The Buck-Passing Account of Value and a False Dichotomy

Consider the following passage from Scanlon:

 

[B]eing good, or valuable, is not a property that itself provides a reason to respond to a thing in certain ways. Rather, to be good or valuable is to have other properties that constitute such reasons. (Scanlon 1998, 97)

 

Thus Scanlon is a proponent of the buck-passing account of value (BPV) who accepts both the following negative thesis and the following positive thesis:

 

BPV- Being good is not a property that itself provides us with, say, reasons to act so as to promote what has this property.

 

BPV+ Something’s being good is just its possessing the purely formal, higher-order property of having some lower-order properties that provide us with, say, reasons to act so as to promote it.

 

Now, in the quoted passage, Scanlon seems to be suggesting that if BPV+ is true, then BPV- can’t be true. [Update 8:40 1/24: There's a typo. Replace 'can't' with 'must'.]  But this seems to be a false dichotomy. As a number of people have pointed out, I believe, BPV- may not follow from BPV+. Something’s being good could just be its possessing the purely formal, higher-order property of having some lower-order properties that provide us with, say, reasons to act so as to promote it while that purely formal, higher-order property itself provides additional reasons to act so as to promote it. Have I got that right? (Admittedly, I’ve only skimmed a small fraction of the relevant literature.)

 

But I’ve always had what seems to me to be a different worry. Even if BPV- did follow from BPV+ as stated above in terms of reasons for action, it doesn’t follow that one can’t adopt a version of BPV+ while rejecting the negative thesis that goodness does not itself provide us with reasons for action. Let me explain.

 

Suppose that it’s true that something’s being good cannot provide us with reasons for having a certain response to it if its being good just amounts to its having other properties that constitute reasons for having that exact same response to it. Nevertheless, the claim that being good is a property that provides us with reasons for responding to what has that property in a certain way (say, by intending to act so as to promote it) is compatible with the claim that something’s being good is just its possessing the purely formal, higher-order property of having some lower-order properties that provide us with reasons for having a different response to it (say, desiring it). That is, the following seem compatible:

 

~(BPV-) Being good is a property that itself provides us with reasons to act so as to promote it.

 

BPV(D)+ Something’s being good is just its possessing the purely formal, higher-order property of having some lower-order properties that provide us with reasons to desire it.

 

It seems to me, then, that one can accept a positive version of the buck-passing account of value while accepting that the reasons we have to desire certain states of affairs (i.e., their goodness) provide us with reasons to act so as to bring them about. Does this seem right?

Comments

[Jussi Suikkanen]

No Doug. I think you got it the wrong way around. What most commentators have pointed out is that the positive thesis does not follow from the negative thesis. This is for instance Dancy's and Vayrynen's view and others have endorsed it too. So, these people think that the reasons an object provide are provided by its first-order properties but that the object still has evaluative properties that are independent of merely having reason-providing properties. They just are not reason-providing. Vayrynen for instance says that the evaluative properties explain the reason-providing relations. But, as far as I know, no-one has held that the implication does not hold the other way round. No-one has even suggested that it would. Scanlon runs the argument the opposite way.

The compination of views you give is really odd too unless you think that reasons to so to act as to promote the thing and reasons to desire it come apart in a radical way. I mean you get an odd double-counting feature. You start from the first-order non-evaluative properties that provide reasons. By that token, you get the second-order property of having first-order properties that provide reasons. And, by having those formal properties, you get even more substantial reasons to do things. Dancy is good in arguing that such reasons are redundant. You already have enough reasons on the first-order level. Whether they are provided by evaluative properties or not is another matter. Anyway, motivating the idea that a purely formal property of having reason-providing properties would in itself be reason-providing seems like an uphill struggle.

[Jamie Dreier]

I can't follow the dialectic.
First, I think there is a small mistake in the entry. Doug writes

Now, in the quoted passage, Scanlon seems to be suggesting that if BPV+ is true, then BPV- can’t be true.

I think the can't is supposed to be a must. Right?

Second, since (as far as I can tell) BPV- does not follow from BPV+, it's hard to think about Doug's question of what else would be true if it did follow. The best sense I can make of this is to ask myself whether BPV+ follows from Doug's BPV(D)+. I'm sure it doesn't! But Doug, is that what you were asking?

(I also think Jussi is wrong, but I'll leave it to Mark Schroeder to say why, since Mark has definitely argued that the negative thesis doesn't follow from the positive one.)

[Robert Johnson]

True, BPV- doesn't follow from BPV+, but (isn't this implicit in what Jussi says??) just add:

That something has a purely formal, higher-order property of having lower-order properties that provide reasons to promote a thing does not itself provide us with an additional reason to desire that thing.

And that seems true, doesn't it? What additional reason could that higher-order property possibly provide?

[Sven Nyholm]

Robert,

Here’s an example of something that people may take to be a reason given by the kind of higher-order fact about reasons that is discussed in this thread.

Derek Parfit imagines that he is told by some friendly and reliable adviser that he has a strong, but confidential reason to immediately go home. Since he does not know what that reason is, he only knows the higher-order fact that something gives him a strong reason to go home. If this gave him no reason to go home, Parfit then argues, he could rationally stay where he is. But, he cannot rationally stay where he is. So, Parfit concludes, the higher-order fact that he has some reason to go home does give him a reason to do so. But, he adds that this is not a non-derivative, independent reason; its normative force derives entirely from the actual confidential reason that he has to go home.

