On Desire 2018. Part 22: Against Doxastic and Perceptual Evaluativism

Continuing on with: Doring, Sabine A. and Bahadir Eker (2017), “Desires Without Guises.” In Deonna J. & Lauria F. (eds). The Nature of Desire. Oxford University Press.

After setting up the case in favor of evaluativism (which Döring and Eker did exceptionally well), they then tackled the view they set up.

Recall that there are two major evaluative theories; doxastic and perceptual. Doxastic evaluativism states that a desire that P is a belief that P is good. Perceptual evaluativism states that a desire that P is an appearance that P is good (or P looks as if it is good). Döring and Eker take each of these in turn and demonstrates what is wrong with each.

Against doxastic evaluativism, the main force of the objections involved providing instances in which people – or other creatures - had a desire that P without a belief that P was good.

For example, a common criticism against doxastic evaluativism concerns the mental states of animals. Animals clearly have desires and aversions. Yet, it seems unreasonable to hold that animals have a belief that P is good – this appears to be beyond their capacity. This creates a problem for the thesis that a desire that P is a belief that P is good.

They also bring up the case that, “You cannot desire that P if you think that nothing you can do would be conducive to P’s being the case.” Yet, as I argued above, I reject this on the ground that if a person goes from believing there is something you can do, to believing that there is nothing you can do, to believing there is something you can do again, that the desire vanishes during the intervening period.

A third type of case involves the agent who decides to do something (e.g., go see a concert) without having a belief that it is any good – somebody deciding to try something they are skeptical about. Though I do not accept doxastic evaluativism, I do think such a theorist has a response to this. Such a theorist only needs to think that it is good to try new things, to experiment, to expand one’s horizons, to explore. It need not be the case that the end result is good if the journey itself is good.

Fourth, there are cases where an agent has a desire that P where one believes P to be bad. Their example is that of a person at a meeting wanting to tell a joke that they know would not be appreciated in the present company yet wanting to tell it anyway. I believe the kind of case they have in mind is that of a person who wants to do something unruly, but thinks that he better not do so, even though he still wants to. The fact that other, more prudential desires override the desire in question does not imply that the desire in question does not exist.

The authors present a fifth type of case concerning instrumental value that I have had difficulty understanding. There may be a typographical error in the example. They write about Thomas, who plans on taking a break in an hour and brewing some coffee. They then ask, “Is it really plausible to say that Thomas now literally believes that going to the kitchen in fifteen minutes is (instrumentally) good?” Did the authors intend to say, “in an hour?” That would imply that the objection somehow compares the desire to take a break in an hour corresponds to the belief that it would be good to take a break in an hour.

I do not see how this creates a problem for the doxastic evaluativist. As I have understood instrumental goods, they are always eliminable. We can split them into component elements of ends plus beliefs about how actions relate to those ends. A desire-as-means to turn the temperature up is a desire-as-end to be warm and a belief that by turning up the envelope to get warm. The doxastic evaluativist version of this holds that the desire is a belief that being warm is good accompanied by a belief that by turning up the thermostat, one can get warm. We shouldn’t need an additional belief to the effect that turning up the thermostat itself is good.

Finally, Döring and Eker assert that doxastic evaluativism cannot really handle the radioman case. Recall the radioman turns on radios even though he does not evaluate doing so in any positive light. Against this, Döring and Eker argue that giving Radioman a belief that turning on radios is good would not make his actions any less bizarre. Recalling the point that desires are not only supposed to explain action but to rationalize them – getting them to make sense – adding a positive evaluation to turning on radios still fails to make any sense of the actions.

The perceptual model of evaluativism is meant to handle the objection that animals and small infants have desires. They may not be able to have beliefs that P is good, but they have the capacity to perceive that P is good in the same way that they can perceive that a feeding bowl is empty (Oddie, 2017) or that an intruder has entered their territory.

However, according to Döring and Eker, perceptual evaluativism cannot handle the other problems that doxastic evaluativism has, plus has a few problems of its own.

They build some of these objections on the premise that somebody cannot desire something that they know to be the case. As I have argued above, I reject the thesis that a person cannot desire that which is already the case. A "desire that P" is a desire that "P" be made or kept true – and keeping "P" true requires that it already be the case.

Döring and Eker argue that the perceptual evaluativism faces a problem is that a desire persists even while the agent’s attention shifts to other things, and then back again. The only type of desire that the perceptual model can handle, they argue, are the occurrent desires that attract an agent’s attention at a particular moment. Yet, the bulk of our desires are standing desires. I have included in this set of standing desires the aversion to pain, which an agent has at all times and not just while he is thinking about pain and how to avoid it.

The perceptual evaluativist response to this is to say that, in the same way that things do not cease to be red when one is not looking at them, something does not cease to be desired when one is not perceiving them. All that is required for something to be desired is that it be such as to appear as good when one looks at them. Döring and Eker describe evaluativism as saying that one cannot desire something without perceiving it as good.

Furthermore, between manifestations of occurrent desire, one is not perceiving the object of evaluation as good. So, they argue, the evaluativist cannot say that, between occurrences of the occurrent desire, the agent cannot desire that thing.

Döring and Eker further argue that perceptual evaluativism does not account for the capacity of desires to rationalize action. After all, on the perceptual model, that which is perceived as good might not be taken to be truly good. If some object of evaluation merely seems to be or is perceived to be good, then it merely seems to be or is perceived to be a reason for intentional action. It is not, in fact, a reason for intentional action unless it is, in fact, good.

Here, we enter into a consideration about the Radioman case we have been discussing – the person with the disposition to turn on radios. If we give Radioman a perception that turning on radios is good, this still seems a bit bizarre, according to Döring and Eker.

So the question is once again whether the assumption that Radioman experiences turning radios on as intrinsically valuable makes his action any les bizarre than it was before, and the answer is once again no.
There is something about perceiving the turning on of radios to be intrinsically good that seems mistaken in some way. One way to account for that mistake is to deny that any urge to turn on radios can make such an act rational.

These points aim to tell against doxastic and perceptual evaluativism. According to Döring and Eker, these are the only two legitimate types of evaluativism so, to defeat these, is to defeat evaluativism itself. Still, they have a few more things to say against evaluativism, which I will examine next.

On Desire 2018. Part 21: Desires as Explanatory Entities

Continuing on with: Doring, Sabine A. and Bahadir Eker (2017), “Desires Without Guises.” In Deonna J. & Lauria F. (eds). The Nature of Desire. Oxford University Press.

Before raising objections to the evaluative theory of desires, Döring and Eker, attempt to motivate evaluativism by presenting an argument in its defense. This, they do quite well.

First, it explains intentional action, not as a mere event, but as a special kind of event. Intentional action seeks an end, and seeks to bring it about by selecting the effective means.

The action is not something that simply happens to the agent but, rather, something that is performed by the agent. . . . Intentional actions are purposeful and goal-oriented; they are performed with a purpose in mind and some sense of how to achieve that purpose – they are, in short, actions performed for a reason.

I mentioned this above in response to Döring and Eker’s own theory of desire – D1. Their account did not seem to answer the question of “why” the agent acted – only that an agent is disposed to act in a particular way. We will get to the authors’ answer to this challenge in a later section.

Second, a theory of desires explains the rationality of (reasons for) intentional action.

An intentional action is motivated by a desire of the agent that encodes a goal, but if such behavior is to be considered rational (even in the aforementioned minimal sense), desires must, so the story goes, also involve an evaluation of that goal as something worth pursuing.

This is a feature of desires that I have not addressed much in this series. Desires provide end-reasons for intentional actions. They are intimately concerned with the rationality of action and the rationality of the agent herself.

As Döring and Eker describe their opponents, they hold that the Evaluationists say that this requires that the object of desire be given some value, that there be some reason to pursue that end, and that the dispositional theorists do not provide this. According to the dispositionists, a desire simply reports THAT the agent is disposed to realize some end without explaining (or accounting for) why they pursue that end. The evaluationist explains this by saying that the agent sees some value in the end.

This, then, is the challenge that Döring and Eker sought to meet. The plan at this point will be to show that both of the key Evaluationist theories are problematic, and then to criticize the argument given to motivate the Evaluationist argument.

On Desire 2018. Part 20: Resurrection of Desire

Continuing on with: Doring, Sabine A. and Bahadir Eker (2017), “Desires Without Guises.” In Deonna J. & Lauria F. (eds). The Nature of Desire. Oxford University Press.

When I considered Graham Oddie’s (2017) paper, I wrote about “the death of desire” – the idea that “a desire for S must vanish once one comes to believe that S already obtains.” I objected to that thesis then, just as I objected to it above. However, some of the comments I found in Doring and Eker suggested a second way of looking at the claim.

Döring and Eker distinguish desires from wishes. A wish (which is not a desire) applies to states that one cannot change through action. What is currently the case is like a wish in that it cannot be changed. So, what is currently the case cannot be desired.

From this perspective, the question of whether wishes are desires is like the question of whether Pluto is a planet.

It is not a question that can be settled in a laboratory.

We could try conceptual analysis. Though, astronomers certainly rejected that option; Pluto had been called a planet since its discovery. Instead, they argued that it was more important that like be classified with like. Pluto was more like a Kuiper Belt Object than a planet, so the International Astronomical Union altered its definitions to classify like with like.

Philosophers do not have anything like the International Astronomical Union to take a vote on the matter. We need to negotiate some sort of agreement.

