The Teleological Conception of (Practical) Reasons (TCR), Part II

My previous formulation of TCR, left something to be desired. Here’s what I hope is a better formulation:

TCR:  (1) S has better reason to do x than to do y if and only if S has better object-given reasons to intrinsically prefer Ox to Oy than to intrinsically prefer Oy to Ox. (2) If S has better reason to do x than to do y, then it is so solely in virtue of the fact that S has better object-given reasons to intrinsically prefer Ox to Oy than to intrinsically prefer Oy to Ox.

From (1), we derive the following: (3) S has a reason to do x if and only if S has an object-given reason to intrinsically prefer Ox to Oø, where Oø is the "outcome" where S does nothing. (4) S has a reason to refrain from doing x if and only if S has an object-given reason to intrinsically prefer Oø to Ox.

Update: Here's a further question: If we accept TCR and moral rationalism (S is morally required to do x if and only if S has a decisive reason to do x) , don't we have to conclude that the correct moral theory has to be teleological (in the way, for instance, that ethical egoism, utilitarianism, and agent-relative consequentialism all are)?

Some clarifications: (a) An act’s outcome (e.g., Ox) includes everything that is the case given that the act has been performed. Thus Ox includes the fact that S did x. (b) S intrinsically prefers Ox to Oy if and only if S desires Ox for its own sake more than S desires Oy for its own sake. (c) S’s reason to prefer Ox to Oy is an object-given reason (as opposed to a state-given reason) if and only if it is provided by facts about Ox and Oy (the objects of S’s preference) as opposed to facts about S’s state of preferring Ox to Oy.

(i) Although Scanlon lumps the two together (1998, 79-80), TCR is independent of TCV (the teleological conception of value), the view that states of affairs or world histories are the primary bearers of value. On TCV, the only kind of value is the kind that is to be promoted, and so the only proper response to value is to promote it, to bring it about, or to preserve it. But there is no reason why the proponent of TCR cannot deny TCV and hold that the primary bearers of value are persons and things and that the proper response to certain things is to love, honor, and/or respect them.

(ii) As Scanlon notes (1998, 81), TCR is compatible with agent-relative reasons. It may be that what I have reason to prefer is not what you have reason to prefer.

(iii) TCR does not imply that the only reasons provided by my wife and our relationship are reasons to promote certain states of affairs. The proponent of TCR can accept that I have a reason to love her, and that this is not a reason to promote any outcome or state of affairs.

(iv) TCR is compatible with the BPV (buck-passing account of value). According to the BPV, “x’s being good/valuable” is equivalent to “x’s having the purely, formal higher-order property of having other properties that provide any evaluator with sufficient reason to respond favorably toward x.”

Now, as I said before, I find TCR extremely plausible. The left arrow of TCR’s bi-conditional seems unassailable: S has better reason to do x than to do y if S has better object-given reasons to intrinsically prefer Ox to Oy than to intrinsically prefer Oy to Ox. If there are better reasons to prefer one end to another, then, given the means-end principle, there are better reasons to take the necessary means to bringing about the one end over the other. Can anyone think of any counter-examples to the right arrow of TCR’s bi-conditional: S has better reason to do x than to do y only if S has better object-given reasons to intrinsically prefer Ox to Oy than to intrinsically prefer Oy to Ox?

Update: Here's a further question: If we accept TCR and moral rationalism (S is morally required to do x if and only if S has a decisive reason to do x) , don't we have to conclude that the correct moral theory has to be teleological (in the way, for instance, that ethical egoism, utilitarianism, and agent-relative consequentialism all are)?

Comments

[Mike Almeida]

Doug,

Quick question. Does 'Ox' refer to the actual outcome of performing x (what would occur were you to do x) or does it refer to what could occur were you to do x (assuming, for instance, that others cooperate in the realization of Ox), ro something else?

[Doug Portmore]

Mike,

I was taking Ox to refer to the actual outcome that would result from S's x-ing.

[Campbell]

Hi Doug.

The following seems like a counterexample.

Let x be preparing a lecture for Monday, and let y be watching TV. I have, let's assume, better reason to do x than to do y. Do I have any reason to (intrinsically) prefer Ox to Oy? Yes, I say. If I prefer Ox to Oy then I'm more likely to do x rather than y. But that's not an 'object-given' reason, because it's provided by a fact about my state of preferring Ox to Oy, the fact that my being in that state makes it more likely that I'll do x rather than y. Do I have any other reason to prefer Ox to Oy? I'm inclined to say no. If that's right, it follows that I have no object-given reason to prefer Ox to Oy, and hence that I don't have better object-given reasons to prefer Ox to Oy than to prefer Oy to Ox. But that contradicts TCR.