Is this an example in which the higher-order fact that something gives us some reason itself gives us a practical reason? Here is an alternative way of looking at this situation. We may instead claim that the fact that some reliable and friendly adviser has told us that we have reason to do something gives us reason to do this thing. Thus it might be that Parfit’s reason to go home is given by the fact that some reliable, friendly adviser has told him that he has reason to do so. But, as before, it seems that this is only a reason insofar as our advisers are right in saying that we have reasons to do these things. That is, something seems to count as good advice only insofar as we actually have reason to do what we are told that we have certain reasons to do. So, if advice provides reasons, these reasons also seem to derive their normative force from other reasons.

[Doug Portmore]

Jamie and others,

Yes, what I should have said (and what I was thinking) was that Scanlon seems to be suggesting that if BPV+ is true, then BPV- must (not can't) be true. Sorry about the mistake.

What I'm asking is whether one can accept both BPV(D)+ and ~(BPV-). It seems to me that one can. And if one can, then one can accept a version of BPV, viz., BPV(D)+, and still be a teleologist in that one holds that the property of being good is itself a property that provides us with reasons to promote what's good.

[Doug Portmore]

Jussi,

You write,

The compination of views you give is really odd too unless you think that reasons to so to act as to promote the thing and reasons to desire it come apart in a radical way. I mean you get an odd double-counting feature.

I don't follow.

The thought is that the fact that I have a reason to desire an end, E, provides me with a reason to perform an act, A, that is a means to bringing it about that E obtains. On BPV(D)+, to say that there is reason to desire E is just to say that E is good, and it's this property, the teleologist supposes, that provides us with reasons to act in ways that will achieve this end. Where's the doubling counting? Where the radical coming apart?

[Jamie Dreier]

Doug, there’s still so much tangle that I can’t get my mind around the question. I’m going to assume that a property provides us with a reason to do something just in case it is a reason to do that thing. (If not, then I need help understanding provision of reasons by properties.)

In that case, I think you are asking whether the fact that there is a reason to desire something could itself be a reason to promote that thing. Hm. Maybe. Can you give an example?

[Doug Portmore]

Jamie,

An Example: I have a reason to want Jones to be happy. My doing X would make Jones happy. Since I have a reason to want Jones to be happy and my doing X is a means to making Jones happy, I have a reason to do X.

“Reasons for me to make something my end are, owing to the hypothetical imperative, equally reasons for me to take the necessary means to it” (Darwall 1983, 16).

[Heath White]

I think the spirit of the buck-passing account is that there are a number of different evaluative properties--goodness, but also desirability, choiceworthiness, admirability, respectworthiness, etc.--all of which are higher-order properties whose analysis is in terms of lower-level properties which give one reasons to take up some attitude or action toward the object. The proposal you have, Doug, raises the question of why you would be a buck-passer about some but not all of these properties. That is, why are reasons to desire lower-level facts, but not reasons to promote? The distinction may be possible but needs some motivation. (Which, for all I know, you can provide.)

On a different note, re Parfit's example: I think the buck-passer's reply should be that Parfit has a reason to go home, he just doesn't know what it is.

[Jamie Dreier]

Doug,

If I understand the main question right, what has to be true in that example is that the reason you have to do X is: that you have a reason to want Jones to be happy.
Is that how you're understanding it?
It's not exactly obvious that that's a reason to do X. It's not obvious that it isn't.

[Doug Portmore]

Heath,

I don't see why I can't be a buck-passer about all evaluative properties. To say that a state of affairs is desirable is to say that it has other lower-order properties that provide us with reasons to desire it. Ditto for admirable and admire, choiceworthiness and choose, etc.

But intention is not an evaluative property. So I'm saying that I can be a buckpasser who accepts BPV(D)+ and the others above and claim that the fact that some state of affairs is desirable provides us with a reason to intend to act so as to bring that state of affairs about.

[Robert Johnson]

Sven,
Thanks for that helpful example, and I'm inclined toward your alternative understanding (and with Heath's point). Advice derives it force from the reasons it's advising you of. Likewise, it's not the higher order property of having other properties that are reasons that is providing the reason to go home. It's the lower order properties that are providing that.

[Robert Johnson]

Jamie & Doug,
Perhaps Doug is thinking this: If for something to be good is for there to be a reason to want it, and there is always a reason to do what is good, then the reason you have to do something will be that you have a reason to want it.

Is that it?

[Doug Portmore]

Jamie and Robert,

Pretty much.

I would say: the reason you have to do X is: the fact that it's true both that you have a reason to want Jones to be happy and that your doing X would make Jones happy.

[Mark van Roojen]

I think that I accept BPV+ without BPV-.

I think that reasons are supposed to be the sorts of things which can both explain and justify/rationalize, and that when you justifiably act on a reason the reason you act on is the one that justifies/rationalizes you in so acting.

I might know that an action has some lower order property which if I knew about it would give me a reason to do it. But I might not know which one that is. The sorts of knowing that one has a reason by testimony mentioned above by Sven seem characterisable in that way.

Suppose I now do the action. I have justifiably acted on a reason, but it doesn't seem that I have exactly acted on the reason my friend knows about but that I do not. Nor do I want just to say that the reason was that my friend told me, since it was also that my friend told me truthfully that goes into my being so justified. So it looks like the only reason in a position to both explain what I did and rationalize it is the one I then had some access to and that was the higher order reason that I knew there was some lower order property the action had which, if I knew it, would give me a reason to do that action.

