bd-from-kg

Alonzo Fyfe:

Here are some comments on your second July 30 post.

1. On desires that do not and never will exist

But your theory already takes into account desires that do not exist and never will exist!

First, it takes into account desires that might (so far as we can tell) exist in the future, weighted by the probability that they will come to exist. But many of these desires never will in fact come to exist. So your �desire-calculus� includes desires that do not exist and never will.

Second, your theory counts only �good� desires, which are defined as desires whose fulfillment tends, in general, to fulfill other desires. Let�s look at what this actually entails. Say we�re considering whether to �count� a desire D. We decide to do so because desires of this type tend to fulfill other desires. But that means that we�re actually taking into account a whole class of desires which have the property that fulfilling D might hypothetically fulfill some or all of them, even though in this particular case it may be that it doesn�t actually fulfill any of them. In other words, we account D a �good� desire on the basis of a class of desires (which could in principle all be purely imaginary) that fulfilling D could in some sense (but in fact won�t) fulfill. Thus desire D gets �counted� in your desire-calculus because of a number of desires that do not exist and never will exist � or at best, have no actual relationship whatever to the act that we are evaluating.

So clearly according to your theory desires that do not exist and never will exist are not irrelevant. They are often included directly in the desire-calculus, and they are routinely used to determine which actual desires get �counted� by virtue of being �good� desires.

Here�s one possibility. If you were actually to come to own such a coin, possessing it might create an aversion to possessing an ounce of 99.99% pure gold. But since you are never actually going to own such a coin, such an aversion does not exist and never will exist. Just the same, this potential aversion should be taken into account in deciding whether you ought to own such a coin. In fact, perhaps it was taken into account when you decided not to buy the coin. If so, you were quite right to take it into account, even though as a result of your decision the aversion never came into existence.

Nope. That�s not what your theory says. In fact, fulfilling a certain desire might not only be compatible with fulfilling other desires that do or will exist; it might actually fulfill a great many desires that do or will exist (far more, in fact, than any alternative), yet still be accounted �bad� (according to your theory) because it is of a type that tends, in general, to frustrate the fulfillment of far more desires than it fulfills.

You can�t have it both ways. Either you do a straightforward summing up of desires actually fulfilled by the various alternatives and say that the one you �ought� to do is the one with the greatest aggregate desire-fulfillment, or you start fiddling with this scheme in one way or another, in which case the alternative that fulfills the most desires will often �lose�. Claiming that you�re doing the one when you�re actually doing the other is false advertising. You�ve modified the original scheme to overcome certain powerful objections. Fine. But the modification has introduced its own problems and is thus subject to other objections. You can�t go back and forth between the two schemes and claim to have met both sets of objections simultaneously.

2. On slavery

There are two problems here.

First, it seems clear that Lincoln was not talking about the results of a desire-summation calculation. After all, it often happens that actions that thwart some desires are nevertheless right. If this were what he had wanted to say, he would have said something like, �All in all, slavery results in the thwarting of more desires than it fulfills�. Or, more in keeping with the language of the time, �Slavery does not produce the greatest good for the greatest number.� What he was actually saying was much more like, �If one imagines vividly what it would be like to be owned by someone else � to be subject to his every command, to have no freedom to do what one wants, live where one wants, marry whom one wants � one will find the idea of subjecting anyone to such an experience, or of willingly allowing anyone to be subjected to it, intolerable.� In other words, he was talking about what one�s attitude would be if one had the kind of vivid, immediate knowledge and understanding of what it means to be a slave that one would attain by actually being a slave. The clear implication is that anyone who clearly understands that he would be opposed to slavery under these conditions should, as a rational being, be opposed to slavery regardless of whether he actually has such a vivid, direct understanding of the matter. Knowing that you would oppose it if you had enough K&U compels you, as a rational being, to oppose it now.

This might be clearer if we imagine that a number of our friends, who are very similar to us in all relevant ways, had been taken into slavery for a time and as a result, upon being freed, were passionately opposed to slavery even though they had been indifferent to the matter before. This should induce us, if we are reasonable, to believe that we too would be passionately opposed to slavery if we had the knowledge and understanding of it that our friends have acquired, and therefore that we too should oppose slavery regardless of whether we fully understand the reasons for our friends� opposition to it.