[Doug Portmore]

Hi Campbell,

When I perform such actions, I have some end that I'm seeking to achieve. This is what makes them intentional actions. For instance, when I prepare for lecture, I do so because I want my students to learn, because I want to fulfill my contractual promise to teach my students, because I want to enjoy the lecture, because I want to get good teaching evaluations, etc. And when I watch TV, I do so to relax and to enjoy myself. Now the fact that I'll be a lot happier in the long run if I prepare for Monday's lecture than I would be if I just watched TV seems to me like a good reason to prefer Ox to Oy. So can you say why you think that you have better reason to do x than to do y, and why some of the obvious candidates for reasons for preferring Ox to Oy (e.g., that I'll be better off in the long run, that my students will learn more, that there will be more overall utility, etc.) are not in fact object-given reasons for preferring Ox to Oy.

[Campbell]

Suppose the following are facts:

F1. If I do x rather than y, I'll be better off in the long run and my students will learn more.

F2. If I prefer Ox to Oy, I'll be better off in the long run and my students will learn more.

Here's my argument. F1 provides a reason for me to do x rather than y, but it doesn't provide a reason for me to prefer Ox to Oy. F1 does provide a reason for me to prefer Ox to Oy, but, since it is a fact about my state of preferring Ox to Oy, the reason it provides is a state-given reason. So neither F1 nor F2 provides an object-given reason for me to prefer Ox to Oy.

[Doug Portmore]

Campbell,

But how is this an arguement that there is no reason to prefer Ox to Oy? It may be that neither F1 nor F2 provides you with any reason to prefer Ox to Oy. But from that, it doesn't follow that you have no reason to prefer Ox to Oy.

Consider F3: If Ox as opposed to Oy obtains, you'll be better off in the long run and your students will learn more. Isn't this a good reason to prefer Ox to Oy?

I find it hard to believe that F1 would be true if you had no reason to prefer being better off and no reason to prefer that your students learn more.

Ask yourself, "Why F1 but not F4 constitutes a reason for doing x rather than y?"

F4. If you do x rather than y, you'll consume more electricity. (Assume that your computer uses more electricity than your TV.)

It seems that the explanation is that whereas you do have a reason to prefer that you'll be better off, you have no reason to prefer that you consume more electricity.

[Campbell]

Consider F3: If Ox as opposed to Oy obtains, you'll be better off in the long run and your students will learn more. Isn't this a good reason to prefer Ox to Oy?

No. Whether I have reason to prefer Ox to Oy depends on the outcome of my preferring Ox to Oy. But the outcome of my preferring Ox to Oy might not be that 'Ox as opposed to Oy obtains'.

Consider F5: if I prefer Ox to Oy, then Ox as opposed to Oy will obtain. The conjunction F3&F5 gives me a reason to prefer Ox to Oy, but that's a state-given reason, because F5 is about my preference. And F3 by itself gives no reason. So I still don't see a non-state-given reason to prefer Ox to Oy.

[Doug Portmore]

Campbell,

You write, "Whether I have reason to prefer Ox to Oy depends on the outcome of my preferring Ox to Oy." That's true only if by "reason" you mean "state-given reason." If, by contrast, "reason" refers to "object-given reason," then what you say is clearly false.

So now I'm wondering if you think that there are ever object-given reasons for desiring or preferring. Is it your view that there are only state-given reasons for preferring, and never object-given reasons for preferring? If so, what is your argument for this view?

If you do think that there are sometimes object-given reasons to prefer one state of affairs (or outcome) to another, could you give me an example and explain why there are object-given reasons in this example but not the one above.

[Mike Almeida]

TCR: (1) S has better reason to do x than to do y if and only if S has better object-given reasons to intrinsically prefer Ox to Oy than to intrinsically prefer Oy to Ox.

This might be similar to Campbell's point. Surely there must be cases where my reasons for prefering Ox pull apart from my reasons for doing x. Suppose it is true that if I prefer Ox and do x the outcome Ox is worse than if I do not prefer Ox and do x. I then have a reason to do x and not to prefer Ox. Maybe the only way to ensure that I produce moral value in doing x is to perform x without prefering Ox. If so, I have a reason to do x and a reason not to prefer to Ox. Maybe that's a little too Kantian. But there have to be other examples; it's just a matter of imagining one.

[Doug Portmore]

Mike, you write

Suppose it is true that if I prefer Ox and do x the outcome Ox is worse than if I do not prefer Ox and do x. I then have a reason to do x and not to prefer Ox. Maybe the only way to ensure that I produce moral value in doing x is to perform x without prefering Ox. If so, I have a reason to do x and a reason not to prefer to Ox.

You have a reason to do x and a state-given reason not to prefer Ox. But this isn't a counter-example to TCR, because TCR says only that reasons to do x have to track object-given reasons to do x.

Of course, you're absolutely right that reasons to do x can come apart from state-given reasons to do x. Here's another example. Suppose that I offer you a million bucks to do x, but that an evil demon threatens to kill you if you prefer Ox to O~x. In this case, you clearly have a reason to do x even though you have a decisive state-given reason not to prefer Ox to O~x.

[Campbell]

Is it your view that there are only state-given reasons for preferring, and never object-given reasons for preferring? If so, what is your argument for this view?

Hmm. I'm not sure I have a view on this issue. I hadn't thought about it before today. But the view you suggest, that there are only state-given reasosns for preferring, doesn't seem bonkers to me. Do you have an argument against it?