So I think I'm with Doug on this.

[Mark Schroeder]

Well, since Jamie advertised that I had something to say about this, I'll leave my two cents.

First off, I don't think the 'higher-order property' talk is very illuminating. As I understand the positive buck-passing thesis, it is the view that facts about what is good are a certain sort of quantificational fact about reasons. That means that for the buck-passer, the fact that something is good is like the fact that there is a reason to do something.

Now lots of philosophers find it obvious that the fact that there is a reason to do something cannot itself be a reason to do it. (The 'provides' talk I find totally unhelpfully ambiguous in sometimes pernicious ways.) If that is so, then if the positive buck-passing thesis is true, then so is the negative one. This appears to be taken for granted in a lot of literature, but I think, as predicted, that it is false.

After all, why think that existential facts about reasons are not themselves reasons? Perhaps the thought is that they can't be _extra_ or _additional_ reasons. They can't _add up_ with other reasons, to make a weightier reason overall. That's true. So if we assumed that if R and S are both reasons to do A, then together they must be a weightier reason to do A than either is separately, then that would give us an argument that the fact that there is a reason to do A is not itself a reason to do it. But this 'additivity' theory about how the weights of reason add up is very rejectable, and I think it is false.

Other than that argument, though, or one resembling it, I've never come across any reason to think that existential facts about reasons are not themselves reasons.

So, for the other side: why think that they are? This follows from some very general theoretical motivations. When someone acts for a reason, and her action is for a _good_ reason, then we expect there to be some close connection between the reason for which she acted - her motivating reason - and the good reason for her to act - her normative reason - such that in acting for that motivating reason, her action counts as grounded in that normative reason.

For example, suppose that Jon, noticing that the child is drowning, steps into the pond to save her. He acts for a good reason. The normative reason for him to act is that the child is drowning, and his motivating reason consists in or involves in some way his belief that the child is drowning. That is why, in acting for that motivating reason, his action is grounded in that normative reason.

Now go back to the case of merely existential facts about reasons. Suppose that Nate's friend tells him that there is a reason for him to go home, but doesn't tell him what that reason is. And suppose that this is true, and that Nate believes his friend, and so he goes home. If Nate is acting for a good reason - if his action is well-grounded - then the preceding principle motivates a connection between his motivating reason and his normative reason. If his motivating reason stems from his belief that there is a reason for him to go home, then the principle requires that that must be a normative reason for him to go home.

Now, of course you can react in different ways. You can say that his motivating reason is that his friend told him that there is a reason for him to go home. But this looks ad hoc, unless you think that if his friend had told him that there is a fire in the building, his motivating reason for going outside would have consisted in his belief that his friend told him that, rather than his belief that there is a fire in the building.

You can also resist my principle. You can insist that Nate's action is able to be well-grounded in the normative reason that his friend told him about (say, that there is a surprise party waiting for Nate at home), even though he does not know what that is. This starts to look like a fairly bizarre view, to me.

The primary considerations that seem to be in play have to do with 1) how we measure the weight of reasons, and 2) the relationship between motivating reasons and normative reasons in cases in which an action is well-grounded. Since I think that adding up the 'weights' of reasons is complicated anyway, I don't see the force of 1). And at the same time I think 2) is under-appreciated.

So that is the answer that Doug wants, I think. At the same time, nothing I've said bears on what I think he cares most about: namely, whether the only reasons for _action_ can either be, or be explained by, facts about what is good, on the assumption that facts about what is good are quantificational facts about reasons to have certain _attitudes_. I think that view is coherent, but I'm still skeptical about how well-motivated it can be.

[Jussi Suikkanen]

Jamie,

Sorry, I misphrased my claim. I didn't suggest that the negative thesis follows from the positive one. I meant to say that no-one has denied the implication because no-one has even suggested that it holds.

They way I see the dialectic is that many people took Scanlon to argue from the truth of the negative thesis to the positive one. Many people pointed out to him that the positive thesis does not follow from the truth of the negative thesis and therefore the argument is not a good one. I agree that the implication doesn't follow to other direction either. My point just was that if the positive thesis is true, then it is more difficult to argue for the negative one.

Doug,

Here's your two theses:

~(BPV-) Being good is a property that itself provides us with reasons to act so as to promote it.

BPV(D)+ Something’s being good is just its possessing the purely formal, higher-order property of having some lower-order properties that provide us with reasons to desire it.

By substitution we get:
Possessing the purely formal, higher-order property of having some lower-order properties that provide us with reasons to desire it itself provides us with reasons to so as to act as to promote it.

I took it to be the case that this view is more plausible if reasons for actions and reasons for desires come apart in the way that reasons for actions are reasons to satisfy desires. If same things were reasons for desires and actions, then the thesis is not as plausible.

Scanlon does not phrase the negative and positive thesis as thesis about reasons for different things - desires and actions. In that way you have changed his view quite radically. He would say that when we talk about reasons for actions we just are talking about our reasons for various attitudes like intentions. If you only talked about reasons for either one then your combination would be odd. You would have, for instance, the claim that good is the property of having properties that provide reasons for actions and having that property provides more reasons for the same actions. That's not quite double counting but it does lead to a regress. So taking apart reasons for desires and reasons for actions is an essential requirement for holding your view.

