Jonathan Dancy

Jonathan Dancy (University of Texas--Austin & The University of Reading ) is "presenting" his paper,"Practical Reasoning and Inference." The invited commentators for his paper are Joseph Raz (Columbia/Oxford), and Candace Vogler (University of Chicago).

Download jonathan_dancy.pdf
Download razs_commentary_on_dancy.pdf
Download voglers_commentary_on_dancy.pdf

Comments

[Jonathan Dancy]

Response to Raz
Many thanks to JR for those comments.
I reject the view that practical reasoning is just ordinary reasoning about a particular subject matter (not on the ground that Raz suggests). If it were, there would be no general distinction between practical and theoretical reasoning. In that sense, to suppose that there is reasoning that is practical, as opposed to being theoretical, is just to suppose that practical reasoning is not just ordinary reasoning with a special subject matter. Anyone who takes seriously the question ‘is practical reasoning possible?’ will be supposing that, to count as practical, reasoning must be distinctive in some way other than in its subject matter. I know of two suggestions here. The first is that practical reasoning is reasoning to an intention, reasoning whose conclusion is (the formation of) an intention; the second is the one that I much prefer, that practical reasoning is reasoning to an action.
Raz thinks that there is no reasoning of either of these forms, and he says that my paper does nothing to show that he is wrong.
The main point at issue between us, I think, is the question whether it is possible to think of an action as the conclusion of reasoning, or whether we are forced to suppose that the reasoning ends before the action begins.
I agree with Raz that if we have to think that there is always a process worth calling reasoning and which ends with some conclusion about what one has most reason to do, or undefeated reason to do, or something like that, then whatever comes after that point will not be worth calling reasoning.
My overall view, however, is that one does not need, in reasoning, to reach a separate conclusion about what one has most reason to believe or do. A theory that demands this, as a separate stage, seems to me to go too far. If we do demand this, we end up with Raz’s picture. And in some of the ways in which I expressed myself in my paper, I unwittingly lapse into the very sort of talk that I am trying to get away from. But the view I was trying to express was that in reasoning we revolve various considerations in our minds, which a view to deciding whether p, or whether to F, and in the light of those considerations (not in the light of their status as reasons, or in the light of some fact about how the reasons come down here) we do so decide. Sometimes the decision to F is not a separate stage; when F-ing is something we can immediately do, we don’t need to decide first and then do the action. On other occasions we decide now to act tomorrow, and then the deciding is a separate stage. We can pass through an intermediate conclusion to the effect that F-ing is best supported by the reasons, or something like that. But we don’t need to. In fact, I would say that one doesn’t need to wield the concept of a reason, or that of the thing to do or of the thing to believe, in order to reason effectively. It is enough if one is dealing with reasons and treating them as reasons.
The difficulty for this view is how exactly to characterize that last bit. What is it to treat something as a reason if it is not to take it to be a reason and to respond to it accordingly? As I see it, there has to be an answer to this question, on pain of over-intellectualising the process we are trying to understand.
Raz does consider, and respond to, the view I am here propounding, on p. 15 of the paper he refers to (‘Reason, Reasons and Normativity’). I may put a separate comment on why I am not persuaded by what he says there.

[Jonathan Dancy]

