Thursday, August 23, 2007

The connection between means and ends and its consequences for the teleological model of reasons-explanations of action.

A third in a series of unknown length. The previous post is here. In the current post, I intend to complete the sketch of the teleological model of reasons-explanations of action that I want to defend. (Of course, I'm going to do so in an incredibly sketchy, promissory note kinda way.) To do so, I have to, as said previously, answer two questions:
  1. What is the connection between means and ends?
  2. What sort of account of value underlies this picture?
Answering these two questions will, I think, fully characterize the sense in which teleological reasons-explanations explain action by showing that the action fit into some efficacious value-increasing pattern.

One move I want to resist is the claim that the connection between means and ends is purely psychological. I want to resist this for a couple of reasons. First, I don't think we gain anything useful by creating a psychological state -- "purposes" -- to play the role of ends. Everything that a purposes-model can capture, I claim, can be captured just as well by a goals-model, where goals are conceived of as states of affairs that agents (attempt to) bring about. Second, I think it will just resurrect the "trying" problem, where I am somehow absolved of practical responsibility if I just try to take appropriate means for my ends. If the connection between means and ends is just in my head, then it seems I cannot be blamed as long as I make my utmost effort to get the connection right. But that strikes me as extremely odd; surely getting the connection right is what matters, not making a really good effort to do so.

I think there's something to this idea that the connection between means and ends is one of propriety or appropriateness. (I like "propriety", so I'm going to use it. Nice-lookin' word, and kinda fun to say.) That is, we can say a given means -- say, φ-ing -- is connected to a given end -- say, ψ -- if the means is appropriate for the end. To make it more technical-sounding, φ-ing is connected to ψ if (and only if?) φ-ing is appropriate for ψ. So, taking my umbrella is connected to not getting wet if taking my umbrella is appropriate for not getting wet. If taking my umbrella is not appropriate for getting wet, then there's no connection between taking it and not getting wet; taking my umbrella must be appropriate for some other goal (maybe it makes me feel comfortable to carry an umbrella around) or taking my umbrella is not something I do for reasons (maybe it's just reflex; as I go out the door, I grab my umbrella).

So, what does propriety amount to? I'd have to be convinced, strongly, that this is a matter of efficacious causal connection between means and ends -- that is, something like a means is appropriate for an end if it causes the end to obtain. For one, this fails to distinguish between what the means is for -- the end or goal -- and what the means just happens to produce -- that is, good ol' epiphenomena. For two, it would also force us into a probabilistic model of causation -- in order to capture the idea that several different means could be more or less appropriate for a given end -- I'm sympathetic to the idea that it's counterfactual. That is, a means is appropriate for an end if the end would not obtain without the means. The obvious problem, though, is that there could be a multitude of appropriate means for an end, so we'd have to set up a fairly complicated counterfactual. I'm not sure that propriety is as complex as all that.

Ultimately, the strategy I think I want to pursue is more radical than either of these alternatives: I want to, basically, (1) collapses (or maybe staggers broadly) into (2). I take the task of finding the connection between means and ends to just be the task of figuring out a workable sketch of value. This is because of the obviously normative connotations of something being "appropriate" for something else. (I suppose my choice of the word "propriety" was not entirely accidental. Heavens.) To be appropriate is to conform to some set of rules, or to exhibit some sort of relevant value. My behaviour is inappropriate if it is, say, rude, or if I'm breaking a law; my behaviour is appropriate if it is polite or in compliance with the law. So, under this view, φ-ing is connected to ψ if φ-ing is appropriate for ψ and φ-ing is appropriate for ψ if either φ-ing for ψ conforms with some relevant rule or φ-ing for ψ is valuable, for some relevant value. The rules and values, since I'm trying to pitch this at full generality, are going to have to be rules and values of practical rationality.

This, I think, will rather neatly solve my problem with (2), as well. It's not quite as nasty a problem as I might have thought; I'm dealing with a pretty broad set of rules and values when it's a matter of practical rationality. (I was worried that the teleological model I want to defend would have to rest on some highly contentious claims about moral value, say, which would have been a whole dissertation in its own right.) So, moving on to (2): the values and rules of practical rationality aren't really all that different from the values and rules of theoretical rationality. That is, when we're trying to figure out what to believe (or, if you like, what propositions are true) we're reasoning in the same way as when we're trying to figure out what to do. The difference is that theoretical rationality has a nice little goal built in -- figuring out what's true -- while practical rationality requires that we agents set the goal, and then reason our way back, in accordance with the rules and values governing practical reasoning, to the means.

What are the rules and values governing practical reasoning? I honestly don't quite know. But I do know that this is a problem which can, in the end, be solved. That I want to rely on them probably accounts for why I'm inclined towards Schueler's model, which also makes heavy use of practical reasoning, and away from Sehon's, which takes teleology to be an unanalyzed blob. Whatever the rules and values are, once they're uncovered, then the characterization of efficacious value-increasing patterns is complete: these patterns are just patterns of practical reasoning.

However, the solution, I think, will come in the third section of the dissertation, where I trace some consequences of adopting the teleological model for areas beyond action-explanation. For next week, I think it appropriate to spend some time showing that other models of reasons-explanation of action are wholly inadequate to satisfy the three basic criteria on a good reasons-explanation of action; that is:
  1. Account for the connection between reasons and action in the agent
  2. Give reasons that the agent genuinely finds worthy
  3. Fit the action within the pattern implied by (a) and (b)

No comments: