Wednesday, March 21, 2007

Oughts and Desires

In the last couple of weeks, I have had three people make comments suggesting that I explain the relationship between ‘ought’ and desire - including a line in a post by eenauk that he is not sure how I link my desires and my oughts.

Therefore, I shall see if I can clarify the issue.

Though I fear that I am cutting my own financial throat with respect to selling copies of that book I wrote. Why should anybody buy the book if I keep giving away all of its arguments?

Actually, I will start with ‘should’. ‘Ought’ is a species of ‘should’ and it is easier to understand the species by first understanding the genus.

Should and Reasons for Action

‘Should’ = ‘There are more and stronger reasons-for-action in favor of doing X than against doing X.’

‘Should’ is intimately related to reasons for action. If there are no reasons for action for doing something, then it makes no sense to say that it should be done. If there are more and stronger reasons against doing something, then it also makes no sense to say that it should be done. In order for it to be the case that something should be done, there has to be a reason for doing it.

Please note that, in writing this, I have ‘should’ on one side of the equation, and an ‘is’ statement on the other side. ‘Are’ is just the plural form of ‘is’. When David Hume made his famous assertion that one cannot derive ‘ought’ from ‘is’, he used the premise that ‘ought’ describes a different sort of relationship, and that one needs to explain the connection between this new type of relationship and ‘is’ relationships to make these inferences valid.

He was wrong. ‘Ought’ relationships are a specific type of ‘is’ relationships. They are ‘is’ relationships – relating actions to reasons for action. As such, there is no mystery in deriving ‘ought’ from ‘is’, as long as we ‘is’ talking about a reason for action.

So, let’s get to these reasons for action.

Reasons for Action and Desires

Desires are the only reasons for action that exist.

As such:

’Should’ (when true) = ‘Is such as to fulfill more and stronger desires in question’.

In my younger days, as a contributor to the Internet Infidels Discussion Board, I said that this equation was true by definition.

I was wrong.

People have, throughout history, postulated the existence of reasons for action other than desires. They have postulated reasons built directly into the fabric of the universe – either by God, or intrinsic to the commands of God, or in nature itself, or in the form of some sort of supernatural property. These other reasons for action do not exist. However, if they did exist they would count as reasons why a certain action should or should not be performed. The word ‘should’ does not exclude them a priori.

As it turns out, desires are the only reasons for action that exist. It is not true by definition, but it is nonetheless true.

Because ‘should’ refers to reasons for action, and desires are the only reasons for action that exist, ‘should’ either refers to desires, or it refers to something that is not, in the real world, a reason for action. If it refers to a reason for action that does not exist, then the ‘should’ statement is false.

The Ambiguity of Should

Now, ‘should’ in this sense is an ambiguous term. It raises the question, “Which desires?”

The answer to this question has to be picked up in the context in which the ‘should’ statement appears.

If I were to say, “The keys are on the table,” this raises the question, “Which table?” In order to answer this question we have to look at the context in which the statement was made. Typically, when I use a phrase like ‘the table,’ I assume that the listener can pick up from the context which table I am talking about. If not, then I may have to be more specific (e.g., “the dining room table”).

When a speaker uses the term ‘should’, he usually uses it in a way where the listener can pick up the desires in question from the context in which the statement is used. If not, then the speaker may need to be more specific.

Example #1, two people are sitting in a car outside of a convenience store. The passenger pulls a gun, checks it for bullets, cocks it, stuffs it in his coat, and opens the car door as if to leave. The driver hands him a ski mask and says, “You should wear this.” In this example, the ‘desires in question’ in this case are the passenger’s desires, and not the desires of the store clerk or other customers.

Example #2: I meet a friend for lunch. She has no idea what to get her husband for his birthday. We discuss options for a while, and I say, “You should buy him a hot tub.” The person listening to the conversation would have heard my friend say how much she wants a hot tub but cannot afford one, in part because she wants to get something special for her husband. They would have heard her say that her husband has also talked about buying a hot tub as well, but was also worried about the expense. My recommendation, in this context, would take the form of fulfilling more and stronger of both of their desires.