From this perspective, I cannot argue that the death of desire thesis is false, any more than I can argue that the “Pluto is a planet” thesis is false. We can debate the findings of conceptual analysis, but even this does not settle the issue. If the conceptual analysis goes against me, I can still, like the astronomers, argue that those concepts are clumsy and in need of revision. They could make the same arguments against me.

In short, against Döring and Eker, my previous response to the death of desire fails because it assumes that the concept of desire is "settled" - and settled in my favor at that. I hold that wishes are desires. Döring and Eker hold that wishes - and anything else that assigns value to something that intentional action cannot change - are not desires. Is Pluto a planet? Or is Pluto not a planet?

In these writings, I am going to stick with the idea that wishes are desires. All affective states that assign motivational value to the realization of a proposition are desires. I am going to hold with Oddie’s suggestion that a desire that P where one believes that P is true is a satisfied desire. I would distinguish this from a fulfilled desire, where P is actually true.

On Desire 2018. Part 19: Desires Without Action

Continuing on with: Doring, Sabine A. and Bahadir Eker (2017), “Desires Without Guises.” In Deonna J. & Lauria F. (eds). The Nature of Desire. Oxford University Press.

In the previous section, I presented Doring and Eker's thesis of desire, which is that it is a disposition to act.

After presenting this thesis, they considered a number of counterexamples. One that they seem to have found particularly bothersome came from Galen Strawson (1994).

The Weather Watchers are a race of sentient, intelligent creatures. They are distributed about the surface of the planet, rooted to the ground, profoundly interested in the local weather. They have sensations, thoughts, emotions, beliefs, desires. They possess a conception of an objective, spatial world. But they are constitutionally incapable of any sort of behavior . . . . They lack the necessary physiology. Their mental lives have no other-observable effects. They are not even disposed to behave in any way.

This conceptual possibility is supposed to break the link between desire and dispositions to behave. These Weather Watchers have desires, but no dispositions to behave.

Doring and Eker respond to this objection by denying that the Weather Watchers have desires. Indeed, to have a desire, there must be something that one can do. Where there is no possibility for action, there is no desire.

In light of this, they add the following criterion to desires:

(D2) Necessarily, for any agent a, any proposition p, any time t, if, at t, a desires that p, then there is at least one act type φ such that, at t, a does not think that her φ-ing not to be conducive to p’s being the case.

In other words, there has to be something a can do that is conducive to p’s being the case. If there is no such thing, then there is no desire.

This lead them to a third criterion of desire:

(D3) Necessarily, for any agent a, any proposition p, any time t, if, at t, a desires that p, then, at t, a does not think p already to be the case.

This is because if, at t, p is already the case, then there is nothing that one can do conducive to making p the case. There is no action one can perform, no disposition to act, and, consequently, no desire.

I have always thought of a desire as being analogous to a force. It has a direction (realizing p) and a magnitude (an assigned value of v), and it applies this motivational force to realizing p. We think of forces as that which explains and predicts the movement of matter through space. Similarly, we use desires to explain and predict the movement of intentional agents – to predict and explain their decisions and the actions they take (or refrain from taking) as a result of those decisions.

There could well be forces in existence that have no effect on matter. However, we have no reason to postulate their existence. For example, there may be a force F where, at any time force F appears, an equal and opposite for F’ appears that counterbalances it. We will never experience this force in any observation of the real world. It may exist. However, we only have reason to postulate the existence of forces that have an effect we can see and measure – that produces effects we have reason to explain and predict.

Accordingly, there could be desires that do not manifest themselves in any way, housed in agents like the Weather Watchers where they cannot result in any course of action of change of state that we can observe. They may exist. However, we have no reason to postulate their existence. So, just as we can imagine the force with the counterbalancing force that never shows up in any observation, we can imagine the Weather Watchers with desires that never show up in action.

However, we when we are considering a desire when the agent thinks that p is already the case, we have a situation where we have evidence of a force, then our evidence of the force disappears, and we are to think that the force itself disappears.

Imagine a compass in a magnetic field. The magnetic field influences the compass, causing the compass needle to point north. We then destroy the compass so that we can no longer see the effects of the magnetic field. A few minutes later, we get a new compass. We can once again see the influence of the magnetic field on the compass. There is no reason to believe that, during the time when the one compass was broken and before the other one arrived, that there was no magnetic field.

Doring and Eker would have us take an agent with a desire that p where the agent believes that p is not the case (and is thus φ-ing to bring about p), where p then becomes the case, and when p ceases to be the case once again begins φ-ing to bring about p, that, during the moments when p was the case, that there was no desire. Yet, that desire returned again, emerging from nothing, the instant that p ceased to be the case. Where did it go during the mean time? And why did it come back when p ceased to be the case?

Note that Graham Forbes (2017) provides another example whereby a person with a desire that God exists firsts questions the existence of God, comes to believe in the existence of God, then is convinced that there is no God. Rather than assume that her desire pops in and out of existence, it makes more sense to believe that it persists throughout. If there is a disposition to act in the desire that p, it is a disposition to begin acting as soon as the agent to start φ-ing as soon as she believes that there is something that she could do to realize p (so long as there are no combination of other desires conducive to not φ-ing).

So, in summary, we can imagine the Weather Watchers having desires though, in reality, we would never assign desires to such creatures. We use desires to explain and predict intentional action. Where there is no intentional action, there is no reason to postulate the existence of desires. But desires, like forces, do not spring into existence only when they have effects and cease to exist when we can’t measure them. It makes far more sense to think of them as persistent entities. The desire that p continues to exist, even while the agent believes that p is already the case.

This does raise a question. The standard understanding of desire is that of a dispositional state – like being soluble or fragile. I am presenting desires as a motivational force. What is the relationship between a dispositional state and a force? Are they the same thing? Or is this a competing conception of desire? I am not sure on that matter.

On Desire 2018. Part 18: Doring and Eker's Agency

Continuing on with: Doring, Sabine A. and Bahadir Eker (2017), “Desires Without Guises.” In Deonna J. & Lauria F. (eds). The Nature of Desire. Oxford University Press.

I have mentioned the view that Doring and Eker are attacking. Now, I want to look at what they defend.

This begins with what they call D1:

(D1) Necessarily, for any agent a, any proposition p, any time t, and any act type φ, if, at t, a desires that p, then a is disposed at t to φ in circumstances where a takes her φ-ing to be conducive to p’s being the case.

Now, I have a question.

Why?

As in: Why is she φ-ing at t? This is where the radioman objections come in. The radioman objections describe the agent φ-ing at t, but without having a reason to φ at t. It just happens. But, then, that seems to miss something about desires.

Doring and Eker will attempt to address this problem later. They are familiar with the objection.

I want to raise a second objection. What does this tell us about cases where an agent is under the influence of two desires at once?

Suppose an agent with an aversion to her own pain (e.g, a “desire that I not be in pain”) and an aversion to causing pain to others (e.g., a “desire that I not cause pain to others”) arrives upon a situation where the only way to avoid pain is to cause pain to others. What does D1 tell us about how that person will act?

We use desires substantially for this purpose. For example, if we know that somebody is particularly kind and concerned about others, we can use this to predict that she will choose to endure pain rather than cause pain to others. At the same time, the selfish individual will more readily choose the option that causes pain to others.

The only thing D1 tells us is that the agent is both disposed to avoid suffering pain and to avoid causing pain to others. D1 tells us nothing about the situation where these desires come into conflict. It tells us almost nothing about one of our primary reasons for attributing desires to a.

In order to answer these types of questions, we have to say more about the agent with the desire than that she has a disposition to act. She needs to have a way of weighing one desire against another – a way of making choices at times of conflict (because we are always at times of conflict). For that, each desire has to be assigned a strength. There must be some sort of value assigned to the agent’s “desire that p” that the agent can refer to when deciding whether the agent is going to φ at t under circumstances where she takes φ-ing to be conducive to p’s being the case, or not φ at t because the agent has a desire that q and not φ-ing is conducive to q being the case.

This assignment of a value – of a degree of importance – to p being the case as opposed to q being the case – is precisely what evaluative theories of desire provide that dispositional theories miss. Not only do these assignments tell the agent how to weigh one desire against another, it answers the question, “Why φ?” The answer can be found in the importance of p being the case.

On Desire 2018. Part 17: The Guise of the Good

I am now moving on to the third article of the summer: Doring, Sabine A. and Bahadir Eker (2017), “Desires Without Guises.” In Deonna J. & Lauria F. (eds). The Nature of Desire. Oxford University Press.

In their contribution to The Nature of Desire”, Sabine Doring and Bahadir Eker take on the thesis that desire is concerned with the assignment of a value to a state of affairs. The title of their chapter, “Desires Without Guises,” comes from the slogan common in the discussion of desire that the objects of desire come to us under the guise of the good. This is a reference to the assignment of value to that which is desired. Rejecting the guise of the good thesis is rejecting the assignment of value thesis.

The thesis held up in contrast to the guise of the good thesis is the dispositional thesis – that desires are dispositions to act in a particular way. This is the thesis that Warren Quinn (1993) famously refuted using his Radioman counterexample, concerning a man with disposition to turn on radios though he has no interest in having the radios on, in turning on the radios, and rather wishes there were no radios around so that he would not turn them on. On this counter-example, the assignment of some value to turning on radios is needed to produce a desire.

Doring and Eker want to challenge assignment of value theories and defend a version of the disposition to act theories.