Then you write that:
"The thought is that the fact that I have a reason to desire an end, E, provides me with a reason to perform an act, A, that is a means to bringing it about that E obtains. On BPV(D)+, to say that there is reason to desire E is just to say that E is good, and it's this property, the teleologist supposes, that provides us with reasons to act in ways that will achieve this end."

Notice that this does not exactly fit the two thesis you give. BP+ account does not say that for there *to be a reason* to desire E is to say that E is good. That is reductive claim about the term 'good' not a buck-passing account of value. Having a reason to desire E is not the same claim as the claim that E has properties that provide reasons to desire it. That there is a reason to desire E may well be a fact that gives reasons to act so as to promote E. However, for your negative claim you need that the fact is something quite different. You need that the fact [that E has properties that provide reasons to desire] provides reasons to act in the appropriate way.

I know Stratton-Lake is working on the view where the latter derivative, instrumental reasons can only be empty of normativity and guide the real normativity of the first-order, basic reasons around.

Here's an example Dancy uses to show why that's something quite odd anyway. Say you have a toothache. That seems like a good reason to desire to the dentist. It also seems like a good reason to go the dentist. In order for us to need more reasons, we should think that these reasons are not sufficient as such for actions - the toothache is not enough reason to go the dentist. It only is reason to desire to go the dentist. That is for the going to the dentist to be good. And, the fact that going to the dentist is good, i.e., there is a reason to desire it, then gives you the reason to go there. I just cannot see the need for that reason over the fact that the tootache suffices as a reason to go there.

[Jussi Suikkanen]

Mark,

great comment.

Doug,

I also wonder about the quote from Darwall you gave:
"Reasons for me to make something my end are, owing to the hypothetical imperative, equally reasons for me to take the necessary means to it” (Darwall 1983, 16)."

Isn't this the opposite view you want to hold? This is the idea that the first-order considerations are both the reasons for desires for the ends and for the actions that are means. Your view would hold that the fact that you have reasons for the desire gives you reasons for the actions. Here it is different considerations that count as reasons for desires and actions.

[Campbell]

I hope you don't mind if I ask a slightly tangential question.

Here's your statement of the positive component of the buck-passing account:

(BPV+) Something’s being good is just its possessing the purely formal, higher-order property of having some lower-order properties that provide us with, say, reasons to act so as to promote it.

I find this rather puzzling. On the standard definition of higher-order and lower-order properties, it just doesn't make any sense.

The standard definition is roughly as follows. First-order properties are properties of individuals. Higher-order properties are properties of properties: second-order properties are properties of first-order properties; third-order properties are properties of second-order properties; and so on. So, for example, greenness is a first-order property, instantiated by individuals like the coffee cup on my desk, whereas being a colour is a second-order property, instantiated by first-order properties like greenness. (Example adapted from one in the SEP entry on properties.)

It follows that higher-order and lower-order properties cannot be instantiated by the same things. For example, it's impossible for one thing to instantiate both a first-order property and a second-order property. However, according to BPV+ the higher-order property, being good, is instantiated by the same thing as the lower-order, reason-providing properties. What's good is not the reason-providing properties themselves, but the thing that has those properties. To be good is to have a reason-providing property; it's not to be a reason-providing property. So, given the standard definition of higher-order properties, BPV+ is just incoherent.

What's going on here? Is Scanlon operating with some non-standard notion of "higher-order"? Or did he simply screw the pooch?

[Mark Schroeder]

Hi, Campbell. That's among the reasons I think the 'higher-order' talk is incredibly unhelpful. The view is better characterized as the view that facts about what is good are quantificational facts about reasons.

[Jussi Suikkanen]

I'm not sure I see the problem. Here's how a difference between certain versions of functionalism is defined in Stanford:

"A Functional State Identity Theory (FSIT) would identify pain (or, more naturally, the property of having a pain or being in pain) with the second-order relational property. Other theorists, however, take a functional theory merely to provide definite descriptions of whichever first-order physical (or other) properties satisfy the functional characterizations and for those properties themselves to be the pains, beliefs, and desires."

On one view then, the property of being in pain is a second-order property that is instantiated by a first order property of being in a certain physical state. Does that make being in pain a property of the first-order property of being in the physical state? No. Being in pain, even on this view, is my property just as is being in the physical state. That doesn't seem to make functionalism of this sort incoherent. That's just a common way of talking.

[Robert Johnson]

Mark vR,
What if you come to know the first order reasons. I want to say then "I knew it! I just knew there was a reason for me to do that". And that fact in turn makes it plausible that, after all, it was that reason, not the second-order property, that was the reason. Or that's now my intuition.

Mark S,
The reason not to think that existential facts about reasons are not themselves reasons is, How do you stop a regress of reasons?

[Robert Johnson]

I meant, "The reason to think..."

[Editor: A missing HTML tag was inserted 9:24 AM, 1/25.]

[Mark van Roojen]

Robert,

I agree that I knew there was a first order reason to do it. But that just falls out of the story since that is precisely what I know on the basis of my friend's having told me I have a reason.

But what rationalizes my action when I act on the reason I have before I have been told what that reason is? It doesn't seem like it is that reason, since I don't know what that reason is. It seems like what rationalizes an action has to be something that can make it make sense to do from my perspective as I act. (If I act on one reason but am only made rational by some other reason then I think what is really going on is that the other reason makes other possible tokens of that general act type rational - those motivated by that reason as Mark S nicely puts it - but not this particular act token because it was not in fact motivated by that reason.)

