Sunday, September 19, 2010

Ought = Value + Can?

I previously thought that the falsity of the parity thesis provided a straightforward argument against value primitivism. After all, I figured, if value is our sole normative primitive, then it must be that any normative claim can be restated simply using evaluative terms. But evaluative terms are 'global', i.e. they apply just as well to mere evaluands as to agential ones. So if every normative claim is like this, that's just to assert the parity thesis -- which, my earlier post argued, is false. So value primitivism is false. We need to start from some other normative primitive (e.g. fittingness) to accommodate the datum that there's a sense of 'ought' that applies to (say) actions but not eye colours.

This argument needs a bit of work, however. For, as Nic Southwood helpfully pointed out to me, one might construct a new normative concept (of 'rightness' or whatever) out of a combination of value plus some non-normative concept, such as 'can be chosen'. In this case, we still have it that value is the only normative primitive, and yet we can make new normative claims that are not reducible to merely evaluative claims (since they also include a non-normative modal component), and so avoid commitment to the parity thesis. Tricky!

Nonetheless, I think there are some important insights in the original argument that can be salvaged. In particular, this move on behalf of the value primitivist can be made to look troublingly ad hoc in comparison to the elegant theoretical framework offered by the fittingness theorist. Consider:

(A) This move only provides the value primitivist with a distinctive normative status to apply to actions. What of other agential features that are subject to rational evaluation, such as beliefs and desires? Here the value primitivist faces three challenges: First, that these mental states are not voluntary in the way that actions are -- hence appeal to the 'can' of chooseable options will not yield a property that applies to these states.

Second, even if they can find a way to fix the scope of each concept's applicability in some other non-normative terms, the resulting collection of concepts as yet lacks any unifying explanation. The Fittingness view, by contrast, systematizes and unifies diverse classes of reasons by understanding each type as corresponding to a particular rationally-evaluable feature of agency. Basically, it tells us that there will be a class of 'reasons for X' for each rational output X (beliefs, desires, actions, etc.) identified by our best moral psychology, since each such X is the kind of thing we can assess as more or less justified/fitting. Can the value primitivist tell a comparably well-motivated story for their chosen collection of normative concepts?

Thirdly, using value as the normative component of these restricted concepts seems to yield incorrect verdicts -- we've reason to believe what's true, not whatever's value-maximizing to believe (of course, it may be desirable to have the latter, unjustified, belief).

(B) What explains the normative significance of the non-normative concept ('can', or whatever) that the value primitivist is appealing to here? The fittingness theorist has a neat answer, as fittingness norms - by their very nature - concern agents' rational capacities. But if value is our sole normative primitive, it seems that we are left without an explanation of the distinctive normative significance of the new (gerrymandered?) concept created by combining 'value' + 'can'.

Arguably, to meet this explanatory challenge we need to take the combination as corresponding to another normative primitive. If we start by recognizing moral rightness (or fitting action) as a normative primitive in its own right, then we can use this to explain the normative significance of 'value + can', if this latter combination turns out to be coextensive with rightness.

This is not to say that we can't in some sense elucidate rightness in terms of value. (I'm a consequentialist, so of course I think there's some important sense in which the good is prior to the right.) But it isn't a conceptual reduction, or anything at such a 'meta-ethical' level. Instead, I see the relevant priority as purely a matter of first-order normative ethics.

In other words, we shouldn't be analytical consequentialists. We need normative primitives corresponding to both the right (fitting choice/action) and the good (fitting desire) -- neither can be done away with or reduced to the other. My consequentialism is instead reflected in what I take to be the relation between these two families of normative reasons. In particular, I take it that our reasons for desiring certain outcomes explain our reasons for action, rather than vice versa. (In psychological terms: a properly functioning agent first identifies what's truly desirable, or worth aiming at, and subsequently conforms their action to the bringing about of these good ends. A deontologist, in contrast, takes the properly functioning agent to first govern their choices by various rules, which subsequently constrains the outcomes at which they aim.)


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.)

Note: only a member of this blog may post a comment.