Derk Pereboom (Cornell) is "presenting" his paper "A Compatibilist Account of the Beliefs Required for Deliberation." The two invited commentators are Joseph Campbell (Washington State University), and Dana Nelkin (University of California--San Diego).
Download derk_pereboom.pdf
Download cambells_commentary_on_pereboom.pdf
Download nelkins_commentary_on_pereboom.pdf
Thanks to Dana and Joe for excellent and challenging comments. Let me first address several of Dana’s points. Dana correctly points out that my characterization of deliberation is in certain respects quite broad: an active mental process whose aim is to choose what to do from among a number of mutually incompatible alternatives. In particular, it does not mention the evaluation of reasons, or the aim of finding the best thing to do – or even a good thing to do. It’s unlikely that there is a canonical characterization of what it is to deliberate, but one worry is that the broadness of the characterization could result in my not talking about the same thing as some other participants in this discussion. I’m open to narrowing the definition, and it’s important for me to watch out for places where I might not be talking about the same thing as some of the other discussants.
On whether deliberation might have a function in a case in which the agent is already certain of what she will do, I agree that she might be certain of what she will do without having yet formed the intention to do it, and at least in one sense, without having decided to do it. As Dana points out, if deliberation’s aim is finding out what one will do, then being certain rules out a process with that aim. But it might be that deliberation still has a role in the agent’s forming the intention to do it. Perhaps, but my thought is that, setting aside my definition, nothing that we ordinarily think of as deliberation would have the role of producing an intention to do A in a case in which the agent is already certain that she will do A. For example, it doesn’t seem to me that consideration of reasons for and against in this situation would have the role of producing an intention in such a case. Rather, at least typically, the agent would simply just form the intention, independently of further deliberation.
On (DE) I am open to Dana’s suggestion that one can deliberate about whether to do A while only believing that deliberation might be effective. In my definition, there is already some similar wiggle room: I say “then (under normal conditions) she would also, on the basis of this deliberation, decide to do A1, and do A1…” For instance, perhaps this needs to be weakened to: “then (under normal conditions) on the basis of this deliberation, she would decide to do A1, and she might do A1…” since the rational agent would be thinking that under normal conditions her choice would not be impeded, but her action might be.
Should the attitudinal condition be weakened from belief to rational commitment, perhaps analyzed as Dana suggests: “if the agent reflected on it, she would believe that her alternatives are open and her deliberation efficacious.” I was thinking of belief (without saying so) in the way Coffman and Warfield (2005, 27) specify, as dispositional, and characterized more precisely as ‘S dispositionally believes p iff acceptance of p is accessible to S by way of a non-generative cognitive retrieval process.’ Maybe Dana’s rational commitment is weaker than Coffman and Warfield’s dispositional belief, since in the case of explication of rational commitment, it’s open that there is no actual belief to retrieve; instead the belief might first come to be by way of agent’s potential but non-actual reflection. Is mere rational commitment in this sense by contrast with dispositional belief, sufficient for rational deliberation? One possibility is that our notion of rational deliberation is not sufficiently determinate to allow us to decide the issue.
Both Joe and Dana suggest competing deliberation-compatibilist accounts. On Joe’s proposal, I definitely agree that the van Inwagenian non-epistemic principle
VIPD**: A person S deliberates about whether to perform exactly one action among a set of mutually excluding actions A1…An only if S believes that it is within S’s power to do each of A1…An
is easier on the eye than the epistemic conditions (S) and (Settled) and ((DE). One motivation for going epistemic is the threat of the consequence argument against ‘within S’s power to do each of A1…An,' or ‘S can do each do each of A1…An,’ or ‘S has the ability to do each of A1…An’ on metaphysical readings of these phrases, given determinism. Even if the consequence argument can be turned back, the compatibilist analysis of ‘within S’s power to do each of A1…An’ (or similar phrases) needs to be of the right sort to fit the salient examples. If the analysis of such notions is a conditional one, then, as I argue in section 4, Kapitan’s point is salient: openness is plausibly a categorical and a not a conditional notion. The idea is that to rationally deliberate, I must believe that each of several options for what to do must in some sense be possible for me to secure, but not merely on the supposition that some condition is satisfied. Epistemic accounts derive plausibility from the fact that we are often not clear on whether in our thoughts ‘possibility’ and related notions are metaphysical or epistemic, and so in the beliefs that accompany deliberation, the epistemic readings are credible. In addition, even though (S) and (Settled) and (DE) are not as easy on the eye as VIPD**, I think that it’s not implausible that rational deliberators will have the corresponding dispositional beliefs, or at least the corresponding rational commitments in Dana’s sense. I also agree with Dana that a simpler account would be better, all else being equal, and I think that her explanatory nexus view is an interesting contender. Let me reserve comment on that proposal for another time.
Posted by: Derk | May 21, 2007 at 01:21 PM
Still thinking about Joe’s subtle and challenging comments, but I thought I’d add a bit to Derk’s response to them for now. I think Joe is right that none of the cases he mentions are counterexamples to Van Inwagen’s principles as Joe sets them out. But, taking him up on one of his challenges, I believe that there are still some reasons to reject the principles (at least when they are read in ways that support the claim that deliberation requires “deliberation-incompatibilist” beliefs).
