Monday, August 13, 2007

Partial Values

Today, I wish to deal with Atheist Observer’s proposed counter-example to my claim that no set of beliefs entails a desire.

Atheist Observer proposed the following:

John believes Mary is a beautiful, sexy woman. John believes Chez Pierre is a good place for a romantic dinner. John believes he enjoys romantic dinners with beautiful, sexy women. John has a desire-as-ends to have a romantic dinner with Mary at Chez Pierre.

If you wanted to change John’s desire as ends for this dinner would you: a) try to condemn, embarrass, or humiliate him about it. b) Reason with him and explain Mary hates men, can’t stand French food, and Chez Pierre is actually noisy, crowded, and ridiculously overpriced.

Unmentioned Desires

In this example, we are to assume that, in addition to the premises that Atheist Observer put in his original argument, that the following premises are also true.

(1) Mary hates men.

(2) The Chez Pierre is noisy.

(3) The Chez Pierre is crowded

(4) The Chez Pierre is ‘overpriced’

However, this account also contains several unmentioned desires-as-ends. These desires-as-ends are necessary for these beliefs to have the motivational power they do. Specifically, Atheist Observer’s example buries the following assumtions:

(5) John has no desires-as-end to form a relationship with somebody who hates men.

(6) John has an aversion-as-end to noisy restaurants.

(7) John hates crowds

(8) John has an aversion to paying too much for food.

Note: John’s aversion to paying ‘too much’ for food is actually redundant. ‘Too much’ is that amount that John has an aversion to paying. Also, more so than with the other examples, the aversion to paying too much can either be a desires-as-end, or it can be a desires-as-means of doing other things with the money.

Whatever the details turn out to be, we have a case where we can assume that John had additional desires-as-ends in play, and that the additional beliefs merely pointed out the relationship between the state of affairs John was creating and his full set of desires-as-ends.

Specifically, note that the proposition “The Chez Pierre is noisy and crowded” by itself does not motivate John to either go to the restaurant or to refrain from going there. His motivation depends on more than a belief about the crowds and noise. It also depends on his attitudes towards crowds and noise.

Perhaps John loves noise and crowds. Perhaps crowds make him feel alive and connected, while he finds quiet, restful places to be depressing. Or, perhaps, he is indifferent. Any belief that the Chez Pierre is noisy and crowded would affect him differently, depending on his desires towards things that are noisy and crowded. The belief itself carries no recommendation to go or to refrain from going.

John might well pick the Chez Pierre because of the noise and crowds.

In fact, every belief that you can feed to John, his attitude towards that belief will depend on his desires. The beliefs themselves are motivationally neutral. The motivation to preserve or to avoid a state in which that belief is true depends on desires.

[Richard: I will get to your points on this matter tomorrow.]

Partial Value

Assume that you have a column of 12 numbers, and you take 3 of those numbers at random and add them up, what are the chances that the sum of those three numbers would equal the sum of all 12 numbers?

It is possible that the sums would be the same (since we are allowing for the possibility of zero and negative numbers), but it is not likely. It is certainly not guaranteed.

This is what is happening in Atheist Observer’s example above. He starts his example by relating a case in which a state of affairs relates to subset of the agent’s desires-as-ends, then he brings in the remaining desires-as-ends and shows that the resulting new sum is different than the original sum. We have shown this to the person by affecting his beliefs – by showing that his original beliefs about the relationship between that state of affairs and his desires-as-ends were mistaken. However, none of this shows that beliefs have the power to entail a change in desires as ends. It simply shows the capacity to learn new facts relating states of affairs to desires that already existed but were not mentioned.

Desire utilitarianism says that a person acts to as to fulfill the more and the stronger of his desires, given his beliefs – and that false or incomplete beliefs may cause a person to act in ways that will not fulfill the more and stronger of his desires. Atheist Observer’s example is one in which an agent had incomplete beliefs and, as a result, considered acting in a way that would not have fulfilled the more and stronger of his desires. By better informing him about how the relevant state of affairs compared to the more and stronger of John’s desires, we have shown John that his original estimate is correct.

The Cost of Precision

I have argued in the past that the cost of information is high, and because of this cost we must sometimes take shortcuts.

Assume that somebody gives you a long list of numbers and tells you that he wants to know the total in, say, 20 seconds. You need to get as close as possible. What would you do?

Personally, I would start adding from left to right, adding the first digits in the largest numbers first and add them to get an estimate. After I got done with the highest column I would move to the next highest. I would give the person asking the question as much precision as I could. Yet, the numbers that I left unconsidered would still have the power to change the total, perhaps from a positive to a negative number.

It makes sense for people to do the same when evaluating whether a state of affairs will fulfill the more and stronger of his desires. He will compare the state of affairs to his stronger desires to get an estimate of its value, and learn not to sweat the small stuff. The estimated cost of error, at some point, becomes less than the estimated cost of acquiring further information. We all act – we all must act, on limited information.