To get started, Doring and Eker define the thesis that they want to attack:

(ME) Necessarily, for any agent a, any proposition p, any time t, if, at t, a desires that p, then, at t, a evaluates p positively (as good).

I have been defending an assignment theory of desire (to desire that P is to assign a value to P being made or kept true), so I am going to want to say a couple of things about his characterization of the evaluation thesis.

First, assigning a value to P being true is not the same as evaluating p positively. The value that I say is being assigned is nothing more than a motivational force being directed at making or keeping P true. The agent might not even be aware of the desire that P. However, making or keeping P true is a goal of his. Recognizing this fact best explains and predicts his behavior.

Second, “as good” is ambiguous. “Good” relates objects of evaluation to desires. However, the there are many ways to relate objects of evaluation to desires. The question arises, “Which desires?” There is a sense of good that, in the thesis laid out above, relates p to the desire that p, in which case the thesis is accurate in this sense. However, this does not imply that p is good – or p bears a positive relation – to all possible desires. Indeed, this is almost certainly not the case.

Somebody reading these caveats may be confused as to whether I am actually offering an evaluative theory of desire or a dispositional theory of desire. After all, I am understanding “assigning a value to P being true” in terms of a motivational force for making or keeping ‘P’ true – a disposition to act. Perhaps this is what we need to settle the dispute between the dispositional and evaluative theorists – a theory that does both at the same time.

I invite the reader to keep these caveats in mind as we proceed.

Also, please recall that Doring and Eker are attacking this view, not defending it.

I will look at the view that they sought to defend next.


Quinn, Warren. (1993). “Putting Rationality in Its Place”, in Morality and Action. Cambridge, UK: Cambridge University Press.

On Desire 2018. Part 16: Felt Need

Continuing with my series on the nature of desire, I am currently commenting on Friedrich, Daniel (2017), “Desire, Mental Force, and Desirous Experience.” In Deonna J. & Lauria F. (eds). The Nature of Desire. Oxford University Press.

In Part IV of the essay, Friedrich begins to describe the feeling tone distinctive of desire in terms of a "felt need". In the sections above I have given reason to question the "felt" part of this account of desire. However, Friedrich makes several claims about the "need" part that fits into the idea that desire is a value assigned to P being made or kept true.

He states that felt needs can "differ in intensity". This is consistent with the idea that different values can be assigned to different propositions P being made or kept true. The value can be low (Yeah, that'd be nice, I guess), or high (STOP THAT! PLEASE! I BEG YOU!).

He states that felt needs can sometimes be given expression and uses the example, "This must become reality." This can be paraphrased as, "P must be made or kept true." Though, please note, there is nothing here about a "felt" aspect, just a need.

He also writes, "These experiences involve a distinctive mental force that is anchored in a distinctive phenomenology that can be articulated in terms of the notion of a felt need inasmuch as in these experiences the desired end is given to the mind as something that has to become reality." Here, Friedrich does mention this "distinctive phenomenology." However, once again, in order to explain and predict your behavior, I do not need to make any reference to this phenomenology. All I need to see is that you are willing to put a lot of work into making P a reality.

What I see is not the Radioman's turning on radios while wishing there were no radios on so that he would not be turning them on and putting up with their noise. For your motivation to be that of a desire, it must be the case that the end has been assigned a value. I must see you working, planning, and plotting to making P a reality. In observing your behavior, it must be the case that my best explanation is, "She wants to make P a reality and, given this and what I know about her beliefs, I can predict that she will do this and that, avoid this other thing, and be willing to make the following bargains." Once again, you may or may not have some sort of feeling associated with this, but that does not have to be true for you to have a desire.

Lastly, Friedrich states:

It is in virtue of the desired end being given to the mind as something that has to become reality that acting in ways one believes to promote the desired end is minimally rational.

Here, too, we can make use of the part of this account that identifies the importance of making something become reality (making or keeping P true) as an end, something the agent is willing to work for, without adding any type of feeling component.

I can well imagine that somebody might say, "The fact that some demon put a chip in my brain making it important to me to realize some proposition P does not give me any reason at all to realize some proposition P." Yet, in fact, I suspect each person will find a reason to avoid pain, to adjust the thermostat to make himself comfortable, to eat, to choose the foods that taste best among what he chooses to eat (consistent with other concerns), and the like – in spite of the fact that his reasons to do so are based solely on the chips that the demons of Evolution, Environment, and Experience have placed in his head.

This account has an implication that some may not like. It has an implication that I do not like. That is, it may be easier to create a robot with desires than we think. Such a robot needs to assign a value to realizing some end, and then go about finding ways to do so, collecting beliefs about the world in order to build a plan to achieve the desired goals. There is quite the risk that, as we did with animals in the 1800s, we will adopt a policy of thwarting the desires of machines while denying that they have real concerns. In fact, they will be giving us all of the evidence they can, and as much as we can demand of them. We are at risk of entering an area where we do to robots what we once did to animals.

On Desire 2018. Part 15: Other People's Pain

Continuing with my series on the nature of desire, I am currently commenting on Friedrich, Daniel (2017), “Desire, Mental Force, and Desirous Experience.” In Deonna J. & Lauria F. (eds). The Nature of Desire. Oxford University Press.

In my last posting, I objected to their thesis concerning of a feeling tone distinctive of desire. In this post, I address the claim that one can launch the same argument against a feeling tone distinctive of pain.

Specifically . . .

In the previous section, I argued against the idea that we can understand desire in terms of a "feeling tone distinctive of desire" because I cannot tell if you ever have a feeling tone distinctive of desire.

The same argument can show that pain is not related to a distinctive feeling either. After all, the only thing I see from you when you are in pain is your observable behavior. I have no access to what you are feeling. It is quite within my capacity to imagine that you feel nothing – that you are just going through the motions without any feeling of pain at all.

Indeed, this was the case in the seventeenth and eighteenth centuries regarding animals. People believed that animals lacked the intellectual sophistication required to actually feel pain. Whimpering, whining, yelping, and attempts to avoid what, in human, would cause pain were merely examples of animals going through the motions of pain, but they did not suffer from any distress. For that, one needed a soul, or a particularly sophisticated type of consciousness.

This part is true: I can imagine an advanced race sending a "pod person" down to Earth with the intention of learning about us. I can imagine that this being has no "feelings". However, it learns how to fit in. It learns that certain types of behavior (I.e., leaking water out of its eye and displaying the mannerisms of somebody who is sad) it is possible to elicit sympathy. Other types of behavior, such as praise and condemnation, mold the behavior in others in such ways that they tend to repeat that for which they are praised and avoid doing that for which they are condemned. Consequently, it expresses anger or joy. It will behave submissively at times to protect itself from harm, and display dominance behavior as a way of getting others to do its bidding. Ultimately, it becomes indistinguishable from a human being in behavior. And, yet, it has no felt qualities of desire, of pain, of joy or sadness. It simply goes through the motions.

Not only can I imagine this, but people have assumed it with regard to animals and, sometimes, other people.

I can imagine it. But what does that tell us?

There is a large body of philosophical literature on this subject having to do with the problem of consciousness and the metaphysics of mind. This has to do with the possibility of p-zombies; creatures that behave exactly like humans to the point of being indistinguishable from humans, but lack consciousness. Hopefully, I can avoid that debate and all of its complications and implications. For my purposes, I am quite content to assume that we all are, in fact, philosophical zombies.

The relevant point for this discussion is that it does not matter morally whether you are a p-zombie or a regular person. Ex hypothesi, I cannot tell the difference by looking at your behavior or through any type of observation. That you have the same sensations I do when you appear to experience pain, and whether you experience a "feeling tone distinctive of desire" when you desire that something be the case, when it comes to explaining and predicting your behavior, if a "desire" or "is in pain" attribution is useful in one case, it is just as useful in the other.

I could be the only conscious person in a world filled with philosophical zombies. As I grow up, I hear them using the terms "desire" and "pain". In learning the language, I would learn to apply those terms to my own experiences. I will learn to say that I want to go camping and I would say that the burn on my hand is painful. I may notice that desires have a distinctive feeling tone (if, indeed, it does), and that there is a felt awfulness to pain. Yet, here, we have to ask whether this is a part of the actual meaning of the word. I learned the words "desire" and "pain" from beings who never experienced such a thing.
So, what does it matter whether other people have a "feeling tone distinctive of desire"?

On Desire 2018. Part 14: Other People's Desires

One question that a theory of desire should be able to answer is: How do you determine what another person desires or, even, if they have desires?

In reading Daniel Friedrich's article, this is a question I had problems with.

Friedrich, Daniel (2017), “Desire, Mental Force, and Desirous Experience.” In Deonna J. & Lauria F. (eds). The Nature of Desire. Oxford University Press.

How do I tell if you have any desires?

This is the question I keep asking as I read Daniel Friedrich’s account of desires.

Freidrich writes:

Non-cognitive evaluation, I have argued, should be understood in terms of evaluative mental force. The notion of evaluative mental force is in turn tied to the distinctive feeling tone of certain experiences. In this section, my aim is to relate these ideas to desire. An extension of these ideas to desires requires that there is a feeling tone distinctive of desire.

But how do I know that you have a particular feeling tone distinctive of desire?

Let us take color by analogy. I do not know whether you see red the way that I see red. What I do know is that, when you point at things and call them “red”, I can pick out what they all have in common and use that to predict what else you may call “red”. You have never seen my heavy winter coat. Yet, I can reliably predict that you would say it was red. Or, at least, the vast majority of people who saw my heavy winter coat would say that it was a red coat.

