples to PAP.
    3) But, now, consider the following *deeper* version of a Frankfurt-style example, one which I think Phil might have had in mind in some previous suggestions he made. Suppose that EM does not merely insure that bad Bob behaves in accordance with his original intention; suppose the control mechanism available to EM insures that once bad Bob forms the intention to x (e.g., embezzle a bank), he could not change his mind. Thus, the present version of the example controls bad Bob’s intention forming/changing mechanisms, not merely his behavior. In this case, it appears that bad Bob could not have acted otherwise because once he formed the intention to embezzle the bank, EM insures that he cannot change his mind, even if he wanted to.
    4) What are we to say about this deeper version of a Frankfurt-style example? Well, since bad Bob cannot change his intention, even if he wanted to, it appears that he could not have acted otherwise. If that is the case, then we do have a counterexample to PAP.
    5) My inclination is to affirm the following principle of intentionality (PIT):
    PIT: an intention that cannot be changed by an agent because of mechanisms beyond the agent’s control is not an intention of his even if it was originally formed by the agent on his own volition.
    If we accept something along the lines of PIT, then we can rebut the version of Frankfurt-style example above as follows: while the original intention to embezzle the bank was indeed formed by bad Bob on his own volition, once the intention was so formed, bad Bob became an instrument of EM and therefore the original intention was no longer bad Bob’s intention. Bad Bob’s mental life, not merely his behavior, became an instrument of EM. Therefore, bad Bob’s subsequent embezzling behavior is not one of bad Bob’s actions, even if the very same behavior would have been performed by bad Bob without EM’s interference. Therefore, the *deeper* version of Frankfurt-style example I proposed cannot be a counterexample to PAP.
    peter

    Reply
  36. Bill Vallicella Avatar
    Bill Vallicella

    Peter,
    That’s excellent, and I would like to agree with you, but I have a question about this formulation of yours: “The very existence of EM [Evil Manipulator] in Frankfurt-style examples is predicated upon the very assumption that the agent could have changed his mind; that he could have acted otherwise; that he had this option.”
    Correct me if I’m wrong, but I think all that the Frankfurt-style examples require is that the agent COULD CHANGE his mind, not that he COULD HAVE CHANGED it, and that he could act otherwise, not that he could have acted otherwise.
    Let the act be the decision at time t to embezzle. We agree that the Franfurt examples will not make sense unless it is possible that the agent act (i.e. decide) otherwise at a time later than t. But why do Franfurt & Co. need to assume that the agent who in fact decides to embezzle at t could have not so decided?

    Reply
  37. Bill Vallicella Avatar
    Bill Vallicella

    Peter,
    The “That’s excellent” remark refers to your comment just before the last one which I just became aware of when I refreshed the page. Now in your second to the last comment you write, >>Thus, the very existence of EM in Frankfurt-style examples represents the fact that and it is an acknowledgement of the very possibility that the agent could change his mind: that the possibility exists that he could act otherwise. And that is all that PAP requires. But, now, if the very existence of EM (or any other equivalent device) represents this acknowledgement, then the very structure of Frankfurt-style examples violates F(ii): i.e., one of the clauses required in order to view Frankfurt-style examples as genuine counterexamples to PAP.<< This seems confused to me. You say that all that PAP requires is that there exist the possibility that the agent act otherwise. I don't think that is right. PAP requires this, but it requires much more than this. To quote Frankfurt from his famous JP article from 1969, PAP states that "a person is morally responisble for what he has done only if he could have done otherwise." What PAP requires, then, is that an agent S who does A at t could have done other than A at t (either by doing nothing or by doing something else). This is a stronger requirement than that the agent act otherwise.

    Reply
  38. Bill Vallicella Avatar
    Bill Vallicella

    >>PIT: an intention that cannot be changed by an agent because of mechanisms beyond the agent’s control is not an intention of his even if it was originally formed by the agent on his own volition.<< I find this very plausible. An intention cannot be mine if it is maintained in existence by someone or something else.

