The Decision-Theoretic Lockean Thesis
2013; Taylor & Francis; Volume: 57; Issue: 1 Linguagem: Inglês
10.1080/0020174x.2013.858421
ISSN1502-3923
Autores Tópico(s)Free Will and Agency
ResumoAbstractCertain philosophers maintain that there is a 'constitutive threshold for belief': to believe that p just is to have a degree of confidence that p above a certain threshold. On the basis of this view, these philosophers defend what is known as 'the Lockean Thesis', according to which it is rational to believe that p just in case it is rational to have a degree of confidence that p above the constitutive threshold for belief. While not directly speaking to the controversy over the Lockean Thesis, this paper defends the general idea behind it—namely, the thesis that there is some threshold such that it is rational to believe that p if and only if it is rational to have a degree of confidence greater than that threshold. This paper identifies the threshold in question—not with the alleged constitutive threshold for belief—but with what I call 'the practical threshold for rational belief'. Roughly, the thesis defended here is that it is rational to believe that p if and only if it is rational to have a degree of confidence that p that rationalizes engaging in certain types of practical reasoning. AcknowledgementsFor their thoughtful input on previous drafts of this paper, I owe special thanks to Yuval Avnur, Andy Cullison, Kenny Easwaran, Paul Hurley, Amy Kind, Heather Lowe, Eliot Michaelson, Susanna Siegel, and all the participants in both the Claremont Colleges Works in Progress Group (WIP) and the Circle of Los Angeles Philosophers (CLAP).Notes1 See inter alia Chisholm Perceiving; Sellars, 'Induction as Vindication'; Kyburg, 'Conjunctivitis'; Foley, 'Beliefs', 'Epistemology of Belief', Working Without a Net; Hawthorne and Bovens, 'Preface'; Christensen, Putting Logic in its Place; Maher, Review; Eriksson and Hájek, 'What Are Degrees of Belief?'; Locke, Essay Concerning Human Understanding, Book IV, Chs. 15–16.2 Henceforth, 'greater than' means 'greater than or equal to'.3 There is an intentional scope ambiguity here: is there one threshold such that to believe that p is to have a degree of confidence that p above that threshold; or is there, for each would-be belief that p, one threshold such that to believe that p is to have a degree of confidence that p above that threshold? Some proponents of the thesis understand it one way, others the other.4 Foley, 'Beliefs'. Some authors use the name 'Lockean Thesis' to refer to CTA. This usage is mostly harmless, given the close relationship between the two principles. Other philosophers use the name 'Lockean Thesis' to refer to what I below call the 'Generalized Lockean Thesis'. If one is not careful, this usage can cause confusion. As I note below, the Generalized Lockean Thesis is neutral with respect to CTA, while the Lockean Thesis is not.5 The argument from the Lockean Thesis to CTA is simple: the former presupposes the latter. The argument from CTA to the Lockean Thesis rests on the following principle: if to be in state x just is to be in one of a possibly infinite number of states y1, y2,..., then it is rational to be in x if and only if it is rational to in at least one of y1, y2,.... From this principle it follows that, if to believe that p just is to have some degree of confidence x greater than T that p, then it is rational to believe that p if and only if it is rational to have some degree of confidence x greater than T that p. Foley attempts to reach the Lockean Thesis from CTA, but he takes a different route.Begin with the assumption that one believes a proposition p just in case one is sufficiently confident in the truth of p. Now add the assumption that it is rational for one's confidence in a proposition to be proportionate to the strength of one's evidence. Together these two assumptions suggest a thesis, namely, that it is rational to believe a proposition p just in case it is rational for S to have a degree of confidence in p that is sufficient for belief. (Ibid., 37)Notice that the 'added' assumption in this passage—the assumption that it is rational for one's confidence to be proportionate to the strength of one's evidence—is doing no work in the argument. The conclusion—the Lockean Thesis—connects the rationality of believing that p to the rationality of having a sufficiently high degree of confidence that p. It is an independent matter whether the rationality of a certain degree of confidence is connected exclusively to evidence. What Foley really needs to get to the Lockean Thesis from CTA is just the principle noted above.6 See: Levi, Gambling with Truth; Lehrer, 'Coherence', Knowledge; Kaplan, 'Bayesian Theory', 'Rational Acceptance'; Maher, 'Irrelevance of Belief'; Lance, 'Subjective Probability and Acceptance'; Frankish, 'Partial Belief'.7 In the Generalized Lockean Thesis we see the same sort of scope ambiguity that we saw in CTA (see footnote 3). But let there not be any surprises: the version of the Generalized Lockean Thesis that I defend below says that, for each particular belief, there is a threshold, but denies that there is one threshold for all beliefs.8 Ross and Schroeder, 'Belief, Credence'.9 See inter alia Armstrong, Materialist Theory of the Mind; Lewis, 'Psychophysical and Theoretical Identification'.10 