Do our reasons depend on our desires?

NOTE: This post assumes an understanding of much of the terminology that I explain in my post, “On Reasons and What They Do?” In particular, this post uses the terms ‘reason,’ ‘objective,’ and ‘subjective’ in accordance with the explanations that appear in that article. Further, there are other important matters (such what it means for a reason to be pro tanto) that my discussion in this current posts assumes a familiarity with. 

I have been writing about reasons, what they are, and what they do. This is an important topic because, as I have argued, reasons play a central role in issues of morality and the meaning of life. The reason for talking about such issues in a philosophy of religion blog is that many religious apologists have argued that, if there is no god, there are no objective moral truths and that if there is no god, life is meaningless. Both of these assertions are false but understanding why they are false requires a good understanding of the nature of reasons and the connection reasons have to morality and meaning.

In a previous post, I argued that at least some reasons are objective. One consideration that is commonly relied on to argue that reasons must be subjective is that reasons are dependent on desires. David Hume famously said, “’Tis not contrary to reason to prefer the destruction of the whole world to the scratching of my finger”[1] Exactly what Hume intended to be saying here about the nature of reasons is a matter of some controversy. He has, however, been widely interpreted as claiming that reason is only a matter of selecting the best means of satisfying our ends. On such a view, reason can recommend no ends. We have our ends on the basis of our desires and passions, but these ends are not rationally evaluable. We have reasons only once we have chosen an end, and then the reasons that we have are to do the things that effectively satisfy that end. This view, regardless of whether it is Hume’s real view, is consistent with a position that I have previously called the Desire-Based Reasons Thesis (DBR):

DBR: For a person, P, to have a reason to engage in some behavior, Φ, P must have some desire (or desire-like state) d, such that Φ-ing tends to promote the satisfaction of d.

I strongly suspect that DBR is false and I want to provide some examples that serve to undermine it. First, though, I think we should acknowledge the intuitive appeal of DBR. DBR fits in very nicely with a certain conception of rationality that is often called “means-ends-rationality.” This conception can be well illustrated via the following example:

Pizza: Sue wants to have a pizza delivered to her house. Given that calling the local pizza place, Pizza Yurt, and ordering a pizza will efficiently promote the satisfaction of Sue’s desire, Sue has a reason to call Pizza Yurt to order a pizza. Further, she only has this reason given that she has this desire. A person who does not have the desire for pizza to be delivered to their home does not have a reason to call for pizza delivery.

This seems right, at least at first glance. Despite this appearance, I will later argue that the above account of why Sue has the reasons she does is completely wrong. Now, though, I want to acknowledge the way in which the account is intuitively appealing. It seems true that anyone who does not want a pizza does not have a reason to call for pizza delivery. Thus, the natural conclusion is that Sue’s reason for calling Pizza Yurt is that she wants to have a pizza delivered. Certainly, if we assume (i) that Sue wants a pizza, (ii) that calling Pizza Yurt is the most efficient (or only) means of satisfying this desire, and (iii) that Sue knows that calling Pizza Yurt is the most efficient (or only) means of satisfying her desire, it seems clear that if Sue fails to call Pizza Yurt, she is being irrational.

It is important to note that we now have before us two different claims: One is that Sue’s reason to call Pizza Yurt is dependent on her desire; the second is that, given this desire plus her belief that calling Pizza Yurt is the most efficient way of satisfying that desire, if Sue fails to call Pizza Yurt, she is being irrational. I think that the second claim is true, but the first is false.

It is important to distinguish claims about what reasons an agent has from claims about the agent’s rationality. These claims are different as can be revealed by a common example:

Snake/Rope: I am walking through the desert; I look down at my feet and see a snake-like object coiled in the path immediately next to where I am walking. I immediately form the judgment that there is a snake in my path. My pulse quickens, I immediately feel fear, and quickly leap away. In reality, the object is not a snake but a coiled length of rope.

Now, let’s ask whether I had any reason to fear and whether I had any reason to jump away. The natural response to such questions is that, given that it was not a snake, I had no reason to fear. Similarly, my daughter has no reason to fear the non-existence monsters under the bed even though she firmly believes in them. Further, I had no reason to jump away in fear since there was nothing to be afraid of. However, given that I believed that there was a snake on my path, my fear-response and avoidance behavior was completely rational. The Snake/Rope example shows that we can behave rationally even when we are not responding to reasons.

In this post, I will be making claims about reasons rather than rationality. In particular, I will argue that reasons are not dependent on desires. I will not be defending any view about rationality. However, the view that I find attractive is that, as Derek Parfit has put it, our reasons are provided by the facts, what is rational for us to do depends on our beliefs.

While DBR seems to make sense of Sue’s reasons in the Pizza example, I think that this appearance is deceiving and that a proper understanding of Sue’s reasons show that they are not at all dependent on her desires. To understand why, I need to explain why I doubt DBR. Let’s look at three cases that present possible counterexamples to DBR.

(1) Gin/Petrol

This example comes from Bernard Williams who is a defender of DBR [2]. Williams’ view is that the reasons that an agent has are dependent on the agent’s motivational set. Your motivational set includes your desires, intentions, positive attitudes, etc., (henceforth, I will use ‘desire’ to refer to all of these kinds of states) which have the tendency to motivate you to act. Importantly, this entails that the existence of a reason for an agent requires the presence of an appropriate desire.