[Mike Almeida]

"You have a reason to do x and a state-given reason not to prefer Ox. But this isn't a counter-example to TCR, because TCR says only that reasons to do x have to track object-given reasons to do x."

I guess I don't offhand see a clear distinction here. Is the distinction between the intrinsic and relational properties of Ox? That can't be it, right? What are the intrinsic properties of an outcome of an action?? The outcome Ox is supposed to include all and only those things what would follow were x done. But some worlds in which x is done it is true that x is prefered. In other worlds where x is done, it is not preferred. Each is an outcome of x, and there seems no non-ciruclar way to count one as providing reasons to do x and the other as not providing reasons to do x. What then is the distinction, exactly?

[Doug Portmore]

Hi Campbell,

I haven’t thought about it carefully, because my main opponent (Scanlon) is committed to the existence of object-given reasons for desiring and preferring; his buck-passing account of value commits him to the existence of such reasons. In any case, below are two arguments against the view that there are state-given reasons for desiring and preferring, but no object-given reasons for desiring and preferring.

(1) Let’s assume that you have a state-given reason to desire Ox, viz., that your desiring Ox will have consequence C. How could the fact that your desiring Ox will have consequence C constitute a state-given reason for you to desire Ox unless C is something that you have an object-given reason to desire? That is, how can the fact that something is a means to an end give you a reason to take that means if the end is not itself something that you have (an object-given) reason to desire? It seems to me that instrumentality can transfer reasons from the end to the means, but it cannot create reasons where there were none in the first place.

(2) Other things being equal, Bob prefers the state of affairs where sentient creatures that live at some distance other than 1,000 light years away from him are benefited to the state of affairs where sentient creatures that live 1,000 light years away from him are benefited. The person who believes in object-given reasons for preferring can say that this preference is contrary to reason because there is no object-given reason to prefer that those who live at some other distance are benefited—whether or not a sentient creature lives 1,000 light years away from him is practically irrelevant. Now the person who believes in only state-given reasons for preferring can hold that Bob’s preference is contrary to reason only if having this preference has undesirable effects. But it may not have any undesirable effects since he never has the opportunity to choose between someone living on this planet and sentient creatures living 1,000 light years away from him. But surely we want to say that Bob’s preference is contrary to reason (and, thus, open to rational criticism) regardless of whether Bob’s being in the state of preferring this has any bad/undesirable effects.

[Doug Portmore]

Mike,

Here’s how Parfit draws the distinction:

Of our reasons to have some desire, some are provided by facts about this desire’s object. These reasons we can call object-given. We can have such reasons to want some thing either for its own sake, or for the sake of its effects…. Other reasons to want some thing are provided by facts, not about what we want, but about our having this desire. These reasons we can call state-given.

Does this help?

Perhaps an illustration will help as well. The fact that if I don't desire that P and believe that Q, an evil demon will inflict great suffering on me is a state-given reason to have this desire and to have this belief. But it isn't a reason to desire that P or believe that Q. Let P be that I possess a saucer of mud, and let Q be that the theory of evolution is false.

[Campbell]

How could the fact that your desiring Ox will have consequence C constitute a state-given reason for you to desire Ox unless C is something that you have an object-given reason to desire?

It could be that you have a reason to bring about C.

[Mike Almeida]

Doug, this does move the discussion along,

"Other reasons to want some thing are provided by facts, not about what we want, but about our having this desire."

But I don't think the distinction is especially clear. Sometimes our having a desire (or not) affects the value of the object of the desire. In those cases our having the desire (or not having it) affects the object-reasons we have. If we know this, then we have a reason to have a desire (or not)--an alleged state reason-- that is based on the knowledge that having the desire (or not) makes the object of desire more valuable--object-given reason. Merlin, let's say, offers you a stone whose value increases the less you desire it.

[Doug Portmore]

Campbell,

Do you think that you could have a reason to bring about C even if C isn't something that is desirable? How do you differentiate consequences that there are reasons to bring about (e.g., more utility) from consequences that there are no reason to bring about (e.g., less electricity)?

What about argument (2)? Do you think that Bob’s preference is contrary to reason only if having this preference has effects that we have reason to avoid bringing about?

[Doug Portmore]

Mike, You write,

Sometimes our having a desire (or not) affects the value of the object of the desire. In those cases our having the desire (or not having it) affects the object-reasons we have. If we know this, then we have a reason to have a desire (or not)--an alleged state reason-- that is based on the knowledge that having the desire (or not) makes the object of desire more valuable--object-given reason. Merlin, let's say, offers you a stone whose value increases the less you desire it.

This is an interesting case, but I'm not seeing what the worry is here? The extent to which Merlin's stone is valuable and thus provides me with object-given reasons to desire it increases the less I desire it. So I have a state-given reason not to desire it. And assuming that I don't actually desire it, I have a very strong object-given reason to desire it. What's the problem?

[Doug Portmore]

Mike,

One more thing: So, as you point out, sometimes our having a desire (or not) for X affects whether or not (or what degree to which) we have object-given reasons to desire X. But I don't see how this is a problem either for the distinction between state-given reasons and object-given reasons or for TCR.

[Campbell]

Do you think that you could have a reason to bring about C even if C isn't something that is desirable?

Yes. You could, for example, have reason to bring about pain and suffering. But I suspect that's not what you meant. You'll need to say more.

How do you differentiate consequences that there are reasons to bring about (e.g., more utility) from consequences that there are no reason to bring about (e.g., less electricity)?

I don't know. I guess one would need a theory of reasons for that. I don't have one.

What about argument (2)? Do you think that Bob’s preference is contrary to reason only if having this preference has effects that we have reason to avoid bringing about?

Yes. (Tentatively)

[Doug Portmore]

Campbell,

In response to this question:

Do you think that you could have a reason to bring about C even if C isn't something that is desirable?

You write: "Yes. You could, for example, have reason to bring about pain and suffering." What reason would that be? The only sort of reason that I can think of for bringing about suffering is that it would have the consequence of increasing net utility for that person on in general. But both things are clearly desirable. Are they not? Suffering can, then, be instrumentally desirable, can it not?

It would help further discussion if, when you claim that there is a reason to bring about C, you state what that reason is. That way, I can better assess whether C is not at all desirable, as you claim.

[Mike Almeida]

". . . So I have a state-given reason not to desire it. And assuming that I don't actually desire it, I have a very strong object-given reason to desire it."

I don't think so. Assuming you do not desire X, you have a strong object-reason to A = {acquire X without desiring X}. You have no object reason to B = {acquire X and desire X}. You have in any case no object-reason to desire X, since X is the sort of object that is valueless upon desire. There is therefore an object-reason to acquire X that is not an object-reason to desire X.
Otherwise we have an inconsistent description: S has object-reason to desire X iff. S does not have an object-reason to desire X. It follows from that that S does not have an object-reason to desire X. And you don't want that conclusion.

[Mike Almeida]

I mean, to complete the thought, you don't want that conclusion, though I do.

[Doug Portmore]

Mike,

Why should I accept this inference: "You have in any case no object-reason to desire X, since X is the sort of object that is valueless upon desire"?

I accept that I have no state-given reason to desire X (no reason to want this desire or to acquire it) since having this desire will render X valueless. But I don't see why I should accept your inference. After all, having an object-given reason to desire X doesn't render X valueless.

Besides, I'm having trouble imagining a case where something is valuable for its own sake iff one does not desire it. You can say that Merlin's stone has this strange property, but I can't imagine a stone being valuable for its own sake, let alone imagine that it has such non-instrumental value iff it is not desired. Part of the problem is that I think that evaluative properties have to supervene on other properties. So it would have to be a case where desiring X would change X's other base properties. So imagine a mental state that was pleasing except when you desired it. When desired, the mental state changed from being pleasing to painful. But isn't this a case where one mental state changes into another onr. It doesn't seem to be a case where the same mental state (with all the same non-evaluative properties) changes from positive value to negative value.

Also, I should note that I accept the buck-passing account of value according to which "S (some state of affairs) is good" is equivalent to "S has the purely formal, higher-order property of having other properties the provide us with sufficient object-given reasons for desiring that S." On this view, it's conceptually impossible for some state of affairs to be good and there not be any object-given reason to desire it. So if my possessing Merlin's stone is good when I'm not desiring to possess it, it has to be that there are object-given reasons for me to desire to possess it when I am not desiring to possess it.

[Mike Almeida]

"But I don't see why I should accept your inference. After all, having an object-given reason to desire X doesn't render X valueless."

Neither does having a state reason to desire X render X valueless. Here's what is strange about the reasoning.

1. I have no state reason to desire X BECAUSE desiring X renders it valueless.

2. I do have an object-reason to desire X DESPITE the fact that desiring X renders it valueless.

(2) is unintelligible to me. The object has lots of properties that make it desirable but one crucial property that makes it overridingly undesirable. Given the nature of the object my ON BALANCE object-reasons (reasons derived from the nature of the object) are to not desire it.
The properties of the object certainly give me reason to believe that the object is good or worthy and reason to acquire it. But it sounds just forced to say that the properties of the object give me (on balance) reason to desire it.

But then you say,
"On this view, it's conceptually impossible for some state of affairs to be good and there not be any object-given reason to desire it."

Conceptually impossible? Isn't it rather impossible *provided that* there is no world in which your other theoretical commitments are false? The proposed counterexamples are supposed to show that there are such worlds.

[Doug Portmore]

Mike,

You say, "The object has lots of properties that make it desirable but one crucial property that makes it overridingly undesirable." But there's nothing about the object that is undesirable. It's the desire for X, not X itself, that has this undesirable property.

Now you might claim, as Rabinowicz and Ronnow-Rasmussen do, that, for any property F, that the desire for X has, there is the following corresponding property that X has: viz., the property of being such that the desire for it has property F. But, as Stratton-Lake has argued, we should not let this doubling up of properties affect the strength of your reasons to desire, or not to desire, X. So if we think that the fact that your desiring X would render it valueless gives you a reason not to desire it, we shouldn't think that, on top of that, the fact that X has the property of being such that the desire for X renders it valueless provides you with an additional reason not to desire X. To sum up, the desire for X certainly has a property that gives us a reason not to desire X (a state-given reason), but we should not think that on top of this there is any object-given reason not to desire X. So, yes, (assuming that the object isn't desired) the object has lots of properties that make it desirable, but it (i.e., the object) has no properties that make it undesirable.