    Reply
  39. peterlupu Avatar
    peterlupu

    Bill,
    I do not believe that we have a substantial disagreement here. However, in order to address your sensitivity to my formulations of the relevant counterfactuals, I will rephrase all of them in a manner that is completely neutral by using throughout the device of ‘it is possible that’ with suitable temporal indexes.
    1) Suppose Bob forms the intention at t to embezzle the bank.
    2) Suppose that Bob acts in accordance with this intention and embezzles the bank at a subsequent time t*.
    3) PAP states that Bob is morally responsible only if it is possible that Bob change his intentions and as a result refrain from embezzling the bank at t*; (where ‘refrain’ is taken to be an act, not mere behavior.
    4) You wonder: “We agree that the Franfurt examples will not make sense unless it is possible that the agent act (i.e. decide) otherwise at a time later than t. But why do Franfurt & Co. need to assume that the agent who in fact decides to embezzle at t could have not so decided?”
    5) Suppose that Bob did decide to embezzle at t and acts upon this intention at t*. According to PAP, Bob is morally responsible for embezzling the bank only if it is possible that between the time Bob formed his original intention to embezzle the bank at t and his actually embezzling the bank at t* Bob changes his mind and forms a new intention to refrain from embezzling the bank at t*.
    Do we agree on this formulation of PAP?
    6) If we do, then any genuine counterexample to PAP must show that no such a possibility exists: i.e., that it is not possible that between t, the time Bob formed his original intention to embezzle the bank, and his actually embezzling the bank at t* Bob changes his mind and forms a new intention to refrain from embezzling the bank at t*.
    Do we agree on this point?
    7) Now, in my previous post I have argued that Frankfurt-style examples do not in fact show that the possibility in question does not exist, as long as we carefully distinguish between behavior and actions. Instead these examples merely show that it is not possible for Bob to behave in a manner that does not embezzle the bank but that, under the conditions in which he so behaves, such behavior in itself does not constitute an action nor does the fact that it is not possible for Bob to behave otherwise entails that it is not possible for him to change his intentions and refrain from embezzling the bank at t*. I think we agree on this claim.
    8) My new argument is stronger than the first one. My new argument is not merely that Frankfurt-style examples are not in fact counterexamples to PAP, but that they cannot be on pain of inconsistency. Why? Because the whole point of EM is to insure that if Bob were to be inclined to change his intention to embezzle the bank between t and t*, then the implanted mechanism kicks in, takes over Bob’s behavior, and induces him to embezzle the bank and, thus, go against Bob’s new intention. But since the role of the implemented mechanism according to Frankfurt-style examples is precisely to insure that Bob embezzles the bank even if Bob changes his mind prior to t*, but after t, it follows that Frankfurt-style examples presuppose that *it is possible for Bob to change his mind prior to t* but after t and refrain from embezzling the bank*. But this is just one of the requirements of PAP. But, now, if we were to interpret the presence of the control mechanism in Frankfurt-style examples as entailing in addition that it is not possible for Bob to act otherwise and change his mind between t and t* and, therefore, refrain from embezzling the bank at t*, then Frankfurt-style examples will entail contradictory statements. The only way to avoid such a contradiction is by interpreting the presence of the controlling mechanism as entailing that Bob could not have behaved otherwise at t*, although he certainly could have acted otherwise by changing his intention and deciding to refrain from embezzling the bank at t*. But such an interpretation renders Frankfurt-style examples harmless; they no longer threaten PAP.
    8) In your second you raise concerns about whether my formulation of what is required by PAP is correct:
    “You [i.e., me] say that all that PAP requires is that there exist the possibility that the agent act otherwise. I don’t think that is right. PAP requires this, but it requires much more than this. … What PAP requires, then, is that an agent S who does A at t could have done other than A at t (either by doing nothing or by doing something else). This is a stronger requirement than that the agent act otherwise.”
    I am not sure I see a significant difference between the two formulations. Surely if an agent S could have acted other than A at t, then it is possible for S to act in a way other than A at t. And, conversely, if it is possible for S to refrain from intentionally doing A at t, then S could have acted other than A at t. So the two formulations are equivalent, so far as I can see.
    But, perhaps, I am misunderstanding something here.
    peter