Response to Vogler.
Thanks to CV for her comments.
It is true that I tried to be conciliatory towards those who try to run things in terms of intention, seeing the conclusion of practical reasoning as something of the form ‘I intend that I F’. I welcome Vogler’s encouragement to be firmer on the point, and her reminder that ‘I intend that I F’ looks too much like a first person version of ‘I intend that he Fs’. I ought to have taken more seriously than I did the difference between ‘I intend that I F’ and ‘I intend to F’.
It is also true that in the paper I grudgingly granted that actions can have something worth calling a content. Vogler encourages me to stand out against this; all I can say is that I will if I can find a persuasive way of doing so. I keep changing my mind on this matter.
I think that Vogler’s main point is that I have written my account of the structure of deliberation in terms of abstract relations, favouring/disfavouring, enabling/disabling and intensifying/attenuating, but that this limited focus tends to hide from view things that we need to bring out if we are to have a full understanding of the distinctive style of practical reasoning, especially moral reasoning. But the things that Vogler suggest that we add don’t seem to me to be aspects of structure. The first is a distinction between what I intend and what I merely foresee. The second is something that is restricted to moral deliberation, namely a list of four ways in which things can go wrong ethically. Then there is a list of four sources of pro tanto reasons for acting: desire, prudence, partiality and morality. Vogler suggests that if we keep all these things in focus, ‘a new structure in deliberation emerges’. I found it hard to understand this suggestion, and the illustrations she goes on to offer did not help. Take the ‘fact’ that when things go wrong through no fault of our own, the fault lies neither with the doer or the deed but with something else. Vogler suggests that this might ‘mark a distinctive specification of the abstract favouring relation in ethical contexts’, and seems to think that I might agree with this. But I don’t. I don’t think that the favouring relation can come in different, more finely specified forms. When we ask what response is favoured to an action that went wrong through not fault of the agent, the notion of favouring that we are using is just the original ur-relation, not some more subtle form suitable for morality.
In fact I don’t think of the favouring relation as abstract at all, but as normatively meaty – an expression which I am sure some people will dislike.

[Candace Vogler]

Thank you, Jonathan.

I didn't mean the final remarks, adopting some of Foot's points, as remarks that need be understood as influencing the structure of a thought process.

I confess that I have some difficulty sorting out how much you mean to be discussing what goes through my mind when I consider what to do or how things are (which could be *any number of things*--e.g., mental images of people involved, practical considerations interspersed with half-remembered tunes, little mental reminders to do this or that tomorrow that are otherwise not to the point, and so on) and how much you mean to be thinking about those aspects of a period of my mental life that can be appropriately construed as belonging to the same process of deliberation/judgment.

I was taking it that you meant "structure" as a way of picking out a group of the many thoughts, half-thoughts, and more fleeting "mental events"--namely the group that *belongs* to the deliberation or the process of judgment. Suppose my practical concern is beinging up a delicate matter with a friend. Suppose I spend two days worrying the question off and on. The sorting/grouping might go like this: being reminded of last spring in Paris [*not* part of the train of thought]; noticing that she might take offense if I put it to her *that* way [*part* of the train of thought]; a slow rehearsal of points to be made in lecture the following morning on the way to school [*not part of it]; the sudden inspiration for a possible re-phrasing that might convey the point without causing offense, coming out the blue as I enter the classroom a day after starting to worry about talking to her [*part* of the train of thought].

If *that's* right, then the intended-foreseen distinction DOES belong to structure, I think, since it will give shape to the emerging course of action in the way that Anscombe discussed in her discussion of the 'A-D' order. I am happy for intentional actions to have content. I think that they have the content that Anscombe plotted for us in *Intention*--this is their rational structure, and it is isomorphic with the structure of intention. But *that* structure depends for its intelligibility ofn the distinction that Sidgwick obliterated for us way back when in treating intention as covering all foreseen consequences of an action.

Anscombe makes the point by imagining that the man at the pump beats out the rhythm of 'God Save the Queen' as he pumps. If this is just a way of amusing himself as he goes about poisoning the inhabitants, then *it* is no part of the arm movement-pump-poison-polish that lot off-get the good ones in A-D order. It has its own means and end, and it could go badly (fail, stutter, what have you) without *in any way* affecting the success of the murder scheme. Evaluation of what he is up to could show a bad musical attempt *and* a highly successful murderous assault on the house water-supply happening at the same time.

She will think that working the pump with an ear toward getting the beat right (as he goes about pumping poisoned water) is deliberation aimed at achieving the ends of his little musical diversion--like taking a bit to find the right key when whistling while you work, only percussively. His sonic adjustment is a kind of deliberation, I take it. Notice that understanding and assessing *it* is understanding it as concurrent with the other thing he is doing. Notice that he will foresee many other things that will happen as he beats and poisons, and that his deliberation is in no way unsound if he pays next to no attention to them.

In short, such matters as calculative order are crucial to understanding what is going on as he works things out in the joint murder-and-music project, and the intended-foreseen distinction is crucial to picking out which kinds of things matter to the soundness or error in his deliberation. In this sense, the distinction, like the 'A-D' order, isolates some of the things that come up for him--those that are salient for one or another of the things that he is doing.

Does that make more sense?

Candace

[Brad C]

Hi Jonathan,

I just taught a section of Michael Williams' *Problems of Knowledge* which got me thinking about this line in your response to Raz: "In fact, I would say that one doesn’t need to wield the concept of a reason, or that of the thing to do or of the thing to believe, in order to reason effectively. It is enough if one is dealing with reasons and treating them as reasons."

Perhaps is is conceptually possible to reason effectively without these concepts, but they - or their cousins - seem to be built into our evaluations and criticisms of decisions, deliberations, and acts. If so, it seems to be a stretch to say we can actually reason effectively (let alone excellently) without these concepts.

It is also unclear whether we can count as responsible in the practical-reasoning sense (the practical reasoning analogue to epistemic responsibility) without at least implicitly deploying these concepts. Perhaps that is, again, why you used 'effectively' instead of 'excellently'?

-Brad

[Jonathan Dancy]

This is response to Candace.
Yes it does make more sense, but my inclination is still to think of things that are ‘crucial to picking out which kinds of things matter to the soundness or otherwise of his deliberation’ (your last paragraph) as not part of the ‘structure’ of the reasoning. Perhaps I am working with too few concepts here. Originally I was thinking in terms of structure and content. But features that are ‘crucial …’ seem to be on neither side of that distinction. They are not themselves part of the content of the reasoning; in fact they aren’t part of it at all, but they are certainly playing some significant role somewhere, if only because we are using those ‘matters’ to determine what is part of the reasoning and what is not.
I am less clear than I would like to be whether to say similar things about the A-D series. That is partly because of my still limited understanding of what Anscombe is up to. What we have is that A is a way of doing B, B is a way of doing C and C is a way of doing D. Since D is my end, I reason backwards down the series, as it were, find something which I can do immediately (pump, say) and so I pump. There is a clear structure here, and a distinction between structure and content. One of my worries about my paper, which I alluded to at the end, was how to relate what EA is saying here, which makes obvious sense, to the sort of picture that I was developing in my paper. The two things don’t seem to be incompatible, but that doesn’t help me see how to put them together. The clearest suggestion I can make, which I associate with some of the things in your excellent book, and with EA too of course, is that the reasoning shows what the (supposed) good is in pumping. Pumping is a way of doing D (getting the good lot in). SO what we have here is a value-based reason, a consideration that favours pumping, and the A-D structure lays out how the value of achieving D makes A worth doing.
I don’t know if I should just stop there. This kind of structure is very different from the sort that I was imagining that I would reveal with my three distinctions (favouring/disfavouring, enabling…)
Jonathan

[Jonathan Dancy]

Response to Brad C
Thanks for those sensible comments, Brad (and congratulations on your position for next year, I was very pleased to hear about that).
You are right in thinking that my use of ‘effective’ was intended partly to address these issues.
It may be that one needs to possess the concept of a reason in order to treat something as a reason in reasoning. But that would not show that that concept need to appear in the reasoning itself. One might need to have the concept of truth as well.
If we say that we are at least implicitly employing the concept of a reason, we are going to get asked what that amounts to. There is no reason for me to object to it, I think, as it stands.
For me the question is in what way the reasoning is improved if one adds premises such as ‘I have considered every relevant issue’, and ‘these are all the reasons there are’ and ‘the considerations together constitute sufficient reason to F’. If those things were not true, my reasoning would be defective in one way or another. But this does not itself earn them a place within the reasoning.
I am having lunch (real lunch, not virtual lunch) with Joseph today in Balliol, and may be able to report back later, since I am sure that we will be discussing these issues.
Jonathan

