On Desire 2018. Part 57: Moore's Paradox of Beliefs

Since the beginning, I have argued that nothing has intrinsic value, not even desire fulfillment.

One of the ways in which I have expressed this fact is to argue that a desire that P gives one a reason to make or keep the proposition 'P' true. However, there is nothing in the fact that "agent has a desire that P and 'P' is true" that automatically generates a reason to create this state of affairs. In a world without desires, there is no reason to bring desire fulfillment into existence. It is only in a world that has desires that the states of affairs that would fulfill those desires have value.

Is this all too confusing?

Here is my standard test case. Alph has a desire that the planet Pandora B into existence. He stands before the button. If he presses the button, the planet Pandora B will exist. He has a reason to press the button.

However, is there a reason to bring into existence a world in which Alph, with his desire that the planet Pandora B exist, himself be brought into existence and set before the button? The answer is: no. If there is no desire that would be fulfilled by this state of affairs, then there exists no reason bring that state of affairs into existence.

David Wall may be trying to bring up an argument against this position. Honestly, I cannot tell at the start. I find his argument a bit perplexing. However, he seems to be arguing that there may be things that a person who merely possesses the concept of "desire" has reason to bring into existence (or to prevent bringing into existence) itself, without evaluating its relationship to another desire. Some desires are just irrational.

Wall, David, (2017), “Desiderative Inconsistency, Moore's Paradox, and Norms of Desire,” In Deonna J. & Lauria F. (eds). The Nature of Desire. Oxford University Press.

Wall is drawing this argument out of Moore's Paradox concerning beliefs.

Moore's Paradox concerns the following propositional attitude. "I believe that (a) it is raining, and (b) I believe it is not raining."

Does it make any sense for somebody to believe this?

Wall grants that there are states in which something like this can be true. That is to say, clearly it has happened many times - that I have confined myself to the basement to work on a blog posting, convinced that there would be no rain today, to have my wife tell me that a storm moved through town within the past hour. So, the proposition: "It is raining and I believe it is not raining" can be true. We can imagine a person having this belief. However, we also have to imagine this person saying to himself, "Hey, something is not right here."

Wall uses this to argue that there is a particular norm of belief - to not have false beliefs. The reason this type of case is odd is because, no matter what one believes, one has a false belief. If one believes the conjunction, then the belief that it is not raining is false. If the belief that it is not raining is true, then the larger set, the "belief that it is raining and I believe it is not raining" is false. The prohibition on having false beliefs means that there is no way to accept this belief.

Ultimately, Wall will use this to argue that there is a similar norm not to have frustrated desires. I will explore that option in a future post. However, I have one more thing to say on this matter of a norm regarding beliefs.

Wall recognizes that there are two ways to get rid of a false belief. If an agent believes that P, and P is false, then one can change one's belief. The other option is that if one believes that P, and P is false, then one can make P true. Either way, one now now longer has a false belief. Wall argues that his proposed norm of belief leaves both of these open.

I am wondering why Wall selects this particular norm. The direction of fit model suggests that the proper norm when it comes to beliefs is that, if one believes that P, and P is discovered to be false, then one ought to change one's belief. It is not that one ought to bring the belief and the world into agreement. If we are going to introduce a norm that makes sense of our use of belief, then we should introduce a norm that actually makes sense of our beliefs.

This, of course, is a series on desires, rather than beliefs. Wall is going to argue that there is a norm of desires as well as a norm of beliefs - a norm that one ought not to have frustrated beliefs. This will imply that there are certain propositions P that "ought not to be desired" regardless of any other desires. Specifically, one ought not to have a desire that will end up being frustrated just as one ought not to have a belief that ends up being false. Yet, this will imply an intrinsic badness to be found in frustrated desires. I do not think that this is the case. We will look at that in the next few posts.

Desires 2008: Summary 01: Daniel Friedrich

One of the things that I want to accomplish in my study of desire is to present a quick summary of a main point in each of the articles.

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

Thesis:

Agent desires that P if and only if Agent has a felt need that P be the case.

Friedrich does not actually put his thesis in these terms, but his writing suggests that this is what he is seeking to defend or something close to this.

In desirous experience, the desired end is given to the mind under a feeling tone of felt need; it is given to the mind as something that must be realized.

Friedrich distinguishes felt needs from biological needs in that felt need is an experience and biological need has to do with avoiding harm. Furthermore, he claims that felt needs have different sense of urgency or intensity, and that, even though one can express the felt need in words (e.g., “this must be the case”), words are often inadequate. Finally, a felt need that something must be the case is not a prediction; one can have a felt need that P even though one knows that P will never, in fact, obtain.

Question:

If somebody were point to a bunch of things and call them “red,” I may not have the same phenomenological experience of redness as that person. However, I can see what they all have in common. Using that, I can then predict what else that person may call “red”. I can be confident, for example, that that he will say that my heavy winter coat is a red coat.

However, when that person points to a bunch of things and calls them “desired,” what do I see that they all have in common that I can use to predict what else might be desired? I do not sense a felt need that they all be the case – not unless I desired all of the same things that the speaker desired. That is unlikely, given the number of desires that are self-referential (e.g., one’s aversion to one’s own pain). I have to find something else that they all have in common.

What the all have in common is the agent’s disposition to realize (make real) those things that the agent calls “desired.” “P is something that I desire” means “I am willing to work to make it the case that ‘P’ is true.”

For the sake of argument, let us speculate that we have discovered creatures on an alien planet. Do these creatures have desires? I believe that I can begin to hypothesize desires simply by discovering that they seem to give some importance to realizing certain states of affairs. It would be rather rash of me to assume that they have something called a “felt need”, but I can observe that they put effort into such things as building nest or caring for their young.

From this, I can recognize that there are also things that I am willing to work to realize, so I get an idea that these things are properly called, “things that I desire”. Here, I may recognize that I have this “felt need” to realize all of the things that I desire. Then, I can infer that he other agent might also have this same “felt need” for all of the things he desires. However, it seems that “felt need” enters the picture rather late in the game. Even then, it contains a bold and unsupported inference that my own “felt need” justifies the claim that others also have this “felt need”. This inference only makes sense if there is some other evidence that requires that I postulate a ”felt need” in others, above and beyond their willingness to work to realize P.

Indeed, the problem of determining that desire fells the same to somebody else as it does to me is substantially the same problem as that of determining whether red looks the same way to somebody else as it does to me. I don’t know how to do that.

Here, it may be thought that this is favoring a dispositional theory of desire – a desire that P is a disposition to realize P. That is an option, but it is not the only option. A desire may be the disposition to realize some end, or it may be the cause of the disposition. If this cause of the disposition includes a “felt need” that P be the case, this is, at best, a contingent fact.

I hold that a desire is an assignment of a value (importance) of a proposition being made or kept true. This manifests itself as a willingness to (disposition to) work to make or keep P true to a degree proportional to the strength of the desire (the greatness of the importance). It may accompany some type of "felt need", but it is not required.

On Desires 2018 - Supplemental: "Actual Value"

The top of my list of things to dislike about academic philosophy is . . .

They choose a language that gets in the way of philosophers talking to real people.

For the past few postings as I have discussed Peter Railton's thesis that desires are like beliefs about the actual value of things I have had to make the claim that "there are no actual values".

This phrase, "there are no actual values," means one thing in the language of philosophers and something quite different among real people. Among real people, saying this is like saying that nothing has real value. It is like telling people, "You might think that your child's health, service to your community, or your own freedom from pain has value, but you are wrong. Nothing has value. Everything is worthless. Literally, everything is worth nothing because there is no such thing as value."

Which, by the way, is not at all true.

Put your hand in a bed of hot coals and tell me that the pain that results doesn't matter.

In writing this post for real people rather than philosophers, let me say that value is quite real. Things really do matter.

What I am saying against Railton is not that "nothing matters" but that "things matter in virtue of the fact that we value them, and not for any other reason." That is to say, pain is bad because we hate pain. The well-being of our children is important because it matters to us. The suffering of people from poverty and disease or from injustice and violence is real suffering. You cannot say, accurately, "nothing bad has ever happened in the world and nothing good can happen." Of course suffering exists. Of course good things are possible. Suffering exists and good things can happen because we are people who care.

So, who is it that decided that "value" that exists independent of desire is "actual value," while value that exists in virtue of the fact that we care about things is something other than "actual value"?

In all honesty, the terminology may have well been adopted by people who wanted their theory of desire-independent value to appear in a positive light and to give opponents to their theory a label that would instantly make it unpopular - as something that nobody would want to accept. We see this tactic used constantly in politics. It is the difference between an "inheritance tax" and a "death tax". Between an "undocumented immigrant" and an "illegal immigrant". Philosophers, after all, are human beings and are just as prone to be protective of their favorite ideas and denigrating of the ideas of others as real people.

Be that as it may, I have regrets over having used the term "actual value" in the posts that I have just written on Railton, and wanted to make a note of this issue. I do not want to be taken as saying that desire-dependent value isn't real. Desires are real, and the values that depend on desires are just as real. We really do care. There are things that are important to us and things that are important for us to avoid. All of this is just as real, just as "actual" as any other real and actual thing we find in the world.