The point here is that this corresponds far more closely to the way people actually think and talk about moral questions than your �desire-calculus� model.

The second problem with your explanation is that it�s not at all clear that a �desire-fulfillment� calculation would in fact come out against slavery in all cases. Your new wrinkle of including only �good� desires doesn�t really help here. A desire to own slaves under certain specified conditions will only count as a �bad� desire if fulfilling it tends to thwart more desires than it fulfills. But of course this is the very question at issue: does owning slaves, under the specified conditions, tend to thwart more desires than it fulfills? If not, then a desire to own slaves under those conditions is a good desire[/i]. And oddly enough, under these conditions a desire not to be a slave is a �bad� desire, and thus does not �count�. (This illustrates a serious problem with your �modified desire-calculus�: There�s a strong tendency to count as �good� just the desires that lead to results that you have decided in advance are �good�. And once you�ve done so, the �desire-calculus� is guaranteed to come up with the desired result, since you�ve removed from consideration all desires that might give the �wrong� result by labeling them �bad� desires.)

Now I don�t want to get into the question of whether there are conditions in which slavery fulfills more desires than it thwarts. The point is that it�s far from self-evident that there aren�t any such conditions, and if there are, your desire-calculus model compels you to conclude that under such conditions a desire to own slaves is a �good� desire (provided that it is limited to a desire to own slaves under those conditions), and a desire not to be a slave (on the part of those who �should� be slaves according to the desire-calculus) is a �bad� desire. And of course, actually owning slaves is �good� and attempting to end the institution of slavery is �bad�.

So your theory compels you to say that the �wrongness� of slavery is contingent on certain very complicated and controversial claims about human psychology, and that if these claims turn out to be false for any given society, slavery is right in that society; that many people in that society will be acting wrongly if they refuse to own slaves, and that slaves will be acting wrongly if they resist being slaves � e.g., by trying to escape or rebel.

Whereas my theory says that if a person with sufficient knowledge and understanding of slavery (for example, by having the experience of being a slave) would be opposed to slavery, then slavery is wrong, period, no matter how some hypothetical desire-fulfillment calculation comes out.

I greatly prefer my theory. But more to the point, I think that this result (and the way I arrive at it) corresponds far more closely to how most people reason about such things.

3. On the �young Archimedes�

�Happiness� is not the name of a specific mental state, but a generic name for all mental states that a person finds (or would find) desirable. If we allow such perfectly generic, unspecific �desires� to figure into the �desire-calculus� it becomes completely meaningless. If, for example, we say that someone who has never even heard of chess �really� desires to become a grandmaster, or that someone who has no idea that stars are anything but points of light in the sky �really� desires to make a breakthrough discovery in the theory of stellar evolution, or that someone who has no idea that there is such a discipline as mathematics �really� desires to discover some fascinating new theorems in algebraic topology, we will have redefined the term �desire� to the point of making it unrecognizable.

Thus, if the �desire-satisfaction� criterion is not to dissolve into total mush, we cannot allow such things as a �desire to be happy� to enter into it. Since we are made happy by fulfilling desires, saying that what we desire is to be happy amounts to saying that what we desire is to fulfill our desires, which is nonsense.

Another major problem with the �desire to be happy� counter is that Huxley�s �young Archimedes� might well have had other potential desires that had nothing to do with his own happiness (aside from the fact, common to all desires, that fulfilling them would have made him happy). For example, he might have wanted to derive elegant proofs of new theorems for the delectation of other mathematicians, or to provide new insight into the structure of plane geometry. And of course, he might have had practical applications in mind such as developing efficient techniques for finding solutions to seemingly intractable computational problems. Unless you�re prepared to defend psychological egoism by arguing that all desires are ultimately desires for one�s own happiness, there is no way to argue that the boy already has these desires.

The point is that there are desires that people do not have, but would have if they had sufficient K&U. In some cases, whether these desires ever come to exist is contingent on All desires are not instances of a single all-encompassing desire, nor is every desire we have a desire that we have always had. A reasonable theory must take the development of new desires based on experience into account.