So it has to be something about my reason for acting, the reason that I act on, or as Mark S says the reason which motivates me to act when I act that rationalizes it. And the thing that can play that role is my belief that there is some other feature the action has which would make it make sense to do if I knew about it (add "absent defeater" clauses or what have you to make this come out right) I have access to it. If you wonder why I did the action, only that will serve the role of making sense of my action and explain why I did it.

It may well be part of the story that there in fact has to be a first order reason for my second level or quantificational reason to be well grounded or count as knowledge or to count as a completely adequate justification or rationalization. But that is just to say we can have an reasonably externalist account of justification, not that we should have one that is so externalist that the only things that count as our reasons for acting when we justifiably act are things that we don't know.

On regresses: they don't go very far. Whenever you know there is some reason to do an action you know that the action has some property which, if you knew it, would make sense of your going in for that action. So all well-grounded reasons will either be of a sort that allows you to cite the particular first level fact or to quantify over these reasons to say that you know there is some reason of that general sort. You don't have to quantify over reasons of this sort again to ensure that your reason is included, because knowing that there was some quantificational reason (to coin a phrase) for doing the action would require knowing that you had some reason which itself required there to be a first level reason so you can just stop at the second level because the claims about higher levels all entail that there is a feature at the first level which would be a reason to act if I knew about it.

[Mark van Roojen]

This message is just to add a close italics code, since an unclosed italics code from the message previous one ne seems to have italicized everything that follows including my whole message. I can't undo that, but I can end it.

[Mark Schroeder]

I agree with Mark vR in response to Robert: first, the 'regress' doesn't go very far. Compare: the fact that the building is on fire is a reason for Jon to step outside. So, there is a reason for Jon to step outside. Is that fact itself a reason for Jon to step outside? Suppose that it is. Since it is, there is a reason for Jon to step outside. Is that fact itself a reason for Jon to step outside? This isn't a new question - it's the one we already settled.

In any case, even if it turned out that if there are infinitely many reasons to do something if there are any at all, there wouldn't be anything 'regressive' about this, and certainly not anything vicious. Compare: if anything, P, is true, then infinitely many things are: that P, that P is true, that it is true that P is true... there's no regress there. This only gets puzzling if you think that all of these extra truths somehow change something important. Similarly, as I suggested above, the existential fact's being a reason only gets puzzling if you think that it somehow makes a difference for what you should do, by 'adding up' with your other reasons. But I insisted that that doesn't happen, and no one has ever articulated to me a defensible and well-motivated principle that would lead us to expect that it would. So even if infinitely many reasons are generated, there doesn't appear to be anything problematic about that.

[Robert Johnson]

Mark vR,
You said that the thing that can play that role is my belief that there is some other feature the action has which would make it make sense to do if I knew about it.
But supposing that is indeed a reason, isn't it different from the supposed reason provided by BV-? The reason, in this case, is that I believe something, that there is a reason of some sort or other, to do something. And that belief is made true by the existence of that reason, whatever it is.

Mark vR and Mark S,
It does go far. If the fact that there is a reason to leave the room is also a reason, then the fact that it's a fact that there is a reason to leave the room is a reason. And so on. That we tire of thinking about these doesn't matter.

I don't know what to think about there being infinitely many reasons to do anything there is reason to do.

[Jamie Dreier]

Robert, is the fact that it's a fact that there is a reason to leave the room really different from the fact that there is a reason to leave the room? If not, then the blossoming of reasons ends there.

In any case, I don't see any problem with there being infinitely many reasons to do one thing, as long as we don't have to assess them indepedently or work out piecemeal how they combine. For instance, suppose your dentist has a new procedure, which is exactly like the old way of filling a cavity except that its duration is ten seconds less. Then the fact that it is ten seconds shorter is a reason to prefer it; so is the fact that it is at least nine seconds shorter; so is the reason that it is at least 9.713 seconds shorter; and so on.

[This entry is not italicized in Preview.]

[Campbell Brown]

Doug,

Small point. Above you give the impression that your goal in all this is to show that the buck-passing account is consistent with teleology. But Scanlon has already explicitly said that these are consistent: "One could accept [the buck-passing account] while still holding a purely teleological conception of value" (WWOTE, p. 98).

Jussi,

Apparently there's another usage of "higher-order property", which is common in the philosophy of mind. For example, here's Jaegwon Kim:

For something to have a second-order property M is for it to have some first-order property or other meeting a certain specification.

The difference between this and the definition I gave earlier might be put like this: according to Kim's definition, to have a second-order property is to have a first-order property that has a certain property, whereas according to mine, it is to be a first-order property that has a certain property.

I confess I'm a little sceptical of Kim's definition. For one thing, I don't see how it can be extended to give a general definition for all orders of properties. This is fairly easy to do for my definition. Say that individuals (particulars, objects) are 0-order properties. Then here's the general version of my definition: for any n>0, to have an n-order property is to be an (n-1)-order property that has a property. But if we try the same thing with Kim's definition, we get this: for any n>0, to have an n-order property is to have an (n-1)-order property that has a property. This doesn't work for the case of n=1, because you cannot have a 0-order property, which by definition is just an individual.

Let me ask a different question. (Sorry Doug, this is still a bit tangential.) What is the point of including the stuff about higher-order properties in the buck-passing account? At times Scanlon seems to think this allows him to dodge Moore's open-question argument. He seems to be saying: "Moore's argument causes trouble only for analyses according to which goodness is a natural property. But on my analysis, goodness can't be a natural property, because it's a higher-order property. So my analysis is immune to the open-question argument." Do you think that's Scanlon's reason for including the "higher-order property" stuff?