On Desire 2018. Part 56: Desires Near and Far

In the last few posts, I have been examining Peter Railton's claim that beliefs aim at actual value in the same way that beliefs aim at actual truth. There is a fact of the matter concerning value, and the desire system collects evidence that seeks to point desire to actual value just as evidence point belief to truth.

Railton, Peter, (2017), “Learning as an Inherent Dynamic of Belief and Desire,” In Deonna J. & Lauria F. (eds). The Nature of Desire. Oxford University Press.

I have gone over four types of cases to show that, contra Railton, there is no "actual value" involved.

These are:

(1) Cases of disappointment - which I described as cases where one discovers that what one desires is not to be found in a state of affairs, and not a case in which one's desire was mistaken.
(2) Malleable desires - where I argued that the fact that a desire can be caused to change does not imply that it is approaching or getting further away from actual value.
(3) Fading desire. Our desire for what is familiar is not the recognizing of actual value resulting from more evidence, but a way the brain has to getting our attention focused on what needs changing.
(4) The utility of desire: The fact that desires can be changed invites see desires as things we may have reason to change, but the instrumental value of a desire does not represent a type of "actual value" that Railton has in mind.

Here, I want to discuss a fifth and final case - the case of near and far desires.

In another quote from Hume, Railton describes the fact that things that are more distant in time and space do not influence our actions as that which is near.

Now as every thing, that is contiguious to us, either in space or time, strikes upon us with such an idea, it has a proportional effect on the will and passions, and commonly operates with more force than any object, that lies in a more distant and obscure light. Tho' we may be fully convinc'd, that the latter object exceeds the former, we are not able to regulate our actions by this judgment, but yield ot the solicitations of our passions, which always plead in favor of whatever is near and contiguous.

There are two types of desires that clearly have no influence on current actions. These are desires that do not exist yet - future desires, and desires that one wishes one had.

In the first case, we must distinguish between a future desire and a present desire for a particular future state. The latter exists, but it is one desire among many - sitting beside all other present desires such as hunger, thirst, and an interest in watching football games on television. As one desire among many, it does tend to be outweighed.

Of course, wishing that one valued exercise. We may well imagine the case where a person is aware of the benefits of exercise and obtains some motivation as a result of these facts, that same agent may despise exercise (all of that effort is just uncomfortable). He may wish that he valued exercise instead. Similarly, one may wish that one disliked chocolate the way one dislikes broccoli, and like broccoli the way one likes chocholate, the desires one has will motivate one's actions, not the desires that one wishes to have.

Closely related to this desire one wishes one had is an interest in that which has actual value. A person may have no desire at all for something that he believes has actual value. However, actual value does not exist, and the desire that one does not have cannot motivate the agent.

However, more to the point, I have mentioned that, while a belief by itself can be wrong (a "belief that p" is false if "p" is false), a desire cannot be wrong. A desire can only conflict with another desire. When a conflict occurs, there is no outside reality to appeal to in determining which is correct. The Hume quote above speaks of being convinced that the value of the distant object exceeds that of the nearer, but the individual pursues the nearer object instead. However, the question still needs to be asked (and answered): What makes the distant object more valuable?

An individual may be convinced that piety is the greatest value, and that this demands a severe form of asceticism. However, the individual finds himself giving in to nearer and more immediate goods, living a live tinged by a bit of guilt, but one in which he does well and does well to others. It does not follow from the fact that a desire is for something more remote and distant that it is for something with a higher actual value. And to take the agent's conviction that something has a higher actual value as proof is to beg the question in this case. I am not denying the possibility of believing that something has an actual value, I am denying that this belief is true.

Another example is illustrated in the movie Mr. Holland's Opus. Glenn Holland takes a job as a high school music teacher to pay the bills as he works on writing a great symphony. He never gets to write that symphony, as his duties as a husband, father, and responsible teacher get in the way. Yet, as illustrated in the movie, his constant yielding to nearer and more immediate concerns proved to have underappreciated value.

In all of this, the question is: How does one determine which desire is linked to the correct value? The fact is: there is no correct value. If two values (or sets of values) come into conflict - whether they be nearer values to more distant values, or different near values, or different distant values, there is no way to say, "this is the correct value" by appeal to the "actual value" of what the desire is for. The only answer is to be found in the reasons there are for promoting or inhibiting that interest - reasons that are found in other desires which it may either fulfill or thwart respectively.

On Desire 2018. Part 55: Fading Desires and the Utility of Desires

A desire is an attitude that aims at the actual value of its object. At least, this is the case according to Peter Railton. (Railton, Peter, (2017), “Learning as an Inherent Dynamic of Belief and Desire,” In Deonna J. & Lauria F. (eds). The Nature of Desire. Oxford University Press.)

The problem is that there is no “actual value.” It does not exist. To show this, I have been examining Railton’s examples of desires chasing actual values to show that there is no actual value.

In our last exciting episode, we looked at the case of a woman who desired to be a partner in a big-name law firm who finds the actual experience to be disappointing. Her desire to be a partner then diminishes to match the actual value of being a lawyer. In fact, she did not desire to be a partner in a big name law firm. She desired something else and believed that being a partner would have those qualities. I compared this to taking a drink out of a cup that I thought contained hot chocolate when it actually contained coffee. It is a mistake to think of this as a desire to drink what was in the cup being corrected by my discovery of the true value of the taste of what was in the cup. Instead, I desired hot chocolate and discovered that the cup did not contain hot chocolate. There was no “actual value”.

I also looked at the case of desire modification through experience - acquiring a fear of bees because of a bee sting. The fact that desires change does not imply that one is discovering an actual value. Gaining weight is a change, but can hardly be described at discovering my true weight. The same is true with aging. Experience adds likes and dislikes as it adds physical scars and develops habits. We can explain all of this without mentioning actual values. A person can change so that his assignment of importance to p being true is increased or decrease without it being the case that the increase or decrease is more or less correct than the original value. Things are different with belief, where a change in credence actually does bring the belief closer to or further away from the truth.

Railton mentions three more cases - two of which I will discuss below.

Fading Desire

There is the case of the fading desire. Railton presents a quote from David Hume that says:

’Tis a quality observable in human nature, and which we shall endeavour to explain afterwards, that every thing, which is often presented, and to which we have been long accustom’d, loses its value in our eyes, and is in a little time despis’d and neglected.

Yes, our desire for that which is familiar tends to fade in many cases. However, it is a mistake to conceive of this in terms of discovering through familiarity that it has less actual value. We are hungry. As we eat, our desire to continue eating diminishes. It fades. Yet, it would be odd at best to describe this change in one's desire to eat as discovering the true value of eating. It is not the case that, when we are hungry, we valued eating above and beyond its actual value - an actual value that we discover as we eat until we realize the truth of the matter, which is that eating never had any value and our original desire to eat was mistaken. Instead, there is no "actual value" of eating that our desire to eat tries to discover.

We seem to be built so that we ignore that which is common, so that we can focus on that which is different. A sound that is always in the background, such as a fan, the wind blowing through the trees, the waves of the ocean or the babbling of a nearby stream, fade into the background of our consciousness so that we can focus instead on any new sounds - sounds that can tell us of a change in our situation. In our earlier evolutionary history, it could be the sound of a nearby predator, or the sound of potential prey.

Similarly, evolution may well have designed us so that once a desire is fulfilled it fades away, allowing us to focus our attention on that which needs changing. As we eat, our desire to eat fades so that, when we have enough food, we can quit eating and focus our attention on the next thing on the list, which may be to discover what is on the other side of the horizon or to manufacture some tools. Once we have built the tools, the desire to build tools fades (assuming the agent actually likes building tools) and, instead, the agent spends some quality time in social activities with other members of the tribe.

Yes, our desire for what we have been long accustomed diminishes, but it is a mistake to think of this as a case where, as an agent becomes familiar with something, the desire comes to recognize and match its actual value.

The Utility of Desire

Referring to "trust" as an example, Railton argues that our sense of trust is an instance of an affection seeking the actual value of things. A person can be too trusting, or not trusting enough. Both of these are mistakes - mistakes where our affections do not match the true value of trust.

But what is the correct value of trust?

I have agreed that some desires are malleable. I then denied that a change in the intensity or object of a desire necessarily counts as an improvement - it is merely a change. However, desires can be more or less useful - can fulfill or thwart other desires. As a result, it can be (and often is) the case that we can have reasons to modify certain desires - reasons to strengthen a desire where making it stronger can be useful, or to weaken a desire where weakening it can be useful.

A person becomes too trusting when her level of trust puts her in situations where there is an undue risk of coming to harm - an undue risk that other desires that the agent has will be thwarted. A person is not trusting enough when her level of trust prevents her from obtaining advantages that may be gained through interactions with others, leaving unfulfilled her own desires that could have otherwise been fulfilled.

In this sense, it us true that trust and other sentiments can have a true value. The same can be said of fear (courage), sociability, and curiosity. Each of these are malleable traits to some extent, and there is a level at which each of these is most useful. We may have reason to increase the intensity of our desire for certain types of states, or increase the intensity of our aversion to them.

However, please recall that one of the differences I specified between beliefs and desires is that a belief, by itself, can be mistaken - can fail to conform to reality. A desire by itself cannot. Another way that I expressed this is by saying that beliefs contradict, whereas desires conflict. There is no way to evaluate a desire except in virtue of its relationship to other desires. Yet this "value in relation to other desires" is how we evaluate all things - from kitchen utensils to blog postings. This does not require that there be any type of "actual value" that desires can pick out more or less correctly. It requires nothing more than value understood as a relationship to desires which, themselves, cannot be more or less correct. It is no mystery that desires themselves can have value in virtue of their relationship to other desires. "Actual value" is not needed.