The AntiChris

Alonzo

I'm sure you haven't forgotten this thread but I I'm interested in seeing your response to bd.

Chris

Alonzo Fyfe

[Shrek]It's on my to-do list.[/Shrek]

The AntiChris

Ok.

Alonzo Fyfe

My formal debate opponent has decided to give me a weekend free, so I am posting a response to some of the issues raised by bd-from-kg.

(You know, bd, you could save me a lot of time if you would, instead, make basic, simple errors that I could respond to in just a few sentences.)


Rationality: Means and Ends

Let me begin by focusing on a subject the subject of rationality. You write:

. . . commitment to a strategy for maximizing knowledge and understanding makes no sense unless it flows from commitment to the principle that what one would do if one had sufficient K&U is "better then" or "preferable to" what one would do if one did not.

As we have discussed before, the only type of rationality that I think makes sense is means/ends rationality. Rationality picks the means for getting to A, but rationality does not pick A. A is picked by desire alone, and rationality has no impact.

Everything you say about rationality is true of means/ends rationality. "Which means should we select to obtain A?" The answer is, I think quite obvious, "We should pick those means that a person with sufficient knowledge and understanding would pick." And I would even fully endorse your definition of "sufficient knowledge and understanding" as "that level of K&U beyond which any additional K&U would not yield a different result.

I can explain why knowledge and understanding is relevant to means/ends rationality. If you are trying to get to A, and you have insufficient knowledge and understanding, you may pick a route that will not get you to A at all. Indeed, you may pick a route that will take you to not-A (thus you are acting as your own worst enemy).

But I do not see why or even how one would use this to select ends. You can only ask the question, "What's the best way to get to A" is you have already picked an A to get to. To ask the question, "What's the best way?" without an assumed A is a nonsense question. So, we have to assume an A before this type of question even starts to make sense.

Your proposal, "The best A is the A that a person with sufficient knowledge and understanding would pick," leaves me to ask, "By what standards is this person picking A?" It is like saying that the best route is the route that a person with sufficient knowledge and understanding would pick, but we have no idea where he wants to go. Without that information, there is no reason for the person with sufficient knowledge and understanding to pick ANY route. And if he does pick a route, it is entirely arbitrary.

In saying this, I must add two clarifications.


The Desire and Its Fulfillment

One clarification is that K&U is important for determining whether a certain state of affairs have the properties that will fulfill the desire. It is not relevant in picking the desire itself, only the object.

So, for example (I think I used this before), grandma wants to preserve grandpa's ashes. K&U is not relevant to the desire to keep the ashes. However, it is relevant to identifying the urn in which the ashes actually reside. If there were a mixup at the funeral home, for example, Grandma may well be preserving an urn that keeps somebody else's ashes, while her husband's ashes were scattered over somebody else's favorite fishing spot.

So, when I say that K&U is not relevant to ends, it is not relevant to grandma's desire to preserve grandpa's ashes. Yet, it is still relevant in determining whether the keeping of this urn with these ashes fulfills her desire to preserve grandpa's ashes.


Ends as Ends vs. Ends as Means

K&U is not relevant to the determination of ends as ends. Yet, it is fully relevant in the determination of ends as means. The desire to smoke, as an end in itself, is not subject to review by the board of K&U. However, the desire to smoke as a means -- its ability to cause fatal or dehabilitating disease -- is fully relevant to the desire to smoke as a means. The desire itself is a bad desire. It is a bad desire insofar as a person with sufficient K&U would opt not to smoke. But it is not a bad desire because a person with sufficient K&U would not WANT to smoke. It is a bad desire because the person with sufficient K&U destroys the means by which he can fulfill other desires.


Summary

Again, I agree with much of what you say. Insofar as we are talking about the rationality of means (and even the rationality of ends as means) I agree with your analysis. However, you wish to go on and say that there is also a rationality of ends as ends. That is where I stop. The rationality of means is the only rationality that exists. There is no way to rationally evaluate something except in virtue of its capacity to serve as a means to fulfilling some other end. There simply is no such thing as a rationality of ends.


Some Ultimate Ends are Better than Others

Let me start by saying that we agree. Some ultimate ends are better than others. No argument here. However, I argue that an ultimate end is 'better' or 'worse' in virtue of its role as a means to fulfilling other ends. Your argument, I believe, has to be taken as an argument that some ends -- those that the person with sufficient knowledge and understanding would have -- are intrinsically better than others.

I believe that you have denied the existence of intrinsic value elsewhere. Yet, do not see how you can square this with the idea that some desires can be 'good' or 'bad' in a sense other than in their capacity to fulfill or thwart other desires (that is, other than in virtue of their value as a means).

You seem to be saying, at the same time, that there is no such thing as intrinsic value and that a person with sufficient knowledge and understanding would know and understand that some desires are intrinsically (independent of their relationship to other desires) better than others.

Note: At times I get the impression that you think that talking about desires means that the concept of intrinsic value does not apply -- that talking about 'the intrinsic value of desires' is self-contradictory. Yet, it is not a contradiction. Even when we are talking about desires it makes sense to ask if a desire has value intrinsically (in virtue of itself) or extrinsically (in virtue of its capacity to fulfill other desires).

My thesis is simply that desires are no different from anything else in that desires (or, more specifically, the desires that a person with sufficient knowledge and understanding would have) cannot have intrinsic value. The value of desires is extrinsic -- that desires themselves only have value in virtue of their ability to fulfill other desires.

This is the only type of value that exists.

Or, in other words, the person with sufficient knowledge and understanding would pick desires based, not on their intrinsic properties (or the mere fact that he has these desires) but based on their extrinsic properties (on the desires' compatibility with other desires).

Because the person with sufficient knowledge and understanding would realize that the only type of value a desire can have is extrinsic value.


Relevant Observations

Elsewhere in your post, you asked, What observations would be relevant in deciding this "claim"?

I get most of this from a study of action theory. This involves attempts to answer the question, "What theory best explains and predicts human behavior?" There are a number of theories available. The theory that I base my claims on is the most widely accepted, the view called "BDI Theory" (which stands for "belief-desire-intention").

It is a part of this general theory of human action that the only form of rationality is means-ends rationality. We can ask how well certain things fit certain desires, but there is no explanatory or predictive necessity for postulating that anything has value, even in part, based on anything other than its capacity to fulfill desires.

The way to refute BDI Theory (the way that its opponents work to refute it) is to say, "BDI Theory cannot handles cases such as X" (where X is a set of observations about human behavior).

[Actually, I have a confession to make. I think that what is known as Script Theory does a better job of explaining and predicting human behavior than BDI Theory. But I am no expert in the field, and the experts still prefer BDI Theory by a wide margin. Who am I to disagree with the experts? I will stick with BDI Theory until the experts themselves decide that it is to be replaced with something better.]


Consistent with this, you interpret my statement But the only 'something real' that exists is the capacity to fulfill desires either directly or indirectly, to mean the only things that are 'relevant' to whether an act is (morally) right or wrong. . .

Actually, combine these two: The only relevant things that are real is the capacity to fulfill desires either directly or indirectly.. Language may determine that other things are relevant (e.g., being favored by God or having intrinsic merit), but the capacity to fulfill desires is the only item on the list of relevant things that is real.


Common Usage

Everything you say in the paragraph on 'common usage', I agree with. Yet, none of this changes the fact that your account has a person with sufficient knowledge and understanding selecting desires based on their INTRINSIC properties, rather than their EXTRINSIC properties, and intrinsic value does not exist.


Later, you write the following:

First, obviously, I do not go with the desires that a person actually has, but rather appeal to the desires that he would have if he had sufficient K&U. And, of course, a crucial part of my argument is that these desires would in large part be determined by the fact that fulfilling them would fulfill other desires. (I have made this argument at length in other threads, so I know that you're familiar with it.

Well, your first point is a bit of a misinterpretation. The passage you quoted does not say the desires that "a person actually has", but the desires that "the person actually has" where "the person" in a paragraph where I specified that "the person" is "the person with sufficient knowledge and understanding."