[Doug Portmore]

Campbell,

Yes, I'm aware of that quote. But I believe that this is one of those areas where Scanlon is being unclear. Scanlon sometimes refers to the "teleological conception of value," which holds that states of affairs are the only or primary bearers of intrinsic value and so the only proper response to value is to bring it about or promote it. At other times, though, Scanlon refers to the "teleological conception of reasons," which holds that reasons for action must appeal to the desirability or undesirability of the result at which the given action aims. So the quote doesn't necessarily show that Scanlon thinks that BPV is compatible with the teleological conception of reasons. Regardless, R. Jay Wallace has a paper that seems to be claiming that the BPV and the teleological conception of reasons are incompatible -- see "Reasons, Values, and Agent-Relativity".

Regarding the lower-order/higher-order properties stuff, I was just following Scanlon's formulation: see p. 97.

Jussi:

I would like to have time to respond to some of your comments, but I don't have time today. I may have time later on.

[Campbell Brown]

Doug, can you fix the italics?

[Doug Portmore]

Hasn't it already been fixed by MvR? Or what exactly do you want me to do?

[Jussi Suikkanen]

I'd like to think that BP is compatible with a version of teleology. I take it that for a teleologist there are two essential claims: states of affairs are bearers of value and the more states of affairs have value the more they are to be desired and promoted. Nothing in buck-passing as such denies this. A buck-passing teleologist can say that the value of states of affairs is their ability to give reasons and these reasons states of affairs have require desiring and promoting. This is compatible with the idea that the reasons are provided on the level of first-order considerations.

Many teleologists make additional claim that goodness is the reason-provider. I don't see why this would be essential for the view. What Scanlon seems to oppose more in the view is the substantial claim that all reasons are for desiring and promoting. He seems to think that there are values that should be responded to in other ways.

Campbell,

I think I agree that nothing in BP hangs on the use of the term 'higher-order' property. If you say merely that value is the property of having other properties that provide reasons I don't see that anything is lost. But, it does seem to be a good way of making the contrast to what the value properties are not. You are right that here we get to the open question argument in making the difference.

[Campbell Brown]

Doug, it's not really fixed. Despite Mark's efforts, the html code is still invalid, and the page displays incorrectly in some browsers (e.g. Safari). To fix it properly, you need to edit Robert Johnson's comment, adding an "</i>" where required.

[Doug Portmore]

Thanks, Campbell. There wasn't a problem on my browser so I didn't understand what needed to be done.

In any case, it appears to be fixed now. Let me know if I'm wrong about that.

[Mark Schroeder]

Robert - reasons, I'm taking it are facts. Or true propositions (I actually take them to be true propositions and think the 'fact' talk is harmless so long as we're clear). Some facts are existential. For example, if there is a reason for Jon to leave the building, then it is a fact that there is a reason for Jon to leave the building, and that is an existential fact.

I claimed that existential facts about reasons are also reasons - to do the same thing as the reason which makes them true. Of couse, they can't be the _only_ reasons to do that thing, because they have to be true to be reasons, but there has to be a reason to do that thing, for them to be true. So there has to be some other reason to do that thing. But I claimed that they are still reasons.

In contrast, I never said that facts about facts are reasons. Maybe the fact that it is a fact that there is a reason for Jon to leave the building is also a reason for him to leave the building, but I didn't say so, and saying so is not required in order to say that existential facts are reasons. So no, it does not follow that there are infinitely many reasons for Jon to do it, let alone that this is vicious in some way.

In fact, suppose that your inference were a good one - that if some fact is a reason, then the fact that it is a fact is also itself a reason to do the same thing. Then we would have your regress without my thesis that existential facts are themselves reasons. Just start with the reason for Jon to leave the building - the fact that it is on fire. By your principle, the fact that it is a fact that the building on fire it also a reason for him to leave the building, as is the fact that it is a fact that it is a fact that the building is on fire, and so on. So if there is anything bad about that conclusion, we can get it without my assumption, anyway.

[Mark van Roojen]

Robert,
You asked:
But supposing that is indeed a reason, isn't it different from the supposed reason provided by BV-?

[Is this supposed to be 'BPV+'?]

The reason, in this case, is that I believe something, that there is a reason of some sort or other, to do something. And that belief is made true by the existence of that reason, whatever it is.

My formulation of an access requirement in the comment was sloppy. I think that a buck passer can say that the reason is the fact, while hanging on to the claim that facts can rationalize only when they are known or accessible.

I agree that the quantificational claim is made true by what's going on with the things quantified over. But that doesn't show that I know or have access to the specific reason which makes it true that the quantificational claim is true. I just know that there is such a reason, and it is this knowledge which explains and rationalizes my action.