[Joseph Raz]

In replying to my comment Jonathan raises two issues, which are quite distinct. I am inclined to agree with him on one, but not on both. I am inclined to agree that in reasoning we use (reflect, consider) reasons, and their credentials and implications, but do not necessarily employ the concept of a reason. We need the concept, and we have to use it (we may say along with Brandom) to make explicit what we do when reasoning, but not in order to reason. (Given that here we are trying to make things explicit I will continue to use the concept). However, whether or not JD and I are right on this point, it does not vindicate his view of practical reasoning. For that he needs to establish – as he himself describes matters - that while when reasoning whether to F ‘(w)e can pass through an intermediate conclusion to the effect that F-ing is best supported by the reasons, or something like that ... we don’t need to.’
JD mentions that I consider & reject this view elsewhere (‘Reason, Reasons and Normativity’ p. 15). Let me indicate why (describing matters as JD sees them), and in doing so I will apply ‘valid’ to reasoning, in a way which will be clear to anyone familiar with its application to inferences. My reasoning – I claim – is not valid if I conclude by F-ing simply because my reasoning led me to a warranted belief that P, and I rightly treat that P as a reason to F. Suppose that my reasoning also included warranted belief in another fact which I reasonably treat as a stronger reason against F-ing, and I conclude by F-ing. You may say that in that case my reasoning is not valid, but when it does not include such undermining belief it is valid. But that would absolve the reasoner from the obligation that his reasoning be a serious effort to come to a right conclusion. The standard for valid reasoning includes that norm. Therefore, to vindicate F-ing as its conclusion, my reasoning must include a warranted belief that there is nothing which negates the reasonableness of treating P as vindicating F-ing. Why so? All I claimed is that (to meet the standard for valid reasoning) my reasoning must be diligent. We know that satisfying that requirement consists in how I reason, rather than merely in the content of my reasoning. The question is whether I can be diligint without that affecting the content of my reasoning as well. To be diligent I must search (if only my memory) for considerations which may negate the appropriateness of taking that P to vindicate F-ing. Can I satisfy that requirement without asking myself: is there anything I am missing? And answering No. Of course raising the question and so answering is not sufficient. I should so answer only if I pursued the matter adequately so that the response would be rational. But it can be argued that I will not be reasoning validly without concluding that so far as I can see nothing negates the view that P vindicates F-ing.
If that is so then JD agrees (as I understand him) that the reasoning stops here. He thinks however that I can satisfy the norm of diligence without anything affecting the content of my reasoning, though he agrees that often we in fact satisfy it in the way I described above, i.e. by reaching the intermediate conclusion which – in his mind – disqualifies the reasoning from being practical.
At this point my doubt about his view is strengthened. I will mention just one other consideration: while the difference between how we reason and what enters into the content of our reasoning is both real and important, the boundary can be rather elusive. Imagine someone who decides to wait a couple of days before reaching a conclusion. That is not part of the content of his reasoning. But suppose that he pauses to give himself a chance to come by additional facts which bear on what the conclusion should be. A couple of days later he resumes and concludes as he would have done had he not paused. Would it be wrong to attribute to him the thought ‘I found nothing which undermines the conclusion that my reasoning so far pointed to’? Must we find him saying this to himself before we can attribute the thought to him? I do not think so. Hence it seems to be that valid reasoning is rightly described (by analogy with inductive reasoning) as including the thought that nothing undermines the conclusion the rest of the reasoning points to, or something along these lines, and that that is one of the facts which undermines JD’s thesis.