Given the trouble I got into defending your thesis from somebody else who made this mistake, I think I deserve a bit more credit than this.

As for your second point, you hold that a person with sufficient knowledge and understanding would, by that alone, disapprove of such things as slavery.

And I countered that a person, with sufficient knowledge an understanding, may know that an aversion to slavery is good relative to all other desires, but may still not have the desire. Just as he may know that a low-fat/high fiber diet would be the best food for him to eat, but still have a desire for high-fat foods.

The knowing of the value of a desire is one thing, the having of the desire is something else entirely, and neither implies the other, either logically or by the laws of nature.

If you agree that these desires "would in large part be determined by the fact that fulfilling them would fulfill other desires," I would simply change "in large part" to "entirely". A person with sufficient knowledge and understanding would come to realize that there is no other part to the value of a desire.


Resolving Differences

Where I write, An appeal to discourse cannot resolve the differences between us . . ., you write, I see no reason why it cannot.

The reason is, ultimately, that the difference between us is whether desires can have intrinsic merit. I hold that all of the value of a desire is extrinsic -- based entirely (not in large part) on its relationship to other desires. When you say in a large part this means that something other than the extrinsic value of a desire exists. I say 'nay'. The extrinsic value of a desire is the only part that exists. There are no other parts to look for.

Now, you do add, First, some desires are desires for ultimate ends. Fulfilling them might not satisfy any other desires at all, but might still be desires that a person with sufficient K&U would have.

Once again, unsurprisingly, I agree fully. This is true. Our desire for sex and our aversion to pain are desires that we fulfill for their own sake. We can acquire others -- a desire to listen to Mozart (for you), or a desire to listen to Jimmy Buffett (for me). But we must distinguish between the value of the desire, and the value of its object.

Because I enjoy listening to Jimmy Buffett music, listening to Jimmy Buffett music has value (for me). But the desire to listen to Jimmy Buffett music has no value. Listening to the music has value (because it is desired). Desiring to listen to the music has no value -- it simply exists. It is an impotant distinction.

The desire, simply exists. If, by some blow to many Margaritas, or a stroke induced by all of those Cheeseburgers in Paradise, my brain wiring were to change so that I now valued Mozart rather than Jimmy Buffett, I would be no better or worse off because of it. The desire to listen to Mozart has no greater or less value than the desire to listen to Jimmy Buffett. Though, after the change, Mozart would acquire the value to me that Jimmy Buffett used to have.

On the other hand, if I were to lose my desire for chocolate, or my aversion to talking to strangers (shyness), these would be good things because these desires are bad.

The existence of a desire gives value to its object. But the desire itself has value only in terms of how that desire stands as an object to still other desires.

If you have a desire that other people have a desire to learn classical Greek (even if it leads them to frustration), then other people having a this desire has value to you, in virtue of your desire that they have it. But, in the absence of your desire, it has no value (to you).

I think, given some of your comments, that you take me as saying that desires are either good (if they fulfill other desires) or bad (if they do not). But this is not a yes/no, good/bad, black/white distinction. A desire, that fulfills no desires but thwarts none, is not 'bad' in virtue of its inability to fulfill other desires. It is 'neutral'.

Furthermore, 'goodness' and 'badness' comes in degrees. That which fulfills more and stronger desires is better than that which fulfills fewer and weaker desires; that which thwarts more and stronger desires is worse than that which thwarts fewer and weaker desires.


Circular Reasoning

And finally (at least for this part) we get to this:

Rating desires on the basis of now well fulfilling them fulfills other desires is ultimately incoherent; it leads to an infinite regress . . .

Yes it does, but not in a problematic way.

It is a form of definition that, in Mathematics, is called a recursive function. Epistemologists call it 'coherentism', linguists call these 'hermeneutic circles', and logicians (when they are not using the mathematical term 'recursive function' instead use the term 'virtuous circles'. In Ethics, John Rawls calls this 'reflective equilibrium'.

Of these, I think, language provides the best illustrative example. Each word in the dictionary is defined in terms of other words, which in turn are defined in terms of still other words, which in turn are defined by still other words, some of which reflect back on the original word being defined. Is this an incoherent infinite regress? No, it is a recursive function.

