DOI: 10.4324/9780415249126-L135-1
Version: v1,  Published online: 2001
Retrieved April 21, 2021, from

4. Fundamental sources of normativity

What this means is that vindicating or debunking arguments, the second way normativity enters into ethics, ultimately require support from a third kind of theory, a more fundamental account of the ‘sources of normativity’ as these relate to questions of rational action (Korsgaard 1996). For example, an important premise in the Humean argument is that practical reason counsels ‘instrumental rationality’, taking the means to one’s ends. This, Humeans argue, is practical reason’s core, which more ambitious accounts of practical reason seek misguidedly to extend. Kantians agree that practical reason includes an uncontroversial instrumental core, but hold that this could not possibly exhaust its prescriptions. The sense in which instrumental rationality is uncontroversial, they argue, is that to have an end and, at the same time, fail to take necessary means is rationally incoherent in much the same way that it is incoherent to believe contradictory propositions or to believe the premises of a valid argument but deny the conclusion (Broome 1999, and see Rationality, instrumental). Each is proscribed, respectively, by norms of coherent practical and theoretical thought. But just as the injunction against believing contradictory propositions does not tell one which to believe, neither does the analogous practical injunction distinguish between taking the means to an end and giving the end up. One can achieve coherence by doing either.

In order to act, however, an agent must have reasons for acting, and Kantians argue that, by itself, instrumental rationality cannot provide these. At best, it provides an account of ‘relative rationality’, of what there is reason to do relative to the assumption that one has reason to have the ends one has (Darwall 1983: 14–17). The sense in which taking the means to an end is ‘hypothetically’ rational, therefore, is that it is rationally prescribed as conditional, not on one’s having the end, but on a ‘hypothesis’ one accepts or is committed to in having the end. Kantians can take a similar approach to formal theories of decision. For example, the injunction to maximize expected utility, where this is a function of an agent’s preferences and subjective probability estimates, can be accepted as a norm of relative rationality (see Rational choice theory; Utilitarianism). It recommends the expected utility-maximizing act, conditional on hypotheses an agent accepts or is committed to in having the preferences and beliefs they have. By themselves, formal theories of decision give no practical guidance either. They simply say which actions are most coherent with our preferences and beliefs. Whether we should take those actions depends also on whether we should have those preferences and beliefs. Like the principle of instrumental rationality, these theories are impotent to guide action without implicitly assuming further premises about reasons for acting.

Kantians argue, therefore, that their position cannot be faulted for being committed to categorical normative propositions that go beyond instrumental rationality’s requirement of practical coherence. A Humean theory cannot provide practical guidance without going beyond this requirement either. Specifically, if it is to maintain that an agent’s desires and interests are distinctively reason-giving, it must defend this as a categorical normative proposition.

This pushes us to a deeper question of the third kind: What can make any fact a reason for acting? What is the ‘source’ of practical normativity? Here Humeans are guided by metaphysical naturalism (see Moral realism; Naturalism in ethics). Any answer to this question, they think, must derive from the natural order, that being all there is. Humeans agree with Kantians, however, that there is no possibility of analytically reducing claims about normative reasons to propositions of empirical psychology. ‘Hume’s Law’ says that no set of non-normative propositions can entail a normative claim (Hume [1739–40] 1978). This leads to one of two conclusions: Either there really is no such thing as normativity, only normative judgements, or normative reasons can consist in natural, specifically psychological, phenomena, despite their conceptual independence. The first tack, however, lends no support to the idea that an agent’s reasons for acting have any special relation to their desires. To maintain that, Humeans must take the second tack. One way of doing so is to argue that although normative and motivating reasons are conceptually distinct, an agent’s motivational susceptibilities nonetheless constrain their normative reasons. Specifically, it is argued, nothing can be a normative reason for someone unless it could be a motivating reason for them, a consideration by which they could be motivated and on which they could actually act, and this requires that it be appropriately related to the agent’s desires (Williams 1975; Smith 1995). Kantians reply that they can accept this ‘internalism requirement’ and claim it to be fully consistent with a vindicating argument for the normativity of morality. The capacity to be moved by moral reasons is part of what it is to be a free rational agent, they argue, since morality itself is the product of the agent’s own ‘self-legislation’ (Darwall 1992; Korsgaard 1986).

Citing this article:
Darwall, Stephen. Fundamental sources of normativity. Normativity, 2001, doi:10.4324/9780415249126-L135-1. Routledge Encyclopedia of Philosophy, Taylor and Francis,
Copyright © 1998-2021 Routledge.

Related Articles