[Brad C]

It seems that there is a slightly weaker position (than the one just introduced by Prof. Raz) that would push Dancy to admit that one needs to be *prepared* to wield the concept of a reason if one is to count as a valid reasonser.

Instead of claiming that it is a norm of valid reasoning that one have a "warranted belief that there is nothing which negates the reasonableness of treating P as vindicating F-ing," one could claim that it is a norm of valid reasoning that one be able to respond to serious challenges to the reasonableness of treating P as vindicating F-ing. This would be the practical reason analogue of shifting from a prior-grounding constraint to a default/challenge constraint on epistemic rationality. You have to be able to respond to challenges, but need not have preemptively thought through how to respond to all comers, so to speak.

This also raises the question of whether norms of valid reasoning go beyond what is necessary for effective reasoning to what is necessary for excellent reasoning or what is a useful means ("practically" necessary for) for effective reasoning.

[Brad C]

Jonathan,

What do you think of this:

Perhaps part of treating a consideration as a reason is being disposed to defend it's status as a reason when given a serious challenge (someone reasonably claims something is operating as a disabler and you need to contest that fact, deny the thing that disables is present, etc). On this view we could still treat it as a reason without having thought of it as a reason - if we have not faced any such challenge.

[Jonathan Dancy]

Thank you Brad, that is just what occured to me after reading your previous comment (May 22, 2.30pm) - so won't that show that your 'weaker' position tends my way rather than Joseph's?
Jonathan

[Jonathan Dancy]

Response to Joseph’s of May 22, 10.43am
The term ‘valid’ should probably not be applied to practical reasoning, on either account. It leads one to suppose that what is not valid is invalid. But we all agree that practical reasoning can be better or worse, and still be reasoning, and still be practical. Something can be reasoning even though it can be improved as reasoning. (The standard reference for this sort of thought is the beginning of Grice’s Aspects of Reason.)
JR sometimes writes as if something is not reasoning if it can be improved. But I think that one can be reasonably diligent – diligent enough to satisfy the relevant norms of diligence, whatever they are – without having done everything that one could have done.
The view I am defending faces several difficulties. The first is that very often people don’t do the action called for by the considerations they recognize as relevant. The second is the fact that sometimes people reason to a decision, when the time of the action comes later. JR introduces a new one: the interrupted reasoning.
In the detective example which I gave towards the end of my paper, one of the ‘premises’ was 4. Our investigation was conducted with great care – scrupulously, as it were’. This consideration can itself affect the confidence with which we draw our conclusion. But that does nothing to show that all good reasoning should include some such self-congratulatory remark.
Supposing then that JR’s reasoner says to himself ‘This is a tricky and important issue. I’ll wait a couple of days and revisit the issue then. Surely by then anything I might have missed will somehow have occurred to me’. It seems to me that, on resuming two days later, he might be expected to say to himself ‘Well nothing has occurred, and so I’ll just stick to the conclusion I would have drawn anyway’. And even if he doesn’t say it to himself, we can reasonably ‘attribute to him that thought’, as JR puts it. But what does this show about good reasoning? Very little, as far as I can see. It is a special case. I just don’t see why this should persuade us to suppose that all good reasoning includes some remark to the effect that nothing undermines the conclusion that the reasoning so far points to.
The scenario suggested seems to be this. I consider relevant matters, and a conclusion suggests itself to me. But before I can ‘validly’ – or responsibly – go with that conclusion I have to ask myself whether I have missed anything out, or something like that. And my answer to that question is an essential part of any piece of responsible reasoning. I just can’t get myself to believe this. Even if things are somehow sounder or more responsible if I do ask myself that question, this does not show that they were in any way defective, or not reasoning, before.

[Brad C]

Hi Jonathan,

Yes. My weak view does incline your way insofar as it only insists on concept possession (not employment). I missed the 'wield' in your initial post; given that instead of 'possess' my weak position can be taken as a friendly amendment.