This does not mean that desires unconsidered will not affect the agent’s choice. Desires (like forces) have an affect wherever they exist. The point that I am getting at here is that, if the desires are weak (or significantly weaker than the other desires in play) it is simply not worthwhile to an agent to find out more about whether a state of affairs will fulfill or thwart that desire – not if it costs him to the opportunity to fulfill some stronger desire.

Yet, in this case, the information that John did not seek becomes available. That information changes his mind, not because it generates new desires-as-ends. It is because the information tells him about relationships between the actual state of affairs and existing desires-as-ends that he had not taken the time to consider.

None of this shows that a change in beliefs entails a change in desires as ends. It only shows that a change of beliefs can change beliefs about how a state of affairs relates to the desires-as-ends that already exist.

Of course, one would use reason to point these types of errors out to a person – to point out, “Hey, John, I know that your desires-as-ends includes and aversion to noisy crowds. Let me tell you, Chez Pierre is a noisy and crowded place. It’s relationship to the full set of desires-as-ends that you already have is not what you think it is.”


Atheist Observer added a second point to his comment that I would also like to address.

My point is that operant conditioning may work in rats, chickens, and dogs, and to some degree in small children, but as a method to change desires as ends in adults it effectiveness is marginal at best. “Clockwork Orange” was just a movie.

The precise degree to which conditioning works on adults is subject to further research. I agree that, as one gets older, its effects weaken. However, I see no reason to assume that humans are any different from animals in this way. It is also true in animals that, “You can’t teach an old dog new tricks.” Or, at least, it takes more effort for little gain to the point that it may not be worthwhile to try. Yet, the teaching of new tricks is not impossible and, if there is a sufficiently large gain to be had, then there is reason to try.

However, I allow that the greatest influence of these social forces rests with their effect on children. This is why I have argued in the past that the lessons of atheists ‘coming out’ and, in particular, in expressing certain moral principles should be done in the presence of or, better, while directly speaking to children.

At some point, it takes so much effort for such a little change in a person’s desires that it is not worthwhile to engage in that activity. We may decide to simply give up and ignore Grandpa’s racist rants, knowing that Grandpa was a product of another age and, with just a few years of life within him, it simply is not worth our while to try to change him. We may decide to quietly put up with Grandma’s religious-based moral attitudes until she passes away, while restricting her opportunity to infect her grandchildren with primitive superstitions.

Yet, none of this argues against the sensibility of using these tools where they do have an effect, and the cost of using them is less than the benefit that can be expected from the changes they bring about.


Martin Freedman said...

Here may not be the specific post this question is for but I think it is relevant to much of the discussion of the last few days.

For the purposes of this question, lets us grant that condemnation, ridicule and punishment are effective in altering malleable desires in the recipient. That is let us focus on where it actually has worked. Now the question is how does this work?

Presumably the recipient experiences something like remorse, regret, guilt, shame or embarrassment and as a result of one or more of these modifies their desires-as-ends?
If so how do these emotions map to belief and desire? First they all appear to be in the class of negative desires, the prototypical example being fear. Fears are alleviated or not. Fear is a propositional attitude to keep or make the proposition false. This is a form of desire negated and trivially reducible to desire. I hope I have this correct.

The issue with my list of emotions above is that they are not only negative desires but backward looking. In a forum someone asked me

To take the Sartrean example if I am aware of being seen spying through a keyhole and I experience shame what exactly is it that I desire?

My answer that I am not entirely happy with was: "In this case shame is an emotion or feeling that comprises a desire and a belief. The belief, B, is that certain of one's actions were blameworthy and the desire, D, is for one's actions not to be blameworthy. To expand on example the initial belief is of the form that one has been apparently (A) caught (C) spying (S) for some purpose (P). (A) allows for the possibility that one was mistaken but one can feel shame due to false beliefs. There are many variants of what B could be, one is that C is blameworthy (and not S or P). Many others can be derived of course. Regardless B is conjoined with D to generate the feeling of shame."

Over to you :-)

Anonymous said...


As I see it:
1) John had a desire-as-end to have a dinner with Mary at Chez Pierre.
2) I provided John with facts and reasoning that caused him to change his beliefs about Mary and Chez Pierre. (Whether I mentioned other desires of his or not relating to these beliefs is not really relevant. I am not interested in changing those desires.)
3) After John considers the new facts and their combined impact on his many desires, he no longer has a desire-as-end to have dinner with Mary at Chez Pierre.
Unless you dispute that the dinner desire was a desire-as-end, or that his ceasing to have this desire does not constitute changing a desire-as-end, then I have shown it is possible to change a desire as end through changing beliefs.
I am not arguing that these beliefs are unrelated to his other desires. Nor am I saying beliefs alone have motivational power. But I am saying virtually all desires beyond those relating to internal states of satisfaction or pain are combinations of other desires and beliefs. And one generally has a better chance of changing the related beliefs leading to a particular desire through facts and reasoning than through the crude tools of operant conditioning.
Most research shows punishment (condemnation, ridicule, or physical punishment) is not a particularly effective form of behavior modification. Why? Because there are so many other responses to it than the one you wish. Agression, denial, counterattack, are very common. If the tool you are using produces more bad consequences than good ones, it’s time to find a better tool.

Anonymous said...

>> After John considers the new facts and their combined impact on his many desires, he no longer has a desire-as-end to have dinner with Mary at Chez Pierre.
Unless you dispute that the dinner desire was a desire-as-end, <<

I don't wish to talk for Alonzo, but as I understand it, I believe this would be the exact dispute. "To have dinner with Mary at Chez Pierre" sounds very much like a desire-as-means. If I were forced to guess, I'd guess that the true desire-as-ends is to form a satisfying relationship with desirable member of the opposite sex. And this dinner with Mary at Chez Pierre is but one of the means among many to eventually achieve that end.

Alonzo Fyfe said...

John did not have a desire-as-end to have a dinner with Mary at Chez Pierre.

John had a desire-as-end to have a romantic dinner in a quiet and reasonably priced restaurant with a beautiful woman who does not hate men.

Unfortunately, he (wrongly) believed that this dinner with Mary at Chez Pierre would be a romantic dinner in a quiet and reasonably priced restaurant with a beautiful woman who does not hate men.

You corrected his beliefs.

You did not correct, or even change, his desires-as-ends.

Anonymous said...


Nice argument, but your position about changing desires by praise and condemnation is just as subject to it. A child has a desire to do something you consider bad, so you condemn him and he stops showing that desire. You may say you have changed his desire, but you can just as easily say he had a stronger desire not to be condemned and you only made him aware that if he did that he would be condemned. Therefore he acted on his stronger desire. You did not change his desires at all.
Even if he does not do the bad behavior when you are not around, that is no proof, since he could still be motivated by the fear of your condemnation. You applied the condemnation to an act, but you did not create or change and desire-as-end, only a desire as means.
You can’t have it both ways. You can’t just say “well if it was changed by facts and reason, it wasn’t a desire-as-end, but if it was changed by praise and condemnation, it is.

Anonymous said...

Refering to how condemnation, praise, and ridicule change desires, Martino said:

"Presumably the recipient experiences something like remorse, regret, guilt, shame or embarrassment and as a result of one or more of these modifies their desires-as-ends."

This does indeed seem to underly what Alonzo argues. The negative emotions induced by ridicule and condemnation serve to weaken desires while the positive emotions that arise from praise strengthen desires. If not this, I'd like to hear what it is about condemnation etc. that influence desires.

The emotional explanation makes sense to me. In fact, I think all people, even Alonzo, the most rational person I know, are mostly motivated by emotions and what I guess would be called "gut" feelings.

However, I would argue that information gained from observation, education, real life experiences, reasoned analysis, etc. also has an emotional impact on us and can therefore also alter desires. I contend that the interplay of the various beliefs we hold, both conscious and subconscious, influence our emotions and desires.

Condemnation is a form of information. It, for example, is information that "Jenny does not approve of what I'm doing." Jenny's disapproval can certainly have an emotional impact, and affect our desires.

But why couldn't this information have similar effects: "Jenny has proven to me that what I'm doing is wrong."?

On a slightly different note, do therapists use condemnation, ridicule, and praise to help their clients change their behaviors? Perhaps to some degree, but I do not think these are the main tools they employ.

Finally, surely there has been research into what methods work to truly change people, just as there is research about how to change the behavior of lab rats. So far, this has all been a lot of armchair reasoning, which is fine, but how does it square with research? Is anyone aware of such research?

Uber Miguel said...

There are also therapists that like to invoke god(s) and even go so far as to prescribe seances and exorcisms.. as you say, research along these lines would be nice. My fear is that most research doesn't approach things from a desire utilitarian perspective but instead rely solely upon traditional concepts of illness and mental disease.

I see this as quite similar to the issue of relying on doctors to properly diagnose and prescribe for suicidal individuals - tending to focus on interventions of chemicals and discussions of their ego instead of directly weighing their scenario and conflicts/confusions in the various desires within their life.

At least, I've never heard of such a perspective ever being mentioned in any medical article. Perhaps we're breaking new medical ground in terms of research potential.. although I'd love to be proven wrong and see some studies that address these topics - to confirm or deny their usefulness.

Alonzo Fyfe said...

All people (even me) are motivated by desires. Beliefs and rationality only choose to pick out the means to those ends. Desires pick out the ends.

The oddity of a 'Spock' who devotes himself purely to logic is that he would have no reason to act. For him to act, he must identify an end, and for every end identified he must care about that end in order to have a reason to realize it. Otherwise, no end survives the shrug test. (The 'So what?' test.)

I am going to leave it to professions to discuss how these forces work. I think that there is enough evidence of them working - from the establishment of roles in a pride of lions or a pack of wolves, to cultural differences among humans, to show that they do work. Even primative humans would have recognized the value (in terms of peace, security, and cooperation) of promoting useful desires in others and inhibiting dangerous desires.