[Francesco]

Doug,
quoting from a previous post of yours, the teleologist says: "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." In this, there is no explicit connection with the good. Indeed, the teleologist, perhaps surprisingly but not obviously incoherently, can express his view without (thin) evaluative concepts. Anyway. If it's true that "S has an object-given reason to intrinsically prefer Ox to Oø if and only if Ox is intrinsically better than Oø", then the principle reads "S has a reason to do x if and only if Ox is intrinsically better than Oø". I take it this the view we are discussing. The view is compatible with different views about what makes the biconditional hold good: 1) the fact that Ox is better is itself the reason to do x; 2) the fact that S has reason to prefer Ox is the reason to do x; 3) the fact, whatever that is, that makes Ox better (or that gives a reason to prefer Ox) is the reason to do x. The truth of teleologism is pretty much independent of which view is correct. BTW, Wallace has clearly in mind a non-buck-passing teleologist, but he's wrong if he thinks that you need to be teleologist to deny buck-passing.

[Francesco]

Mark S.,
I do agree that the connection motivating/normative reasons is underappreciated in the buck-passing debate. But that there is such a connection doesn't mean that everything that is an acceptable motivating reason is thereby an acceptable normative reason. If P is an acceptable motivating reason for F-ing then this means that the relation between P and F-ing tracks some (say) favouring relation. However, it doesn't mean that the favouring relation is between P and F-ing. It may be between what P implies (or what makes P true, etc.) and F-ing.

[Jonas]

Robert and Mark S,

I am puzzled by the regress discussion.

Suppose that it is a fact that the building is on fire. Call this fact F. Suppose that F is a reason for Jon to leave the building. By existential generalization we get, It is a fact that [F is a reason for Jon to leave the building]. Call this fact F*. Is F* itself a reason for Jon to leave the building? That is to say, does F* have the same property as F, namely the property of being a reason for Jon to leave the building? I am not sure what to say but I am inclined to answer no. In general, anyway, it seems that higher-order facts need not have the same properties as the lower-order facts that they are facts about. Suppose that G is an interesting fact. Is the fact that G is an interesting fact itself an interesting fact? It may be, but it may also not be.

Have I gone wrong? I find reason talk terribly puzzling.

[Jussi Suikkanen]

I think I can handle reasons but what I have really hard time with is the facts. I have no idea how they are individuated or what they actually even are. I can make sense of Strawson's minimalism about facts but if they are something REAL and FACTS, then I'm lost what they are and how to count them.

[Heath White]

I, like Jonas, am a little puzzled, but without a firm intuition here. Suppose the fact that there is a fire is a reason to leave the building; I don't know there's a fire. Someone tells me, however, that there is a reason to leave the building. Now, how many reasons to leave the building are there?

1. there is a fire.
2. there exists (at least) one reason to leave the building, namely #1.
3. there are (at least) two reasons to leave the building, namely #1 and #2.
4. ...

I just don't know what to say here; maybe there is something wrong with counting reasons. We might get some light on this from epistemological discussions of evidence, about which I know nothing. E.g., suppose there is smoke in the building, but I don't know this. However, someone tells me that there is evidence (=reason to believe?) there is a fire in the building. Now, how many pieces of evidence for fire in the building are there?

1. there is smoke in the building
2. there is evidence that there is fire in the building.

Are claims about evidence, evidence themselves? Does anyone know?

[Robert Johnson]

Mark & Mark, et. al.,
I think I too am getting confused by the fact-talk, but I'm fine with MarkS's view as stated. But let's go back to property-talk, which MarkS finds unhelpful. For Scanlon, it is properties that are reasons. But not the second-order property of having a property that is a reason. That's what goodness is.
So, I think there is still an point in the area here to make. Just to recap that bit of the discussion, that I had said that all that you need to add to BPV+ and BPV- to get an inference from the one to the other was:

That something has a purely formal, higher-order property of having lower-order properties that provide reasons to promote a thing does not itself provide us with an additional reason to do that thing.

That seemed right to me. But then Both Sven and Mark vR suggested this:

My knowing that an action has the higher-order property of having some lower order property which is a reason to do it is itself a reason to do it.

I agreed that my knowing that it has this second-order property is a reason, but I denied that the second-order property itself is a reason.

Now I'm inclined, in response to Mark vR's last comment, to think that the reason I doubt that the second-order property is a reason is that my knowing about it gives me an epistemic connection to the first-order reason -- the real and only reason.

In any case, if the property of having a property that is a reason is also a reason, then it just follows straightaway that the property of having the property of having a property that is a reason is itself also a reason, and so on.

[Jussi Suikkanen]

Robert,

I think that's right :-) I'm not sure though that Scanlon would go as far as to say that reasons are properties. To make things very confusing he actually says that reasons are propositions but that's another story. My take on him is that he'd more probably say that when some object has certain property then the object having that property is the complete reason to do something. Properties provide reasons in as much and when they are instantiated in some objects. As a result, an object is valuable by having properties which thus instantiated provide reasons.

[Mark Schroeder]

Francesco - I didn't say that any acceptable motivating reason is an acceptable normative reason. That's clearly false. One obvious reason why it is false is that acceptable motivating reasons can involve false beliefs. For example, Jon may leave the building because he believes that it is on fire, even though it is not. In such a case, he has an acceptable motivating reason, but his action is not well-grounded. The reason he does it for is not, intuitively, a normative reason for him to do it.

I was also trying to be neutral about the ontology of motivating reasons. According to Smith, for example, they are belief-desire pairs. According to others, they are beliefs. According to Dancy, they are states of affairs. I happen to think that they are propositions. But the point I was trying to make was supposed to be neutral between these conceptions.