A new worry: In most situations in which we engage in practical reasoning, there is some puzzle or problem that provokes the episode of reasoning (as the Pragmatists emphasize). When so, that suggests that we engage in reasoning because we are unsure about what the thing to do is. And that seems to involve wielding the concept of a reason or the thing to do (at least in some sense).

Even if that worry is avoided, the fact that one is puzzled or faces a practical problem seems to indicate that one is aware of purported reasons that favor in different directions (so to speak) - so there is some challenge that we must answer by engaging in *reason* employing thinking if we are to count as responsible.

You could avoid the second worry here by pointing to examples where we engage in practical reasoning but are not impelled to do so by puzzlement - or not by puzzlement of that sort. Perhaps an example of specificationist reasoning - in which there is only one way of specifying but one does not see it right off - will do the trick?

[Jonathan Dancy]

In reply to Brad's last

I'm rather of the view that Joseph suggested, that we need the concept of a reason, and that of the thing to do, in order to make explicit what we are about in deliberation, but those concepts need not appear as part of the deliberation itself.

Consider theoretical reasoning? One might start 'I wonder what one can infer from these premises'. After all, the question addressed by the inferrer is what follows from what. But I would not wish to see that question as part of the inference/reasoning.

In general, I think I can allow that the concept of a reason is wielded in various ways as the deliberator goes along, so long as that concept is not required as part of the sort of intermediary conclusion that Joseph was urging on me.

[Ray Cheung]

One thing that marks the distinction between theoretical and practical reasoning is what kind of thing the end-product is. According to Prof. Dancy, for practical reasoning, the end-product is action, while for theoretical reasoning, it is belief. What if we try to understand the latter as a special case of action, namely a mental act to believe? Or to understand the former as decision to act, and the latter decision to believe?

Certainly, even if we are to understand the two in terms of decisions, there are still differences between them. For example, decision to act is usually more transparent than decision to believe. If you decide to change your sitting posture, you can easily do it right away. However, it is unclear whether it is as easy to change one’s beliefs. When one suspects that his wife is unfaithful, even though he has found plenty of reasons against that she is and decided to believe that she is faithful, he may still have to fight against his suspicion for some time.

Further, on the face of it, the normative aspects of the two kinds of decisions are different. If I understand Prof. Dancy correctly, the following passage also touches on this issue:

“The real question is whether we can find a relation on the practical side that plays the explanatory role played by the two truth-relations (entailment and probability-related) on the theoretical side. . . . There is, unfortunately, only one candidate for this explanatory role on the practical side, as far as I can see. This is some appropriate configuration of the notion of value.”

Roughly, a decision is good for an agent if it brings about things valuable to the agent. If we can find other kind of value that plays the explanatory role on the theoretical side, we may have a more unified account of reasoning. If the epistemic norm is such that what makes a belief good is for it to be true, or at least to be of high degree of confidence or certainty, truth, confidence, and certainty are possible candidates for such kind of value. Of course, a lot still has to be said about the operating relations. I am just wondering what you think about this. Thanks.

[Jonathan Dancy]

Well, there are going to be all sorts of difference between different actions, on any account; different actions can differ from each other in many respects. I don’t know that if belief is treated as a sort of action, as I am rather suggesting (strongly influenced by Raz on this point), all beliefs need be the same in all such respects. Perhaps beliefs differ from each other much as actions do.
But I agree that my general line has been to pull practical and theoretical reasoning together, so that at least the similarities between them are far more prominent than any differences. The fact that one concludes in action and the other in belief is not much of a difference, on my showing. So suppose that we said that truth is a good-making feature of a belief. Still, there would be a difference between the roles of considerations of value on the two sides. A reason for action is a consideration that reveals some value to be gained by doing the action. If we thereby show some value in the action itself, this may be a derived value, a value derived from that of the outcome, shall we say, or of the goal. But we can’t say anything like that in the belief case. Even if what makes a belief good is for it to be true (which in fact I doubt), still that value is not ‘derived’ in the same way: it does not derive from the value of something else.
One way to put this point is that the good of action may be instrumental, while the truth-based good of belief looks intrinsic.