On Desire 2018. Part 54: Learning to Desire

Does the type of change that a desire undergoes as a result of an encounter with reality constitute “learning?”

No. Not really.

There is a sense in which a change of desire may be spoken of as "learning" to like or dislike something. However, the type of learning relevant to this discussion is that learning the "correct value" of things.

Etter Railton writes:

In supporting feed-forward action-guidance through expectation and reliance, and in supporting thereby a process of feedback from experience by assessing discrepancy with expectation, desire exhibits an inherent learning dynamic.

Railton, Peter, (2017), “Learning as an Inherent Dynamic of Belief and Desire,” In Deonna J. & Lauria F. (eds). The Nature of Desire. Oxford University Press.

More importantly, this learning has to do with determining the "actual value" of something. He writes that his methods of learning show "ways in which desire can go astray . . . from the actual values" and "ways in which 'one's affections' can fail to correspond to actual values."

There are no actual values for desires to go astray from, or for affections to fail to correspond to.

So, let us take a look at how Railton thinks we can learn the correct value of things.

The Relationship of Parts to Whole

One example that Railton gives us to illustrate how desire can deviate from actual value concerns a woman wanting to be a partner in a prestigious law firm. She becomes a partner, only to discover that it fails “to yield the pleasure and other benefits it promised.” The value of becoming a partner, in this case, was not in agreement with her desire to become a partner. Her experience tells her that her desire to become a partner was mistaken. This disappointment reduces the desire, the way that new evidence may reduce a person's confidence in a belief, providing an example of a person learning the true value of being a lawyer. In Railton's terms, "This, in turn, would weaken her desire that p going forward."

I think that this value is mistaken. Once upon a time, I took a drink from a cup that I thought had hot chocolate, only to discover that it was coffee. I was greatly disappointed in what I discovered to be the taste of what was in the cup. Here, it would be a mistake to say that I had a desire to drink what was in the cup. I had a desire to drink hot chocolate. What I discovered when I took a drink of what was in the cup was not that my desire was mistaken. It was my belief about what was in the cup that was mistaken.

Railton takes panes to distinguish cases in which what we learn about is a mistake between means and ends. He recognizes that, in some cases, what we learn is that something we thought to be useful is, in fact, no use at all. He recognizes that these represent cases where we acquire a new and more accurate belief. However, these are not the types of cases we are concerned with here. The woman's case is not one in which she discovered that being a lawyer serves as a poor means to some other end. Instead, she has discovered that she really does not like being a lawyer - period - independent of its usefulness.

However, the analogy to the cup of coffee still stands. It is quite reasonable to say that the woman thought that being a lawyer would "taste" different than it actually would. She knew what she wanted it to taste like. However, what she discovered upon sampling the contents of this particular cup that she was mistaken about what she thought it would taste like. This is not a matter in which her desire was mistaken. The mistake, instead, is to be found in her beliefs about that which she was tasting.

We could say that the discovery of a mistake in the relationship between means and ends is that the agent desired that Q, the agent believed that P would bring about Q, and therefore desired P as means to bringing about Q. The discovery that P will not bring about Q is the discovery that P does not have the value of a means that the agent thought it had. This is not a matter of improving one's desires. It is a matter of improving one's beliefs.

Railton's case is one in which the agent desired that Q because he believed that P would be included in Q - a part of Q. P's relationship to the fulfillment of the desire is not one of means and ends. It is a relationship of the part to the whole - of whether that which one actually desires can be found within a larger and more complex state of affairs. When one discovers that what one desires is not there, it is not a case in which one's desire was incorrect. It is a case in which one's belief that a desire would be fulfilled within a particular state of affairs was incorrect. So, what we learn is not that our desire was mistaken. It was our belief that was mistaken.

If the contents of the cup had been what I believed it to be, I would have enjoyed the content, and enjoyed it because it would have been that which I desired. If the contents of being a lawyer had been what the woman believed it to be, she would have valued it, and valued it precisely because it would have been that which she desired.

Acquired Tastes

This is not to say that the type of learning that Railton talks about does not take place. I have mentioned several examples. An encounter with poison ivy might give one an aversion to running through the bushes a preference to staying in the open. An encounter with poisonous mushrooms might give one an aversion to eating food with mushrooms. An encounter with a bee might give one an aversion to the proximity of bees. In these types of cases, the value of a state of affairs actually changes through experience.

We can easily explain how we evolved these dispositions. Pain, for example, is a sign of damage to the body - the type of damage that we evolved a disposition to avoid. However, when it gets to the point that we are in pain, it is already too late. The damage has been done.

What we truly need is a system that will cause us to acquire an aversion to those things that would likely result in a state of pain. We can do this if we can process the complex relationships between means and ends that will tell us, in advance, what might lead to pain. However, we evolved from creatures that lacked this capacity. For those creatures. it would have been more efficient to learn an aversion to that which would tend to bring about pain.

If action A results in a state of being in pain, one does not need to learn that "action A results in a state of being in pain, so I have reason to avoid action A." One can simply acquire an aversion to action A. Thereby, one acquires an aversion to and a disposition to avoid that which causes pain, not as a means, but as an end in itself.

However, when this happens, one has acquired a trait, which is quite different from learning a fact. If I should fall and cut my arm which leaves a scar, I have acquired a scar as a result of my interaction with the environment. However, this change in the shape and structure of my skin is not to be described as learning a particular fact. It is simply the effect of a particular type of interaction. The same is true of acquiring a food preference against mushrooms or a dislike for the proximity of bees.

Note that if somebody else has different experiences and acquires different preferences, there is no sense in which we can say that one set of preferences is correct and another set incorrect. This is no more the case than if one acquires a scar in his arm and another does not that the correct state to be in is that of having a scar on one's arm or not. These are simply states brought about through one's interaction with nature.

This is not to say that we cannot talk about the value of things. Insofar as one has an aversion to pain, then one has a reason to avoid being in a state of pain. Insofar as one has acquired an aversion to eating mushrooms, one has a reason not to eat mushrooms. Insofar as one has a dislike of bees, one has a reason to avoid being in the company of bees. However, more work needs to be done to show that this represents the correct state to be in and that being in a different state is wrong, rather than just being different.

There is more to be said on this issue. Railton brings up other examples such as learning the correct amount of trust to put in people that seems to suggest the correct content and force of a particular desire. We have blocked some of Railton's attempts to get to the conclusion that desires represent a type of knowledge of actual values, but he has options we have not yet examined.

On Desire 2018. Part 53: Beliefs, Desires, Sentiments, and Learning

In our previous episode, we looked at Peter Railton’s theory of belief.

(Bel p) A belief that p is a compound state consisting in (1) a degree of confidence or trust in a representation, p, that (2) gives rise to and regulates a degree of expectation that things are or will be as p portrays them, and (3) this degree of confidence or trust is disposed to strengthen or weaken in response to the extent to which this expectation that p is met or violated in subsequent experience.

Railton, Peter, (2017), “Learning as an Inherent Dynamic of Belief and Desire,” In Deonna J. & Lauria F. (eds). The Nature of Desire. Oxford University Press.

He then makes a similar claim about desire.

(Des p) A desire that p is a compound state consisting in (1) a degree of positive affect or favoring toward a representation, p, that (2) gives rise to and regulates a degree of expectation that p will be satisfying or beneficial and a corresponding degree of motivation to bring about or sustain p, and (3) this degree of positive affect or favoring is in turn disposed to strengthen or weaken in response to the extent to which this expectation that p is met, exceeded, or disappointed in subsequent experience.

His overall goal is to claim that beliefs and desires are species of the genus “sentiment”.

(Sent p) A compound mental state in which (1) a degree of sentiment toward a representation, p, that (2) gives rise to and regulates a degree of action-guiding and potentially motivating expectation with respect to p, and (3) this degree of sentiment is in turn disposed to be modulated by whether, and to what extent or in what direction, this expectation with respect to p is met or violated.

With this in mind, let’s go back to (Des p) - which is short for “Desires p” if you did not catch that - and (Sent p) refers to having a particular sentiment towards P where beliefs and desires are both types of sentiments.

The key component relevant to this discussion is going to be Railton's third component. This identifies what Railton calls "learning". In the case of belief, one's encounter with evidence increases or decreases the strength of one's sentiment that the belief is true - its credence. Railton's claim is that there is a component of desires that is also called "learning" that makes desires like beliefs. This is what he describes in (3) the degree of positive affect or favoring is in turn disposed to strengthen or weaken in response to the extent to which this expection that p is met, exceeded, or disappointed in subsequent experience.

Railton is providing an accurate account of how the reward mechanism works. The reward system actually measures a "deviation from expectation". This is shown in experiments. If, as a result of some action, the results are exactly what the agent expects then there is no reward signal. Everything remains steady at its baseline. However, if the agent obtains an unexpected benefit - a pleasant surprise - there is a reward signal. This reward signal increases the "positive affect" of that which causes the reward.