    Reply
  40. Bill Vallicella Avatar
    Bill Vallicella

    Peter,
    >>5) Suppose that Bob did decide to embezzle at t and acts upon this intention at t*. According to PAP, Bob is morally responsible for embezzling the bank only if it is possible that between the time Bob formed his original intention to embezzle the bank at t and his actually embezzling the bank at t* Bob changes his mind and forms a new intention to refrain from embezzling the bank at t*.
    Do we agree on this formulation of PAP?<< I don't think we agree on PAP. PAP is simply the principle that a person is morally responsible for what he has done only if he could have done otherwise. How is what you just wrote a version of PAP? More tomorrow.

    Reply
  41. peterlupu Avatar
    peterlupu

    Bill,
    The question is whether there is a difference between the formulation in terms of ‘could have done otherwise’ and ‘it is possible that’. I maintain that the two formulations are equivalent: the agent could have done other than x at t just in case it is possible for the agent to refrain from doing x at t. I do not see how these are different in meaning. If they are equivalent, then we do not disagree on PAP because the two formulation are equivalent and, therefore, the two formulations of PAP are are going to be equivalent.
    peter

    Reply
  42. Philoponus Avatar
    Philoponus

    Peter,
    Thank you so much for this wonderful discussion of Frankfurt-style examples. As Bill reminds us, this debate is already 40 years old, and I suspect many people still don’t recognize the import of these cases.
    Let us agree that what you call the “deeper version” of Bad Bob is exactly what I have in mind. My view is that Bad Bob and his friends force us to abandon, or shall we say, reformulate the PAP just as you are now doing. I want to emphasize that this is a NEW principle you are now giving us – morally responsible only if he could have changed his mind. The PAP has always been about ACTING differently and Frankfurt-style examples like Bad Bob try to show that someone is responsible though he could not have ACTED differently. I will admit at once that it is not clear that any Frankfurt-style counterexamples pertain to your new principle.
    Besides a new version of the PAP, you also present another new principle, PIT, and that this where I want to turn. I see what you are trying to do with it, turn Bad Bob’s actions into mere behaviour again, but I don’t find PIT credible at all. Whatever do you have to support it? Because at some point I cannot change an intention I have formed does not make it not my intention. I find it hard to understand the claim that it does. Some people become obsessed with certain thoughts. I fall in love with some young lady and cannot control myself in pursuing her. Is it no longer MY intention to possess her? I think it is. Someone takes drugs or alcohol or both and become fixated on doing someone bad. We don’t excuse his misconduct on the grounds that it ceased to be HIS intention to do what he did. A psychiatrist may testify that he could not control himself, but no-one would testify that he did act from HIS intention to do what he did. In general, Peter, I wonder how much inner control I have over any strong intentions I have formed. Maybe little or none, but they are and remain MY intentions.
    In the case of Bad Bob, the EM, as you call him, does not have any way to directly alter Bad Bob’s intentions. Bad Bob’s bad character means that in planning the embezzlement he is getting ready to act very much as he desires to act. The only thing that could stop Bob at the last minute is a jolt of fear, fear of being caught. Fear arises because another part of Bad Bob’s greedy little brain begins to see disaster ahead and sounds the alarm (fear). Let’s simplify and say that the key brain center for fear is the amygdala. Here the EM has one of his sensors, and as soon as he senses the beginning of a fearful arousal, he does someone to stop or drown it. He stimulated the nucleus accumbens, for example, and dopamine floods Bad Bob’s brain and he feels wonderful. The fear that could have stopped him is instantly overwritten with a joyful superconfidence in his criminal plan.
    What the EM does is in effect act as an ally of Bob’s greedy conscious brain and give him the “self-control” over his fear that he naturally lacks. The EM does nothing to manipulate his thinking. Bob is throughout the whole experience conscious of only acting as he wishes to act with confidence. Because the EM helps quiet Bob’s nascent fears, there is absolutely no reason for declaring that Bob is not ACTING AS HE WISHES & INTENDS TO ACT.
    I think PIT is implausible. It would “alienate” a great deal of our inner life, as I suggested about, and I don’t like the general idea that what I cannot change or lose control of overautomatically becomes not mine. Very counterintuitive. Bad Bob and the EM act in concert. If PIT condemns acting in concert as “mere behaviour” that is just another reason to reject PIT.