In logic and mathematics, there is a great deal of very complex work being done on the properties of recursive functions. Not only are they not viewed as a problem, they are considered an essential part of science, math, and logic.

Alonzo Fyfe

bd-from-kg

This entry discusses the purpose of moral discourse and its compatibility with the laws of nature, including the best theory of human action.


Here, it is time to address the function of morality. Yes, it is to influence action. But we must pay attention to the ways in which it is possible to do these things.

I mentioned in yesterday's post that the theory that I defend is compatible with what is called the BDI Theory of human action. BDI theory states that actions are a part of the following causal chain:

(BELIEFS + DESIRES) -> INTENTIONS -> ACTIONS

With my wife gone for the weekend, last night I went to the store to buy something to eat. I had a desire for a particular food, and a belief that I could get it at the store. I formed an intention to go to the store, which caused the muscle contractions that carried me across the bridge to the store to buy the food and then carry it home.

I also had a slight aversion to paying the money, and a slight aversion to the health-effects of the food, but it did not stop me. The latter aversion modified my selection somewhat; I selected seafood partially because I believe it is healthier for me. And the former aversion caused me to purchase the smaller (and less expensive) package of seafood.

This is BDI theory in action.

According to BDI theory, beliefs are motivationally neutral. The belief system is a database of supposed facts -- 'supposed facts' because some beliefs may be incorrect. My belief that seafood does not produce the same health risk as, say, barbequed chicken wings (my first choice in last night's shopping trip) may not be true.

Only desires have motivational force. But desires have no truth value. Desires cannot be correct or incorrect. It makes no sense to say that my desire for barbequed chicken is false.

This next point is crucial in understanding the difference between us. Because desires have no truth value, knowledge and understanding have no relevance here. Knowledge and understanding are relevant only in the realm of belief. Indeed, 'knowledge' in philosophical circles is simply defined as 'justified true belief'. 'Understanding' also implies a truth value - it implies 'getting things right'. As such, it too is limited to the realm of belief. With desires, there is simply nothing to get right.

There is a common point of confusion here. The word 'desire' is ambiguous. It identifies both ends (my desire to eat barbequed chicken) and means (I want to go to work today so that I can afford to get a new computer). Everything I say here pertains only to desire-as-ends. Desire-as-means is just a shortcut way of speaking about what are, in fact, bundles of desires-as-ends-plus-beliefs.

This does NOT imply that desires cannot be changed. Desires can be learned, and unlearned, through the process of conditioning. If you punish a person for doing X, he begins with an aversion to being punished and a belief that doing X will lead to punishment. But, over time, he acquires an aversion to X that is quite independent of his belief that he may be punished. He simply does not want to do X, even when he goes that he can do so without being punished. Using conditioning techniques, we teach children to have an aversion to lying and to taking things that do not belong to them.

Actual punishment is not necessary. We can condition humans by calling to mind unpleasant thoughts. "How would you like it if somebody did that to you?" This is a conditioning technique that associates a thought "doing unto others" with an unpleasant mental event -- an imagining "others doing unto you," as a way of conditioning an aversion-response to the former.

The difference between argument and conditioning is that argument concerns logical relationships between premises and conclusions, and conditioning concerns causal relations between stimulus and response.

"You would not like it if somebody did this to you, therefore you ought not to do this to somebody else," is not a logical entailment. Heck, it violates, 'is'/'ought' for one. But it is a causal relationship. "If I can associate doing X unto others, with the negative associations of imagining others doing X unto you, then the effect may be an aversion to doing X."

These are a few of the basic, relevant elements of BDI Theory.

Beliefs have no motivational force, and desires have no truth value. These two elements of BDI theory are crucial.

From here, I will agree with you that much of 'the logic of moral discourse' assumes that desires have a truth value. (This, by the way, is the same thing as saying that either desires, or the objects of desires, have intrinsic value.) The best of our theories of human action say that this is a mistake.