That seems problematic when it comes to desire. When you point to things and call them “desired by me,” I look for what they all have in common so that I can come to reliably predict what else might be desired by you. However, I have different desires. Let us imagine that you like sports, while I prefer to play computer games. You drink coffee in the morning; I prefer cherry flavored Diet Dr. Pepper. You like Jazz; I prefer classic rock-and-roll.

When I look at the set of things that you desire, I am not going to pick up on some "feeling tone distinctive of desire" as being what they all have in common, which I can then put to use to determine what to get you for Christmas.

Perhaps I can infer the feeling tone distinctive of desire. I look at the things that I desire, discover a common “feeling tone”, and infer that you must feel the same thing.
There are at least two problems with this inference.

First, how do I know to attach the term "desire" to playing computer games and the like? You are not pointing to these things and calling them "desired". And I do not find "desiredness" in the things that you point to. In order to know that these are the objects of my desire, I have to find something else in common in what you are pointing to in using that term, and then find that in playing computer games and the like. From there, I may be able to discover that they all have a distinctive feeling tone. However, I first must identify them as objects of desire. Then, I can discover the common tone.

Second, even after I discover the things that the things that I desire have this distinctive feeling tone, how am I justified in inferring that you experience the same feeling tone with the objects of your desire? There is the question of whether your distinctive feeling tone is the same as mine. Then there is the question of whether you have a distinctive feeling tone, or if, for you, there is a collection of feeling tones. If you associated a different feeling tone to watching sports, drinking coffee, and listening to Jazz, but, so far as I can see, you treated them the same way (as objects of desire), I could not tell that they had one tone or three. Nor would it matter.

Whatever a desire is, I need to be able to base my claim that you have a desire that P based on what I can observe, and I cannot observe your private mental experiences. What I can observe is that you seem to be concerned with realizing certain states. You seem to be willing to put some stake in them, devoting effort to discovering and removing obstacles. They are important to you.

Upon observing that there are states that are important to you, I can take this assignment of importance and call it a desire that P. Recognizing that there are states important to me, I can recognize that I must have some desire that Q. I know nothing about the feeling tone of your desires, not do you know anything about mine. However, we know enough to negotiate. “If you help me to realize Q, I will help you to realize P”.

On Desire 2018. Part 13: Objections to Dispositional and Doxastic Desires

We are moving on to another article on desire:

Friedrich, Daniel (2017), “Desire, Mental Force, and Desirous Experience.” In Deonna J. & Lauria F. (eds). The Nature of Desire. Oxford University Press.

“Desire, it is often said, is a pro-attitude”. This is how Daniel Friedrich begins his article.

He begins by looking at a couple of ways of understanding this, and providing reasons to reject this. The dominant views are: (1) desiring p disposes us to act in ways designed to realize p, and (2) desiring p entails a positive evaluation of p. These are the “doxastic” and “evaluate” theories of desire.

As a point of contrast, I have been arguing that desiring p entails an assignment of value to p being made or kept true. This is distinct from both of the theories presented above, which can be seen in the way it handles the objections.

Dispositions to Act

Friedrich begins by mentioning the same objections to the “dispositions to act” thesis discussed in Oddie (2017). We can imagine Radioman, who simply has a disposition to switch on radios. This mere disposition – something more of a habit or a tic – does not fit what we are talking about when we talk about desires.

Of course, what is missing is that there is no assignment of value to “the radio is on” being made or kept true. Add this, so that Radioman sees the radio being on as an end in itself – something to be done for its own sake – and we get something nearer to a desire.

However, the claim that a desire is an assignment of a value to p being made or kept true is not the same as believing that p is good or perceiving that p is good. To see this, we need to examine the doxastic view.

Beliefs that P Is Good

An alternative to the view that to desire that p is to be disposed to bring about p is the view that to desire that p is to believe that the realization of ‘p’ is good.
Friedrich brings up several objections to this doxastic view.

First, Friedrich brings up nihilists who believe that nothing has value yet who still has desires. He also brings up people who have a “desire that p” but who believes that p is bad. We can include in this the desires of the addict.

On the assignments theory, there is nothing problematic with a person believing that nothing has value yet still having a brain that assigns negative value to being in pain or positive value to having pumpkin pie with whipped cream. Nor is there any problem with a person having a desire that p and, at the same time, an aversion to having a desire that p, recognizing that his desire that p motivates him to act in ways that thwarts his other desires, that others have an aversion to people having a desire that p, or that others have many and strong reasons to condemn and punish those who are disposed to make or keep ‘p’ true.

Second, Freidrich brings up the objection that a believe that ‘p’ is good seems to require more intellectual sophistication than it is reasonable to assign to animals and small children who, nonetheless, have a desire to eat or an aversion to being in pain. There does not seem to be much sense in saying that one’s pet has a belief that being in pain is bad or that protecting one’s offspring from harm is good.

None of this sophistication is required of an animal that simply assigns being in pain a negative value – and thus works to prevent such states to the degree that she can recognize what might cause them.

Third, Friedrich reports that “beliefs are subject to the norm of truth”. To believe that p is to believe that ‘p’ is true. So, to desire to have pumpkin pie with whipped cream is to believe that having pumpkin pie with whipped cream is good, which means believing that “pumpkin pie with whipped cream is good” is true. Now, we need a theory to explain what it is for p to be true. And, if it is true for our agent Alph, is it true for everybody?

If you imagine there is an elephant in the room, you do not fall short of any inbuilt standard if there isn’t. But if you believe that there is an elephant in the room, your belief falls short of an inbuilt ideal if the closest elephant is in the zoo three miles away.

If one believes that p is good then one is “falling short of an inbuilt ideal” if p is not, in fact, true. So, now we need to determine what it is for p to be, in fact true.

When it comes to assigning a value to ‘p’ being made or kept true, there is no inbuilt assumption of truth. If ‘p’ is true, the desire motivates the agent to keep it true. If ‘p’ is false, the desire motivates the agent to make it true.

A common move to answer this kind of objection to the doxastic view is to say that value is like perception – that desiring that p is like perceiving that p is good.

Yet, perception also seems to have an inbuilt standard of truth. If one perceives that there is an elephant in the room then this perception is suspect if there is no elephant in the room. We still need to ask what is required to make, “There is an elephant in the room” true.

To be fair, one of the objections to the assignment view is that the assignments seem arbitrary. There is no fact of the matter to back them up, so one assignment is as good as another. This is true. Though, it is also the case that the relationships among desires are not merely a matter of opinion. It is a matter of fact that addiction tends to thwart other desires and believing this is not the case will not prevent it from being true. People, who have this arbitrary aversion to pain that evolution happened to give them, as a matter of fact have a reason to cause others to have an aversion to causing pain.

Of course, these are the theories that Friedrich rejected as well. The interesting part will be to look at the theory that Friedrich supports.

On Desire 2018. Part 12: The Desires of Young Children and Animals

Young children and animals have desires. They have hunger, thirst, and an aversion to pain at the least. Cats have a desire to chase and catch things that are like prey. Herd animals have an aversion to cats.

Oddie addresses a concern that says that this is a problem for his “appears to as being good” thesis of desire. “Appears to as being good” seems to be beyond the cognitive capacities of infants and animals – which would leave them without desires. Specifically, it is unlikely that any animal or infant has an understanding of the concept “being good” that would be necessary for anything to appear as being good.

In response to this, Oddie suggests two possible answers.

For his first possible answer, Oddie notes things can appear a certain way to us even though we do not have a concept to describe it. He uses color as an example, noting that, “We experience a far richer palette of colors, for example, than we have the conceptual tools to characterize.” In fact, we cannot even ask the question, “What is that?” unless we had a prior ability to pick “that” out so that we can investigate and think about it.

The second possible answer, he draws on the ideas of Friedrich and Lauria that something can “appear round” in many different ways. It can look round. It can feel round. Using the example of a bat he claims that something can also sound round – though he could also use the example of rolling a marble around in a box.

Similarly, one can argue that there are different modes of presentation of a state of affairs. In the perception of S, S is presented as being the case. In the desire for S, S is presented as being good. One and the same state can be presented in these two different ways. The perception that S and the desire that S take the same object but present S in different ways. (p. 51).

This defense still leaves me with two questions.

The first question springs from noting that, nowhere in this section, did Oddie mention “fittingness”. It is possible for something to “appear good” without its goodness being, in any way, fitting, in the same way that something can appear round or appear red without any claim of roundness being a fitting shape or redness a fitting color. The idea that the brain, in assigning a negative value to “I am in pain” makes it “appear bad” can simply be a basic description that this is how the brain works. From here, survival of the fittest will determine if this particular assignment of value (or this particular way of drawing an assignment of value out of the environment and experience) will get passed to the next generation.

The second question deals with the fact that I do not know what “appears good” is supposed to mean. Specifically, rather than introspecting on my own desires, I am curious to know how I understand that somebody else has a desire. I cannot see how some particular thing “appears” to them. All I can see is their observable behavior and, from that, try to infer whether a desire provides the best explanation.

With respect to colors, such as red, I cannot tell how “redness” appears to other people. However, I can look at what other people point to and call “red” and, from that, make predictions regarding what other things people will call red. I can get pretty good at it – predicting what other people will call “red” with exceptional reliability, without having the slightest idea of how “red” appears to them.

When it comes to desire, I have a problem. People are in substantial agreement concerning what they call “red.” There is no such substantial agreement with respect to what they desire. It would make my job easier if everybody pointed to the same thing and called it “good” or “desired,” but they do not. This is in spite of the fact that, when two people point to the same thing and give it two different evaluations, every other appearance is (quite nearly) the same.

When I turn that knowledge inward, that is where I learn to explain and understand my own behavior as the pursuit of certain ends. I may discover that those ends have something in common, but the word is attached not to this appearance, but to what I can know that I share with other creatures who have desires – a disposition to pursue certain ends or goals. We may not have the same goals, but we do have goals.

This is now I know that young children and animals have desires. It is not by knowing how things appear to them – something I cannot know. It is because the best method I have for explaining and predicting their behavior is to understand them as agents who are disposed to perform goal-directed action. They act with a purpose – an end – to realize (or to prevent the realization) of certain states of affairs. This represents more than just a disposition to act. It represents a disposition to plan – to alter one’s behavior in ways that will realize an end even in environments that provide different means.

On Desire 2018. Part 11: The Death of the Death of Desire Principle

The “Death of Desire” principle notes that a desire, once fulfilled, ceases to exist. Another way in which it is phrased is to say that a person cannot desire that which they know to be the case.

Here, I must admit, this simply seems wrong. It is easiest to see with respect to aversions. My desire that I not be in pain does not cease to be exist simply because I am currently in a state in which I am not in pain. My fear of deep water does not vanish when I am not in deep water. In fact, the persistence of these “desires that not-P” even when not-P is true provides the motivation to make sure that not-P does not become true. It is my aversion to pain when I am not in pain that causes me to make sure that I avoid future pain. It is my fear of deep water even when I am not in deep water that keeps me out of deep water.

In the case of positive desires – desires to realize a state rather than to prevent the realization of a state, it makes sense that evolution would equip us with desires that fade when they are realize. After all, desires command action. It makes sense that evolution would equip us with desires that fade when they are fulfilled so that we can move on to the next project. We eat until we have obtained the nourishment we need, then we go on to do something else. We are thirsty until that point at which we have consumed enough water to restore a healthy balance. We desire sex until we have reached an end that makes reproduction possible (at least males do), and we explore until we have discovered whatever it is we were exploring to discover.

This provides some understanding of where the idea that a desire ends when that which is desired has been realized. However, it is a mistake to attribute this to all desire.

Even in the case of some desires persist. The desire that one’s offspring is healthy and happy persists even when one knows that one’s offspring are healthy and happy. One’s desire to be a novelist persists through the writing of several novels.

Oddie brings up as an example Hillary Clinton’s desire to become president. Then (in his hypothetical alternative universe) Hillary does become the first female president of the United States. She can no longer become the first female president of the United States because she is the first female president of the United States. The desire disappears. However, being the first female president of the United States still appears good to her. This argument creates an objection to Oddie’s thesis, since this is an example where an agent can no longer desire that P (to become the first female president of the United States), but this still appears good to her. If a desire is an appearance of something as good, then there can be no appearance of good if the desire is dead.

Oddie answers this objection by stating that there is a thin desire that persists through the election, but we give different names to the different parts. At the start, Clinton has a perspective desire (a desire for a perspective state) of being the first female president of the United States. Then, she wins the election, and the perspective desire becomes a satisfied desire that she is president of the United States. Indeed, if the desire did not continue to exist, then she could not be experiencing the satisfaction of the desire the day after the election – not if the desire no longer existed to be satisfied.

The fact that Clinton can be satisfied with winning (if she wins) and disappointed with losing (if she loses) suggests that something of the desire survives the election. It does not, in fact, die. It simply changes its name.

The assigned value theory of desire would have the same response. The brain assigns a particular value to being the first President of the United States. This motivates the agent to make or keep the proposition true. When Hillary wins the election, the desire changed from making the proposition true to keeping it true. The desire did not die. It simply shifted to a new, appropriate object.

On Desire 2018. Part 10: Unexperienced Value

I found this part of Graham Oddie's paper difficult to write on. I think it is because I found a hard time getting my thoughts into the correct context.

That paper, by the way, for anybody who may have forgotten, is: Oddie, Graham (2017). "Desire and the Good: in search of the right fit." In Deonna J. & Lauria F. (eds.), The Nature of Desire. Oxford University Press.

So, here is my attempt to understand this part of the paper:

The “admirable” demon example, discussed in the previous section, showed that it was possible for it to be the case that people ought to (or that it would be good to) admire someone (who otherwise threaten to do great harm) to prevent great harm even though that someone was not admirable. At least, this was true in the "Deontic" and "Axiological" versions of FA - but not in the Representational version.

(Representational FA) X is V if and only if it is representationally accurate for one to take attitude F(V) to X.

The “solitary goods” problem is meant to show that the left-hand side of such a biconditional is true, then statement cannot be state-entailing or belief-entailing.

The biconditional we are going to work with here is Oddie's conception of "good".

S is good if and only if favoring S is fitting.

Oddie wants to show that this is false if "S is good" is state-entailing ("S is good" implies "S exists"), or belief-entailing ("S is good" implies "Agent believes 'S is good'")

He will then show that the appearances thesis meets these criteria.

So, what are these “solitary goods”?

Solitary goods are those that exist without anyone’s being around to respond to them fittingly.

I mentioned that I found this difficult to understand. Does this mean that nobody exist who can respond to them fittingly? Or does this mean that such a person exists, but is unable to respond to them fittingly (e.g., because the object is at the center of the Sun where nobody can experience it)? If the former, then does the person have to exist at the same time as the object that has value? For example, what would we say of a situation where I respond fittingly to something that will not exist until 10 years after I die?

In this biconditional, “favoring” is to S being good what “admiring” is to X being admirable or desiring is to D being desirable.

So, the solitary goods case asks whether it is possible for “S is good” to be true, and “favoring S is fitting” to be false.

I would say “yes” to this and present as my examples the object of every desire that evolution, the environment, and experience has planted as a chip in my brain. The awfulness of that sore throat that results when my body is fighting off a flu, the taste of pumpkin pie with cool-whip, sex, Jimmy Buffett music, and a long, hot shower. All of these are good. Favoring these are not fitting – they are simply what the chips that evolution, the environment, and experience have planted in my brain.

However, for the sake of discussion, let us limit our focus to the same types of goods we discussed in the previous section – the admirable, the desirable, and the moral. These are goods that people generally have reasons to promote universally. I will bring forth my example from the previous section – the aversion to causing others pain (under the assumption that everybody has an aversion to pain).
Does this have a problem with solitary goods?

Oddie gives us an example:

Consider an apparently good state, E, that happy egrets exist. Conjoin E with the state F: that there are no past, present, or future favorers. Suppose that the conjunctive state E & F is also good.

Well, when I am asked to suppose that there are no past, present, or future favorers, I have to ask, “What about the happy egrets?” If happy egrets exist, then there are present favorers. If there are no present favorers, then happy egrets do not exist. Imaging such a universe in which E & F are good is like imagining a married bachelor named Jim or a round square that is pink.

Perhaps I think I can make this work if I consider an apparently good state – that G.E. Moore’s beautiful planet exists. Though it is beautiful, it contains no evaluative creatures. It has flowers and rainbows clean mountain streams, but no animals. In fact, in this universe, no evaluative creatures exist, have existed, or will exist.

Now we have a situation in which E (a beautiful world exists) & F (there never has been, is, or will be an evaluative creature) are both true. Combining E and F does not create a contradiction.

I would argue that it would be false to say that E & F (or E alone, for that matter) is good. For it to be good, there must be a creature with a reason to bring it about – an evaluating creature. However, this is not a logical requirement. It is a contingent fact about how value actually comes about. I can imagine – even if it is not real – an intrinsic value property attached to E alone and E & F combined that makes this combination logically possible.

However, this clearly does not entail a state in which somebody favors E & F. I already stated that we are imagining that value is an intrinsic property, and value as an intrinsic property does not imply an evaluator. Only value as a relational property between objects of evaluation and valuers requires a valuer, and this is not logically necessary. It is only metaphysically necessary.

So, “good” is not state-entailing.

And, if we can do without he evaluator, “good” is not belief entailing either.

I can agree that “S is good” is not state-entailing on the grounds that much of what we are concerned about in evaluating something as good concerns reasons for bringing it about – and bringing it about might not even be possible. For example, it would be good to be 30 years younger. However, my being 30 years younger does not obtain. So, “my being 30 years younger” is good does not imply “I am 30 years younger”.

To support Oddie’s claim that goodness is not belief-dependent, I can return to our village filled with people who have an aversion to pain. For them, a universal aversion to causing pain would be good – they certainly have reason to bring about such an aversion. However, it is good regardless of whether anybody in the community believes that this is the case. They may be totally in the dark concerning the merits or even the possibility of promoting an aversion to pain. Perhaps a malevolent demon has falsely informed them that condemning those who cause pain will bring divine wrath or bad luck. Yet, given the facts of the case (they have an aversion to pain and a reward system that makes it possible to promote an aversion to causing pain in others) this universal aversion to causing pain is good.

I am not certain that anything I wrote here makes sense of the original argument. I struggled with it. I have given it my best shot and this is what I came up with. Something can be good without anybody believing that it is good. Something can be good without anybody favoring it (though, perhaps, like “causing pain”, it may be something they should favor or, in this case, disfavor). Nothing can be good without somebody valuing something, but his is not a logical entailment. This is just how the universe works.