    Reply
  43. Philoponus Avatar
    Philoponus

    Here’s an exercise that may sharpen our intuitions in the case of Bad Bob. Imagine yourself sitting as a juror in his embezzlement trial. At first everytihng seems straightforward and normal. Bob has confessed. Bob has a record of criminal schemes. Etc. Then a surprise defense witness takes the stand under a grant of immunity. A CIA agent admits he secretly had electrodes implanted in Bob’s brain and did in fact stimulate the pleasure centers in Bob’s brain when his sensors detected nascent amygdala arousal (fear). The pleasurable sensations that Bob experienced blocked out any anxieties and let Bob feel extremely good about his criminal scheme and carry it out without hesitation.
    Does the CIA agent’s testimony, which we shall accept as fact, exonerate excuse Bad Bob’s embezzlement? I see no legal grounds for this defense succeeding. Bob was not induced or tricked into doing something he did want to do. He chose and planned his crime unassisted. He knew what he was doing and knew it was wrong. The electrodes did not impair him cognitively in any way. The electrodes merely gave him the self-control (over his fears) that he did not have naturally. The electrodes merely helped him act the way he wished to act. The CIA agent was his accomplice in the crime, to be sure, but I see no grounds to excuse Bad Bob. Do some of you have different intuitions in this case? ,

    Reply
  44. peterlupu Avatar
    peterlupu

    An Objection against Frankfurt-style Examples,
    In this post I will state my objection against Frankfurt-style examples in as a concise form as I can.
    1) Frankfurt-style examples are intended to be counterexamples to PAP.
    PAP: A person S is morally responsible for intentionally doing X at t only if S can intentionally refrain from doing X at t.
    Assumption: I assume that intentionally refraining from doing X is identical to intentionally doing some Y, where Y is not identical to X.
    2) A Frankfurt-style example is going to be a genuine counterexample to PAP just in case it entails the antecedent of PAP as well as the negation of its consequent: i.e.,
    (I) S is morally responsible for doing X at t;
    (II) It is not the case that S can intentionally refrain from doing X at t.
    3) The following are two claims I shall prove:
    Claim 1: Any Frankfurt-style example that is interpreted to entail (I) and (II) is inconsistent: i.e., it also entails the negation of either (I) or (II).
    Claim 2: Any Frankfurt-style example that is interpreted as entailing (II*) instead of (II) is consistent but is not a counterexample to PAP.
    (II*) It is not the case that S can behave in a manner other than X at t.
    But, (II*) is not the negation of the consequent of PAP. Hence, (I) and (II*) do not refute PAP. Therefore, standard Frankfurt-style examples are either inconsistent or they are not genuine counterexamples to PAP.
    4) Suppose a Frankfurt-style example (choose your favorite example) entails (I) and (II).
    4.1) Then such an example includes a backup mechanism that is capable of directly causing S to do X at t in the event S intentionally refrains from doing X at t. But the very existence of such a backup mechanism entails
    (III) S can intentionally refrain from doing X at t.
    Because if (III) were false and S could not intentionally refrain from doing X at t, then there would be no need for a backup mechanism.
    4.2) (II) and (III) are contradictories.
    4.3) Therefore, the assumption stated in (4) must be false.
    4.4) This proves Claim 1 above.
    5) Suppose that a Frankfurt-style example entails (I) and (II*).
    5.1) As before, such an example entails (III) as well. But, now, notice that as long as we maintain a sharp distinction between behavior and action, (II*) and (III) are perfectly consistent.
    5.3) Therefore, the supposition stated in (5) does not lead to a contradiction. Frankfurt-style examples that entail (I), (II*), and (III) are perfectly consistent.
    5.4) But notice that none of these propositions; i.e., (I), (II*), and (III) contradict the consequent of PAP. Therefore, consistent Frankfurt-style examples are not counterexamples to PAP. This proves Claim 2.
    6) Since this holds for any arbitrary Frankfurt-style example, we can state the following:
    (*) Every Frankfurt-style example is either inconsistent or it is not a counterexamples to PAP.
    7) The only potentially vulnerable move that I can see in this argument is the claim that the existence of a backup mechanism entails (III): S can intentionally refrain from doing X at t.
    7.1) But, how can the proponents of Frankfurt-style examples deny such an entailment? The very point of Frankfurt-style examples is that the existence of such a backup mechanism (however it is described) is feasible and that its sole purpose is to insure that in the event S intentionally refrains from doing X at t, then the backup mechanism induces S to do X at t. Thus, the rationale of such a backup mechanism presupposes that S can intentionally refrain from doing X at t; but this just is (III). Hence, any Frankfurt-style example entails (III).
    peter

