Monday, October 05, 2009

Smith on Parfit: 8 of 15: Desires as Means and Beliefs

I hate being wrong. Yet, in writing this blog, I think I have to say that I have been wrong about the relationship between desires-as-means and beliefs.

In this series of posts I am commenting on an article sent to me by a member of the studio audience. These are, in a sense, notes written in the margin as it were as I highlight passages in the article and explain my agreement or disagreement.

In that article, there is a place in which Smith attempts to, classify as procedural or substantive a range of principles that various theorists have thought qualify as principles of rationality governing the formation of desires (in what follows 'RR' = 'Reason requires that'):

He does not endorse all of these principles. Instead, he mentions them so as to classify them, and h classifies them so that he can raise questions for Parfit’s distinctions between those classifications.

From my perspective, this dispute over classifying the things that reason requires is purely academic. I do not see a reason to get involved in that dispute. However, I am interested in what these people say about what reason does and does not require.

However, since Smith went ahead and presented them, I have a chance to scribble some notes in the margins regarding each of these principles.

If someone has an intrinsic desire that p and a belief that he can bring about p by bringing about q, then he has an instrumental desire that he brings about q.

I scribble in the margins:

What is an intrinsic desire?

I would use different words – words that may or may not mean the same thing that Smith intends them to mean.

As I have already argued, there are two types of desires; desires-as-ends, and desires-as-means, where desires-as-means are bundles of desires-as-ends and beliefs.

Applying this distinction to Smith’s principle yields something like this:

If someone has a desires-as-ends that P and a belief that she can bring about P by bringing about Q, then she has a desire-as-means that she bring about Q.

I have sometimes used this formulation in the past. However, as I think about it here, I believe it is mistaken. I have said that desires-as-means are bundles of desires-as-ends and beliefs. However, at this point I am going to change my mind and say that desires-as-mans are bundles of desires-as-ends and facts that relate means to ends, independent of beliefs.

Though I would love to say that I had this right all along, I suspect that this might not be the case.

Specifically:

(1a) If a person has a desires-as-ends that P, and can bring about P by bringing about Q, then she has a desires-as-means that she bring about Q. There is nothing about beliefs in this formulation.

(1b) If a person believes she has a desires-as-ends that P, and a belief that she can bring about P by bringing about Q, then she has a belief that she has a desire-as-means that she bring about Q. However, this belief can be mistaken. To say that it is mistaken is to say that she does not have the desires-as-means that she thinks she has.

A person can have false beliefs about whether Q has value as a means by having false belief about whether Q can bring about P, which she does desire.

Where we assume that a woman who is thirsty is about to drink from a glass of water that, unknown to her, contains poison, then she thinks she wants to drink the contents of the glass. In fact, she does not want to do so.

A stranger, knowing what is in the glass, can well be imagined telling her, as she takes the glass and raises it to her lips, You do not want to drink that.. The stranger, in this case, is correct. She does not want to drink the contents of the glass. She only thinks (believes) that she wants to. And her belief is mistaken.

5 comments:

Luke said...

I wish I could keep up with you. I need a 15-year education in moral theory, pronto...

Marc said...

Like Luke, I've also missed out on that 15-year education in moral theory. :-)

But I've been reading your blog for quite a while now and this is probably the first time that I think you may really be mistaken.

I can't really see how you can say that a desire can be mistaken. Someone either has a certain desire or she doesn't. Maybe she wouldn't have this desire(-as-means) IF she had perfect knowledge about the facts of the matter, but she has the desire nevertheless.

You said:
If a person believes she has a desires-as-ends that P, and a belief that she can bring about P by bringing about Q, then she has a belief that she has a desire-as-means that she bring about Q.
I think you should rephrase that last part as: "... about Q, then this will cause her to have a desire-as-means that she bring about Q."

A person can have false beliefs about whether Q has value as a means by having false belief about whether Q can bring about P, which she does desire.
This is most obviously true, but I don't think it warrants the following statement:
If, in the example above, the glass she thinks contains clean water contains poison, then she thinks she wants to drink the contents of the glass. In fact, she does not want to do so.
I think she really DOES want to drink the contents of the glass. The fact is that doing so and thus fulfilling this desire-as-means will not fulfill her desire-as-ends.

So I think the stranger's statement will not make much sense to her. Her reply will probably be: "How do YOU know what I want?". And rightly so.

This being my first serious attempt at a contribution, I may of course be very wrong...

Anonymous said...

A> Where we assume that a woman who is thirsty is about to drink from a glass of water that, unknown to her, contains poison, then she thinks she wants to drink the contents of the glass. In fact, she does not want to do so.

She is acting in accordance with her strongest desire, given her belief.

Her belief is that there is water in the glass.

If she is about to drink the contents of the glass, what desire is giving her the motivation to do so?

Martin Freedman said...

(1a) If a person has a desires-as-ends that P, and can bring about P by bringing about Q, then she has a desires-as-means that she bring about Q. There is nothing about beliefs in this formulation.

How can one "bring about P by bringing about Q" without having the relevant belief? It might be tacit and difficult to explicate and state but it is still a belief.

It is in this sense that we can say animals have beliefs, even if they cannot state the propositions which are the content of the belief. The belief corresponds to the facts or does not. You are saying we can somehow operate directly on the facts (states of affairs) without having any beliefs (true propositions about those states of affairs) about these facts. Without such beliefs in this sense such facts would be invisible and so incapable of influencing actions.

Now I am trying to see why Smith's point made you think this. All I can see is the same thing "he can bring about p by bringing about q", but how does one know without a belief? (knowing being a type of belief in the BDI sense).

She does want to drink the liquid, given her beliefs is correct.

The stranger is also correct in saying she does not want to drink the liquid - if she knew what he knew, then she would not want to drink the liquid, given the belief that it is poison.

AFAICS there can be no sufficient reason without at least one belief and one desire. Or having a desire without a belief is not a reason (to act). Is this what you are challenging?

Dan Doel said...

I think I have to agree with the others here. Isn't your definition[1] of desire something like:

An agent has a desire that P iff he/she will act in ways that (they believe) will realize states of affairs in which P is true.

If an agent has a desire that P, and believes that realizing Q will realize P, then one would expect them to act in ways that they believe will realize Q, and so, by the above definition, they would have a desire (as means) that Q.

Also, as faithlessgod notes, even if the agent has a desire that P, and realizing Q would in fact realize P, if the agent does not believe that the above connection exists, then they presumably wouldn't act to realize Q (given only the above as motivation), and so would not qualify as having a desire that Q.

In the case of the water/poison, they may be mistaken that 'the stuff in the glass' is water, but they still seem to have a 'desire to drink the stuff in the glass.' If someone is able to tell them that it's actually poison, and change the agent's beliefs, then they would presumably no longer have the desire (as means) to drink the stuff in the glass.

Of course, this doesn't make sense of utterances like "you don't want to do that," but everyday language isn't always perfectly consistent. One can always be unambiguous (presumably) by only talking about beliefs, facts and desires-as-ends, but the above notion of desire seems to come down on the desires+beliefs end for desires-as-means.

"You don't want to do that," can probably be sensibly interpreted as "you practical-ought not do that." In the water/poison case, the agent practical-ought not drink from the glass, despite his desire to do so.

1: I'm not disputing the definition, just noting that that's the one you typically use.