[Doug Portmore]

By the way, I retract the conceptual impossibility stuff. That was stupid of me.

[Mike Almeida]

Doug, you write this,

"But there's nothing about the object that is undesirable. It's the desire for X, not X itself, that has this undesirable property"

But X itself has the property of being devalued by desires. It's not an intrinsic property, it's a relational or structured property of X. But that's ok. It is a property relevant to our assessment of the value of X.
Compare: I would consider the property of being-such-as-to-collapse-when-sat-upon as one that makes a bicycle B undesirable. That too is a relational property. Similarly I consider the property being-valueless-when-desired as one that makes X undesirable. Since being-valueless-when-desired is a property of X, it provides an object-reason not to desire X. That's the idea.

[Doug Portmore]

I accept everything you said but your last sentence. As Stratton-Lake has argued, if we're going to say that the property of the desire provides a reason not to desire X, then we had better not say that this property of X also provides a reason (see above for the reason why). So I admit that the property of being-valueless-when-desired is a property of X, but I'm denying that is a reason-giving property.

[Campbell]

I accept everything you said but your last sentence.

You don't accept 'that's the idea'?

[Campbell]

Doug, I think I'm losing track of the dialectic. So let me back up a little.

I want to say that there are certain things -- e.g. happiness, let's say -- that we have reason to bring about. So, if our having a certain desire would bring about happiness, then that provides a reason for us to have the desire. But, I assume, it would be a state-given reason.

Now you say: if we have reason to bring about happiness, then (a) there must be something that makes it the case that we have reason to bring about happiness, and (b) nothing could make that the case except our having an object-given reason to desire happiness.

But there seem to be other options here. Some might say our desiring happiness makes it the case that we have reason to bring it about. Others might say the fact that happiness is good makes it the case that we have reason to bring it about. Still others might say it's just a brute fact that we have reason to bring about happiness. I don't know which of these things, if any, I would want to say. But none seems less plausible, prima facie, than the view you've suggested.

Admittedly, it is natural to say that happiness is desireable. But the term 'desireable' seems to ambiguous to draw any conclusions from this.

Finally, a question of clarification. Suppose it's a fact that if I desire that my students do well, then they will do well. That fact seems to be both about a certain state of mine, and about the object of that state. So, if this fact provides a reason for me to desire that my students do well, it's not clear whether it's state-given or object-given, or both. Which is it?

[Doug Portmore]

Sorry, I should have said. I accept everything Mike said but these two sentences: "Since being-valueless-when-desired is a property of X, it provides an object-reason not to desire X. That's the idea."

[Doug Portmore]

Campbell,

Originally, you offered what seemed to you to be a counter-example to TCR: a case where there was a reason to do x rather than y but no object-given reasons to prefer x to y. I suggested, since you didn't specify, that the reason for doing x rather than y was that you would be happier in the long run. You seemed to accept this suggestion. So you have to say that the fact that you'll be happier if Ox obtains than if Oy obtains is no object-given reason to prefer Ox to Oy. So you need to claim that happiness isn't something that we have object-given reasons to desire and object-given reasons to prefer more of. This is what I find implausible.

Now you note that there are a number of possibilities about why you might say there is a reason to bring about more happiness for oneself. I admit that there are those possibilities. But what is it that you want to say given that you claim that there is more reason to do x than to do y? My response will depend on what your view is.

Regarding the clarification question, it's a state-given reason.

[Adam]

Hi Doug,
This is a broader criticism than those above, but seems worthwhile to add.
It strikes me that "object-given reasons" are rather too abstract for agents to consider in a way that we can actually use. It'll work well for thought experiments and post-hoc examinations, but not so well for acting. Perhaps to improve the theory we may need to find a method to exclude other considerations (TRV, inclinations, narrative up to the point of deliberation of doing x, etc) in order to isolate the object-given reasons, so one can practically practically reason about doing x.

[Doug Portmore]

Hi Adam,

If I understand your criticism, it's not a criticism of TCR as a criterion for what we have reason to do, but of TCR as a decision procedure for determining what we have reason to do. But TCR isn't a decision procedure; it's a criterion. So TCR may be true even if, as you claim, it isn't of much practical use.

[Adam]

hi Doug,
I do understand the distiction between them.
Perhaps I should write longer posts more fully explaining myself.
What troubles me is that decision procedure for a plausible moral theory must get the criterion right, most preferably all of the time, at second best most of the time, and at the very least sometimes.
I've been thinking of your TCR, and wondering how I would (or could) do things if it were true, that is, I've been trying to formulate a decision procedure. But the TCR seems too abstract: there seems to me to be (what in epistemology is called) a "problem of access".

[Doug Portmore]

Hi Adam,

So you accept that TCR is a criterion and not a decision procedure. And you accept that a decision procedure must conform to the correct criterion (trying to mimic, as much as practically possible, its verdicts), not vice versa. And you want to claim that it is going to be a difficult to formulate a decision procedure for TCR. Do you think that this calls TCR into doubt in some way? If so, how so? It seems that if it's difficult (or even impossible) to come up with a useful decision procedure for a given criterion, this doesn't support the conclusion that the criterion isn't the correct one. The only conclusion to draw from this is that if it is the correct one, we're going to have a hard time deciding what to do. That's a problem for us, not for the view (TCR).

[Campbell]

Doug,

So you need to claim that happiness isn't something that we have object-given reasons to desire and object-given reasons to prefer more of. This is what I find implausible.

I doubt I'll succeed, but let me try to convince you that it's not so implausible. Here, briefly, are my reasons for favouring the view that there are no object-given reasons.

1. It's intuitive. Imagine that your desiring happiness makes no difference to anyone's life. It doesn't affect how happy anyone is, or how well satisfied their desires are, or how successful in pursuing their goals they are, or anything like that. With respect to people's lives, your desiring happiness is entirely inconsequential. Now, ought you to desire happiness? My intuition says, 'who cares?' There's no reason to desire happiness, nor any reason not to do so.

2. It's parsimonious. We shouldn't multiply kinds of reasons beyond need, and we don't need object-given reasons in addition to state-given reasons. You, of course, will deny this. You claim that, without object-given reasons, we can't explain how we have any reasons at all. But I'm not convinced. As I said earlier, there are other ways to explain reasons, and I don't see why these are all hopeless. (And, sorry, I can't tell which of these alternatives I favour, because, like I said, I don't know.)

3. It gives a unified account of reasons for desire and reasons for action. Any reasons you have to perform an action depend on the outcome of the action (where outcome is defined broadly, as you suggest). I assume you would accept this. It would be nice, then, to have a unified account according to which any reason you have to desire something depends on the outcome of your desiring that thing.

[Doug Portmore]

Hi Campbell,

Regarding (1), do you feel the same about the belief that P? Suppose that you have good object-given reasons to believe that P (i.e., the evidence supports P) but no state-given reasons to believe that P (your believing or not believing that P will have no positive or negative effects). So assume that your believing that P is entirely inconsequential. Would you say that there's no reason to believe that P, nor any reason not to believe that P? If not, why do you treat desires and beliefs differently. Are they both "judgment-sentitive attitudes" in Scanlon's sense?

Regarding (2), you don't think we need object-given reasons in addition to state-given reasons? So you think that there are only state-given reasons to believe? In that case, the only way I'll convince you of TCR is to show that your being in the state of believing TCR will have good consequences. I'm not sure that I can show that.

Regarding (3), I want to give a unified account as well. I want to say that there are only object-given reasons. You say that the fact that your desiring X will have good consequences is a state-given reason for desiring X. I say that this fact isn't really a reason to desire X; rather, it's an object-given reason to want to desire X and to intend to do what will bring it about that you desire X.

[Campbell]

Regarding (1), do you feel the same about the belief that P?

Very good question. I do concede that the intuitive support for object-given reasons is stronger in the case of belief than in the case of desire. Still, I say, there are no object-given reasons for belief. That may sound radical, but it's not really. Many purported object-given reasons can be assimilated as state-given reasons. For example, it is in your interests, and in the interests of others, that you believe in accordance with the evidence. Hence, if the evidence supports P, that provides a state-given reason for you to believe that P.

So you think that there are only state-given reasons to believe? In that case, the only way I'll convince you of TCR is to show that your being in the state of believing TCR will have good consequences. I'm not sure that I can show that.

Don't be so pessimistic. As I say, believing in accordance with the evicence has good consequences. So, to convince me of your view, you need only show me that it is supported by the evidence. In that respect, I'm no harder to convince than a believer in object-given reasons.

You say that the fact that your desiring X will have good consequences is a state-given reason for desiring X. I say that this fact isn't really a reason to desire X; rather, it's an object-given reason to want to desire X and to intend to do what will bring it about that you desire X.

This is an interesting suggestion. But it doesn't lead to a unified account of reasons for action and reasons for desire, at least not in the sense I intended. On your proposal, reasons for action and reasons for desire differ in the following way. A reason for desire is independent of the outcome of that for which it is a reason. For example, according to you, the fact that your desiring X would, say, make you happy provides a reason for you to desire that you desire X. But that reason is independent of the outcome of your desiring that you desire X. (It's not independent of the outcome of your desiring X, but your desiring X is not that for which it is a reason.) A reason for action, on the other hand, is not independent of the outcome of that for which it is a reason.

I wonder if the following is an argument against your view. Consider self-defeating desires. Imagine, for example, that you have an important appointment tomorrow for which you need to be are well-rested. So it's very much in your interests to sleep well tonight. Suppose, however, that desiring to sleep is self-defeating. If you desire to sleep, you won't sleep, whereas if you desire not to sleep, you will.