    Reply
  45. Philoponus Avatar
    Philoponus

    Peter,
    Thank you for restating your argument. It is still (necessarily) a little compressed, but I think I see where I dissent. Nothing about the EM or how he acts entails (III). Bad Bob can embezzle with or without the EM’s help. There are alternative way the same criminal plan can be brought to fruition, but either way Bad Bob happily and deliberately embezzles. Without (III) your argument doesn’t work. I also reject, for reasons I have explained in several posts, your view that the EM CAUSES Bad Bob to do anything. If Bad Bob is the EM’s puppet, then we have behaviour, not action. But I have been energetic in explaining why helping Bob to do what he wants to do is not causing him or coercing him to do anything.

    Reply
  46. peterlupu Avatar
    peterlupu

    Phil,
    I have anticipated and responded to the line of criticism in your last post. Nevertheless, here are a few comments.
    1) You are correct that without III my last argument does not go through. My previous weaker argument, however, still holds even if one rejects III because it only depends on the distinction between behavior and action.
    2) “Nothing about the EM or how he acts entails (III). Bad Bob can embezzle with or without the EM’s help. There are alternative way the same criminal plan can be brought to fruition, but either way Bad Bob happily and deliberately embezzles.”
    The fact that bad Bob can embezzle and even that he intentionally does so, or that there are “alternative ways” the same result of embezzling can be intentionally accomplished is simply irrelevant to my argument. The only relevance it has to Frankfurt-style examples is to entail that bad Bob was indeed morally responsible. But, establishing that bad Bob is morally responsible is only half of what a Frankfurt-style example must demonstrate in order to constitute a counterexample to PAP. The other part Frankfurt-style examples must entail is that it is not the case that under the circumstances bad Bob could have acted otherwise. This brings me to the first sentence I quoted above and to the crux of your objection.
    3) You claim: Nothing about EM entails III.
    If so, then let us inquire how exactly does it follow from a Frankfurt-style example that bad Bob could not have behaved (or acted) otherwise?
    I suppose the right answer here is this: the examples are set up so that a control mechanism is implanted in Bob. Now comes the pivotal question:
    What exactly is the purpose of the control mechanism?
    The only answer I can think of that is consistent with the Frankfurt-style examples is this: the purpose of the control mechanism is to take control of Bob’s behavior *in case he changes his mind and decides not to proceed with embezzling the bank*. This answer, however, presupposes that Bob can intentionally change his mind and decide to refrain from embezzling the bank. And this in turn entails that he can refrain from embezzling the bank (i.e., III). Of course, were that to happen, the control mechanism will take over and induce Bob to embezzle the bank anyway. But causally inducing Bob to embezzle the bank establishes only that he could not have behaved otherwise; not that he could not have acted otherwise. That is how Frankfurt-style examples profess to produce a counterexample to PAP.
    Since the very set up of Frankfurt-style examples presupposes III, it entails it. The only way to avoid this consequence is to give a different answer to the pivotal question above: i.e., give me an answer to this question that does not presuppose that bad Bob could have intentionally refrained from embezzling the bank. If you can do that within the parameters of Frankfurt-style examples, then my latest argument collapses.
    4) “I also reject, for reasons I have explained in several posts, your view that the EM CAUSES Bad Bob to do anything. If Bad Bob is the EM’s puppet, then we have behaviour, not action.”
    We agree on the last sentence, but we disagree in many ways about the first.
    I did not say that EM causes bad Bob to embezzle the bank because I did not need to say any such thing. Ex hypothesis, and by the very set-up of Frankfurt-style examples, Bad Bob does proceed to intentionally embezzle the bank. This point is not questioned by ANYONE. The issue is this: Does the presence of EM prove that bad Bob could not have intentionally refrained from embezzling the bank? And if so, how does the presence of EM achieve this task?
    The answer to the first question is this: EM’s purpose is to insure that if bad Bob were to change his mind and decide to refrain from embezzling the bank, then EM *would cause him to do it anyway* (go against bad Bob’s change of mind). And my point in the first argument was that since EM achieves this result by a purely causal mechanism, the presence of EM only proves that bad Bob could not have behaved otherwise. From this it does not follow that he could not have acted otherwise. Hence, Frankfurt-style examples do not refute a properly formulated PAP (i.e., in terms of action, not mere behavior).
    5) Now, this argument addressed your original example and it is the original set up of Frankfurt-style examples. In these examples EM is set up so as to cause the agent to behave in a certain way in the event the agent reneges upon his original intentions to do such-and-such. The description of these examples are not mine. They are yours and Frankfurt’s; I simply respond to the details of the example. If you decide to change the examples, then I shall respond to the new one.
    6) “But I have been energetic in explaining why helping Bob to do what he wants to do is not causing him or coercing him to do anything.”
    I agree. But, now, how does an example which substitutes *causally induce Bob to embezzle the bank* with “helping Bob embezzle the bank” constitutes a counterexample to PAP? After all, if you “help” Bob by persuasion etc., and that contributes to his actions, then according to PAP he still could have resisted your persuasions and “help” and thus refrain from proceeding with the embezzlement. This scenario is not going to produce a case whereby Bob is both morally responsible and could not have acted otherwise. So I do not see how replacing “persuasion” or “help” for “cause” salvages Frankfurt-style examples.
    peter