Another friend complains about her job. I tell her that she should just tell her boss, “I quit,” and walk out. She knows that I am not serious because, even though this act would fulfill some of her desires, it would not fulfill the more and the stronger of her desires (e.g., the desire for food, clothing, and shelter).

It is important to note that these are clearly not moral statements. The robber does not have an obligation to wear the mask, my friend is under no obligation to purchase the hot tub, and my other friend has no duty to keep her job. These are all non-moral senses of the word ‘should’. So, we still need to look for what is special about moral senses of the word ‘should’.

But first, we have a complication.

Ought and Can

‘Should’ lives under a restriction that it implies ‘can’. For example, it is not the case that an agent should teleport a child out of a burning building unless the agent can teleport a child out of a burning building. It is not the case that he should fire his boss unless he can fire his boss.

However, at the moment of action, the only act that an agent can perform is the act that will fulfill the more and the stronger of the agent’s own desires, given his beliefs. In other words, the only act that an agent can perform is the act that an agent does perform.

Does this imply that an agent always should do only what he does?

Not really.

Typically, when a person asks, “What should I do?” it is possible to interpret this as the question, “What would I do if my beliefs were true and complete?” We act so as to fulfill the more and stronger of our desires, given our beliefs. However, we seek to fulfill the more and stronger of our desires. When our beliefs are false or incomplete, there is a gap between these two that often causes an agent to do what she should not have done.

This does not change the fact that an agent can only act so as to fulfill the more and the stronger of her desires, given her beliefs. The agent, in asking, “What should I do?” is seeking a set of relevant and true beliefs (or a closer approximation to what is relevant and true) that would actually allow her to fulfill the more and stronger of her desires, when she acts so as to fulfill the more and stronger of her desires given her beliefs.

Future Desires

In addition to the gap between what an agent should do and will do caused by false and incomplete beliefs, there is a problem of future desires. Desires are not capable of backwards causation. Therefore, there is no way for a future desire to directly cause a present action. Whenever an agent acts, she acts so as to fulfill her current desires, given her beliefs. Future desires are left to fend for themselves.

It is still sensible to ask, “What should I do?” in a sense that considers future desires. The hypothetical question, “What would I do I had all true relevant beliefs and future desires had the power to influence present action,” is still a valid question, and an important question to answer for some people.

There are two ways that present actions can bring about the fulfillment of future desires. This is important, by the way, because I would estimate that a majority of the objections to what I write then I cover this subject comes from a failure to distinguish these three different relationships. People assume (wrongly) that I am talking about one of these relationships and raise all sorts of objections against that straw man, when in fact I talk almost exclusively about the other relationship.

So, let me cease to be cryptic and answer the question.

One way that a present act can fulfill future desires is if the agent simply has a desire to fulfill future desires. If I know that I will have an aversion to pain 10 years from now, and I have a present desire to avoid future pain, then I have a ‘reason for action’ for avoiding future states in which I will be in pain. If I know that I will want to eat 10 years from now, and that this will take money, my present desire to see that my future desires are fulfilled will cause me to save money.

The problem of drug addiction exists because future desires have no direct affect on present action. If a person has a particularly strong desire for nicotine, alcohol, or cocaine, for example, they will act so as to fulfill the more and stronger of their current desires, given their beliefs. Their future desires cannot mediate their actions. Therefore, they continue to use the drug, even though they know that their actions will thwart future desires.

The other relationship, and the relationship I always use when I talk about desire utilitarianism, is that a desire can be a desire that tends to fulfill future desires. This is not a desire to fulfill future desires. This is a desire for something else, where the pursuit of that “something else” tends to bring about future states of affairs that will fulfill future desires.

For example, a person can acquire a desire for exercise. She does not exercise because she will live longer. She exercises because she truly enjoys exercise. As a result, she is healthier, and her future desires stand a much greater chance of being fulfilled. Yet, the fulfillment of future desires is not what she is after when she exercises. It is, instead, a side effect.