This is easier to illustrate in the case of unexpected pain or some other averse reaction. As one goes about one's business, one encounters a particular type of plant with white berries, or a mushroom, or an insect with black and yellow stripes. The result is a rash (poison ivy), vomiting (poisonous mushroom), or a bee sting. The effect of this encounter with the environment is not just to cause one to learn to avoid this type of plant as a means of avoiding a rash, or to avoid mushrooms as a means to avoid vomiting, or to avoid bees as a means to avoiding bee stings. One acquires an aversion to running through the bushes and prefers to say in open areas. One acquires an aversion to mushrooms and will not touch a mushrooms - picking the mushrooms out of one's food if one should find them. One acquires a fear of bees and simply wants to be away from them "for its own sake".

Railton describes another type of case - an individual who desires to become a lawyer. Upon becoming a lawyer at a top law firm, she discovers that she does not like it as much as she thought the would. The effect of the persistent disappoint is to cause her desire to become a lawyer at a top law firm to diminish. Her interaction with the state of actually being a lawyer, according to Railton, causes her desire to diminish as she learns the "true value" of being a lawyer.

It is certainly the case that interaction with the environment can change one's desires in this way. However, the question remains whether this counts as "learning" in the relevant sense. More specifically, do these types of changes count as learning the true value of that which one is evaluating. In the next post, I will argue that it does not.

On Desire 2018. Part 52: Railton's Beliefs

There is an element of Peter Railton’s theory of desires that concerns me. Like Gregory, Railton (and this is my interpretation) handles certain objections to his theory of desires by removing key desires - hunger, thirst, sex, pain - from the set of desires. As such, his theory of desires is actually a theory of “something else” - something quite different from desires.

Railton, Peter, (2017), “Learning as an Inherent Dynamic of Belief and Desire,” In Deonna J. & Lauria F. (eds). The Nature of Desire. Oxford University Press.

Alex Gregory made a similar claim in the previous article.

Gregory, Alex, (2017), “Might Desires Be Beliefs about Normative Reasons for Action?” In Deonna J. & Lauria F. (eds). The Nature of Desire. Oxford University Press.

Whereas Gregory seems to be providing a theory of beliefs, desires-as-means, and “drives”, Railton seeks to distinguish between beliefs, malleable desires, and drives.

An important part of Railton’s theory is that desires are like beliefs in that desires can be learned. That is to say, through experience, desires change. However, we need to answer the question of what this “change through experience” implies. Railton wants to compare this to updating beliefs by “encountering” evidence. I do not think it is correct. Rather, it is more like changing one’s weight by encountering a large quantity of chocolate cake. The fact that something changes does not imply that it is seeking to match some independent truth.

This corresponds to what I have said is a distinguishing characteristic between beliefs and desires. A single belief can be wrong - it can be in need of being corrected. A single desire can be in need of being corrected if it does not correspond to the world. A single desire cannot be mistaken. It can only be in conflict with another desire. Railton wants to conceive of changes as potential corrections.

That is the overall view. I need to get into specifics.

With respect to belief, Railton’s thesis is:

(Bel p) A belief that p is a compound state consisting in (1) a degree of confidence or trust in a representation, p, that (2) gives rise to and regulates a degree of expectation that things are or will be as p portrays them, and (3) this degree of confidence or trust is disposed to strengthen or weaken in response to the extent to which this expectation that p is met or violated in subsequent experience.

The important point that Railton wants to make is that “confidence” is a sentiment - a feeling - attached to a proposition. Furthermore, the nature of this sentiment - it’s strength - is altered by interaction with the environment. Specifically, interaction with evidence or counter-evidence strengthens or weakens that belief.

He is going to relate this feature - being a sentiment altered through experience - to what is true of desires to show that desires are like beliefs. Desires, after all, are also sentiments that can be altered through experience. I will get to this in my next post.

While the focus is still on beliefs, I want to draw attention to what I have called the defining characteristic of belief insofar as belief is distinct from desire. A belief, by itself, can be incorrect. A desire, by itself, cannot. A desire can only be in conflict with another desire.

There is nothing in Railton’s statements about belief that denies this feature. The sentiment that measures the certainty of belief is a sentiment that measures the agent’s confidence that a belief, by itself, matches the world. It may be a sentiment that can change as a result of interaction with the world, but that change is still a measure of the confidence (or lack of confidence) in whether the world is as it is believed to be.

Now, this is supposed to be a series on desire, not on belief. So, let’s look at what Railton says about desire and how this view of belief is relevant.

On Desire 2018. Part 51: Details on the Direction of Fit

Remember a few episodes back when I mentioned in passing this idea that beliefs and desires were quite alike in one respect.

I suggested that beliefs and desires both assign a value to a proposition. That they are both propositional attitudes has been a main theme in the philosophy of mind for a long time. The idea is that, more specifically, they both assign a value to a proposition.

In the case of belief, the value is the credence or credibility of the proposition - the chance that it is true.

In the case of desire, the value represents the importance of the proposition to the agent - the effort one is willing to go through to make or keep the proposition true.

Well, look at this:

This contrast between belief and desire has powerful intuitive force. Moreover, it seems to fit well with contemporary rational decision theory, which divides the basic components of intentional action into two fundamental categories, credences and preferences, taken to be fundamentally different in kind (cf. Lewis 1988, 1996).

This s from Peter Railton’s contribution to The Nature of Desire: Railton, Peter, (2017), “Learning as an Inherent Dynamic of Belief and Desire,” In Deonna J. & Lauria F. (eds). The Nature of Desire. Oxford University Press.

Railton’s is Chapter 8 in the anthology. He is going to argue that advances in psychology are more alike than different. This can be held in contrast to Timothy Schroeder’s thesis, discussed in the previous three postings, that beliefs and desires are different in that they are connected to action in different ways.

Before we get into Railton’s argument, I want to specify the main problem I have with attempts to reduce desires to beliefs. It has to do with the general topic that has come up several times - direction of fit. But the concern is not over its conformity to some abstract philosophical concept. There is a very real and important issue at stake.

If I believe that P, and you believe that not-P, then one of us must be mistaken. One of us is in need of changing our beliefs. Beliefs contradict regardless of who holds them. This is what it means to have a “mind to world” direction of fit. There is only one world. The belief matches the world, or it is wrong.

Desires do not contradict, they conflict. If I desire that P, and you desire that not-P, we do not have a contradiction, we have a conflict. I am trying to make or keep the proposition P true, while you are trying to make or keep it false. There is no sense in which one of us is right and the other is mistaken. While I can try to prove to you that the correct belief to have about P is that P is true, I cannot prove that the correct desire to have about P is that it is desired.

Another relevant difference is that a single belief can be mistaken. If a person believes that the Earth is the center of the universe, then this belief can be mistaken if, in fact, the earth is not the center of the universe. However, a single desire cannot be mistaken. There is no way to evaluate a desire unless and until there is a second desire with whom it may be in harmony or conflict. There is a reason to change a belief if the belief does not correspond to the world, but there is no reason to change a desire until and unless that desire comes into conflict with another desire.

Railton is going to argue that beliefs and desires are more alike then different. He is going to argue that there is a particular sentiment associated with belief, and that desires can be learned. He may be right about these facts, but the real test is whether he can erase the distinction whereby beliefs yield contradiction whereas desires yield conflict. I am going to agree that desires can be learned, but the different directions of fit between beliefs and desires represent two different styles of learning.

I am going to agree with him on these points. However, I will argue that they can be (and must be) made consistent with the idea that beliefs contradict while desires conflict, and a single belief can be mistaken but a single desire cannot be judged good or bad without a second desire.

On Desire 2018. Part 50: Neo-Humean Theories of Desire and the Brain

In discussing Timoty Schroeder's article on the neurobiology of intentional action (Schroeder, Timothy, (2017), “Empirical Evidence against a Cognitivist Theory of Desire and Action", In Deonna J. & Lauria F. (eds). The Nature of Desire. Oxford University Press), I have so far (1) presented his summary of evidence regarding the neurobiology of intentional action, and (2) used this information in presenting his arguments against cognitivist theories of desire.

The next thing to do is to look at what Schroeder says this information has to say about neo-Humean theories of desire.

The theory that I defend is something of an odd duck. Typically, theories of desire are divided into two camps, evaluationist (desires are evaluations that something is good or ought to be the case) and dispositional (theories of desire are statements about how people are disposed to act). So, if there are problems with evaluationist theories, then that leaves dispositional theories. Schroeder has so far used his empirical evidence against evaluationist theories of desire. This leaves the dispositional or, what Schroeder calls, "neo-Humean" theories of desire.

On this simplest approach, one holds that desires characteristically dispose us to perform actions that seem likely to bring about (or make progress towards) the ends we intrinsically desire.

The theory that I support is an evaluationist theory. However, a "desire that P" is not a belief or judgment that P has value. It is an assignment of value to "P" being true. It requires no belief or judgment, so it does not require a pathway through the parts of the brain associated with beliefs and judgments. An action, such as writing a blog posting on desires, does look at the propositions P that are believed and judged to be true in that state of affairs, but the assignment of a value to those propositions being true can all be done behind the scenes without any contribution from beliefs or judgments. The weighing of the value of those propositions being true against the value of different propositions being true if I were, for example, to go fix myself something to eat, keeps me at the computer writing this blog post.