It was simply supposed to be that whatever they are, you have to have a belief or be in some similar state in order to have one. Then, I invited us to examine cases in which actions are intuitively well-grounded. In those cases, I maintained - or at least, in all of the paradigm ones - there has to be a connection between the belief that is involved in your having the motivating reason, and the normative reason in which it is well-grounded. On Dancy's view or on mine, the motivating reason and the normative reason have to be identical, but there is a wide family of views that will do the same work. If you think that motivating reasons are beliefs, then the claim is that the action is well-grounded only if the content of the motivating reason is a normative reason. Etc.

Then I described Jon's case, in which he acts on testimonial knowledge of the existential fact that there is a reason for him to leave the building, and I claimed that this is a well-grounded action. Given that assumption, it follows from any of the views in the class I considered that the existential fact is a reason.

Jonas - Just to reiterate - I never claimed that existential facts always have whatever properties their truth-makers have. That's obviously false - for example, there is a banana in my kitchen. It is yellow. But the fact that there is a banana in my kitchen is not yellow. QED. All I claimed was that it is not obvious that existential facts about reasons are not themselves reasons, and I gave an argument in favor of thinking that, in fact, they are. I just repeated it again in the first part of this comment.

Heath - I agree. The numbers don't matter unless they make a difference in how the reasons 'add up' to determine what we ought to do. That's been one of my key points. So how do we judge the issue? Just based on intuitions? I proposed a test. If something can be the reason for which you act in a case in which your action is well-grounded, then it should get classed as a reason. Then our account of how reasons 'add up' to determine what we ought to do should be constrained so as to not make those further considerations which are, additionally, reasons, turn out to be 'additional reasons', in the sense that they contribute further to fixing what we ought to do.

Your point about the epistemic reasons is a good one. One way to think about literature that bears on it, is to compare it to the Principle of Reflection. The Principle of Reflection basically says that your degrees of belief now should reflect your best judgment about the quality of the evidence that you expect to encounter in the future, even if you haven't encountered that evidence, yet. Of course, Reflection is a much stronger principle, but it's the sort of thing in the neighborhood of the idea.

[Jussi Suikkanen]

Mark,

can you accept both of these claims at the same time:

"The reason he does it for is not, intuitively, a normative reason for him to do it."

and

"On Dancy's view or on mine, the motivating reason and the normative reason have to be identical'.

I know Dancy would not accept the first claim just because it doesn't fit the second.

If there are motivating reasons that are not normative reasons, I could accept that in Jon's case he has a motivating reason that is the existential fact. This fact and his knowledge about it helps us to understand Jon's actions. I could also accept that his action is well-grounded because there is a good normative reason for it (even he doesn't have direct access to that reason).

I still don't know whether I would want to say that that existential reason is a good normative reason that comes with its own favouring force.

[Mark Schroeder]

Jussi -

It's not enough for your action to be well-grounded, that you do it for a reason, and there be a good reason to do it. Suppose that Jon leaves the building because he believes that it is on fire, but it is not. Yet, unbeknownst to him, there really is a reason for him to leave the building - say, that there is about to be an earthquake. In that case, Jon's leaving the building is not well-grounded.

So there has to be match between the reason for which you act and the normative reason for you to act, in order for your acting to be grounded in that normative reason. My claim is that the most natural principles about how match works will lead us to conclude that existential facts about reasons are themselves normative reasons for action.

If you feel the need to resist this totally natural view, then explain why. So far in this thread, I haven't really seen any argument that the existential fact that there is a reason for Jon to leave the building is not itself a reason for him to leave the building. You seem to want to resist my principle in order to be able to say that. But why?

And what principle about well-groundedness will you defend, instead? Given the earthquake case, there has to be some connection between the motivating reason and normative reason in order to get well-groundedness. So what is it? Is it that the normative reason must be part of what makes true the consideration that you bear in mind when you act?

If that's the proposal, then we can test it on other cases. Suppose that Jon leaves the building because he thinks that it smells like roses. And suppose that it does smell like roses to Jon, because there is a fire, which has caught the florist downstairs, among other things. Jon's belief is true, and it is true because of the fire. But his leaving the building because it smells like roses is not well-grounded in the normative reason for him to leave the building - that there is a fire.

In general, if you want to say that Jon's action can be well-grounded in the fact that there is a fire, when he acts in light merely of the information that there is a reason for him to leave, and that is not itself a reason for him to leave, then you have to be able to defend a principle about how well-groundedness works that gets you that result but doesn't lead to the wrong results in other cases. I don't know how to do this, and I don't see what the motivation is for trying to do it. The picture looks totally simple, to me. You'd need some kind of motivated reason to deny that existential facts about reasons are themselves reasons, in order to feel the need to resist these kinds of principles and come up with your own. But I still haven't seen where that motivation is supposed to come from.

[Mark Schroeder]

Oh. And again in response to Jussi's question about consistency. I only think that your motivating reason is identical to your normative reason in the case in which your action is well-grounded. Non-well-grounded action is action for motivating reasons that are not normative reasons. Simplifying somewhat, to be a motivating reason, a proposition has to be believed. To be a normative reason, it has to be true.

[Campbell]

Jussi

Another question. Scanlon says explicitly that, on his buck-passing account, goodness is a non-natural property. Is that sheer stipulation? Or is it supposed to follow from some other feature of the account. As I said before, it's tempting to read Scanlon as saying that goodness is non-natural because it's higher-order. But you say it's inessential to the account that goodness is higher-order. So, is it also inessential that goodness is non-natural? Or is its being non-natural not supposed to follow from its being higher-order? (Okay. That was more than another question.)