Yet, I am rejecting the claim, “S is good if and only if favoring S is fitting.” This makes sense for a certain kind of goodness, but not for all goodness. There is still the goodness that evolution, environment, and experience simply assigns to certain states, where there is no fittingness.

On Desire 2018. Part 09: The "Admirable" Demon

This commentary on Graham Oddie's paper is turning out longer than expected. Still, I have come to value the technique of creating commentaries.

For reference, so that you do not need to go hunting for it, the previous nine posts have all had to do with: Oddie, Graham (2017). "Desire and the Good: in search of the right fit". In Deonna J. & Lauria F. (eds.), The Nature of Desire. Oxford University Press.


In a universe apparently filled with demons, Oddie postulates that, “an evil demon threatens the world with some terrible outcome unless you admire him.” In this case, there is a sense in which you ought to admire him, but that the demon is not admirable.

This creates a problem for “Deontic FA”, which Oddie defined as:

(Deontic FA) X is V if and only if one ought to take attitude F(V) to X.

Because, here, one ought to admire the demon (to prevent the terrible outcome), but the demon is despicable. The right side of the biconditional is true, but the left side false, so the biconditional does not hold.
Oddie identifies a similar problem for “Axiological FA” where, “the demon threatens to bring about the worst outcome unless you desire that outcome,” thus, “it is clearly better for you to desire the worst outcome than not.” Yet, it is still the worst outcome.

To answer these problems, Oddie considers a type of response that comes from Olson (2009) and Ewing (1959) that suggests that there are multiple definitions of ‘ought’. It is like the claim that “Georgia is one of the United States” is true when talking about the region north of Florida, but false when talking about the country on the east side of the Black Sea bordering Russia. The biconditional does fail under the definition of “ought” that appears in the objection, but there is another definition where the biconditional still holds.

Ewing presented some additional detail by claiming that one sense of “ought” refers to what people generally have reason to condemn (they have reason to condemn the person who fails to admire the demon). He distinguishes this from the ‘ought’ that is fitting to admire. It is in the first sense that the biconditional is false, while it remains true in the second.

As Oddie argues, “Representational FA” does not have this problem since, regardless of the merits of what an agent ought to do or it would be good for the agent to value, it remains true both that the demon is not admirable, and that it is not representationally accurate for one to take the attitude of admiring the demon (though it may be prudent or even obligatory to do so).

However, I still do not know what “representational accuracy” is.

We could be working under an assumption that representational accuracy requires representing the admirable quality as an objective, intrinsic property of “deserving-admirationness”. This could make the most sense of how we use the term, but it could lead us straight into an error theory. All claims of admirability would then be false since we are representing things as having a property that nothing actually has.

The tension found in Deontic FA and Axiological FA would be minor compared to this error.

I am not saying that representational accuracy requires this and that we must reject Representational FA as a result. I am saying that this is one way it can go. Another alternative is that representational accuracy is found precisely in Deontic FA – that to accurately represent admirability one represents it in terms of what people ought to admire.

Furthermore, I do not see reason for concern in the responses from Olson and Ewing mentioned above. The fact that the word “Georgia” refers to both a state and a country may generate some confusion, but it does not provide a reason to prefer a theory of “Georgia” that holds that some propositions are true of “Georgia” in the one sense and false of “Georgia” in the other. That is not a problem – it is simply a fact about the language we have invented.

Given uncertainty over what “representational accuracy” consists in and that the ambiguity of a term like “ought” need not be much of a problem, I would like to look more closely at what Oddie called “Deontic FA”.

In the previous section I described a community containing individuals who all had an aversion to pain and a capacity to create in others an aversion to causing pain by using rewards such as praise and punishments such as condemnation. The members of this community have reason to call “admirable” those who go out of their way to avoid causing pain to others, and to call “deplorable” those who do not. These are terms of praise and condemnation and, as such, are useful in creating a community where individuals have this aversion to causing pain that people generally have reason to promote universally.

Let us add the admirable demon to this community. He apparently has a desire to be admired. To get what he wants he is threatening to harm others unless they admire him. Given that others have an aversion to pain, he threatens to cause others pain unless they admire him (and not necessarily limit that pain to those who do not admire him).

This demon does not have a trait that people generally have reason to promote universally. To admire this demon is to promote universally the trait of being willing to harm others unless he is admired. In fact, the agent (not the demon) in holding that such a trait is admirable would have to also believe that she herself should adopt this trait – that she should also be disposed to cause pain if she is not admired. The same can be said of all her neighbors.

At this point, I need to admit to a shift in what I have called “admirable”. In the original example, we were talking about admiring a demon who wishes to inflict pain if he was not admired. Here, I am talking about admiring a trait. More precisely, we can combine the two by saying that one is admiring a person in virtue of a trait. We cannot simply admire the demon. We must have a reason to admire him – something we admire him for.

The demon’s demand, if not carefully worded, would leave us with a loophole. While the demon is deplorable in virtue of his being willingness to inflict pain unless he is admired, perhaps he is also an extremely gifted painter who can be admired for what he can put on a canvas. Thomas Jefferson owned slaves and raped Sally Hemmings repeatedly – her non-consent being grounded both on the fact that she was a slave and, at the start of the affair, a young teenager. Insofar as this was true, Jefferson was not admirable. Yet, he may still be admired for his skill at eloquently presenting principles of and enlightenment.

In terms of examining Deontic FA, this would be cheating. The counter-example assumes that there is no other trait for which the demon can actually be admired. However, the fact that our admiration is focused on traits, and on individuals only in virtue of the fact that they exhibit admirable traits, does require that we specify what is going on with the demon.

With these considerations, I would offer an alternative to Deontic FA as follows:

(Deontic FA) X is V in virtue of X having trait T if and only if people generally ought to promote universally T by taking attitude F(V) to X in virtue of X having T.

To have a genuine counter-example to this version of Deontic FA, one would need a case in which the demon exhibited a trait that would not be counted as admirable, and yet for it to be a trait that people ought to promote universally by praising those who exhibited it and condemning those who did not. The demon’s trait of being disposed to cause pain unless he is admired is both despicable, and not a trait that people ought to promote universally. It is not a counter-example.

This defense of a form of Deontic FA does not defeat Representational FA. Recall that the objection raised against Representational FA concerned its lack of specificity when it came to cashing out “representational accuracy.” We can now cash out representational accuracy in terms of representing a person as having a trait that people generally ought to promote universally using praise and condemnation. The demon is deplorable, and it is representationally accurate to deplore the demon.

We can apply the same analysis to “delightful.” There are things we have reason to want people to take delight in – the laughter and the accomplishments of one’s children. There are things we do not want people to take delight in – the suffering and failure of one’s children. People generally have reason to encourage delight in some states and not in others.

At the same time, people sometimes use the term to refer to things that people do not have reason to promote delight in – a delightful meal or concert where there is no fault in others who not only take no delight in the but find them horrible. The use of “delightful” in these cases generally represents an error. In some cases, it may be an exaggerated compliment, “This is so good that those who do not delight in it are somehow defective.” In some cases, it is snobbery and prejudice, “Though there is no reason to promote a delight in this universally, those who do not delight in it are inferior beings – defective in some way.” These uses of the term do not obligate us to come up with a theory in which these uses report facts.

Oddie ends his discussion by drawing some lessons for the theory of the good.

This delivers a constraint on fitting attitudes (namely that they be capable of being representationally accurate) that will narrow the range and nature of the fitting responses to evaluative attitudes in general and to the thin evaluative attribute of goodness. The fitting response to a state’s being good must be a presentation of that state as good.

I have not given any reason to reject Oddie’s analysis of goods for which there is a fitting response – a response that people generally have reason to encourage universally. However, I am including under the concept of “goodness” those states that fulfill desires – the aversion to pain, hunger, thirst, certain food preferences, basic environmental comfort (temperature preferences), and the like – that evolution, environment, and experience have planted in our brains. In fact, I am using these desires as the foundation for the fittingness of such things as the aversion to causing pain.

, and that evolution and experience has planted in the brain. am using a broader definition of desire, and of good, than Oddie. I am including as desired the value chips planted in our brains by evolution and the regular course of biological development. In my description of the community of individuals with an aversion to pain, the evolutionarily acquired brain chip of aversion to pain, the admirability of honesty, and the delightfulness of a child’s achievements.

In the sample community seeded initially with people who have an evolutionarily acquired aversion to pain and a reward system, these are what make the aversion to causing others pain admirable. Without the evolutionarily acquired aversion to pain, promoting admiration of those who avoid causing pain would be pointless. Without a reward system, it would be useless.

On Desire 2018. Part 08: Desirable, Admirable, and Delightful

Continuing my series on Oddie, Graham (2017). "Desire and the Good: in search of the right fit." In Deonna J. & Lauria F. (eds.), The Nature of Desire. Oxford University Press.

Oddie’s key focus in his theory of "desire" is on the idea of “fitness”.

This will require that we look more closely at the idea of fitness.

The fitting attitude account tells us that the delightful is not just what people happen to take delight in or what people typically take delight in, but in what it is fitting to delight in.

Oddie provides us with three conceptions of this which he generated by “commandeering” a similar topology presented by Toppolet (2011):

(Deontic FA) X is V if and only if one ought to take attitude F(V) to X.

(Axiological FA) X is V if and only if it is good to take attitude F(V) to X.

(Representational FA) X is V if and only if it is representationally accurate for one ought to take attitude F(V) to X.

In these sentences, FA stands for a fitting attitude such as admirable. We can read Deontic FA as saying, for example:

Honesty is admirable if and only if one ought to take an attitude ‘admiring’ to honesty.