Of course, assigning a value to a proposition being true also determines the motivational force of making or keeping that proposition true, and so disposes one to act to the degree that one can find a relevant action to perform. However, the action is caused by the desire. The assignment of a value to a state of affairs is causally prior to action and is the reason for the action, when there is action.

With this distinction in mind, let us look at what Schroeder says about the empirical data with respect to neo-Humean dispositional theory and see if we can apply it to the assignment theory of desire.

As it turns out, Schroeder's interpretation of the causal map does not actually involve dispositions to act. It is more consistent with the assignment of value to "representations of perceivable or conceivable contents".

If intrinsic desires are, or are realized be, or subvene on, perceptual and conceptual representations that are connected in the right way to the reward system (such that their contents are constituted as rewards), then instrinsic desires will turn out to have the properties and play the roles attributed to them by the simplest form of neo-Humeanism.

So, we are looking at "perceptual or conceptual representations" (think of these as "perceives that P" or "believes that P"), where their content (P being true) are constituted as rewards (are assigned a value V, which is the strength of the reason to make or keep P true), then they will generate dispositions to act to make or keep P true.

Desires, as Schroeder describes them, are not mere dispositions to act. They are assignments of value to propositions being made or kept true that provide agents with reasons to act, if a viable act can be found. And this "if a viable act can be found" can, perhaps, be related to the acts queued up as a result of the belief/perceptual portion of the brain looking for permission to continue.

I am, so far, quite pleased about the way that the evidence Schroeder presents on the biology of intentional action fits in with the assignment theory of desire used in desirism. However, the book The Nature of Desire contains a second article discussing the empirical evidence. I will be examining that next.

On Desire 2018. Part 49: Cognitivist Theories of Desires and the Brain

In my last exciting episode I looked at what Timothy Schroeder had to say about the biology of intentional action.

Schroeder, Timothy, (2017), “Empirical Evidence against a Cognitivist Theory of Desire and Action", In Deonna J. & Lauria F. (eds). The Nature of Desire. Oxford University Press.

Now, Schroeder used this as an argument against cognitivist theories of desire. His argument is that cognitivist theories are not consistent with biology.

By cognitivist theories, Shroeder is referring to theories that reduce desires to believes or judgments. These include theories that say, for example, "Agent desires that P" is to be understood as "Agent believes that P is good" or "Agent judges that P ought to be the case."

He provided evidence that, in the brain, intentional action has two main inputs.

One input is the cognitive part of the brain (the unimodal perceptual cortex and the multimodal association cortex). This is the part that we may associate with beliefs and judgments.

Another input comes along pathways that Schroeder says is "not yet interpreted in cognitive or non-cognitive terms". However, he identifies it as "the reward (and punishment) system."

Then, Shroeder looks at three ways to try to make sense of a cognitivist theory of desire.

Option 1: A pathway directly from the cognitive portions of the brain to intentional action.

This exists. The standard course of action is that the cognitive parts of the brain queue up actions, but they are all inhibited. Input from the second pathway selects among the queued up actions which will go through. So, it is possible for the cognitive portion of the brain to queue up an action that goes through without any input from this second system. It just goes straight through, without "obtaining permission" from the second system.

However, Schroeder argues that the types of cases we know about where this happens is Tourette syndrome - the involuntary tics and utterances that those with this system feel compelled to do. Significantly, people with Tourette syndrome do not take these actions to be theirs - they are not "from me". Schroeder also reports, "This same sort of overriding activation, when induced by direct electrical stimulation of the brain, likewise induces movements from which experimental subjects feel alienated."

(I would like to note, as a theory of personal identity, I have played with the theory that what "I am" is a particular set of desires. To say that I am responsible for an action - to say that an action is mine - it must have its source in my desires. If it does not come from my desires, then it is not mine. Schroeder's claims here would be consistent with that thesis. This second system "choosing" among the queued up actions which to let through and turn into actual action is what makes the actin mine. It is what makes "me" the cause of the action. However, I have not worked with this thesis enough to actually endorse it.)

There is a second direct route that need not involve the second (reward/punishment) system. It takes a detour through the motor basal ganglia. Schroeder identifies this as the region of habit formation. I think that a good way to summarize Schroeder's description is to think of the motor basal ganglia as a place where the input of dopamine from the reward/punishment system creates channels of low resistance for certain actions. When queued, these actions "pass through" with little resistance. So, if one has a habit of, say, turning left at a particular intersection, the "turning left" habit may pass through the motor basil ganglia even though, on this one occasion, one actually wanted to turn right. Those bad habits of, for example, biting one's nails are found here.

However, habits are not judgments. Habits, in fact, often go against our better judgments. In this way, we can have good habits and bad habits. So, equating habits with beliefs or judgments that something is good or ought to be the case is problematic.

Option 2: Beliefs or judgments about what is good or ought to be the case are formed in the reward/judgment pathway.

We need not interpret the functioning of the reward/punishment pathway as desires as distinct from beliefs. It may be used to form beliefs of a particular kind - of the kind that motivates action.

Schroeder provides three objections to this interpretation.

First, Schroeder states that such a person must interpret damage to the reward/punishment system of the type that prevents action (extreme Parkinson's disease) as damage to the ability to form beliefs/judgments about what one ought to do. Yet, as Parkinson's disease gets increasingly severe, sufferers do not report increasing difficulty in judging things as good or as something that ought to be done.

I do not see how Schroeder makes a good case here. To make a parallel case with desires, one would have to argue either that Parkinson's disease has an effect on what patients desire. If the effect of Parkinson's disease on desires is the same as that Parkinson's disease would hypothetically have on judgments of goodness or oughtness, then Schroeder hasn't come up with anything that distinguishes one from the other.

So, I don't see Schroeder's argument here.

Second, the functioning of the reward/punishment system is isolated from memory and consciousness.

Apparently, there are no direct connections from the reward/punishment system to the parts of the brain associated with memory or consciousness.

Once one begins to look specifically at the reward system, one sees an absence of projections from this system to regions of the brain that seem involved either in consciousness or in episodic memory, at least, organized in a manner that would support being conscious of or remembering specific judgments of what one has most reason to do, all things considered.

Only the parts of the brain typically associated with traditional beliefs and judgments have these types of connections. I find this quite interesting. It appears to be consistent with the thesis that we do not have very good direct knowledge of our desires - that some of them are unconscious, and that people can be wrong about what it is they desire. We are clearly able to form beliefs about our desires, but we derive those beliefs (like all beliefs) from observing our own behavior and monitoring our body (recognizing symptoms of fear or longing). Like all beliefs, our beliefs about our own desires can be mistaken. Like the rest of the world in general, our desires will influence our action independent of our beliefs in just the same way that the shape of the earth is independent of our beliefs. But the reason we think we have, and the reason we actually have, for doing something may not agree.

Anyway, this draws a genuine distinction between beliefs and judgments about what is good or what one has a reason to do from desires in that unconscious beliefs and judgments are problematic. A person who acts because he judges something to be good is relying on a mental state - a judgment - that has connections to consciousness and memory that desires, apparently, do not have.

Third, pleasure must be "causally upstream" of judgments about reasons, or we cannot judge that we have a reason to do something on the basis that it will be pleasurable. In contrast, pleasure is "causally downstream" of desire. It is in virtue of the fact that one desires something that one gets pleasure from obtaining it. It is not in virtue of the fact that one judges it to be good that one gets pleasure from obtaining it. This represents another distinction between desire and judgment.

So, to me, at least, two of these objections to relating activity of the reward/punishment system to beliefs or judgments about what is good or ought to be the case show that there are problems with this option.

Option 3: The reward/punishment is a component of, rather than the whole story behind, judgments about what is good or ought to be the case.

This option answers the problems about connections to memory and consciousness (since there is a part of the judgment located in the parts of the brain associated with beliefs and perceptions with its strong connections to memory and consciousness). However, it does not answer the objection of the direction of causation. It would still be the case that pleasure must be causally prior to the judgment that a reason for doing something is that it is pleasurable. One must first judge that one has a reason to do it (that one desires it, in the cognitivist sense of desire) before it can be pleasurable.

More importantly, the reward/punishment system can be activated along pathways that have nothing to do with those parts having to do with beliefs and judgments. It can be activated through sensory input directly, as when something tastes good, or one is pleased by the beauty of a sunset or a person. If desire requires both a reward/punishment component and a belief/judgment component, then we cannot prefer the taste of butterscotch to chocolate or prefer Beethoven to Bach. Or, at least, these preferences are not desires.

Conclusion

So, using evidence of brain structure, we can see that there are problems with the thesis that a desire can be understood in terms of a belief that or a judgment that something is good or ought to be the case. This creates a problem for cognitivist theories of desire. But, can a neo-Humean theory of desire do better?

On Desire 2018. Part 48: The Biology of Action

*begin Jaws theme here*

We have reached the scariest part of the series for some desirists.

*pause for dramatic effect while the Jaws theme becomes more intense*

Particularly for those people who value some empirical evidence in their philosophy.

*Jaws theme nears its climax*

Chapters 8 and 9 of The Nature of Desire bring forth empirical evidence.

*Jaws theme reaches its climax*

The topic of discussion for the next few posts is: Schroeder, Timothy, (2017), “Empirical Evidence against a Cognitivist Theory of Desire and Action", In Deonna J. & Lauria F. (eds). The Nature of Desire. Oxford University Press.)