Derk notes some of them in his response, and here’s one more related thought. As Van Inwagen notes, there are and have been lots of deliberating determinists. (Actually, Van Inwagen says something similar, but not quite the same, namely, that there are deliberators who sincerely deny the existence of free will, understood by Van Inwagen in a sense that he takes to be inconsistent with determinism.) Now these people are not obvious counterexamples to Van Inwagen’s principles, because one might say that they simply have inconsistent beliefs—in virtue of being deliberators, they believe they are free (or undetermined), but in light of their evidence, they believe that they are not free (or determined). And, as Joe points out, Van Inwagen seems to suggest that inconsistency isn’t so rare or bad, especially if you just can’t help believing something false. (Compare Van Inwagen’s Japanese astronomer case.) But even if deliberating determinists aren’t counterexamples to Van Inwagen’s principles, they do seem to pose a challenge for them. For if there is a plausible way of avoiding the attribution of inconsistency to otherwise rational people who don’t appear to be involved in believing inconsistent things when they acquire a certain sort of evidence, then that seems to be the preferable choice. So a lot hinges on whether alternative belief conditions can do all the work the VI principles do when it comes to deliberation.
Posted by: Dana Nelkin | May 22, 2007 at 12:12 PM
Great question, Dana!
I don't think that the problem that you note arises. For one thing, I don't think that van Inwagen's principles entail that deliberating determinists hold inconsistent beliefs. (Note that van Inwagen says that his argument is similar to Richard Taylor’s – which does attempt to reach this conclusion – but that “it is importantly different from Taylor’s argument” (154).)
There is a temptation to suppose that, in light of van Inwagen's principles, the situation for a deliberating determinist is a lot like the situation for the person in van Inwagen's locked room example, where one "is in a room with two doors and that he believes one of the doors to be unlocked and the other to be locked and impassable, though he has no idea which is which." Van Inwagen suggests that in such a situation one cannot deliberate "about which door to leave by." (1983, 154)
Here one cannot deliberate because it would be in violation of
VIPD**: A person S deliberates about whether to perform exactly one action among a set of mutually excluding actions A1…An only if S believes that it is within S’s power to do each of A1…An.
It is NOT the case that the person in the locked room believes that it is within his power to leave by either door. He believes it is only in his power to leave by one of the two doors yet he doesn’t know which.
But this is not the situation if determinism is true. Suppose I'm in a room with two doors but that I believe that both doors are unlocked. I'm also a determinist. Thus, I believe that if I open door 1, that I opened door 1 was entailed by some suitable conjunction of true propositions. And I believe that if I open door 2, that I opened door 2 was entailed by some suitable conjunction of true propositions. But – if I am a compatibilist – I might also believe that if I open door 1, it was within my power to do so, and if I open door 2, it was within my power to do so. If I am a compatibilist, then I might also believe that whatever I do is within my power even if determinism is true. In this case, I don't see any inconsistency and I don't see any violation of VIPD**.
The difference is that in the locked room example, I might try to open door 1 and it might turn out to be locked. Believing that possibility is what gets in the way of my deliberating about door 1, for I do not believe that it is within my power to open a door if it happens to be locked. The compatibilist determinist, though, need not hold a similar belief.
Of course, if I were both a determinist and an incompatibilist, then matters would be different. But this is just van Inwagen’s point that deliberation about two or more mutually excluding actions is inconsistent with the belief that no one has free will, which I would believe were I both a determinist and an incompatibilist.
Posted by: Joe Campbell | May 22, 2007 at 09:16 PM
Here’s a comment of Joe’s that I haven’t yet addressed:
“According to VIPD*, no one can deliberate unless she thinks that she has the power to perform certain actions, that is, unless she thinks that she has free will in a sense that is amenable to one who endorses the source view. Furthermore, we often deliberate between actions and this suggests, according to VIPD**, a belief in free will in a sense that is amenable to one who endorses the leeway view. Derk thinks that deliberation is possible but he also asserts that no one has free will in either sense. This is a problem that requires an explanation. According to van Inwagen, those who deny free will, like Derk, have two choices: either they “do not mean what they say” or they are “inconsistent” (1983, 154). So I ask: Which is it? Giving Derk the benefit of the doubt, I might ask: Why is van Inwagen wrong?”
My proposal, the (S) component in particular, has it that an agent who rationally deliberates about whether or not to do A, must believe that she might or can, in an epistemic sense of ‘might’ or ‘can,’ do A, and that she can in this sense refrain from doing A. These beliefs are consistent with the agent’s belief that she lacks source and leeway freedom. She might think: “I do not believe that I can, in a metaphysical sense of ‘can,’ do A and refrain from doing A, I don’t believe that I have the power to do A and refrain from doing A, but still I believe that I can, in an epistemic sense of ‘can’, do A and refrain from doing A – that it’s epistemically possible that I do A and refrain from doing A.
Let’s agree that when a rational agent deliberates about whether to do A or refrain from doing A, she will believe that it’s possible for her to do A and refrain from doing so. One philosophical interpretation of these beliefs is that the notion of possibility employed is metaphysical, and this would give rise to Joe’s ‘power’ interpretation, as specified in the VIPD conditions. But it is very often unclear whether the notion of possibility employed in a belief is metaphysical or epistemic. It may turn out (1) that for most rational agents, the notion of possibility that figures into these beliefs is epistemic. If that’s so, then an epistemic condition such as (S) captures the actual openness beliefs that most rational deliberators have. Or it may turn out that (2) that most rational agents have beliefs that employ a metaphysical notion of possibility. If that’s so, then the advocate of (S) might argue (2a) that while these agents have the metaphysical-possibility beliefs, they typically will also have the epistemic-possibility beliefs, and the latter is all that’s required for rational deliberation by way of beliefs in openness. Then, while most rational agents, while deliberating, actually have beliefs that would conflict with a belief in determinism, all that’s required for rational deliberation are openness beliefs that they also have that would not so conflict, to which they could retreat. Or else, the advocate of (S) might argue (2b) that while most rational agents have the metaphysical-possibility beliefs, they do not also have the epistemic-possibility beliefs, but adopting the epistemic-possibility beliefs is all that’s required for rational deliberation by way of openness beliefs. Then, while most rational agents, while deliberating, actually have beliefs that would conflict with a belief in determinism, all that’s required for rational deliberation are beliefs that they do not have, but might adopt, that would not so conflict. I would like it best if (1) were true, (2a) is my second choice, and (2b) my third.
A question for Joe: what’s your compatibilist account of ‘S has the power to do A’?
Posted by: Derk | May 23, 2007 at 10:49 AM
Correction to the last post: In the third and fifth to last sentences, I should have said something like:
Then, while most rational agents when deliberating actually have beliefs that might or might reasonably be thought to conflict with a belief in determinism ...
I don't want to suppose that compatibilist metaphysical versions of openness beliefs, such as those Joe favors, are ruled out.
Posted by: Derk | May 23, 2007 at 04:43 PM
Derk writes:
"Dana correctly points out that my characterization of deliberation is in certain respects quite broad: an active mental process whose aim is to choose what to do from among a number of mutually incompatible alternatives. In particular, it does not mention the evaluation of reasons, or the aim of finding the best thing to do – or even a good thing to do. It’s unlikely that there is a canonical characterization of what it is to deliberate, but one worry is that the broadness of the characterization could result in my not talking about the same thing as some other participants in this discussion. I’m open to narrowing the definition, and it’s important for me to watch out for places where I might not be talking about the same thing as some of the other discussants."
Characterizing "deliberation" in this way seems to make contact with the van Inwagen/Taylor discussions about what deliberation presupposes. You say you're open to "narrowing" the definition, where one "narrowing" would be to require that "evalulation of reasons" be involved in deliberation. But this change would drift out of contact with the PvI/Taylor discussion -- they think of deliberation as something that happens *only after* the reasons for and against the various alternative courses have been evaluated. There are, to be sure, other notions of deliberation worth talking about, and as Coffman and I noted in our article that you kindly refer to, the PvI/Taylor notion is odd in many ways. But drifting away from it will leave you out of contact with their concerns.
Nothing wrong with that, of course: one might well be interested in whether other senses of "deliberation" presuppose a 'belief in ability' or other notions. But giving a negative answer on this issue won't be or imply *any* conflict with the Taylor/PvI thesis which is, after all, a thesis about what their notion of deliberation presupposes and not what some alternative notion of deliberation presupposes.
I hope to say a bit more about the substance of both the paper and good commentaries later.
Posted by: Fritz Warfield | May 23, 2007 at 05:20 PM
Derk,
Just a quick point since I have to pick up my son from school; I'll elaborate tomorrow.
I think that van Inwagen's original argument was intended to be neutral with regard to any particular analysis of 'free will' or 'can.' I intended to do the same. I don't want my considerations to hinge on acceptance of the leeway view of freedom, for instance, or some 'metaphysical' analysis of 'can' rather than an 'epistemic' analysis. In fact, I'm trying to avoid talking about the word 'can' at all since it only complicates matters. But perhaps I do need to say more about 'power.'
Suppose that I endorsed an 'epistemic' analysis of 'power.' Where would that get us?
(I'll explain tomorrow why I put 'metaphysical' and 'epistemic' in quotes. Hint: I'm not sure I understand the difference.)
Posted by: Joe Campbell | May 23, 2007 at 06:00 PM
Suppose I believe I will A tomorrow. I'm sure of it; no doubts. Still, my belief isn't an intention. Suppose I don't yet have an intention to A tomorrow. And I think I won't A unless I come to intend to A. I prefer to form intentions only on the basis of an (at least cursory) appraisal of my reasons. Why can't I deliberate about whether to A tomorrow?
This is something that has puzzled me. Here's a suggestion. Deliberating is engaging in practical reasoning aimed at making up one's mind what to do. (Sorry, Fritz.) Making up one's mind is making a decision. But deciding requires prior uncertainty about what to do. And I'm not uncertain.
Actually, what I'm certain about is what I will do. Can I be certain about what I will do while being uncertain about what to do?
Derk: How would you spell out 'might--in an epistemic sense' in (S)?
Posted by: Randy Clarke | May 23, 2007 at 06:34 PM
Joe says: "I think that van Inwagen's original argument was intended to be neutral with regard to any particular analysis of 'free will' or 'can.'"
I think Joe's right about this.
Randy writes: "Here's a suggestion. Deliberating is engaging in practical reasoning aimed at making up one's mind what to do. (Sorry, Fritz.)"
No need to be sorry Randy -- that sounds like one thing worth calling "deliberation". And one might wonder whether deliberating in that sense presupposes "open-ness" and if so in what way and of what sort. I'm not at all opposed to people discussing such issues. I do, however, continue to be puzzled as to why people who deny that deliberation in that sense presupposes belief-in-ability think they are disagreeing with some claim Taylor and van Inwagen have endorsed. That these philosophers think that their sense of deliberation requires belief-in-ability is silent on whether others senses of deliberation require this. Not that you, Randy, would make that mistake.
Posted by: Fritz Warfield | May 23, 2007 at 07:28 PM
I'm tempted to view an intention as weaker than a belief, such that having the belief that I will A tomorrow entails an intention to A tomorrow. After all, what constitutes an intention to A tomorrow other than being committed (in the here and now) to performing A tomorrow? And why would I *firmly* believe I'd do A tomorrow if I'm in no way committed to doing A tomorrow?
(Possible counterexamples to a belief entailing an intention include beliefs where "what I do" is actually passive, such as "I believe I will get sick tomorrow" -- but that doesn't seem to be what we're discussing here.)
Posted by: Benjamin Kelsey | May 23, 2007 at 07:39 PM
If I may interject with another concern, here are my thoughts:
According to (DE), a deliberating agent must believe that if her deliberative process yields a decision to do A, she will do A, and she will do A because of her deliberative process. Perhaps this is a misguided intuition on my part, but I can't help thinking that the hard incompatibilist must assume more than the compatibilist to embrace (DE). That is, it seems the compatibilist could claim that she will do A by way of her free will, but that she would not do otherwise because her deliberative process simply wouldn't have led her to any other decision. The hard incompatibilist, however, must simply trust that the same facts about the past and the laws that will guarantee a decision to do A will also necessarily bring about the agent's performing A. In that case, it seems the phenomenological decision to do A is superfluous. If the agent's decision to do A is extraneous to the agent's doing A, then the hard incompatibilist should not assume that a decision to do A will necessarily coincide with the performing of A. That is, one might decide to do A but find oneself doing B. But if the phenomenological decision to do A does play a causal role in the agent's performing A, it seems the compatibilist could easily claim this is what free will is, and I personally would not see why the hard incompatibilist would want to disagree. Perhaps this all comes down to a gut instinct on my part, which is that the compatibilist and the hard incompatibilist must somehow interpret (DE) differently, though I don't obviously see how.
Part of my problem may be the fact that I can't help viewing a decision (or choice) as anything but a certain kind of phenomenological event that is then put into action. For the hard incompatibilist to suppose that her decision to do A is efficacious in her actually doing A is, on my reading, to suppose a certain kind of mental causation that (again, in my view) one could easily argue constitutes free will. My point here is not to argue for or against such a notion of free will, but to ask if the hard incompatibilist can really hold her decision to do A as efficacious without assuming this kind of mental causation and—so it seems to me—thereby begging the question. If the question must admittedly be begged, then it will be useful to delve into the topic of mental causation and why it does or does not constitute free will, but that may be outside the scope of this comment thread.
Posted by: Benjamin Kelsey | May 23, 2007 at 07:45 PM
Derk writes (in the first comment above):
"...my thought is that, setting aside my definition, nothing that we ordinarily think of as deliberation would have the role of producing an intention to do A in a case in which the agent is already certain that she will do A. For example, it doesn’t seem to me that consideration of reasons for and against in this situation would have the role of producing an intention in such a case. Rather, at least typically, the agent would simply just form the intention, independently of further deliberation."
I would like to hear more about this, since I am inclined to agree with Dana on this point. Consider the case Derk cites from Randolph Clark, concerning Edna's deliberation about where to spend her vacation. She discovers, from the reliable testimony of her friend Ed, that there exists a decisive reason for her to spend her vacation in Edinburgh. However she does not yet know what this reason is. At this point it seems right to describe Edna as not yet having the intention of visiting Edinburgh, but having the expectation that she will later form such an intention. Moreover, It seems that Edna now has a choice. She could go ahead and immediately form the intention to visit Edinburgh based on the secure testimonial knowledge that there exists a decisive reason to do so; or she could wait until she discovers the decisive reason herself, and then form the intention on that basis. In both cases Edna would be forming her intention by considering her reasons, even though she was certain that she would form just that intention.
Is the claim that neither of these count as deliberation (in the sense of what "we ordinarily think of as deliberation")? Or that the first does not but the second does, and that the first would be the "typical" choice of an agent? Or is the claim that these both count, but that there is a third and more "typical" possibility, where an intention is formed as the result of an expectation, without being mediated by reasons?
Posted by: Brad Weslake | May 24, 2007 at 12:16 AM
Joe and Benjamin: ‘free will,’ as I apply it in the context of the philosophical debate, designates the sort of free will required for moral responsibility in the “basic desert” sense, or perhaps more precisely, whatever sort of control in deciding and acting is required for moral responsibility in this sense. (What my view claims is that human actions are never freely willed in the sense required for moral responsibility in the “basic desert” sense.) This technical use of ‘free will’ is not intended to correspond with exactness to the use of the term ‘free will’ in ordinary language -- or in philosophy generally speaking, for that matter. It is consistent with hard incompatibilism that there are correct applications of the ordinary language term ‘freely willed’ to human actions. Frankfurt, for example, specifies that a person acts freely and of his own free will just in case he wills X and wants to will X, and wills X because he wants to will X (1971). The arguments for hard incompatibilism pose no challenge to the proposal that we sometimes act freely and of our own free will in Frankfurt’s sense, and the hard incompatibilist position is consistent with it. And it may well be that the meaning of our ordinary language term ‘free will,’ and our ordinary thinking relevant to this concept, allows for Frankfurt’s use.
So now consider this principle:
VIPD*** A person S deliberates about whether to perform exactly one action among a set of mutually excluding actions A1…An only if S believes it’s possible for her do each of A1…An.
On an epistemic or compatibilist reading of ‘possible,’ my view has no worries about endorsing the claim that for a set of mutually excluding actions A1…An, (an actual human agent) S can truly believe that it’s possible for her do each of A1…An. For, in my view, the kind of freedom involved here would not be enough for moral responsibility. I can also endorse the claim that what Benjamin is calling ‘phenomenological decisions’ are causally efficacious, since that won’t be enough for moral responsibility either.
Randy asks: How would you spell out 'might--in an epistemic sense' in (S)? How about: the possibility of S’s choosing A is not closed off by any credence that the relevant agent has; that is, S's choosing A has a non-zero credence for that agent.
Later today I’ll comment on Brad’s challenge and Fritz's points about the characterization of deliberation.
Posted by: Derk Pereboom | May 24, 2007 at 10:49 AM
Dana points out that there is an ambiguity in the technical use of the term ‘deliberation’ that trades on two senses of ‘determine.’ Here are two possible readings of the characterization I assume:
(D1) deliberation is an active (or actional) mental process whose aim is to determine – in the causal sense of produce – a choice or intention for what to do from among a number of mutually incompatible alternatives.
(D2) deliberation is an active (or actional) mental process whose aim is to determine – in the epistemic sense of settle on – a choice or intention for what to do from among a number of mutually incompatible alternatives.
Now to Brad’s alternatives: Edna could (a) go ahead and immediately form the intention to visit Edinburgh based on the secure testimonial knowledge that there exists a decisive reason to do so; or (b) she could wait until she discovers the decisive reason herself, and then form the intention on that basis.
It seems to me that in both cases we don’t have deliberation in sense (D2), since she is already certain of what she will do, so there is no choice or intention yet to be settled on epistemically. As for (D1), on (a), if this immediate formation of the intention counts as an actional mental process whose aim is to produce an intention, then there would be deliberation in this case. On the application of (D1) to (b), maybe it is clearer that forming an intention on the basis of the reason would count as the right sort of actional mental process.
Fritz – my sense from re-reading the key paragraph on p. 28 of Coffman and Warfield is that (D1) is closer to your meaning – is this right? Would (D2) not join the issue with van Inwagen, for example? Would there be any objection to applying this notion of deliberation to the immediate forming of an intention that Brad describes – I’m imagining that such an immediate forming might not count as robust enough of a process to count as deliberation.
Posted by: Derk Pereboom | May 24, 2007 at 12:08 PM
Randy writes: "Suppose I believe I will A tomorrow. I'm sure of it; no doubts. Still, my belief isn't an intention. Suppose I don't yet have an intention to A tomorrow. And I think I won't A unless I come to intend to A. I prefer to form intentions only on the basis of an (at least cursory) appraisal of my reasons. Why can't I deliberate about whether to A tomorrow?"
I don't have a problem with this.
I also think that you can be certain at some time T1 that you will A tomorrow and that this doesn't prevent you from deliberating at some later time T2 about whether to A or some alternative action B tomorrow.
What I doubt is that at T2 you can be both certain about doing A tomorrow AND, at that very same time, be deliberating about whether to do A or B tomorrow. I don't see how considering the genuine possibility of doing B wouldn't defeat your certainty of doing A.
I'm not sure that the question was directed to me but I wanted to answer it whether it was or not.
Posted by: Joe Campbell | May 24, 2007 at 02:44 PM
Thought I'd jump back in on the certainty question. I see the appeal of the idea (defended here by Derk, Joe, and Benjamin) that if you’re certain of what you are going to do, you can’t deliberate. But I wonder if at least part of the appeal of the idea isn’t due to the fact that most (or all) of the time, when we human beings are certain of what we are going to do, we don’t deliberate. In turn, that could be because in most real life cases, we tend to be certain of what we are going to do (if we ever are) only because we base our belief on the evidence that we’ve already decided what to do. Still, I don’t see that it is incoherent to think of someone deliberating while being certain of what they are going to do. But I gather that people have different views about this (and different reactions to the cases that Randy and Brad described). So I wonder whether those who think certainty does preclude deliberation can more about why they think it does. I see that if Derk takes D2 as his understanding of deliberation, then he has a clear answer in terms of the aim of deliberation. But on other conceptions of deliberation, including D1, I don’t see why certainty should preclude deliberation. This is really a request for you to say more!
Posted by: Dana Nelkin | May 24, 2007 at 05:03 PM
One further clarification concerning a point I made yesterday about the lack of a difference between 'epistemic' and 'metaphysical' views of 'can.' This is a VERY minor point, so folks interested in the main points in this thread might want to skip this post.
Here is a view of 'can,' that I call 'the relevant facts view' ("Compatibilist Alternatives," Canadian Journal of Philosophy 35: 387–406 (September 2005)). (This is not my view but the name is mine and I discuss it in detail in this paper.)
S can do a iff that S does a is compossible with the relevant facts.
What counts as the relevant facts? Well, you could take the relevant facts to be the set of hard facts about the past together with the laws of nature. In which case, the relevant facts view leads to a metaphysical analysis of 'can' that is favorable to an incompatibilist theory.
You could also understand the relevant facts to be a set of more local facts, which would be favorable to a compatibilist theory.
Both of these accounts seem to be pretty metaphysical. What might an epistemic account of 'can' look like? Well, if I understand Derk correctly it too begins with the relevant facts view, only now instead of restricting the relevant facts to, say, the set of local facts, we restrict them to the set of facts that are known, or believed, by the agent S. (At least this is how I understand Derk.)
The epistemic account looks pretty metaphyscial to me, no less metaphysical than the others. I would say that the epistemic analysis is one kind of metaphysical analysis.
Posted by: Joe Campbell | May 24, 2007 at 05:30 PM
I'll stop after this.
I think that deliberation about two or more options requires a belief that those options are equally possible. This, in turn, would defeat any knowledge claim such as the claim one of the options was certain. You can be certain before or after you deliberate about whether to do either of two or more alternative actions but not during the deliberation.
I'll close with an example and a change of perspective.
Imagine Sally who is dating two men -- Allen and Ben. Sally has been dating Ben for 10 years and they are engaged to be married. Recently, by chance, she met Allen.
Suppose that Sally says to Ben, "I'm certain that I want to marry you Ben but recently I've been dating Allen. I'd like to take the time to deliberate about with whom I'd like to spend the rest of my life -- but I am REALLY CERTAIN that I want to spend the rest of my life with you!"
What would Ben think? Would Ben think that Sally was certain?
Posted by: Joe Campbell | May 24, 2007 at 09:14 PM
Thanks, Joe, cool case! Here’s one answer. First, I guess I'm still not sure why in general you think certainty precludes possibility in the relevant sense (unless possibility is already understood as epistemic possibility). I would think I could be certain of things that at least in some senses are not necessary. Second, here’s what I’d say about your case: Ben should feel unhappy because Sally hasn’t committed to him (at least I would!). That’s a real problem for Ben. Certainty should be cold comfort if the commitment isn’t there (yet). Again, I think that in real life, our certainty (if we ever have it) tends to derive from our having already deliberated. So this would indeed be an odd case. The question is whether it is incoherent that someone not yet have committed, even though she is certain that she will.
But I also have some question about how your case is set up. First, is the object of deliberation whether she wants to marry Ben or whether to marry Ben or whether to commit to Ben? (This wasn’t entirely clear to me from the set-up.) I assume it isn’t the former, since that is a different kind of deliberation than we are thinking about here, I think, but correct me if I am wrong. If the second option, then her dating someone else while engaged to Ben could undermine the possibility of actually marrying (since it takes two to get married and Ben might have had enough), and this might get in the way of our seeing Sally as certain of marrying. If it is the third option, then this might be a bit confusing, too, in the context of the case, since she is described as already engaged (are we supposed to think the engagement—that is, the commitment to marry—has been broken)?
Posted by: Dana Nelkin | May 25, 2007 at 02:06 PM
On Fritz’s first comment, which is about how to characterize deliberation: Coffman and Warfield say that Searle and van Inwagen conceive of deliberation as occurring “after reasons for various actions have been weighed and evaluated,” and this is part of how C&W propose deliberation be understood for the purposes of this debate (2005, 28). A concern about this proposal is that seems to leave too little to count as deliberation. It would appear to leave the forming of an all-things-considered judgment about what it is best to do, and the forming of an intention to perform an action. But ordinarily we think of the weighing and evaluating of reasons for the options for what to do as central to deliberation. Suppose we now characterize deliberation as an active mental process whose aim is to choose what to do from among a number of mutually incompatible alternatives, a process understood as one that can include the weighing and evaluating of reason for the options for what to do.
Now Coffman and Warfield are concerned that the issue with van Inwagen and Searle will not be joined on a characterization that includes more than they specify, and Fritz reiterates this worry in his comment. But, at least initially, adding in the weighing and evaluating of reasons does not appear to have this upshot. The quest at hand is for necessary conditions for rational deliberation. If having the belief that I can do A and can do B is a necessary condition for my instantiating an active mental process whose aim is to choose whether to do A or to do B, understood as one that does not include the agent’s weighing and evaluating of reasons for the options, it would seem that having this belief will also be a necessary condition for my instantiating the active process understood as one that can include weighing and evaluating of the reasons for the options -- and vice versa. The same would appear to be true for the belief that I am not certain that I will do A and not certain that I will do B. More generally, let me suggest that any belief-condition at issue in this debate would be necessary for both the broader and the narrower processes, if it were necessary for one of them.
In addition, the relevant quotation Coffman and Warfield (2205, 27) cite from van Inwagen is: “serious deliberation occurs when one is choosing between alternatives and it does not seem to one (once all the purely factual questions have been settled) that the reasons that favor either alternative are clearly the stronger” (2004, 217) This characterization of serious deliberation would appear consistent with the following scenario: all the purely factual questions have been settled, and it does not at this point seem to the agent point that the reasons favor either alternative. But now genuinely serious consideration and evaluation of the reasons, and hence genuinely serious deliberation, can begin. Coffman and Warfield cite the following quotation from Searle: “there is the gap of rational decision making, where you try to make up your mind what you are going to do. Here the gap is between the reasons for making up your mind, and the actual decision that you make” (2001, 14). It’s consistent with this quote that the gap is between the mere noting of or apprehending of reasons and the actual decision, and what happens in the gap is serious consideration and evaluation of reasons.
So perhaps on this proposed amendment on the characterization of deliberation, the issue is still joined with Searle and van Inwagen. And it has the virtue of characterizing deliberation in a more natural way.
Posted by: Derk Pereboom | May 25, 2007 at 04:24 PM
Joe writes:
"I think that deliberation about two or more options requires a belief that those options are equally possible."
If "equally possible" just means "possible" I can see the force of this on the sorts of grounds that have been discussed here, but if it means "equally likely" I think this is false, since I think we can have reasonable expectations about what the results of our deliberations will be. Of course, these expectations cannot themselves serve as reasons to be taken account of in the process of deliberation.
For example, suppose I'm trying to decide whether to have dessert. I know that there is a high likelihood that I will have dessert, since I nearly always do so. In this situation I can still deliberate about whether or not to have dessert, though when I do so I do not include as a relevant reason the fact that I expect to decide one way rather than another.
If this is right then the cases involving certainty we've been talking about are just the limiting case of these sorts of expectations. What makes the limiting case puzzling is that the two aspects of deliberation Derk distinguished above here come apart. We think of deliberation as typically resulting both in the formation of an intention (D1 above) and in the reduction of our uncertainty concerning how we will act (D2 above). Both of these will be present even in cases where our expectations are heavily biased towards one action (as in the dessert example). But in cases where expectations are certain (as in the Edna example) there is no role for (D2), and so we are less inclined to call a process just involving (D1) a case of genuine deliberation.
I think it's important to remember however that the type of certainty we are talking about is certainty about what deliberation will deliver, not certainty independent of whether deliberation occurs. In being certain that she will go to Edinburgh, Edna is relying on the expectation of her deliberation turning out a certain way. She does not think she will go to Edinburgh no matter whether or not she bothers to form an intention to do so. That would be the kind of fatalist certainty that I think would be inconsistent with Edna taking herself to be free. One virtue of Dana's explanatory nexus account is that it is well-placed to handle this distinction.
Posted by: Brad Weslake | May 25, 2007 at 05:06 PM
Derk, I'm glad you brought up the characterization of deliberation offered by Searle and van Inwagen and taken up by Fritz and EJ, that it is what happens "after reasons for various actions have been weighed and evaluated." I'm glad you brought it up, because I just don't get it. I don't understand my own deliberation in this way. I think of deliberation as the process of looking for reasons for the various actions and weighing and evaluating them. And I think of the *goal* of deliberation as reaching a decision based on this process. And I think of cases where I don't know how to decide what to do *after* considering and weighing reasons as much as I know how as cases where deliberation has *failed* and I have to just make a decision anyway. And I think of those decisions as a lot like flipping a coin. And I don't feel particularly free and responsible for those decisions.
Maybe my phenomenology (and conceptualization) about deliberation here explains why I'm a compatibilist. (It might also explain why I drive my wife crazy by over-deliberating to try to find a clear-cut decision about what is best to do.)
But can people who think of deliberation the way Searle, PvI, Fritz, and EJ describe it please explain more about what they have in mind.
Posted by: Eddy Nahmias | May 25, 2007 at 05:17 PM
Good points about my example, Dana. (I haven't had time to look at the posts that follow Dana's.)
Let me make one clarification. I meant to say, in response to Randy’s example, that I have NO problem with it, contrary to what I wrote! In Randy’s example, we have a case of what I’ll call ‘single-alternative deliberation,’ and the relevant principle to appeal to is VIPD*: A person S deliberates about whether to perform an action A only if S believes that it is within S’s power to do A. Someone may be certain that he is about to do A and also believe that it is in his power to do A.
What I have a problem with is certainty in cases of what I’ll call ‘multiple-alternatives deliberation,’ cases where one deliberates about whether to perform exactly one action among a set of mutually excluding actions A1…An.
I also agree that the engagement confuses my last example unnecessarily, so let's drop it; likewise with the fact that Sally had been unfaithful to Ben. Let me try to fix some other problems with the wording, as well. (Clearly I had a bad day yesterday but I’ll skip those details, too.)
In the new example, Sally and Ben have been casually dating. Ben asks Sally to go steady and she says in reply to him: "I'm certain that I want to have a committed relationship with you, Ben, but I haven’t even tried to decide to have that comment yet. Furthermore, there is this other man, Allen, who has recently asked me to go steady. I don't know Allen at all and he seems rather dull and unattractive, so there is no reason to be worried. Nonetheless, I'd like to take the time to deliberate about whether to make a commitment to you or about whether to make a commitment to Allen. Don’t worry, though, for I am REALLY CERTAIN that I will decide to make a commitment to you!"
My take is that the suggestion that Sally needs to deliberate about the matter creates some level of doubt in Ben's mind, enough that he would question whether Sally really was certain that she will decide to make a commitment to him.
But it isn't just Ben who has these doubts. If I were Ben's friend, and he told me what Sally said and asked me what he should believe in light of that information, I'd tell him that it didn't seem to me that Sally was certain. If I were Sally's friend and she told me what she said to Ben and asked me what I thought, I'd tell her that I didn't think she was certain either. Maybe Sally feels that she is certain but she’ll have a hard time convincing anyone else that she is.
Perhaps the example is still problematic, though, since as you say: “in real life, our certainty (if we ever have it) tends to derive from our having already deliberated.” And this is just one case, so it is difficult to see how one can jump from this to the claim that the situation that you describe is incoherent. I just wanted note that it is one thing to say that someone is certain in the kind of situation that you describe yet another for them to be certain. It would be difficult to image anyone believing that, say, Sally was certain once she reports that she’s about to deliberate on the matter. This seems important.
Lastly, I wanted to make one further clarification in response to some points in Dana’s first paragraph, directly above. I didn’t mean to suggest that “certainty precludes possibility” or that one could not “be certain of things that at least in some senses are not necessary.” I didn’t intend to make any general claims about certainty but just a particular claim about certainty in the context of multiple-alternatives deliberation. If we understand such deliberation in the way suggested by van Inwagen, then “To deliberate is to try to decide between [or among] various incompatible courses of action” (1983, 154).
Apart from the worry already expressed in this post, there is a further worry that I have about certainty in the context of multiple-alternatives deliberation. Suppose that S is trying to decide between two courses of action: to do A or to do B. Given VIPD**, S believes that it is within his power to do A and S believes that it is within his power to do B. Thus, while deliberating about whether to do one or the other, these appear to S to be live options. I think that S’s beliefs about the genuine liveliness of, say, option B defeats any knowledge, or certainty, that S might have that he is going to do A.
Thanks for your patience!
Posted by: Joe Campbell | May 25, 2007 at 05:23 PM
I’m definitely not sold on certainty stemming from deliberation that has already taken place. I am certain that I will not choose to abandon my wife and son and flee to the mountains to live in the wild by myself. But do I consider it possible for me to do this? Yes. Does the fact that I consider it possible for me to do this allow me to deliberate about it? I don’t feel like it does. I do not doubt that I can think of all the upshots of taking to the mountains and leaving my family behind (however few those upshots may be), but I genuinely don’t believe there could be an existent and discoverable fact that would change my mind about the matter. Without having given it extensive thought, this latter belief seems, in my mind, to be necessary for deliberation.
All of this makes me ask: if someone believes it is within his/her power to do X, and the thought occurs to him/her that he/she could do X, even if X is readily dismissed, is it fair to say that this person has necessarily deliberated about doing X? Even fleetingly? If so, then it could be argued that, in my fleeing to the mountains scenario, my certainty *is* the result of deliberation—an incredibly swift and effortless deliberation, but a deliberation nonetheless. I’m not sure what I think about this. If the thought occurs to me to do something, even if I reject it immediately, it seems like I have thereby made a choice of some kind or another. But to call this deliberation might be stretching things a bit. If nothing else, it seems horribly misleading to say I’ve now deliberated about running away from my family. And there are probably a thousand other things I’d rather not mention that, under this reading, I would have to say I’ve deliberated about.
Posted by: Benjamin Kelsey | May 25, 2007 at 09:11 PM
I want to comment on Derk’s conditions (S) and (DE). Concerning (DE), suppose I am deliberating about whether to do A1 or whether to do A2. I deliberate in order to make up my mind, and I hope that my decision will accord with my judgment about what is best to do. However, I am well aware of my occasional akrasia, and I am not convinced that my so judging will bring about my decision, much less my action. I don’t think that lack of belief blocks my deliberation, though it would if (DE) were correct and I am “rationally deliberating.” Now, perhaps I am not rationally deliberating. Well, then, what is it to rationally deliberate? And even if I, qua akratic, am not rationally deliberating, I am still deliberating and, as such, I need to have a belief in the openness of each of the alternatives I consider. So, either (DE) is incorrect, or it does not provide a rich enough account of the contingency involved in the sense of openness or freedom.
In my 1986 paper on deliberation, I noted two problems with principles similar to Perebom’s (S). First, epistemic possibility—consistency with what one knows—is not enough to account for the sense of openness in all cases of deliberation. For example, I do not deliberate about flying to Copenhagen tomorrow because I am already committed to staying in Chicago and I regard it as regard as “settled” that I will remain in Chicago and not choose to fly to Copenhagen. However, it is consistent with what I know that I will choose to fly to Copenhagen tomorrow and I believe that if I make that choice then I will fly to Copenhagen tomorrow (perhaps some unforeseen emergency will call me to Copenhagen). Now I don’t know whether I am “certain” about remaining in Chicago and not choosing to fly to Copenhagen, but I believe that I will so remain and I deliberate about other things on the basis of that belief. So, flying to Copenhagen tomorrow is impossible relative to what I believe and so does not appear open to me, though it is contingent with respect to what I actually know. Accordinly, the more comprehensive interpretation of the sensed possibility (or “might”) of an option during deliberation is doxastic, not epistemic.
Second, one can easily deliberate about doing something even though it is not consistent with what one regards as settled. Suppose that at 10 a.m., Mr. Hawkins, having decided to take his son bowling at 3 p.m., acquires the belief that he will take his son bowling then. He takes it as settled and asks his secretary to remind him of his commitment at 2:30. At 2:29 p.m., temporarily overlooking his earlier resolve, he deliberates about playing golf at 3 p.m. Given all that he believes at 2:29 pm. or regards as settled, it is not true that he assumes it possible that he play golf at 3 p.m. One can forget something one believes without ceasing to have that belief. When Hawkins is reminded by the secretary at 2:30 pm of his commitment he does not acquire a new belief. Instead, his attention is refocused upon a belief he already has. In responding to this example in my 1986 paper, I suggested that a quantifier phrase like “every proposition she, in the present context, regards as settled” should be placed inside the scope of a belief operator, not outside as in (S). Unless I missed something, I do not see that what Perebom says in his note 7 or section 5 succeeds in retaining an external reading of that quantifier.
Posted by: Tomis Kapitan | May 27, 2007 at 11:52 AM