I say you have a state-given reason to desire not to sleep: if you desire not to sleep, you will be well-rested tomorrow. But you disagree. On your view, you have no such reason. Rather, you have an object-given reason to desire that you desire not to sleep.

But suppose this higher-order desire would also be self-defeating. If you desire to desire not to sleep, then you won't desire not to sleep. Then, according to your view, the only desire that is supported by a reason (i.e. the desire to desire not to sleep) will have the outcome that you are not well-rested. Whereas, according to my view, the only desire that is supported by a reason (i.e. the desire not to sleep) will have the outcome that you are well-rested. According to your view, if you desire as you have reason to, you'll not sleep a wink. Whereas according to my view, if you desire as you have reason to, you'll sleep soundly.

I like my view more.

[Campbell]

What is a 'judgement-sensitive attitude', by the way?

[Doug Portmore]

Hi Campbell,

You write,

Many purported object-given reasons can be assimilated as state-given reasons. For example, it is in your interests, and in the interests of others, that you believe in accordance with the evidence. Hence, if the evidence supports P, that provides a state-given reason for you to believe that P.

Do you think that this is always true? Can't there be cases where the evidence supports P, but there isn't any state-given reason for you to believe that P? On your view, is it the evidence that provides the reason as you suggest in the last sentence quoted above? I would have thought that the fact that believing P had good consequences would be the state-given reason. To say that the evidence for P is a reason to believe that P sounds like an ascription of an object-given reason to believe that P. So isn't it true that on your view the evidence never provides any reasons for believing that P? On your view, it's always the consequences of believing that provide the reasons for believing P.

Regarding (3), admittedly my view isn't unified in your sense. (It is just as parsimonious, though.) But I wonder why we should expect it to be unified. After all, the intentional object of an intention is an action (an event) whereas the intentional object of a desire is a proposition. Propositions don't have effects, only events do. So if the reasons stem from their objects, we should expect a disunified account. Thus I think that your suggestion that reasons for desiring should be unified (in your sense) with reasons for intending begs the question. It's plausible to expect a unified account only if there are only state-given reasons.

Judgment-sensitive attitudes are attitudes for which normative reasons can be asked for and offered. But, as I understand you now, you're not Humean about desires. You accept that we can rationally assess ends.

[Campbell]

So isn't it true that on your view the evidence never provides any reasons for believing that P? On your view, it's always the consequences of believing that provide the reasons for believing P.

Yes, admittedly, what I said before was a little sloppy. Perhaps I should say that the evidence is part of what provides the reason. The reason is provided by the evidence plus the fact (if it is one) that it's in your interests to believe in accordance with the evidence. Does that work? In any case, I think I can say that, in the usual case, if there is good evidence for P, then we have reason to believe that P, without thereby committing to object-given reasons.

Propositions don't have effects, only events do. So if the reasons stem from their objects, we should expect a disunified account.

Actions have effects, and so do desires. So I don't see why we shouldn't expect a unified account of reasons for action and reasons for desire, which is the only kind of unified account I have suggested.

But, as I understand you now, you're not Humean about desires. You accept that we can rationally assess ends.

What did I say to give you that impression?

And what did you think of my objection, the one about sleeping?

[Doug Portmore]

Perhaps I should say that the evidence is part of what provides the reason. The reason is provided by the evidence plus the fact (if it is one) that it's in your interests to believe in accordance with the evidence. Does that work?

But what if it’s in your interest (and contrary to no one else’s interests) to believe what’s contrary to the evidence. In that case, you have to say that the evidence against P provides no reason for believing that not-P, right? So if my believing that God exists or that I have an immortal soul would have better consequences than my not believing these things, then I have good reason to believe these things even if all the evidence suggests that these things are false. The evidence itself (in absence of there being good consequences associated with my believing in accordance with evidence) provides no reasons to believe or not to believe on your view. Is this supposed to be intuitive?

Actions have effects, and so do desires. So I don't see why we shouldn't expect a unified account of reasons for action and reasons for desire, which is the only kind of unified account I have suggested.

We should accept a unified account of reasons for intending to X (an action) and reasons for desiring that P (a proposition) if all reasons are state-given reasons, because, as you note, both states--intentions and desires—have effects. But if reasons are object-given, then we shouldn’t expect a unified account of reasons for intending to X (an action) and reasons for desiring that P (a proposition), because whereas the objects of intentions (i.e., actions) do have effects, the objects of desires (i.e., propositions) don’t. So it seems to me that you’re just begging the question in presupposing that there should be a unified account of the two.

What did I say to give you that impression [the impression that you think that we can rationally assess ends/desires]?

Well, you think that there are reasons for and against having a certain desire/end, right? You think the reasons are state-given reasons, but they’re reasons. So I assume that you would hold that if I acknowledge that my desiring that P would have good consequences and no bad consequences, then I have reasons, reasons of which I am aware, for desiring that P. So if I fail to respond to these reasons (if I fail to desire that P), then presumably I would be irrational. But this is where I think that your view is very odd. State-given reasons are not the sorts of reasons that I can respond to. I don’t see how my knowing that my desiring that I eat broccoli would have good consequences or my knowing that my believing that I have an immortal soul would have good consequences can lead me to have this desire and this belief. It seems that what I respond to when my beliefs change is the evidence (the object-given reasons for belief) and that what I respond to when my desires change is my assessments of desirability of their objects (again, the object-given reasons for desire). And if you asked me to justify my belief in evolution, I would cite the evidence I have, not the fact that my believing in evolution has good consequences. So your view seems to get our practice of both responding to reasons and giving reasons all wrong.