We can see the difference between the person who exercises out of a desire for future health and a person who exercises for current enjoyment. Assume the news were to report that a comet will slam into the Earth in two months and end all life. The first person – the person who exercises for future health – no longer has a reason to exercise. He will stop. The second person – the person with a desire to exercise – still has a reason to exercise (her current desire to exercise), and will continue to do so, until the comet hits (unless other concerns force her to abandon her favorite past-time).

Now, we do, in fact, have the capacity to choose our desires. A person knows that cigarettes are designed to give the user a particularly strong desire to smoke cigarettes. That is to say, they are addictive. To prevent himself from acquiring a desire to smoke cigarettes, he refuses to smoke. This person is actually choosing a desire by knowing what causes the desire and choosing not to do that which would cause a desire he does not wish to have.

We also choose our current (or near-future) desires by deciding to acquire new habits. For example, a person may start to exercise out of a desire for future health. However, after she does this for a while, she discovers that she has come to value exercise for its own sake, and not for the sake of future health. She has acquired a new desire.

Because we have the capacity to modify our desires, we have reason to ask (and to answer) not only the question, “What should I do?” We also have reason to ask (and to answer) the question, “What should I want?”

Of course, the answer to the question, “What should I want?” has the same general answer as all other “Should” questions. “What reasons exist for and against my wanting X?” This is the same as asking, “Is it the case that wanting X would be such as to fulfill the desires in question?”

Other People’s Desires

Future desires have no capacity to directly influence current actions, or current desires. However, there is another set of desires out there that can have an effect on our current desires – the desires of other people.

The same two relationships that exist between current desires and future desires also exist between current desires and the desires of other people. I can have a desire to fulfill the desires of other people. Or (actually, ‘and’) I can have desires that tend to fulfill the desires of other people. My future self has no capacity to reach back in time and mold my present desires towards their fulfillment. However, other people can easily reach across space and affect my current desires. Furthermore, they have reason to do so. If they have a desire that Q, then they have reason to cause me to have those desires that will bring about or maintain a state of affairs in which Q is true.

So, in addition to my questions, “What should I do?” and “What should I want (given future desires)?” there is a question, “What should I want (given the desires of other people)?” Or, in other words, what desires do other people have reason to cause me to have? And, correspondingly, what desires do I have reason to cause other people to have? Of course, they have reason to cause me to desire that which will fulfill their desires, and I have reason to cause them to desire that which will tend to fulfill my desires.

There are, in short, desires that people generally have a lot of strong reasons to cause others to have, and desires that people generally have a lot of strong reasons to cause others not to have. If there are actions that one can perform (e.g., praise, condemnation, reward, and punishment) to cause people to have stronger desires that people have reason to strengthen, or to cause people to have weaker desires that people have reason to weaken, then there are reasons to perform those actions.

These posts tend to get very long, so I will end here.

However, I would like to note that nowhere in this essay did I use the word ‘moral’ – other than to make some distinctions near the introduction. I started off with a notion of practical ‘should’, and ended up talking about desires that people have reason to promote in others – in some cases, a lot of very strong reasons.

I will leave it up to the reader to decide if, at any point in this essay, any ‘should’ that I spoke about started to sound suspiciously like a moral ‘ought’.

1 comment:

BlackSun said...

Love it! Great post.

Two things:

1) When we break down the brain into competing centers (Marvin Minsky - Society of Mind), we will have an even better way of establishing how desires are formed. I really liked your addiction metaphor, short term vs. long term desires. But I suspect it gets more complicated than that with competing brain centers, one saying "don't take that hit of crack, you'll regret it" but more powerful primal ones (with the dopamine receptors) saying "go for it."

2) When looking at desires, and the possible modification of our evolutionary programming, it does pose a dilemma: How should we be evolving. Should we try to get away from animalistic and pugilistic desires and embrace more communitarian values? Or should we value individualism and desire fulfillment strictly based on our evolutionary heritage. The larger question is what happens under desire utilitarianism when two competing desires such as these are in conflict?