Kaplan, Decision Theory as Philosophy, 4.11 For convenience, I henceforth suppress the parenthetical qualification acknowledging the existence of cases in which there is not a unique act that is most preferable. But everything I say should be understood with that qualification in place.12 Ibid., 6.13 This view has some affinity with the view defended by Frankish, 'Partial Belief'. Unfortunately, space does not permit me to discuss Frankish's view here. See, however, footnote 28.14 Ross and Schroeder, 'Belief, Credence'.15 Ibid., 6.16 Note the similarity between SRD and the Act View discussed in Section III. The reasons to reject the former are essentially the reasons to reject the latter.17 Ibid., 9.18 See inter alia Fantl and McGrath, 2009, Knowledge in an Uncertain World; Hawthorne, Knowledge and Lotteries; Stanley, Knowledge and Practical Interests; Hawthorne and Stanley, 'Knowledge and Action'.19 Ross and Schroeder, 'Belief, Credence', 3.20 Ibid.21 Ibid., 13–15.22 Brown, 'Knowledge and Practical Reasoning, 1144.23 See Schwitzgebel, 'Belief'.24 Here is the argument, a bit more slowly.It is not rational for Liz to take the bet (assumption).Thus, it is not rational for Liz to treat b as true in her reasoning about whether to take the bet [by (1) and the Treating as True Principle].Thus, if Liz's believing that b essentially involves a disposition to treat b as true in her reasoning about whether to take the bet, then Liz is not justified in occurrently believing that b [by (2) and the Justification Condition on Occurrent Attitudes].Thus, Liz is not justified in believing that b [by (3) and URD].25 Locke, 'Pragmatic Sensitivity'.26 This account of pragmatic encroachment is what DeRose calls a 'warranted assertability maneuver' or 'WAM'. However, as I discuss elsewhere, the MRD-based account is importantly different from earlier WAMs, such as those defended by Rysiew and Brown. DeRose, 'Contextualism'; Locke, 'Pragmatic Sensitivity'; Rysiew, 'Context Sensitivity of Knowledge Attributions'; Brown, 'Contextualism'.27 Cohen, 'Why Acceptance that p', 60.28 This is essentially the same question as that asked by Keith Frankish. Whereas my answer looks to the particular way in which an agent treats a proposition as true in various choice situation, Frankish's answer concerns the particular sort of choice situations in which one is disposed to treat p as true. Frankish, 'Partial Belief'.29 See Goldman, 'What is Justified Belief?'. And please do not be misled by the Latin meanings of these phrases. Here they mean precisely what they are stipulated to mean.30 The distinction is perhaps more commonly drawn with respect to the notion of justification than rationality, but there is no reason why we should not draw the distinction with respect to both of these normative notions. I here set aside the vexing question of just how the two notions are related.31 Stanley, Knowledge and Practical Interests; Hawthorne and Stanley, 'Knowledge and Action'.32 As Neta convincingly argues, these two notions are not equivalent. Neta, 'Treating Something', 685.33 Adapted from Littlejohn, 'Must We Act'.34 Adapted from Neta, 'Treating Something'.35 Brown, 'Knowledge and Practical Reasoning'.36 In defense of these principles, some have appealed to the idea of pragmatic encroachment, which was introduced in the previous section. The idea here is that, contrary to appearance, Liz is not justified in believing that she was born in England, or does not know that she was born in England, or is not justified in believing that she knows that she was born in England, and this is precisely because Liz is in a situation in which it is not rational for her to act as if she was born in England. This response, however, is not very convincing, for cases like Liz's would seem to also be counterexamples to the strong forms of pragmatic encroachment needed to save the principles.37 Or, for causal decision theorists, the agent's confidence that S2 would be the case, were S1 the case. See Gibbard and Harper, 'Counterfactuals'.38 More precisely: 'the act or set of acts'. As indicated above, I suppress this sort of qualification throughout this paper.39 Locke, 'Practical Certainty'.40 The proof assumes that what S has most reason to do is whatever maximizes expected value as calculated by the degrees of confidence that S is rational to have. So calculated, the expected value of any action x is equal to V(x&p)Conf(p) + V(x&~p)(1-Conf(p)), where V(-) is the value of - and Conf(p) is the rational degree of confidence that p. Because this function is linear, the graphs of how the expected values of any two actions A and B depend on the rational degree of confidence that p, respectively, can cross only once. This means that, for whichever action x has highest expected value when the rational confidence that p is one, there is some degree d of confidence that p such that x has highest expected value if and only if the rational degree of confidence that p is greater than d. It is this d that I refer to as T(C) above. Thanks to Kenny Easwaran for suggesting this simple geometric proof.41 In any case, the prediction of indeterminacy is no reason to reject the DT Lockean Thesis, unless one assumes that there is always a determinate fact about whether an agent's belief is rational. That assumption is dubious at best.
Referência(s)