Jason Bridges

University of Chicago


Phil 31410, A03—lecture 15 notes

Scanlon on reasons and desires

(What We Owe to Each Other, chapter 1)


Recall the view of reasons explanations we ascribed to Scanlon (cf. pp. 33-34):


A reasons explanation begins: S φ-ed because ______.

What fills in the gap?

Smith’s Humean view: an ascription of a desire to ψ and a belief that in φ-ing S would ψ.

Scanlon’s non-Humean view: a specification of considerations that S took to count in favor of φ-ing, perhaps coupled with an account of why S took these considerations to count in favor of φ-ing, of how S took these considerations to weigh against other competing considerations, etc.


This is a somewhat roundabout formulation of the disagreement.  I put it this way to avoid begging a question we’ll discuss next time.


Thus, to return to our example, the explanation for my washing the car may be that I took my having so promised my wife to count decisively in favor of my doing so.


In the case of a rational action, the subject does what she takes herself to have most reason to do.

But an action needn’t be rational to be susceptible, on Scanlon’s account, to a reasons explanation.  A person can φ because she believes there to be something that counts in favor of her φ-ing, without her believing that φ-ing is what she has most reason to do—indeed, without even believing that her reason for φ-ing is a good one.

What cannot be the case, on Scanlon’s view, is that a person φ’s without her taking there to be anything at all that counts in favor of her doing so.

This disarms a familiar Humean objection to “cognitivist” accounts of reasons for action, to the effect that their view of human nature is too optimistic (e.g., Velleman’s complaint in “The Guise of the Good”).


Scanlon acknowledges that there are worries one might have about his non-Humean view of reasons.  Before we talk about his response to those, I want to discuss his objections to weak Humeanism.  (Remember that weak Humeanism is the view that every agential reason, every reason for which a person acts, includes as a component a psychological state or attitude that is not a belief—i.e., a “desire”.)


IV. Scanlon on why desires don’t motivate

To place Scanlon’s objections we need first a preliminary point.


Weak Humeanism, as we have understood it, is a view about motivation.  It is a view about agential reasons—the reasons for which people act.  Weak Humeans hold that agential reasons must contain desires.


Typically, weak Humeans ascribe a further role to desires: desires are what justify actions.  One way to put that idea is thusly: to show why an agent took an action to be justified, it is sufficient (and perhaps necessary) to ascribe to the agent a desire for something, something that she believes will result from the action in question.


Scanlon will argue both against the claim that desires motivate and against the claim that desires justify.

Why should we care about the second of these claims, given our topic?  Well, recall Davidson’s normativity thesis about reasons explanations.  He holds that one requirement a reasons explanation of an action must meet is that it justify the action in a certain minimal respect.  That minimal respect is that it show how the action seemed justified, worth doing, to the agent.


We have seen that this requirement has some appeal.  And if we accept it, then one way of arguing against a particular view of the composition of agential reasons is to show that the posited elements cannot in fact justify actions in Davidson’s sense.


So Scanlon has in effect two ways of arguing against weak Humeanism.  One depends upon the normativity thesis, and one does not.  We will discuss both, beginning with the latter.


To begin, let me repeat what I said when we were discussing Smith about talk of motivation.  It seems to me that the question of what motivates an action should be understood as identical to the question of what are the agential reasons behind an act, the reasons for which one acts.  To give the motivation of an action is not to give any old cause of the doing.  It is to give the reasons upon which or for which the person acted.  To see a doing as motivated is thus to see it as an action.

That is why I equated (barring some confusion in Smith’s discussion) motivating reasons with what we are calling agential reasons.


It has been taken for granted by most people in the philosophy of action that desires at the very least can motivate.  The question, if there is one at all, is thought to be over whether every action must have a desire as part of its motivation.  Thus we can see McDowell as arguing that in some special cases, a belief can by itself move a person to action.

Scanlon’s view is that not only is it not the case that every agential reason must include a desire, but moreover that they almost never do.  Desires are simply not the kind of thing, except in perhaps very special and uninteresting cases, that can be something for or upon which people act.

Moreover, Scanlon does not attempt to defend this view by attempting to expose, à la McDowell, some deep metaphysical assumption that stands in the way of its appreciation.  He just asks us to think about what desires are, and to thus be brought to see that they lack this power.


We might lay our some of the central considerations he appeals to as follows:


Scanlon against weak Humeanism: [reiterate the weak strong contrast]

1.  If by “desire” we mean “pro-attitude”, then agential reasons certainly involve desires, but this lends no support to weak Humeanism.


A “pro-attitude”, recall, is any propositional attitude that involves seeing something in a favorable light or being disposed toward it or what have you.  Sometimes philosophers, for example in decision theory, use the term “desire” in a thin sense as just meaning pro-attitude.

That desires so understood are inevitably involved in agential reasons, though, lends no support to weak Humeanism, the view that agential reasons always have non-belief components.  For of a course a non-Humean will hold that beliefs can be pro-attitudes, as for example a belief that one has a compelling reason to do something.

(Compare our criticism of Smith’s argument for weak Humeanism.)