Schroeder's argument is that the biology of intentional action is more compatible with neo-Humeanism than it is with the various theories that interpret desires as cognitive states - beliefs or judgments that something is good or that one has a reason to perform some action. Specifically, the biology of action suggests that there are two inputs into intentional action - a cognitive input from the parts of the brain that are associated with beliefs and judgments, and an affective input that words through the reward system and dopamine production. This is more compatible with the neo-Humean system.

So . . . how does the brain work?

Schroeder provides us with a causal map. Note that he presents this map after a long discussion of the empirical research. This is not a map drawn while sitting at the philosopher's armchair. It is drawn while standing in the neuroscientist's laboratory.

One reasonable interpretation of the neuroscientific account might go as follows: one senses the world and makes judgments about it. These sense experiences and judgments are realized in the unimodal sensory and multimodal association cortex. They cause possible basic or primitive actions to be primed, with the priming of more complex possible actions (say, strumming a guitar) causing the priming of the less complex possible actions that are required (say, shaping one’s hand in a certain way, moving one’s arm, etc.). These primed actions can be commanded by clusters of neurons found in the motor hierarchy, naturally. Primed possible actions existing at a given time are typically more numerous than the actions one actually performs or would wish to perform. Thus, primed possible actions are all chronically suppressed, this being the role of the motor BG. At the same time as possible actions become primed, one’s experiences and judgments cause a response in one’s reward system. This response combines, in the motor BG, with one’s experiences and judgments. And this combination of experiences, thoughts, and reward signals in the motor BG causes a release of some of the primed possible actions.

Fine. What does this mean in English?

Here you area, reading this post . . . I assume . . . perceiving the world, assessing what you read, saying things to yourself like, "Hey, that finally makes sense?" and "How can anybody believe something as stupid as what this guy is putting on his blog?" But, are you going to keep reading? Are you going to go to the kitchen and get a piece of that chocolate cake? Are you going to write a comment?

As you perceive the world, your brain cues up a whole gaggle of actions compatible with those perceptions. It cues up a set of actions that would involve continued reading, a set of actions associated with going to the kitchen for some chocolate cake. A set of actions that are involved in making a comment. All of these are nominated for the action in question.

If all of these actions go through the gate, then you end up doing a whole lot of different actions at once. You can't do them all at once - it is physically impossible to do so. Normally, your brain will pick one of the nominees and say, "That one." And then you do it.

Sometimes, there is a specific type of action that leaks through the gate. It is usually a simple action. In most cases it is a part of an action - the utterance of a particular word or a simple muscle action such as a wink or a twitch. This is Tourette Syndrome. The behavioral oddities are commonly called "tics". The person afflicted with Tourette Syndrome report being "alienated" from these actions - they come from someplace else. They are not chosen. Whereas the action that is selected to make it through the gate is the action that is chosen - and, thus, it is the action that belongs to the agent.

So, what about this "choosing" of various queued up actions? How does that work?

Well, that comes by means of a dopamine signal that enters through the reward system. This signal selects one of the various act alternatives and says, "Do that one." If this signal is weak, an individual will choose an action, by find it difficult to execute the action. If the signal is stopped altogether, then the agent cannot act - no act is chosen - even though the agent wants to act. This is Parkinson's Disease.

Schroeder's description of the reward system and its effect on actions lacked detail. He wrote:

The [orbitofrontal cortex] has fairly stable, though not immutable, long-term dispositions to discriminate between [its] input signals. In response to some, nothing happens. In response to others, a signal is sent along the affective [basil ganglia].

The signal that is "sent along" is the signal that is relevant to selecting the action that gets let through the gate.

Schroeder does not say anything to this effect, but it seems that these "stable, though not immutable, long-term dispositions" seem to have a close relationship to what I have been calling desires. Desires, as I have described them, are certainly stable and long-term but malleable, and are used in selecting actions. However, nothing in Schroeder's description suggests assigning a value to a proposition being true. So, this may mean absolutely nothing. This is simply something to be filed away - perhaps it will be useful later.

One of Schroeder's comments that I am particularly curious about is this:

In addition to a reward system, the brain almost certainly also contains a distinct punishment system. Its existence may be inferred from such facts as that intuitively rewarding stimuli and intuitively punishing stimuli cause distinct but similarly located responses in OFC and distinct but similarly located responses in the affective BG, and from the fact that there are special brain chemicals particularly released in animals like us under punishing conditions and causing, e.g., freezing behavior in rats. Unfortunately, the effects of the punishment system on behavior can only be inferred by analogy to the effects of the reward system, as there is as yet no clear identification of the punishment system’s output structure.

I have long wanted to know something about the punishment system but, so far, have only been able to find information on the reward system. This is my first confirmation of the hypothesis that it works in a way similar to the reward system.

This actually is not a lot to go on - but it is more than I had. Schroeder used this to argue against cognitivist theories of desire by showing that the way the brain actually works a signal from the affective parts of the brain - the reward system - are used to choose action. This is not a part of the brain that is associated with beliefs and perceptions. So, theories that claim that a desire is to be understood as a belief or a perception are going to have problems matching their theories to the empirical fact. A neo-Humean, such as myself, will not have such problems.

So, it seems, I can breathe a sigh of relief . . . so far.

On Desire 2018. Part 47: Animal Desires

Alex Gregory proposes: To desire to φ is to believe that you have normative reason to φ.

(Gregory, Alex, (2017), “Might Desires Be Beliefs about Normative Reasons for Action?” In Deonna J. & Lauria F. (eds). The Nature of Desire. Oxford University Press.)

If this is true, then it seems to follow that if Fido has a desire to go for a walk, then Fido has a belief that he has a normative reason to go for a walk. Yet, attributing such a belief to Fido seems absurd. Thus, it cannot be the case that Fido's desire to go for a walk is a belief that he has normative reasons to go for a walk. Meaning that Gregory's thesis must be rejected.

This, in my opinion, is a knock-down objection to Gregory's thesis.

Gregory makes two attempts to defend himself.

Gregory's first attempt is to suggest a rather loose attribution of beliefs and desires. He argues that a cat can have a desire to drink milk even though the cat can have but a very rough idea of what "milk" is. Fido can lament that his food bowl is empty with only the most basic understanding of "food bowl" and "empty". So, it seems conceivable that an animal can have a rude type of proto-belief about having a reason for action.

However, we have to draw a line somewhere between what the animal can understand and what humans can understand. As Fido jumps up to catch the ball thrown for him, we can say that he understands something about the relationship "Force = mass * acceleration". However, it this is mostly due to the development of instincts and habits that take advantage of the relationship. We would be making a mistake if we implied that Fido was a cognitive expert in Newtonian physics. Understanding the desire to go outside in terms of a belief that one has normative reasons to go outside is simply over-intellectualizing what is, in fact, a basic desire.

It may be difficult to know exactly what it is a dog believes, but "that I have a normative reason to go outside" hardly seems like it is even close to the line.

Gregory's second attempt is to narrow the concepts of desires and beliefs. Here, he tries a move that I discussed in the previous section, distinguishing "basic drives" from "desires". Here, he argues that animals do not, in fact, have desires. Instead, animals have "basic drives" that are distinct from desires, and thus are not to be understood in terms of "having a belief that one has a normative reason to φ."

In this case, as I said previously, Gregory himself defeats this option. Gregory wrote:

Hunger, we might think, is the paradigmatic desire, and if a theory of desires excludes hunger from its purview, then it is no longer a theory of desire at all.

Yet, when Gregory makes this move, he is removing hunger from the category of "desire" and putting it in the category of "basic drives" - something different from desire. Gregory, then, may be offering a legitimate theory of what he is calling "desire," but he is no longer offering a theory about what the rest of us call desire. Such a theory is, explicitly, a theory about such things as hunger, thirst, aversion to pain, desire for sex, along with the more complex desires that humans are capable of acquiring.

Indeed, Gregory is offering a theory of "desires-as-means". That which is desired as a means to an end can be understood as being desired in virtue of the belief that one has a reason to realize these means, because it serves the end. The squirrel's desire to climb the tree to get to the food can be understood as understanding that he has a reason to climb the tree (to get to the food). But his desire to eat cannot be understood in this way. Gregory's thesis fails the instant he switches from talking about the value of means to the value of ends.

The reason a cognitive understanding of "desires-as-means" makes sense is because "desires-as-means" are mixtures of "desires-as-ends" and "means-ends beliefs". And means-ends beliefs are subject to cognitive analysis. There is no mystery here.

But the true source of desire - that point at which all inquiry about desire ends, because it is has reached the things that we desire for their own sake and not for the sake of something else - is simply outside of the scope of his theory.

On Desire 2018. Part 46: Addiction

First, I rewrote the previous post before posting this one. I did this to improve the fit between the two posts. They discuss related topics. While the previous post discussed cases where a person apparently believed that he had a normative reason to φ without having desire to φ, this one has to do with an agent having a desire to φ without believing that she has a normative reason to φ.