Where these types of conflicts crop up, language must yield to the theory. The fact that the word 'atom' literally meant 'without parts' gave us no good reason to throw out our best theory of the atom, which is that they do have parts. The fact that the word 'malaria' literally meant 'bad air' provided no meaningful objection to our best theory of that disease, that it is caused by a bacteria.

And the fact that the logic of moral discourse assumes that desires-as-ends have a truth value is not a meaningful objection to the most widely accepted theory of human action, which tells us that desires-as-ends have no truth value. (Which means that the concepts of knowledge and understanding are irrelevant to desires-as-ends.)




Mabel's Life's Savings

Does morality make sense in a universe where desires have no truth value?

Yes.

Even though desires do not have a truth value, they still have a value-as-means. (That is to say, their effect in fulfilling or thwarting other desires does have a truth value.) Plus, it makes sense for each of us to want to condition others to have those desires that best aid in the fulfillment of our own desires. So, we, as a society, 'condition' each other into having 'good' desires -- those desires that aid in the fulfillment of other desires.

We condition each other to have an aversion to lying, to taking things that do not belong to us, to killing innocent people, because there is some truth value to these desires being good-as-ends, but because there is some truth value to these desires being good-as-means.

So, we don't want Fred to take Mabel's life savings. How do we prevent this?

Well, we should start young. We would want to condition Fred while he is growing up to have an aversion to taking things that do not belong to him. We let him know of our disapproval of people who do such things. (Fred, like most children, likely has a fear of disapproval that can be used to condition him into having an aversion to taking things that do not belong to him.) We invite him to imagine, "How would you like it if somebody did that to you?" And, we punish Fred when we catch him stealing things.

Perhaps Fred, as an adult, does not have this aversion to taking things that do not belong to him. We may be able to find some substitute aversion that we can use to prevent him from taking the money. Perhaps he has a more general aversion to doing something wrong. In which case, if we can convince him that taking the money would be wrong, we can cause him not to take the money. By "wrong" we mean, "Something that a person with good desires would not do," which in this case it is something that a person with an aversion to taking things that do not belong to him will not do.

But, to succeed, we must tie the taking of the money in with some existing desire. Otherwise, we will not be successful.

If �moral discourse� suggests otherwise, then the fault lies, not with the BDI theory, but with the false assumption included with moral discourse that desires have a truth value.

Failing this, we have a third way to tie not-taking the money with an existing set of desires. We tell Fred, "If you touch that money, I will make sure that you regret it." But, then, we're not talking about morality any more. We are talking about the law.



Ought implies can

Your statement that "a being who is completely impervious to moral arguments . . . is simply not a moral agent" ultimately begs the question. A person impervious to moral arguments may still be pervious to moral conditioning.

The standard compatibilist definition of "could have done otherwise" is "would have done otherwise if he had wanted to," and that he 'could have wanted to'. The 'could have wanted to' does not impose any limits on HOW he could have wanted to -- it gives no priority to moral arguments as opposed to moral conditioning. So, to use this as a mark in favor of moral arguments would beg the question.

It is not legitimate to blame a person for failure to do something that even a person with good desires could not have done. You can't blame a person for failing to teleport a child out of a burning building, for example. Because even if he had wanted to, there was nothing he could have done.

As for humans being moral agents, and not cats, I believe this is ultimately tied to the habit of thinking of humans as having a soul, but not animals. And having a soul is essential to being a moral agent. Many of the arguments against human cloning make reference to the idea that the clone would not have a soul, and thus not be able to tell the difference between right and wrong.

Get rid of the concept of a soul, and we can start to question the habit of not attributing moral statements to animals. My cat, Tsunami, is evil. At this very moment he is tormenting his sister, chasing her around the house. This is desire-thwarting, and not something that a cat with good desires would be doing. My wife and I have tried those methods of moral conditioning available to us -- which I like to think have been successful to some extent.

Still, I cannot condition my cat by saying, "How would you like it if your sister did these things to you?" Regardless, to whatever degree my wife and I can alter his desires through moral conditioning, to that degree (and no more) it still makes perfectly good sense to call him a "bad kitty."

Now, please excuse me while I engage in some moral conditioning in order to alter the desires of my bad kitty.