Williams presents the following example as a reason to amend his view: Suppose Jim is sitting at a table on which is a bottle of clear liquid. Jim wants to have a gin and tonic and believes that the stuff in the glass on the table is gin. Suppose, though, that the stuff in the bottle is not gin but gasoline. Does Jim have a reason to mix the stuff in the bottle with tonic and consume the concoction? Williams points out that we are pulled in two different directions. On the one hand, it seems natural to say that since the stuff in the glass will make Jim sick, he does not have a reason to drink it and has a reason to not drink it. On the other hand, if Jim does drink the gasoline, there is a natural explanation for why he did so: Jim thought that the stuff in the bottle was gin.

Williams’ conclusion is that we shouldn’t think that the fact that we would have an explanation for why Jim would drink the stuff entails that Jim would have a reason to do so. In fact, according to Williams, even though he believes that the stuff is gin, Jim does not have a reason to drink it. This requires that Williams amend his view. He does so by claiming that an element, d, of an agent’s motivational set will not provide a reason to the agent if d is based on a false belief (see Williams, 293).

Williams is right, in my view, that Jim does not have a reason to drink the stuff in the bottle despite the fact that he has a desire to drink it.  However, I think that the example poses a bigger problem for his view than he realized. The problem is that it is difficult to square the claim that desires that are based on false beliefs do not generate reasons with the view that desires generate reasons.

Williams’ view seems to be that some desires generate reasons and some do not. But, having asserted that all reasons depend, for their existence, on desires, his view provides no basis for claiming that some desires do not generate reasons. Let’s call the capacity to generate reasons, which, on Williams’ view, at least some desires have, “reason-generating power” (or “rg-power” for short). On Williams’ view desires have rg-power unless those desires that are based on false beliefs. But what is it about this class of desires that makes them impotent to generate reasons? Why would a false belief interfere with a desire’s rg-power? Why does a desire lose its reason-generating capacity just because it is based on a false belief? Williams provides no answer.

This example does not refute DBR, but despite Williams response, I think it should cast some doubt on the thesis. Unless we have some basis for thinking that desires that are based on false beliefs cannot generate reasons, the Gin/Petrol example suggests that desires do not generate reasons.

(2) Agony

This example is inspired by Parfit’s Agony Argument.[3]

Suppose Sally want to experience excruciating pain. She realizes that stabbing herself in the eye with a metal fork will satisfy this desire. Does Sally therefore have a reason to stab herself in the eye with a fork? I don’t think she does. The fact that Sally will experience horrendous and needless suffering is a reason for her to not stab herself in the eye with a fork regardless of her desires.

Suppose that Ryan does not have a desire to experience excruciating pain but also does not want to avoid it. Does Ryan have a reason to avoid stabbing himself in the eye with a metal fork? On a view according to which an agent only has reasons to engage in actions that satisfy his desires, Ryan does not have such a reason. This is implausible. Ryan has a reason to avoid stabbing himself in the eye with a fork (and to avoid any activity that will cause horrible suffering) regardless of whether he has any desire to avoid pain/agony.

(3) Radio

This example is an adaptation of one originally provided by Warren Quinn in his paper, “Putting Rationality in its Place.” [4]

Suppose you have a friend, Tom, who engages in the following behavior: When Tom enters a room with a radio that is turned off, he immediately turns the radio on. He does not tune the radio to a specific station; he seems content merely to have the radio on, even if it is playing static. Tom does this consistently. When you ask him why he does this, he says that he wants radios to be turned on. He does not cite a desire to hear music or sound of any kind. He merely says that he wants that radios are turned on.

Does the presence of such a desire make Tom’s behavior reasonable?  I think it is natural to say that Tom’s answer to the question of why he is always turning on radios makes his behavior seems even more unreasonable since the desire itself is irrational. Quinn’s point, I take it, is that the mere presence of this desire cannot give Tom a reason for turning on radios. What would rationalize Tom’s behavior, according to Quinn, is Tom’s belief that by turning on radios, he is achieving something good. If Tom was turning on the radio to listen to good music or to hear the news, this would make his behavior reasonable. But the mere presence of a desire that radios be turned on does nothing to make this action reasonable.

These three examples, taken together, cast a great deal of doubt on the DBR thesis. I will not claim that they effectively refute the thesis, but merely that they strongly suggest that it is false. There are many other arguments against DBR. If you are interested, I highly recommend Jonathan Dancy’s book, Practical Reality, which mounts a sustained criticism of DBR.

Let me return, briefly, to the Pizza example. If we reject DBR, how are we to account for Sue’s reasons? More importantly, if we reject DBR, how are we to account for the fact that Sue does have a reason to call Pizza Yurt, but a person who has no desire for pizza (apparently) does not have such a reason?

Here is my answer: Sue does have a reason to call Pizza Yurt, but this reason is not generated by and does not depend on her desire to have a pizza delivered to her home. What gives Sue a reason to call Pizza Yurt, I believe, is the following collection of facts: Sue is hungry, eating pizza is a good way of satisfying that hunger, Sue likes the taste of pizza, and calling Pizza Yurt is an efficient way of getting a pizza delivered to her home. What about those who lack a reason to call for delivery? For such people, it (currently) will not be good to have a pizza. Either such people are not currently hunger or do not enjoy the taste of pizza or else having a pizza delivered would in some other way be bad.

What reasons we have depends only on what is good and what is bad. It does not depend on our desires.

[1] Hume, D. A Treatise of Human Nature (

[2] Williams B., “Internal and External Reasons” reprinted in Shafer-Landau and Cuneo (eds.) Foundations of Ethics (Oxford: Blackwell, 2007).

[3] In Parfit, D. On What Matters, Volume 1 (Oxford: Oxford University Press, 2011). See pp. 73-82 for Parfit’s Agony Argument.

[4] Quinn, W. “Putting Rationality in its Place” in his Morality and Action (Cambridge: Cambridge University Press, 1993).