Monday, May 02, 2005

A Taxonomy of Reasons

First let's distinguish motivating and normative reasons. Someone's m-reason for acting is simply whatever motivated them to act. These explain actions. I take it they consist of psychological states, or more precisely, belief-desire pairings. But I'm more interested in normative reasons, i.e. things which count in favour of an action. These justify actions. (For the moment I'll put aside the question of what these consist of.)

We might next distinguish between objective and subjective normative reasons, where o-reasons appeal to facts, whereas s-reasons appeal to beliefs. An example will make this clearer. Imagine an unwilling drug addict, who wishes he could be free of his cravings. Suppose he comes across a needle which he believes contains the drug, but in fact it contains a special serum which will cure his addiction. The addict's cravings become overwhelming, and in a moment of weakness he injects himself. We can now say:
  • His m-reason for acting was that he craved the drug and believed that the needle contained the drug.

  • He had s-reason to act differently and not inject himself.

  • He had o-reason to do as he did and inject himself, though he was unaware of it and thus not acting on this reason.

This is a good start, but we need to clarify the notion of an s-reason, and how it relates to o-reasons (which from now on I'll simply refer to as "reasons", unprefixed). We might then say that an s-reason is simply the belief that a reason obtains. We have subjective reason to do whatever we think we have (objective) reason to do. That is, we have s-reason to perform act A iff we believe that P, and we believe that P constitutes a reason for us to A.

But this raises some new possibilities. What if we have only one of the above two beliefs? Suppose that S, a libertarian opposed to "positive rights", sees a child drowning. S does not believe that she has any reason to intervene and jump in the pool to save the child. But she does believe the child is drowning -- a fact which, though she refuses to recognize it, is a reason for S to jump in the pool. Should we say that S has subjective reason to jump in the pool? Not according to the above account. But there is a sense in which she does. Let us call this a "partial reason", or p-reason, defined as the belief that P, where the truth of P would, in fact, constitute a reason. (Note that p- and s-reasons will also come apart in cases where an agent falsely believes that something constitutes a reason. Such a case is exactly opposite to the above example, as she will have s-reason but not p-reason to act.)

Lastly, we might consider a case where we believe that P constitutes reason to A, and P is true (though we may not realize this). We might call this a "quasi-subjective reason", or q-reason, to A. So if, say, an evangelical thinks he has reason to convert people to Christianity, and - unbeknownst to him - the best way to achieve this end would be to nuke Mecca, then we say he has q-reason to nuke Mecca. Note that it's not really an objective reason (assuming that evangelism is ungrounded), and it's not a subjective reason either, since he doesn't realize that nuclear war would achieve his unsavoury ends. It's a variety of reason in a class of its own: quasi-subjectivity.

To summarize, the 'objectivity-level' of normative reasons depends on the status of two factors:
(1) a proposition P
(2) P's constituting a reason.

Normative reasons (of various types) are yielded when each is either true or believed. Objective reasons are when both factors are true. Subjective reasons are when both are believed. Then we have the two 'hybrid' varieties of reasons. 'Partial' reasons are when (1) is believed and (2) is true; 'Quasi-subjective' reasons are when (1) is true and (2) is believed.

These types of reasons may fully or partially coincide. If each of (1) and (2) are both believed and true, then all four types of reasons will result! But we can also have four varieties of partial coincidation, i.e. reasons that are {either objective or subjective} and {either partial or quasi-subjective}. That is, one factor will be both believed and true, and the other factor will be either believed or true, but not both.

I think that's a full account of the various possible 'objectivity levels' of reasons. I'm not yet sure whether such an analysis actually serves any useful purpose, but it's kind of cool to have ready, just in case ;)

So we've distinguished motivational from normative reasons, and within normative reasons we've identified some different levels of objectivity. The last thing we need to do is distinguish between different normative frameworks. We can have prudential reasons, moral reasons, epistemic reasons, and so forth.

Suppose that General Jim has command over some nuclear warheads. He believes that if he nukes the Middle East then this will kill only bad people, and end terrorism forever, but cause him to face severe punishment. He thus has subjective moral reason to send the nukes, but subjective prudential reasons not to. Further suppose that he is mistaken on both counts. Sending the nukes would screw up the world, but Bush would give him an official pardon and blame it on "a few bad apples" further down the ranks somewhere, whereas Jim himself would get promoted and glorified. Thus Jim has (objective) prudential reason to send the nukes, but (objective) moral reason not to.

I hope that's all reasonably clear. Now, the really tricky questions are: what constitutes an objective reason? and which frameworks are genuinely reason-giving? This post is too long already, so I'll have to discuss that another day.


Post a Comment

Visitors: check my comments policy first.
Non-Blogger users: If the comment form isn't working for you, email me your comment and I can post it on your behalf. (If your comment is too long, first try breaking it into two parts.)