Oddie does not specify any preference for either of these three formulations here. He saves that task for when he discusses objections, where he sides with the representational view. I fear that I am going to have problems with the representational view because I do not know how to cash out the phrase “representationally accurate.”

At this point, he seeks only to specify the options. I will do the same. However, I want to say a bit more about these formulations as seen from the assignments perspective.

Again, there is nothing here that we can see as an objection – just a clarification.

For illustrative purposes, allow me to take Deontic FA.

The initial examination of the assignments theory of desire would seem to suggest that honesty is admirable if and only if people admire it. However, this is clearly problematic. There have clearly been cases in the past where at least some people have admired cruelty or ruthlessness in getting what one wants, yet that did not make these admirable qualities. However, there is a way of getting something that fits more closely to what Oddie has in mind out of the assignments theory.

Consider a hypothetical community whose members have an evolved aversion to pain; evolution planted wiring in their brains such as to assign a negative value to states of being in pain.

Let us further assume that the beings in this community have what we may call a “mesolimbic pathway” – a reward system. By means of reward and punishment (including praise and condemnation) inflicted on agents, people can create in those agents an aversion to causing pain. In addition, people can acquire these aversions by observing others being rewarded or punished, or even hearing about them in a story where those who avoided causing pain were praised and those who did not were vilified.

Now, we have this fact: People generally have a reason to promote in others universally an aversion to causing pain. Nobody at the start of this community has an aversion to causing pain. However, this does not prevent it from being trued that they have a reason to create such an aversion. In fact, it may be the case that nobody has even yet figured out, "You know, if we all were to reward and praise those who refrain from causing pain, and punish and condemn those who do not refrain, we can promote universally an aversion to causing pain." Yet, this will not change the truth of the claim.

We can understand honesty in this way.

Honesty is admirable if and only if people generally have on balance many and strong reasons to promote an attitude of ‘admiring’ to honesty.

For the same types of reasons that the people in my hypothetical community have for promoting an aversion to causing pain, we can make an argument that people generally have many and strong reasons to promote, universally, an admiration for honesty. In this sense, honesty ought to be admired, even where the fundamental desires are evolutionary designed “brain chips” that assign values to such things as avoiding pain, caring for one’s offspring, hunger, thirst, sex, and environmental comfort.

This gives me a way of making sense of the claim I made in the previous section that maintaining a fiction of independent goods is a bad thing. People generally have many and strong reasons to promote, universally, an aversion to promoting the fiction of independent goods. The reasons come from the fact that people get criticized for having brains that make value assignments different from those of others. Recall that the problem is that those others see these value assignments as reliable indicators of a type of goodness that does not exist. This is true even though many of those reasons come from the “brain chips” evolution and experience has planted in people’s brains.

Nothing here so far provides a reason to Oddie’s fittingness thesis. It provides a useful analysis of terms like “admirable”.

Oddie does go on to consider three objections. The first of these will lead him to accept "Representational FA" as defined above. I wish to consider his objections and suggest a problem with Representational FA in the next section.

On Desire 2018. Part 07: Radioman and Radiowoman

The story of Radioman and Radiowoman is used to argue for an evaluation theory of desire as opposed to a disposition-to-act theory of desire.

An evil demon has given Radioman a disposition to act. By means of a chip planted in Radioman's head, every time he walks near a radio he will reach out and turn it on. It is something like a habit – something the agent simply does in a given circumstance. A radio is within reach, out goes his hand, on goes the radio. If we ask him if he wants the radio to be on, he would say he does not. He has no interest in the radio being on and, in fact, he hopes that there is no radio nearby so that he will not end up turning it on.

Radioman has no desire to turn on radios.

The purpose of this example is to illustrate that our concept of desire is not a concept that attaches itself merely to a disposition to act. Here is a disposition to act, but it is not associated with a desire, so the concept of a desire points someplace else. It is a mistake to associate the concept of a desire with a mere disposition to act.

In this, the thought experiment does its job. Whatever desires are, Radioman does not have one.

Radiowoman represents the evaluative theory of desire. She also reaches out and turns on a radio when one is near. However, she does so because she sees the radio being on as something good.

She has the same behavioral disposition as Radioman, but this is because the radio's being on seems good to her, she feels drawn to the prospect, it is alluring. And when she hears the radio come on she feels satisfied by that.

Radiowoman, in contrast to Radioman, has a desire to turn on radios. Thus, demonstrating the merits of the evaluative theory of desire.

Now, we are asked to consider what would be the case if Radiowoman discovered that she has these sentiments about turning on radios because a demon put a chip in her head.

Suppose Radiowoman were to find out about the etiology of her desires. Then she would know that they are not reliable indicators of goodness. Rather, they are systematic illusion of goodness. They are like the Mueller-Lyer illusions that, once you know about them, give you no reason at all to believe that the lines that appear unequally really are unequal. And even if she knows nothing of the peculiar etiology of her desire, Radiowoman’s desires are defective.

This is one way to look at it, but I think it runs into problems.

The aversion to pain can quite accurately be described as a chip in the head foisted on me, not by an evil demon, but by evolution. The wiring that causes me to assign negative values to certain sensations caused by damage to my body was not an intelligent designer.

Rather, the wiring came into existence because that wiring kept my ancestors alive and helped them to produce viable offspring. Random mutation, natural selection, and luck dictated the specifics of the wiring. None of this requires any mention of an external 'good' (or 'bad') of which the pain is a reliable indicator. The only thing it is a reliable indicator of is that which was a part of an evolutionary package that caused my ancestors to have viable offspring.

It does not follow from this that, upon recognizing this fact, and denying the existence of any type of 'good' for this aversion to pain being a reliable indicator of, that I lose my reasons to avoid pain. The aversion to pain – the awfulness (the negative value) assigned to states of affairs in which the proposition 'I am in pain' is true – simply is a reason to avoid pain all by itself.

My hunger and thirst have a similar etiology. Not only did evolution plant in my brain hunger and thirst chips, those chips are programmed with preferences for those kinds of food that helped keep my ancestors alive. That was a function of their environment – which foods were healthy, which were poisonous, which provided enough calories to survive, and which provided other necessary nutrients.

A different evolutionary history would have resulted in different tastes. However, knowledge of this etiology and that there is no “good” out there for it to be a reliable indicator of does not make a pumpkin pie with Cool Whip heaped on top lose any less delicious or remove my reason to eat a slice if I can.

One can still postulate an external good for these evolution-designed brain chips to be reliable indicators of. The problem is that they are not needed. As Street (2005) argued, the best scientific theory scientific theory has evolution modifying these brain chips – modifying their assignments of value – using only random mutation and natural selection.

Whether there is an external good for our desires to be a reliable indicator of is an important philosophical question. However, like some other philosophical questions such as the existence of God or free will, some people strongly desire that the “reliable indicators of the good” hypothesis is true. Radiowoman sounds like a person with a particularly strong desire that her desires track some sort of external good. If this is true, she may be very upset to discover that this is not the case – as upset as others are when confronting arguments against the existence of God or of free will.

At the same time, maintaining the idea that our desires are reliable indicators of some external good has its own undesirable consequences. Some people get the idea that their value assignments match up with some external good. From this, they infer that those who do not perceive this goodness are defective. They denigrate such people, calling them “sick” or “perverse,” and dismiss their interests as concerns that cannot only be ignored (since they are interested in no real external good), but intentionally frustrated (since, in their perversion, they are motivated to realize that which is bad). Such has been the fate, for example, oh homosexuals whose brains simply attach positive value to same-sex relationships. This is a problem.

At this point, somebody may be tempted to accuse me of an inconsistency. In the previous paragraph, I used the term “undesirable” as in “bad to desire.” A critic may point that this seems inconsistent with the idea that there is no truth for desires to reliably refer to – no “ought to be desiredness” in the universe for our desires to track. If this criticism was sound, this would be a problem.

I will look at that question next.

On Desire 2018. Part 06: Desiderata of a Theory of Desire

Graham Oddie claims that there are three things to ask for from a theory of desire. He argues that such a theory must be non-belief entailing, desire entailing, and ubiquitous.

See: Oddie, Graham (2017). Desire and the Good: in search of the right fit. In Deonna J. & Lauria F. (eds.), The Nature of Desire. Oxford University Press.

Not Belief-Entailing

One of the “desiderata” that Oddie defends for desires I’d that they not be belief-entailing. One of his arguments for this is that, analytically, it simply seems to be the case that it is possible to believe that something can appear good without being good. In an analogy to other types of perception, he compares this to the way in which something can appear round but not be round or appear red and not be red.

The assignment thesis would meet this standard. In fact, I would argue that we have no special access to our own desires. Instead, we theorize as much about our own desires as we do the desires of others. Consequently, we can be mistaken about what we desire – about what our brains assign value to – just as we can be mistaken about what other people desire.
Granted, we have much more information about our own desires. We have not only a long and continuous history of observations to draw upon, but we also have access to additional information such as the emotions stirred within us upon contemplating and realizing certain states of affairs. Consequently, we usually know our own desires better than we know the desires of others. But that knowledge is not without error.

At the same time, our awareness of our own desires can be muddled by the fact that they are ours. There are certain desires that we do not want to have, and others that we do. This encourages us to favor interpretations of our own conduct that allows us to assign the best possible motives to ourselves and blind us to our faults.