To address the previous problem, Alex Gregory suggested the possibility of desire without motivation. Thus, it is possible to have a belief that one had a normative reason to φ (and, thus, a desire to φ) without being motivated to φ. I argued that this was unnecessary. The only thing we needed were cases where the desire to φ was outweighed or overpowered by other reasons, or the desire to φ was a future desire that did not have the capacity to reach back in time and motivate current action. When this happens, even though all desire is intimately connected to motivation to act, it does not always generate action. The reason (and, with it, the motivation) to φ may simply be overwhelmed by the reasons not to φ. (Gregory, Alex, (2017), “Might Desires Be Beliefs about Normative Reasons for Action?” In Deonna J. & Lauria F. (eds). The Nature of Desire. Oxford University Press.)

Now we are going to address a different kind of case.

Amy has been a heroin addict for many years. She believes that she has very little reason to take the drug. But she strongly craves it nonetheless.

Recall that Gregory's thesis was: To desire to φ is to believe that you have normative reason to φ. Here, we have a case in which Amy has a desire to φ without having a belief that she has normative reason to φ. So, this would be a problem for Gregory's theory.

Gregory responds to this case by suggesting that Amy's weak desire to φ over-motivates her to φ. Recall that he argued for breaking the association between desire and motivation in the previous case. If we can have desire without motivation then it seems plausible to argue that we can have motivation beyond that which is warranted by a given desire.

However, I argued that we can handle the case without breaking the link between the strength of a desire and its motivational force. Moreover, I argued that we need the link between desire and motivational force since if desire has nothing to do with giving an agent a strong or weak motivation to act, then we are left to puzzle over the question of what a desire actually is supposed to be doing.

Furthermore, we can raise the objection against Gregory that his way of handling this objection certainly conflicts with our common understanding of the term. The reason Amy cannot give up the drug is because she has an overwhelming desire to take the drug - a desire too strong to resist.

In fact, this is exactly the way that the assignment theory of desire analyzes addiction. The addiction is a very strong desire that overwhelms all other desires. The effects of the addiction on the brain is to create a desire against which other present desires cannot compete. While the addiction can overpower all present desires, it also can easily ignore all future desires as well - given that a future desire cannot reach back in time to motivate present action. The result is a desire so strong that it causes the agent to sacrifice all other present and future concerns.

Another option that Gregory suggests is to distinguish desires from "some more primitive compulsion or drive".

I would suppose that if Gregory were to go this route, there would be no better candidate for "more primitive compulsion or drive" than the aversion to pain and the desire to eat. Yet, to go this route would mean that the aversion to pain is not an aversion, and the desire to eat is not a desire. Neither is the desire for sex, or the interest that one has in being within a comfortable temperature - not too hot and not too cold.

I have argued previously that Gregory's theory works for desires-as-means. This is because desires-as-means are combinations of desires-as-ends and beliefs about how the means will serve those ends. Desires-as-means respond to evidence because the belief component of desires-as-means responds to evidence. Desires-as-means appear to be beliefs about reasons for action since they are beliefs about how certain actions (φ-ing) relate to certain ends.

Gregory removes ends from his theory by claiming that desires-as-ends are not desires. The aversion to pain, hunger, and desire for sex, are not desires. Gregory himself admitted that if his theory ever reached this point then it was no longer a theory of desire. In discussing appetites, he wrote:

Hunger, we might think, is the paradigmatic desire, and if a theory of desires excludes hunger from its purview, then it is no longer a theory of desire at all.

At the time, Gregory kept hunger in the realm of desire. However, here he is at risk of removing hunger from the category of desire and placing it in the realm of "more primitive compulsion or drive". There is, perhaps, no more primitive compulsion or drive than the aversion to pain, or the desire for sex, or the desire to eat - all necessary for sustaining life and biological reproduction.

So, Gregory's first response suggests an implausible break between desire and motivation. His second response removes hunger, thirst, lust, and aversion to pain from the category of "desire".

His third suggestion is to postulate a combination of the above reasons.

Real-life addicts might be partially motivated by a genuine desire to avoid withdrawal symptoms, partially over-motivated by a very weak desire to take the drug, and partially compelled to take the drug by some drive.

However, if the break between desire and motivation does not work (because what else are desires supposed to do if they do not motivate), and the distinction between drives and desires does not work (because hunger, aversion to pain, desire for sex, and other basic drives would no longer be desires), then a hybrid of the two will not work.

A theory of desires is, first and foremost, a theory of the very things that Gregory is now calling drives. These are things that we value, not because they serve some further end, and not because we believe we have a reason to do them. We value these things in spite of the fact that we have no reason to do so. They simply are things that have value - that are assigned the values they have, not by reasons, but by evolution, environment, and experience. Addictions are very strong - overpowering - desires. They are desires that people have no reason to have, but which evolution, environment, and experience have caused them to have anyway. Anything that considers itself a theory of desires must be a theory of ends, and not just of means.

On Desire 2018. Part 45: Desires Without Motivation

In our last exciting episode, Alex Gregory distinguished hunger from feelings of hunger - hunger pangs and the like. He argued that the latter are not subject to evidence like beliefs. However, that does not imply that the former are not subject to evidence like beliefs. Thus, hunger and other appetites do not provide counter-examples to his thesis: To desire to φ is to believe that you have normative reason to φ. (Gregory, Alex, (2017), “Might Desires Be Beliefs about Normative Reasons for Action?” In Deonna J. & Lauria F. (eds). The Nature of Desire. Oxford University Press.)

As a part of my response, I agreed with the distinction. After all, the assignment theory of desire states that a desire is the assignment of a value (importance) of a proposition being true. The desire to eat is an assignment of a value of importance to "I am eating" being true. This does not imply any type of feeling or sensation. One can assign a value to eating without hunger pangs, just as one can assign a value to being with one's children without hunger pangs. So the distinction is sound.

However, Gregory's thesis still has problems with the idea that a desire to eat can be understood in terms of a belief that one has motivating reason to eat.

In my response, I asserted that desire is closely related to motivation. I then mentioned in an earlier post in this series addressing the appetites, the desire to eat is related to motivation. Motivation is related to (among other things) the concentration of the hormone ghrelin. I argued that it is not difficult to understand concentrations of ghrelin being related to the assignment of a particular value to the proposition, "I am eating". However, it is difficult to understand how to relate concentrations of ghrelin to the "belief that I have a normative reason to eat".

Well, now that I am on the record as relating desire to motivation, I need to note that Gregory, in the next section of his article, denies this relationship. According to Gregory, a person can have a desire to φ without at all being motivated to φ.

Gregory is making this claim in an attempt to defend his thesis from counter-examples such as the following:

Sally is a smoker. She knows full well that she has very good reasons to quit: smoking is costly and unhealthy. But she is weak-willed and continues to smoke. Such a case might seem to threaten DAB. Isn't Sally's problem that while she knows she should quit, she doesn't want to?

The objection, put in terms more closely related to Gregory's thesis, is that Sally believes that she has normative reasons to quit, but she does not have a desire to quit. If a desire to φ is a belief that one has a normative reason to φ, then this is a problem.

To answer this objection, Gregory seeks to assert that Sally does have a desire to quit smoking. However, a desire to quit smoking does not imply motivation to quit smoking. According to Gregory, desires have the potential to motivate, but this potential is not always actualized.

There are a number of problems with Gregory's response. I want to start with his claim that desires do not motivate. Gregory has already separated hunger from the feeling hungry (a distinction I agreed with). Now, he wants to separate hunger from a motivation to eat. If he does that, then what is left of hunger? It is neither "feeling hungry" nor is it a drive to find something to eat. What is it? I argue, as I did above, that it is not reasonable to relate hunger to purely contingent physical sensations such as hunger pangs. Instead, it is a motivation to eat - a motivation partially under the influence of concentrations of ghrelin in the system.

Desires are a part of a theory that aims to explain and predict intentional action. Desires provide the motive to act - the reason to act. A desire that does not motivate is as odd as a force that does not provide acceleration. Saying that desires merely have the potential to motivate is like saying that gravity merely has the potential to provide a downward force on those objects that are near to the body - something that might, somehow, fail to be actualized.

There is a way that a desire can exist without resulting in any intentional action. This happens when there are more and stronger reasons to do something else. For example, a book sitting on a table does not move. The reason that it does not move is because the upward force from the table precisely counters the weight of the book. While the force of gravity still exists, drawing the book towards the center of the earth, it is not moving anything.

Similarly, a desire, countered by other desires, need not result in any action. Sally has desires to quit based on her desires to spend her money on things other than cigarettes and to preserve her health. However, she also has an even stronger desire to smoke, so the desire to smoke overrides her weaker and fewer desires to spend money on other things and preserve her health.

Also, there is, in fact, a type of desire that does not motivate. A future desire has no ability to reach back in time and motivate present actions. The agent may well believe that smoking will thwart future desires - e.g., the desire to be watching her grand children graduate from college. The agent can know that her present desire to smoke is thwarting her future desire. However, the future desire has no ability to motivate her to stop smoking. This would require a present desire that her future desire be fulfilled. The motivational force this and other present desires may exist, but they may not be strong enough to outweigh the desire to smoke. They are like the sheet of paper in the path of a bullet. They provide some resistance, but they do not stop the bullet.

Please note that this way of understanding Sally the smoker does not require that we disconnect desire from motivation in any meaningful way. It disconnects future desires from present motivation, but it would be more strange to link motivation in such a way that desires - unlike everything else in the macro universe - has the power of backwards causation (the capacity to cause things to happen in the past, when it did not exist).