    Reply
  47. Charlie Payne Avatar
    Charlie Payne

    Professor:
    Greetings from Hershey, PA. I found your blog while looking for info on Nietzsche. I got into an arguement with a Phd. friend of mine who is a fairly thorough-going Nietzscheian (as best as i can tell). It started with my notion that we’re responsible for our own actions; not victims of our own urges; despite feelings can choose our response (ie S. Covey, Seven Habits of highly Effect. People). This seemed completely unsophisticated to my friend; “un-interesting” way to see human behavior.
    Not having much mileage with the German, my question is: If N is basically correct (he seems to be saying we can make truth what we want) Wouldn’t N have to admit that ultimately his position doesn’t matter? N could be correct, but if I can decide my own truth, can’t I just completely disregard him, even if he’s logically correct?
    I’m not a philosophy student (as i’m sure you can tell), but N seems to lead to complete hopelessness. Just wondering. Need to spend more time on your blog, which i’ve enjoyed.
    You live in tremendous part of the US; i go to Phoenix on biz a couple times a year; riden a bicycle through the entire AZ. Unbelievably spectacular.
    Thanks,
    Charlie Payne

    Reply
  48. peterlupu Avatar
    peterlupu

    Charlie,
    1) “This seemed completely unsophisticated to my friend; “un-interesting” way to see human behavior.”
    First, respond that it may be “uninteresting” but true. Then ask him to tell you which view he finds “interesting”; and in what sense of “interesting” is that view more “interesting” than the one you proposed.
    2)Second, ask yourself whether N can hold the position you ascribe to him regarding the very proposition that expresses it; namely, this proposition:
    “we make up all truths any way we want.”
    If N holds that this proposition applies also to itself, then he made up this proposition too and, therefore, it does not represent the way things are;
    if this proposition is exempt and N did not make it up, then there is at least one truth N did not make up. But, if there is at least one truth N did not make up, then perhaps there are others.
    peter

    Reply

Leave a Reply to Philoponus Cancel reply

Your email address will not be published. Required fields are marked *