[Ray Cheung]

Thanks Prof. Dancy for the response. I agree with what you said, and would like to add a little bit more on actions' and beliefs' being valued in itself and in virtue of other things, and see what you think. Actions are very often valued in virtue of its outcome or goal. For example, that was a good thing that he did because what he did made his friend happy. I guess actions could also be valued in itself in some special cases. Suppose a patient has been paralyzed in bed for some time, yet he is alert. One day he could suddenly move his arm again (and he repeats the action in excitement). This may be a case in which an action is valued in itself.

On the other hand, beliefs could also be valued in virtue of other things. For example, it may be good for a patient to believe that the chance of recovery is high (regardless of whether it really is) because that belief would make him have a higher chance of recovery. Yet, like actions, there are some properties of beliefs that we value in itself, as you have also pointed out in the comment.

If these are correct, one may ask: is there anything that underlies all these values? I am tempted to think that pleasure is what really is intrinsically valuable, and that it underlies the goodness of, say, an altruistic act, a sound argument, and chocolate ice-cream. But then I know this is controversial, and I will need more arguments for it.

[Constantine Sandis]

Hi Jonathan,

Thanks for a great paper. This is changing direction from the comments so far, but my interest was piqued by your discussion of beliefs and their purported objects and contents.
On p.9 you identify what I am afraid of with the object of my fear, but I would have thought that what I believe/fear/suspect and the object of my belief (fear, suspicion, etc.) are two different things e.g. the object of my belief may be tomorrow’s weather whereas what I believe would not be the weather but something like ‘that it will be sunny’. At least this seems to be the case if we understand the object of a belief as its intentional object viz. what the belief is about/directed at. I'm not sure what else we can mean by an object of belief…
On the same page you write ‘what I believe is a putative state of affairs, something capable of being the case but not of being true’. But the (putative) state of affairs of the sky falling on my head is not something one can believe (though one might believe in it, in a mystical/religious sense, as one might believe in the putative state of affairs that is judgement day). Also, what I believe (that the sky will fall on my head) is capable of being true (in which case it will also be the case that the sky…). So perhaps ‘that the sky will fall on my head’ is not a putative state of affairs but, at best, a purported fact.
I say at best because of the obvious worry here: we talk about the fact that the sky will fall on my head, the belief that that the sky will fall on my head, the proposition that the sky will fall on my head, and so on (suspicion, fear, hope etc.). ‘What I believe’ cannot be all of these things, and there might even be good enough reason to think that it can be none of them.
As you suggest, what I believe can be the case, whereas facts, beliefs, and propositions cannot be the case. In addition, facts are problematic here because it makes no sense to talk of true and false facts yet what I believe may be true or false. Belief is problematic because I don’t believe my beliefs anymore than I suspect my suspicions, desire my desires, fear my fears etc. (which is not to say that the term ‘belief’ doesn’t sometimes denote a thing believed). This probably explains why most philosophers go with proposition here, though I think that you and Alan White are right to dismiss this option.
Indeed I’m not sure that we need to give in at all to the pressure to reduce the notion of what one believes to that of an entity of any sort, be it a fact, state of affairs, or proposition. Why suppose that what I believe is an entity at all? My belief that the sky will fall on my head may be something that exists but I don’t see any reason to think that what I believe (‘that the sky will fall on my head’) is something that exists, even if my belief turns out to be true. If this is right, then we should resist the temptation to identify ‘things believed’ with entities of any sort. There are things people believe, but these things are neither universals nor particulars (though they may fall under types in the ordinary sense of the word).
I don’t see what the content of an action, or indeed a perception, could be. If this is right then we might also worry about what the content of a belief is meant to be. Unless ‘content’ is a technical term that just means what is believed, perceived, etc. in which case it cannot be a proposition. There are things done, things perceived, and things believed etc. What is added by introducing the misleading notion of content? (Misleading because beliefs cannot seriously be thought to contain the thing believed/feared/suspected even if we think of this as a proposition, let alone a state of affairs or some feared creature).