So, the objection to Gregory is that he cannot sensibly disconnect hunger both from feeling hungry and the motivation to eat; there would be nothing else. At the same time, we do not need to disconnect desire from motivation to handle cases like Sally. We only need to recognize that the motivation of some desires may outweigh others, motivating the agent to choose the option that will fulfill the most and strongest of her desires, and overpowering her fewer and weaker desires to realize an incompatible end.

Gregory brings up another supposed problem case that he calls "Teething Tabatha".

Tabatha knows that she has good reason to go to the dentist: her teeth are in an awful state. But she will quite keenly insist that she doesn't want to go to the dentist. Who does?

This is supposed to create a problem for Gregory's theory because Tabatha has no desire to do something (go to the dentist) that she believes has normative reason to do. Thus, a belief that one has a normative reason to φ is not, in this case, a desire to φ.

Gregory begins by charging that this is also a challenge for the Humean theory. I consider the assignment theory of desire to be a Humean theory, so, if true, this would also challenge me.

This kind of case seems as much a problem for the Humean theory of motivation as for DAB, since it is natural to presuppose that Tabatha might go to the dentist even though she has no desire to do so.

But this is no problem for the Humean theory of motivation. The Humean would distinguish desires-as-ends from desires-as-means. Tabatha's statement that she does not want to go to the dentist can be accurately taken to be a (true) claim that she does not value going to the dentist as an end in itself. Indeed, she may even have an aversion to going to the dentist. Yet, her reasons for going to the dentist - her present desire to avoid future pain and to prevent the loss of her teeth - outweigh her reasons not to go to the dentist. So, even though going to the dentist is not valued as an end, this is more than made up for as a means to her other ends.

In defense of DAB, Gregory could offer the same type of analysis.

Tabatha doesn't think that she has good reason to seek out the pain at the dentist. This is the sense in which she doesn't want to go to the dentist. But all the same, Tabatha does think she has good reason to visit the dentist, all things considered.

The response here still does not require any break between a desire and motivation. We can still allow that the aversion to going to the dentist has motivational force. It is just outweighed by the many and strong reasons to go to the dentist, providing an even stronger motivation. Like a rocket leaving the Earth's atmosphere, the reasons for going to the dentist obtain enough power to overcome the "gravity" of staying away from the dentist.

Separating desires from motivation simply is not necessary.

On Desire 2018. Part 44: Appetites as Beliefs About Reasons

In considering objections to his thesis, Gregory considers an objection near to the objections I have raised. Specifically, he considers the objection that it makes no sense to think of appetites (such as hunger) as beliefs about reasons to eat.

Recall, Gregory’s thesis is:

To desire to φ is to believe that you have normative reason to φ.

(Gregory, Alex, (2017), “Might Desires Be Beliefs about Normative Reasons for Action?” In Deonna J. & Lauria F. (eds). The Nature of Desire. Oxford University Press.)

So, on Gregory’s account, to desire to eat is to believe that one has reason to eat. It is absurd to think that a desire to eat is a belief that one has reason to eat. Therefore, we should reject a Gregory’s account.

Gregory attempts to defend his thesis be distinguishing the feeling of hunger from the desire to eat. We can have the desire without the feeling, he argues, as when one eats to be polite. And we can have the desire without feeling hungry, as when we have desert. The feelings of hunger are not beliefs about reasons to eat or anything else. However, the feelings are not the desire to eat, so this leaves open the possibility that the desire to eat is a belief about reasons.

I wonder if Gregory ever wanted to lose weight. That experience alone demonstrates that the desire to eat is independent of any and all beliefs about having reasons to eat. There are a great many obese and overweight people who are fully aware that they have no reason to - and many reasons not to - eat, but are driven to eat by an overwhelming desire. There are also a lot of people of normal weight who constantly struggle against a desire to eat that does not go away just because one rehearses the reasons.

We know something about hunger. For one thing, eating behavior is associated with the hormone ghrelin. Ghrelin is released by the stomach when it is empty, and stops being released when the stomach is full. Empirical research shows that the concentration of this hormone determines eating behavior. Note that the experiments that researchers have conducted on animals do not associate amounts of ghrelin with feelings of hunger. Researchers, so far as I can tell, have no way to determine what subjective sensations their non-human research subjects are feeling as a result of the concentration of ghrelin in their system. They only measure the disposition to eat, and their findings show that the disposition to eat is not only associated with ghrelin concentrations in the bloodstream, but with the amount of those concentrations. The more ghrelin, the stronger the disposition to eat.

On Gregory's model, the desire to eat is the belief that one has a normative reason to eat. The disposition to eat is influenced by the concentration of ghrelin in the system. Consequently, we seem to be required to imagine a system where the concentration of ghrelin somehow influences the belief that one has a normative reason to eat.

One of the ways that this can happen, of course, is that the concentration of ghrelin determines one's reason to eat which, in turn, determines one's beliefs about one's reasons to eat. However, if we get a reason to eat from the concentration of ghrelin itself, there doesn't seem to be anything else for the belief that one has a reason to eat to do. We can simply go with the reason to eat (the desire to eat) that the ghrelin produces.

Imagine if there was a drug where, with a small dose of the drug, you would believe that the surface of the sun has a temperature of 3000 degrees. However, with higher concentrations of the drug, you come to believe that the surface of the sun has a higher temperature. At double the concentration, you believe that the temperature of the surface of the sun is 6000 degrees.

There is another problem with this thesis, given Gregory's defense of his thesis. The concentration of ghrelin in the system is not subject to evidence. There is no logical proof that entails, as a matter of deductive reasoning, a different concentration of ghrelin and, with it, a different belief about the importance that eating has to the agent.

These, then, three problems with Gregory's account of appetites. First, there seem to be a lot of large gaps between what an agent believes he has reason to do (has reason to eat or to drink) and what he desires to eat and drink. Second, Gregory needs an account of the relationship between concentrations of ghrelin in the system and the belief about how much of a reason one has to eat - which is an odd type of relationship to have. Third, concentrations of ghrelin in the bloodstream is not subject to evidence.

An alternative option - the option that I would favor (the assignment theory of desire) - says that the desire to eat is the assignment of a value indicating the importance to the agent of the proposition "I am eating" being true. The amount of ghrelin in the system can simply be associated with the value assigned to the proposition being true. The more ghrelin, the higher the assigned value. The desire itself provides the reason to - and the motivation to - eat.

Appetites cannot, in fact, be easily reduced to what an agent believes he has a normative reason to do.

On Desire 2018. Part 43: Dual Direction of Fit

For the next few posts, I would like to address what Alex Gregory took to be the main objections to his thesis:

To desire to φ is to believe that you have normative reason to φ.

(Gregory, Alex, (2017), “Might Desires Be Beliefs about Normative Reasons for Action?” In Deonna J. & Lauria F. (eds). The Nature of Desire. Oxford University Press.)

It will be instructive to see if those objections are also objections to the assignment thesis that I am defending:

To desire that P is to assign a value V to the importance of ‘P’ being made or kept true.

The first objection Gregory confronts is that of belief theories generally - that they have the wrong “direction of fit”. Recall that the standard distinction drawn between beliefs and desires is that, “If you believe that P and ‘P’ is false, change your belief. But if you desire that P and ‘P’ is false, change the world.” Beliefs have the wrong direction of fit for desires.

Gregory’s response to is to say that desires have both directions of fit at once. It is a belief that one has a normative reason to change the world - that the world ought to be changed. This has both a mind-to-world direction of fit (if one believes that one has a normative reason to change the world, and one does not, then change the belief), and a world-to-mind direction of fit (if one believes that one has a reason to change the world, and it is true, then one should change the world).

As I have been arguing, this works well for means, but means are mixtures of ends and beliefs. A belief about means is a belief that an action would serve an end. If one believes that an action will serve an end, and it is false, one should change one’s belief. If one believes that an action will serve an end, and it is true, then one should perform the action.

However, the desire that picks out the end cannot be understood in this way, and this is the type of desire that provides the foundation for all reasons for action. What is it to believe that the end is something one has reason to bring about? What is it to believe that spending time with one’s children is something one has reason to do? What does it take for the belief that one has this type of reason to be true?

I want to stress how important this is. All means-statements are statements are “ends plus beliefs” statements. In the distinction between desires and beliefs, all desires are ends. All value, all reasons for acting, all motivational force comes from these ends. If one does not have a theory of ends - if all one has is a theory of means - then one has an account of only the belief side of the equation, and is saying nothing about the desire side. An actual theory of desires must be a theory of ends - that which gives value, reasons for acting, and motivational force to means.

Assignment theory of desire is a theory of ends. Evolution, environment, and experience combine to assign values (importance) to certain propositions being made or kept true. These assignments are basic biological facts. Once an assignment is given to an end, the agent has a motivating reason to realize that end. Beliefs combine with ends to determine means - the route to take, the action to take, to realize the ends to which evolution, environment, and experience has assigned a value.

When it comes to ends, there is no mind-to-world direction of fit. There is nothing in the world that an end needs to match to be correct or incorrect. Desires can be good or bad - just as anything can be good or bad - according to whether it tends to fulfill or thwart (other) desires. But desires cannot be correct or incorrect like a belief can, because there is nothing in the world for them to match.

The fact that means have a mind-to-world element (in virtue of the mind-to-world element of the beliefs that make up means), does not imply that ends have such an element.