To sum up: (1) Your view doesn’t seem at all intuitive. I don’t see how you can say that the view that the evidence for P is not itself a reason for believing that P is an intuitive position. (2) Your view isn’t any more parsimonious than mine. You believe there are only state-given reasons, and I believe there are only object-given reasons. (3) The fact that your view offers a unified account of reasons for intending to X (an action) and reasons for desiring that P (a proposition) doesn’t constitute any independent grounds for favoring your view for we should expect a unified account if and only if all reasons are state-given reasons. As you predicted, then, I remain unconvinced. But I don’t think that I’m just being stubborn, but then maybe I’m just being stubborn in think that.

[Doug Portmore]

Regarding your sleeping example, I believe that I have an object-given reason to intend to do whatever will make it less likely that I will have the self-defeating desires. This reason is an object-given reason. So I don't see any problem. And, in fact, this seems to be the reason that I can respond to. For instance, if there's a pill that I can take to ensure that I won't have the self-defeating desire, I'll take it. By contrast, I don't see how knowing that my desire to sleep is self-defeating can get me not to desire to sleep. It's not a reason that I can respond to.

[Adam]

Hi Doug,
My comments were not made in an effort to "shoot down" your suggestion. Although picking apart an argument does have some useful role, I also think that its not the main aim and driving force behind thinking about morality.
You have made a suggestion (the TCR) - great!
Now, lets look at how that fits in with, or what the consequences are for, the other areas of morality, and philosophy - or, *a* philosophy - generally.
I wanted to follow the argument where it lead (a decision procedure), and I found it rather awkward, and I told you why.

[Campbell]

For instance, if there's a pill that I can take to ensure that I won't have the self-defeating desire, I'll take it. By contrast, I don't see how knowing that my desire to sleep is self-defeating can get me not to desire to sleep. It's not a reason that I can respond to.

I don't understand. Why wouldn't taking the pill be a way to 'respond' to the state-given reason?

[Campbell]

Just to follow up my previous comment:

Let's say that a reason is satisfied iff the thing for which it is a reason obtains. Ane let's say that a person reacts to a reason iff she recognises the reason and her recognition of it moves her to do what is required to satisfy it.

Now, suppose you have a state-given reason to desire not to get to sleep, because that will have the good consequence that you sleep soundly. Surely, you could react to this reason by taking the pill you suggested.

So, we must ask, is reacting to a reason what you mean by 'responding' to a reason? If it is, then the example shows that one can respond to a state-given reason. If it isn't, then I don't see why it matters whether one can respond to state-given reasons. What's important is that one can react to them.

[Doug Portmore]

Hi Campbell,

Let's say that a reason is satisfied iff the thing for which it is a reason obtains. An[d] let's say that a person reacts to a reason iff she recognises the reason and her recognition of it moves her to do what is required to satisfy it.

I would change the second sentence to: "a person reacts to a reason iff she recognizes the reason and her recognition of it moves her to have the attitude that the reason supports her having." On your view, I react to my reason to believe that P iff I recognize the reason I have to believe that P and my recognition of this reason moves me to do/intend what is required to ensure that I believe that P. This doesn't seem right to me. I can react to my reason just by believing that P. I don't have to perform any intentional act. E.g., I see the cat in the backyard and believe that there is a cat in the backyard. I’ve reacted to the reason without having to have done anything. (I’m assuming that believing that there is a cat in the backyard is just a reaction, not an intentional/voluntary act – at least, not in this kind of case.)

Now I don’t know whether you would accept this revision. But if you would, then I think that you can see why I think that one can’t react to state-given reasons. For instance, suppose that you are an atheist (if you’re not already) and suppose that you’re persuaded by Pascal’s wager. Do you think that such pragmatic considerations would ever move you to believe in God? I think that you might be moved to desire to believe in God and to intend to do whatever will make you more likely to believe in God (perhaps, go to Church), but it won’t ever by itself (or directly) move you to believe in God. At least, that’s the way things seem to me. State-given reasons to believe never move me to believe anything. I might recognize that I would be better off believing that I have no immortal soul, but that just doesn’t seem capable of moving me to believe that I do. As with belief, I think the same goes with desire and intention. The reasons that move me are object-given reasons.

[Doug Portmore]

Campbell,

Also, note that what I say above comports with our practice of offering justification for our beliefs, desires, and intentions. If someone asks me to justify my belief that P, I offer object-given reasons for believing that P, not state-given reasons for believing that P. For instance if you asked me justify my belief in evolution, it would be quite odd for me to give as my reason for believing this that my believing this has good effects. Thus, when asked I point to evidence, the object-given reasons for believing in evolution.