The role of all things considered judgements in practical deliberation

Abstract

Suppose an agent has made a judgement of the form, ‘all things considered, it would be better for me to do a rather than b (or any range of alternatives to doing a)’ where a and b stand for particular actions. If she does not act upon her judgement in these circumstances would that be a failure of rationality on her part? In this paper I consider two different interpretations of all things considered judgements which give different answers to this question, one suggested by Donald Davidson, the other by Paul Grice and Judith Baker. I argue that neither interpretation is adequate. However, a third interpretation that combines features of the Grice/Baker view with the Davidsonian view is possible. In the final section of the paper I defend this interpretation against two objections.

Keywords:

• practical irrationality
• defeasibility
• second-order reasons
• Davidson

Acknowledgements

I would like to thank Bill Child and David Charles for helpful comments on earlier drafts of this paper. Thanks also to two anonymous referees for Philosophical Explorations.

Notes

1. For this way of characterizing the general view, I am indebted to Greenspan (2005). One version of this view is sometimes called ‘judgement internalism’ and is often associated with Kant. According to this view, a judgement of value provides motivation for a rational agent to act accordingly. Kant applies this view to the moral sphere, where he claims that those who recognize the ability of their maxim to serve as a law for all rational beings are imperfectly rational if they fail to act on it. See Kant (1997). But the view may also apply more broadly to non-moral value judgements. Thus, in setting up his sceptical argument against the possibility of weakness of will, Davidson assumes what he calls ‘a mild form of internalism’ according to which ‘a judgement of value must be reflected in wants (or desires or motives)’ (1980, 26). For further details of different kinds of internalism, see Brink (1989, 37–43), Darwall (1997) and Koorsgaard (1986).

2. Other authors who have done important work relevant to this distinction include Edgington (1995), Lewis (1986) and Stalnaker (1994), to name a few.

3. In their own words, to facilitate exposition, Grice and Baker use the form ‘x should do a’ rather than ‘doing a is better than doing b’. Thus, they claim that ‘the kind of judgement whose status is in question is one such as “given the fact that on my evidence I should do A, and that my judgement that this is so is optimal, I should judge that I should do A”’ (see Grice and Baker 1985, 38). However, for reasons I hope will become clear, I shall stick to the Davidsonian schema and use the relational predicate ‘is better than’.

4. It is important to note that the sense of ‘conditional’ in question here is not the truth functional sense of classical logic, that is, not the sense in which a judgement of the form ‘if p then q’ is a conditional judgement. Clearly, the latter judgement need not be a defeasible conditional. The sense of ‘conditional’ Grice and Baker have in mind is the sense in which a probability judgement is conditional on the evidence. Being a conditional judgement in this ‘softer’ sense is usually taken to imply that it can be upset without falsification of any of its premises. It is this view of ‘conditional’ I shall assume throughout this paper (see Grice and Baker 1985, 39).

5. Davidson's own response to the objection seems to be along these lines (see Davidson 1985, 201–6). Thus, he argues that the objection is harmless because it assumes one particular understanding of ATC judgements, while in fact his own distinction between conditional and unconditional judgements is consistent with many different construals of ATC judgements, some of which (presumably) do not give rise to troubles of the kind Grice and Baker describe. However, Davidson does not provide any plausible alternative account of ATC judgements to back up this claim. Both of the two construals I discuss in the text can, I think, be found in Davidson. Unfortunately, neither of them seem to work against the unconditionality objection.

6. For those unfamiliar with Davidson's notation, some remarks of clarification are in order. Davidson introduces the ‘pf’ as a prima facie sentential operator. Prima facie in English means something like ‘in so far as’. The idea is that the ‘pf’ makes a complex sentence out of two sentences, one stating that some action a is better than some action b, the other stating in which respects a is better than b, that is, relative to what reasons a is better than b.

7. See also Grice and Baker (1985, 35) for a similar objection to this construal.

8. Raz's distinction has generated much discussion. For some recent contributions, see Pettit et al. (2004).

9. For a recent defence and extension of Raz's distinction, see Bratman (2004). Bratman distinguishes between ordinary reasons and what he calls ‘framework’ reasons, i.e. ‘self-governing policies’ of discounting or bracketing certain reasons. Dancy argues in a similar vein for a distinction between what he calls ‘enticing’ and ‘peremptory’ positive reasons, the first of which contribute to making actions worth performing, but not to requiring them (see Dancy 2004).

10. Of course, this is not to say that it is not possible to decide to do something without believing that one has sufficient reason to do it. Sometimes we clearly decide to do things even if we believe we have very little reason to do it. Weakness of will is one example. The point is that in the cases where we do rationally decide to perfom some action on the basis of practical deliberation, deciding to perform that action may be seen to involve the belief that our set of reasons is, in the circumstances, sufficient to act on. When I speak of ‘decision’, I shall have in mind only rational decision in the context of practical deliberation.

11. This example is a variant of a similar example suggested by Raz (1978, 129). However, Raz's concern is not ATC judgements or the principle of continence.

12. Raz also argues that promises and rules can be second-order exclusionary reasons in this sense (see Raz 1978). This claim might be more difficult to defend. For a recent critical discussion of Raz's use of the concept of exclusionary reasons to explain the authority of law, see Mian (2002).

13. I borrow this way of putting the distinction between first- and second-order reasons from Greenspan (2005).

14. It can be added here that the distinction between recognition and motivation also puts pressure on Davidson's mild form of internalism since it shows that any conceptual connection between judging something best and being motivated to do it, must be a defeasible one (see Note 1). In fact, there seems to be a growing awareness of the importance of this distinction in the recent literature. For example, Fisher and Ravizza draws a similar distinction between what they call ‘receptivity’ and ‘reactivity’ as forms of ‘reasons-responsiveness’ (1998, 69). Greenspan defends this distinction in her paper, ‘Adequate reason’ (2005).

15. For a discussion relevant to this issue, see Frankfurt (2005).