This gives us both sides of Oddie’s non-belief entailing claim. Our brains can be assigning values to states of affairs without our believing that they are good, and we can believe that something is good even though our brains assign no value to it.

Desire-Entailing

To be desire-entailing it would have to be the case assigning a value V to the realization of proposition P entails desiring that P. There would not be a case in which the brain assigned a value to the proposition ‘P’ being true where the agent did not have a desire that ‘P’. Nor would there be a case in which the agent had a desire that P where the brain did not assign a value to ‘P’ being true.

Of course, this is exactly what needs to be shown. In looking at the various objections that Oddie addresses to such theories, I hope to bring out this fact.

What is needed here is an example of some state of affairs that meets the conditions described here that would not fit the common-sense account of an agent having a desire. My inability to think of one can just be my inability to think of the account I am defending being false.

Though, even if somebody finds one, I still have a response. I could counter with the claim that the counter-example includes a false assumption and that we would be better off modifying our concepts than rejecting the thesis. This would be a reductionist/revisionist response.

Ubiquitous

We have evolved with dispositions to assign value to many things that sustain life and promote evolutionary fitness. From the aversion to pain to hunger and thirst, to what tastes good and tastes bad, to the comfortable of a room, to the company of friends, to the desire for sex and concern for our offspring, assigned values to ends govern our lives.

We also have a reward system whereby we acquire new desires and aversions, and mold existing ones, based on our experiences. This is how we come to like or dislike philosophy, Jazz, poetry, and pushpin. Here is where we learn our prejudices and, I would argue, many of our moral sentiments such as a disapproval of slavery and a fondness for virtue.
All of these involve assigning values to ends.

Summary

When it comes to meeting Oddie’s desiderata for a theory of desire, the assignment theory can hold its own. It is not belief entailing, since the fact that the brain has assigned a value to a proposition ‘P’ being true implies nothing about what the agent believes. It is desire-entailing directly from the fact that if the mind attaches a value to a proposition ‘P’ being true does imply that the agent desires that P. And it is ubiquitous since the whole of our intentional behavior is directed to realizing propositions ‘P’ to which our minds have assigned value.

On Desire 2018. Part 05: Objects of Evaluation

I have argued that desires are assignings of value.

Assignings of value to what?

Oddie considers two options: propositions vs. states of affairs – thought they are closely related.

See: Oddie, Graham (2017). Desire and the Good: in search of the right fit. In Deonna J. & Lauria F. (eds.), The Nature of Desire. Oxford University Press.

Despite the apparent diversity of types of widely presumed that the objects of desire, like the objects of belief, all hale from some uniform ontological category. And the prevailing view is the objects of desire (and of belief) are propositions, or closely related entities like states of affairs.

Yet, he asserts that the difference between these are not relevant to his thesis, but chooses states as the ultimate object of evaluation.

I am inclined to go along with this. More specifically, I would argue that a “desire that P” assigns value to a state of affairs in virtue of ‘P’ being true. Thus, a person with an aversion to pain (a desire that I not be in pain) attaches a value to any state of affairs in which ‘P’ (I am not in pain) is true.

However, there is a point that seems a bit confusing that I would like to clarify. Oddie writes, “Whenever a desire seems directed at something non-propositional—like a hokey-pokey ice-cream, Kyoto, or happiness—what makes it true that one wants this or that is that one wants to stand in some appropriate relation to this or that.”

This is certainly true for a large and important set of desires. Yet, this seems to be saying that this is true of all apparently non-propositional desires. On that interpretation, the statement would not be true. A person who cares about another person, a species, or an ecosystem often has no interest in her relationship to that object of evaluation. Rather, she desires that the person is well, that the species not go extinct, or that the ecosystem persists in its current state.

This is not an objection to Oddie’s thesis. What matters is that there some proposition that serves as the object of the desire. THAT the person is healthy and happy, THAT the species not go extinct, and THAT the ecosystem be preserved in something near its current state are adequate propositions for the propositional account. These go along with desires such as a desire THAT I am eating hokey-pokey ice cream, THAT I am in Kyoto, or THAT I am happy.

Ultimately, I would argue that the desire assigns a value to some proposition being true. That proposition could (and quite often does) describe a relationship between the agent and something else (that I am eating hokey-pokey ice cream). It could simply describe some state of affairs without referring to the agent (that Antarctica be preserved). The desire motivates the agent to make or keep the proposition true.

What else does a good theory of desire? Oddie lists three desiderata. I will turn to those three next.

On Desire 2018. Part 04: Perceiving vs Assigning Value

What is a desire?

According to Graham Oddie (2017):

Desire = The fitting response to the good.

On the way to this destination, Oddie defends an evaluative theory of desire – whereby to desire something is to evaluate it as being good. This, he argues, has two main options:

There are two possibilities within this approach: that desires are value judgments (doxastic value seemings) and that desires are value appearances (non-doxastic value seemings). I defend the second of these, the value appearance thesis. To desire something is for it to appear, in some way or other, good.

Of these, Oddie defends the second option.

To Oddie's account, I want to divide this second option into two sub-options.

Desires are value-perceivings:. A desire that P perceives ‘P’ being true a value as an end in itself, which motivates the agent to act so as to realize states in which ‘P’ is true.

Desires are value-assignings. A desire that P assigns a value as an end to ‘P’ being true and, thereby, motivates the agent to act so as to realize states in which ‘P’ is true.

I am going to be seeking to defend the second option. To do this, I would like to run the assignings option through the same obstacle course that Oddie set up for the appearance option generally to show that it can handle these problems, as well as list some additional advantages for the former.

I do not want this thesis to be confused with the theory of projectivism. Projectivism holds that we project qualities onto an object as if the object has those qualities. This assumes that the projection is some type of error or mistake.

The situation is more like that which happens when you are in a car that is supposedly stopped. You see a vehicle in the next lane move backwards relative to your car and, for a moment, you are confused. Both descriptions are equally accurate, and the situation can only be resolved by calculating movement relative to some third thing, such as the Earth.

Similarly, when something appears to have value all we see is that it has value.

All we perceive is that something has value. We do not see that it has value as an intrinsic property, nor do we see that it has value in virtue of it being assigned. If the value seems to be an intrinsic property, it is because the perceiver is adding assumptions that are not found in the appearance. To work out these answers, we must look at other facts to determine which explanation makes the most sense.

So, something appears to have value. Is this value perceived or assigned?

I find it difficult to match the perceivings options to my understanding of how the rest of the world works. It seems to require that there is a good independent of desire that is either known through reason alone or seen directly. It is a property of things that commands that people have a particular reaction (motivation) towards that good. J.L. Mackie (1977) labeled this “objective, intrinsic prescriptivity” and argued that it is such a strange entity that there is reason to doubt that there is such a thing.

Sharon Street (2005) adds strength to Mackie’s objection by pointing out that that the theory of evolution creates problems for any theory that places values “out there” to be known about or perceived. Evolution has molded our motives to whatever produces biological fitness. Our biological ancestors had more to gain by being disposed to act in ways that produced evolutionary fitness than to correctly perceive some good independent of fitness.

The assignment view would have it be the case that evolutionary pressures selected assignings of value that promoted genetic fitness. Our aversion to pain evolved to cause is to avoid states of affairs that would result in the types of injuries that would have prevented our ancestors from having and raising children. The desires for sex, for food and drink, our food preferences, all came about because the random genetic mutations and survival of the fittest selected some candidates over others.

If our genetic history had been different, if we had evolved in different environments, and if the fortunes of fate had worked out differently, we would have evolved into creatures that assign values differently.

We see evidence of this in the fact that animals, too, assign values to certain states. A dog’s aversion to pain and a cat’s desire to hunt and catch anything that flitters about are examples of natural selection creating brains that assign value to these ends. Each preference that we find in nature is a preference that we could have had if we had evolved along the same lines.

Oddie will confront the issue of animals and infants having desires later in his paper, so we will get back to this issue.
Of course, not all assignings of value are genetic. In fact, among humans in particular, many are learned.
We know some things about how values are learned and they, too, seem to support the assignings view over the perception view. Rewards and punishments (including praise and condemnation) act on the mesolimbic pathways in the brain – the reward system – to encode rules of behavior in, primarily, the pre-frontal cortex. Our experiences cause us to like that which produces a positive reward, and to form aversions to that which produces a negative reward or, in psychological terms, a punishment.

In effect, this system takes as input the experiences of the agent and produces as output alterations in the value assignments that the brain makes, thus altering behavior.

In this way, a person can come to like philosophy or opera, can hate sushi or card games. The different experiences of different individuals results in different assignings of value for different agents.

In contrast, the perceivings theory seems to require that different agents should all perceive the same value in things – the value that it actually has. Those who do not perceive something correctly has a perversion or some other sort of defect. Mackie (1977) argued that the different assignments among individuals and among cultures is evidence against objective intrinsic prescriptivity. Though there are ways in which it can allow two different people can assign different values to the same thing and neither be wrong, these are far more complex than the simple explanation of the assignings view – that people with different histories (evolutionary and experience) simply come to assign different values to things.

These give reasons to support the assignings view over the perceptions view. The next question that I want to address is: assign value to what?


Mackie, J. (1981). Ethics. Inventing Right and Wrong. Penguin Press.

Oddie, Graham (2017). Desire and the Good: in search of the right fit. In Deonna J. & Lauria F. (eds.), The Nature of Desire. Oxford University Press.

Street, Sharon (2005). A Darwinian dilemma for realist theories of value. Philosophical Studies 127 (1):109-166.