Constantine

[Jonathan Dancy]

This is in response to Ray Cheung's last.

Yes indeed some beliefs may be valuable for their effects, and some actions may be valuable as ends. But we are looking here for a fully general account of what explains the favouring relation when what is favoured is an action, on the model of the truth-relation that explains favouring when what is favoured is a belief (at least in inferential cases).

I am afraid that I am among those who reject all forms of hedonism. The main attraction of hedonism, as I see it, is that it is simple. But I have always liked Ross's remark that it is better that our theory be true than that it be simple.

JD

[Jonathan Dancy]

Thanks for those thoughts, Constantine. The issues you raise are ones to which I am very sensitive. You have read Alan White's wonderful paper 'What we Believe', which discusses these issues. White's conclusion seems to be the one that you prefer, namely that the things we believe are not things at all, and that all attempts to find a metaphysical slot for them are misguided. I would believe this if I could. Perhaps I am misled by starting from other so-called propositional attitudes, such as fear, doubt and suspicion. I can be afraid of him and afraid that he will hit me. I can believe him and believe what he says, that he will hit me. If the that-clause specifies an object (rather than a content) for the attitude, it needs to specify something of which one can be afraid, in the first case, or believe, in the second. We are agreed that no proposition can fit the bill in the first case, fear, and there is a strong temptation to say the same of the belief case, contrary to established philosophical practice. Then we ask what sort of thing is 'that he will hit me'. It sounds to me like a prospect, that is something that may or may not turn out to be the case - and this is very like what I said of belief, that the thing believed is a putative state of affairs.
However I confess that this entire area is one in which, though I see part of my way, there remains much that is very murky. I would not be surprised if I ended up where you are urging me to go.
Yours, Jonathan

[Jonathan Dancy]

Thanks for those thoughts, Constantine. The issues you raise are ones to which I am very sensitive. You have read Alan White's wonderful paper 'What we Believe', which discusses these issues. White's conclusion seems to be the one that you prefer, namely that the things we believe are not things at all, and that all attempts to find a metaphysical slot for them are misguided. I would believe this if I could. Perhaps I am misled by starting from other so-called propositional attitudes, such as fear, doubt and suspicion. I can be afraid of him and afraid that he will hit me. I can believe him and believe what he says, that he will hit me. If the that-clause specifies an object (rather than a content) for the attitude, it needs to specify something of which one can be afraid, in the first case, or believe, in the second. We are agreed that no proposition can fit the bill in the first case, fear, and there is a strong temptation to say the same of the belief case, contrary to established philosophical practice. Then we ask what sort of thing is 'that he will hit me'. It sounds to me like a prospect, that is something that may or may not turn out to be the case - and this is very like what I said of belief, that the thing believed is a putative state of affairs.
However I confess that this entire area is one in which, though I see part of my way, there remains much that is very murky. I would not be surprised if I ended up where you are urging me to go.
Yours, Jonathan

[Constantine Sandis]

Thanks Jonathan. The view you propose sounds a little like that found in Hornsby, McDowell (in Mind & World), the early Wittgenstein (Tractatus 3.02), Frege, and Aquinas, where 'what we think' is literally a part of the actual world (or of a possible world when we get contingent things wrong).
One worry here (motivated by Alan White) is that, assuming that we can believe things that people say, you seem to end up with a view where what is said (agreed with, etc.) is a putative state of affairs. And that just doesn't sound right to my ears.

Yours, Constantine

[Cheri]

clap clap Candace wins