Clearly, we need a richer notion of desire, one that excludes beliefs.


2. If by “desire” we mean what we mean in ordinary discourse, namely, desire in the attention-directed sense, then although the presence of a desire may sometimes (although not always) be relevant to the explanation of an action, it is not part of the agent’s reasons for acting.

(Again, Scanlon would say “not part of the agent’s motivation for so acting”.)


A person S has a desire in the attention-directed sense that p = S’s attention is insistently drawn to considerations that appear to count in favor of bringing it about that p


Here’s an example.  Suppose my attention is constantly drawn to various considerations that strike me as counting in favor of eating donuts.  Thus I keep thinking about how good they would taste, how satisfying it would be to eat them, how they would satiate my hunger, how it’s such a cool hipster thing to eat such unsubtle, trashy food, etc.  Maybe some of these considerations are genuinely reasons that count in favoring of eating donuts, maybe they aren’t.  Regardless they strike me as reasons in favor of that, and my attention is continually drawn to them.  Then I have a desire in the attention-directed sense that I eat donuts.


To have a desire in this sense involves having views, perhaps fleeting, about what counts in favor of doing something or something’s coming about.  And then if one acts, one’s reasons for acting are those considerations to which one’s attention is being drawn, not the fact that one’s attention is so drawn.


Here’s an analogy.  Suppose I have a disposition to have my attention drawn to things moving in the periphery of my vision.  On some occasion something so moves and I turn to it and it’s a bat.  This causes me to startle in fright.

Now the source of my startling is not my disposition, but the presence of the bat.  The disposition can be cited in a richer explanation of why I startled: it’s an important enabling condition for the bat’s presence to impinge on me in that way.  But what impinged on me, what startled me, is not my disposition but the bat that I see (or if one wants to say it’s my seeing the bat, that makes no difference to the force of the analogy).

Similarly if I’m so disposed at the moment to have my attention drawn to considerations that present themselves as speaking in favor of my eating donuts, then if I do eat donuts that disposition might be cited in the explanation of what I do.  It’s an important enabling condition.  But my reasons are those considerations, or, if one likes, my taking those considerations to count in favor of what I do.  Either way, it ain’t the disposition of attention, but what I attend to that counts as my reason.

That is why Scanlon says desires so understood are not part of one’s reasons.  One’s reasons are the considerations to which one attends.  Or, if one wants to remain within the confines of psychologism, one’s reasons are one’s taking, believing, those considerations to count in favor of whatever is in question.


We shall return to one issue raised by this, that there can be more to be said in the explanation of an action than what reasons one acted for.  This may seem to present a problem for Scanlon’s sort of account; we’ll discuss why and whether that’s so in a bit.


3.  If by “desire” one means a bare, inexplicable urge to do something, not explained by or associated with any perception of considerations that count in favor of so doing, then one doesn’t know English.  And moreover such “desires” are vanishingly rare.


One might try to meet the point about attention-directed desires by arguing that there’s associated with the perceptions of reasons counting in favor of eating donuts, an “urge” to eat donuts.  And this is the real driving force behind one’s doing what one goes on to do.


For this to support weak-Humeanism, it must be possible to have some independent grasp on what these “urges” are, which would require our being able to conceive in a robust way of the presence of an “urge” in the absence of any attention-directed desire.

But can we do this?  Certainly there are very few cases in fact in which anything like an “urge” is present in the absence of any corresponding perception of reasons for that which one feels the urge.  Such a case would presumably have to be like the case Scanlon cites of the person inexplicably driven to turn on radios.  He sees nothing positive about doing so: he is not interested in hearing anything on the radio; he has no aversion to silence that would be alleviated, etc.


Even if we think that we can make sense of such bare urges, it’s clear that we wouldn’t ordinarily conceive them as desires.  Indeed, the most natural thing to say about this case is that the poor sap has no desire to turn on radios, but he’s none the less pathologically compelled to do so anyway.  And the reason we would say that he has no desire seems precisely that he sees absolutely nothing good or worthwhile or appealing about doing so.


The upshot of all this seems to be that to the extent that something we might want to call “desire” is part of one’s reason for doing something, it is simply the taking of something to count in favor of one’s doing that.  And that is a belief or at least belief-like state.  It is a matter of accepting something’s being so.


IV. Scanlon on why desires don’t justify

Now, consider again the normativity thesis about reasons explanations: that in giving a person’s reasons for acting one shows why the action might have seem justified to her.

Suppose we accept this view, as Davidson urges we must.

It follows the reasons for which a person acts, the motivating reasons behind the action, must either be identical to, or at least give rise to, potential normative reasons in favor of the action.  The very thing that we cite in the explanation of an action must also, at least to the agent, justify the action.  And presumably, if the agent is not misled, then what we cite will actually justify the action.


It’s clear on Scanlon’s view how that connection is to be achieved: what explains an action is precisely the agent’s view about what counts in favor of, justifies, the action.


What about on a Smith-style Humean view?  Is there anything to be said here?  Remember, on anyone’s view a normative reason, a justifying reason, will be a consideration, a fact.  What sort of fact could that be on the Humean view?

Well, Humeans tend to think that there definitely is something to be said here.  If they accept the normativity thesis, as many do, then they think they can reconcile this with their Humean view of agential reasons as follows:


The Humean view of normative reasons for action: The considerations that justify an agent’s action are, at bottom, facts about what would satisfy the agent’s desires.


An action satisfies a desire if it brings it about that the desired state of affairs obtains.  Thus if I desire that I eat some donuts, an action on my part satisfies that desire if it brings it about that I eat some donuts.

It is the satisfaction of desire that is the ultimate source of our justification for everything we do.  If, for example, actions are justified on moral grounds, it is ultimately because we have moral desires, desires, for example, that everyone be treated equably or what have you.


This may seem a disturbing, deflationary view.  And there has been endless argument, especially with respect to morality, about whether this can be so.

Scanlon’s tack, at least here, is again not to engage in any deep disagreement about the nature of reality that one might take to lie behind this issue.

Rather, he suggests that if we look closely at actual cases, we’ll see that it is never the satisfaction of a desire that we take to be the justification for anything we do.  To the extent that as philosophers we purport to deny this, our own ground-floor view of the reasons we have is opaque to us.


One way he tries to make this case is by reverting to the various notions of desire he distinguished in discussing the motivation question.  And so he argues, for example, that what justifies my eating of donuts, if anything does, is the reasons to which my attention is drawn, not the fact that my attention is drawn to them.

Once again, consider our analogy.  What justifies my fear, if anything does, is that I see a bat; it is not the disposition that puts me in a position to see that.


But Scanlon does think there is a further wrinkle that gets introduced when one focuses on the justification issue.  He thinks there’s a certain mistake here that we’re very liable to make.


The mistake is this: to infer the Humean view of normative reasons for action from the view that what justifies our actions are often or always facts about what we would enjoy or would give us pleasure, or on the other hand what would enable us to avoid pain.  Some philosophers think that all justification of action ultimately lies in facts about what would bring us pleasure or prevent pain, but we don’t have to go that far to see how the inference might go.

Consider the fact that people vary in what they enjoy or what brings them pleasure.  I hate opera; you love it.  It seems eminently plausible then to say that you have a reason for buying opera tickets that I lack.  And so it’s not a fact about the world but about you that gives you the reason.  And it seems open to construe this as the fact that your going to the opera would satisfy a desire you have, whereas my going to the opera would satisfy no desire I have, and in fact bring about a state of affairs I actively desire to avoid.


But Scanlon thinks this last step is simply a confusion.  It may well be that the fact that you would enjoy the opera gives you a reason to go to the opera.  But it doesn’t follow that the reason is that your so going would satisfy a desire you have.


We can see his point here without adverting to his view that desires involve perceptions of what one has reason to do.

The point is only that your enjoying the opera and your desiring to go to the opera are two different things.  And it’s the former state of affairs that provides the reason.

Suppose you’re deciding now whether to buy tickets to go to the opera tonight.  You reflect that you would enjoy it, and you have no other pressing engagement, so you decide to go.

Now, it’s true that in going you will satisfy your desire to go.  Does it follow from the fact that your enjoying opera counts in favor of your going that its satisfying a desire of yours counts in favor of your going?

No.  That you will enjoy the opera is a fact about something that will happen in the future.  That you desire to go is a fact about you now.  Your future enjoyment must be a different state of you than your present desire, for they obtain at different times.  So it’s just a confusion to equate them.


The fact that different people enjoy different things, have different talents and capacities and interests, certainly provides for differences in what reasons we have.

This shows only that:


Many reasons (in the standard normative sense) have subjective conditions.  That is, it is often the case that a given consideration’s counting in favor of a person’s doing something depends on facts about that person: her enjoyments, her familial and other personal ties, her skills and interests, etc.


Thus I have reasons for taking care of my son that you all lack.  And an opera lover has reasons for doing things that I decidedly lack.  And so on.


Scanlon thinks that the Humean view of normative reasons gets illicit support in virtue of being somehow thought to follow from this fact.  But it simply doesn’t.


(Scanlon notes a difference in this regard between desires and intentions: he agrees with Bratman that forming an intention to do something in and of itself gives one additional reason to do it.  So intentions give rise to normative reasons but desires don’t.

Don’t have time to get into this, but the basic point is that the fact that one has a certain intention can be a relevant subjective condition on the reasons one has.  Since we need to plan to get anywhere with our projects, and since our intentions reflect our plans, that certain intentions are already in place can matter a great deal to determining what one now has reason to do.)


Also: Scanlon says various interesting things about how talk of desires obscures various structural features of our reflection on what we have reason to do.  Desires are naturally understood as overriding each other only in virtue of one’s having a stronger weight than the other.  But the metaphor of weight, says Scanlon, doesn’t capture the diverse ways in which we might take reasons to conflict with each other.  And the diversity of these conflicts is a central feature of our practical reasoning.

Unfortunately we don’t have time to get into this interesting stuff.  I recommend reading that section closely.