![Sample our Humanities journals, sign in here to start your FREE access for 14 days]({"type":"image" , "src" : "/sda/5939/TOC-Subject-Sample-2021-Humanities.jpg"})
ABSTRACT
One important distinction in the debate over the nature of epistemic justification is the one between propositional and doxastic justification. Roughly, while doxastic justification is a property of beliefs, propositional justification is a property of propositions. On a rather common view, which accounts for doxastic justification in terms of propositional justification plus the so-called ‘basing relation’, propositional justification is seen as the prior notion, and doxastic justification is explained in terms of it. According to the opposing view, the direction of explanation needs to be reversed, and doxastic justification should be seen as primary. I distinguish between two notions of priority, and I argue that they give different verdicts with respect to the issue of which notion of justification comes first. The lesson may be taken to be that propositional and doxastic justification are in a relation of intertwinement.
Notes
1 Just a blunt example to drive home the point: Inspector Rebus can only find out about the details of the murder by cognitively engaging with the circumstantial evidence available, but the facts that constitute the details of the murder did not obtain in virtue of the circumstantial evidence.
2 For example, facts about biology obtain at least in part in virtue of facts about physics, and yet one can cognitively engage with biology without having to cognitively engage with physics.
3 See Pollock and Cruz [Citation1999: 35–6], Feldman [Citation2003: 46], and Lasonen-Aarnio [Citation2010: 205–6] for just some of the many examples of this way of drawing the distinction. The distinction, and indeed the choice of the terms ‘propositional’ and ‘doxastic’, goes back at least to Firth [Citation1978: 218].
4 I use the preposition ‘to’ to refer to propositional justification (as in ‘S is justified to believe that p’), and the preposition ‘in’ to refer to doxastic justification (as in ‘S is justified in believing that p’).
5 As an anonymous referee helpfully observed, one can have propositional justification to believe p and yet fail to be doxastically justified in believing p, in at least two different ways: one might fail to form the belief that p, or one might fail to base it on the relevant reasons.
6 Turri [2010: sec. 2] argues that propositional justification plus the basing relation is not sufficient for doxastic justification. In addition, he contends, the agent has to use the basis in the right way. While this much seems right, there is no need to discuss Turri's examples for the present purposes. For more on the debate raised by Turri's influential paper, see footnote 12 below.
7 I take it that this is also what grounds the distinction in ordinary discourse: for example, it would seem that when one criticizes a creationist for failing to believe that dinosaurs existed, despite all of the fossil evidence available, one is criticizing the creationist for failing to be doxastically justified in believing that dinosaurs existed despite having propositional justification to that effect.
8 The reasons for my choices in the next two sections are explained in notes 13, 22, and 23.
9 A suggestion along these lines can be found in Fumerton [Citation2006: 36]. Kvanvig and Menzel [Citation1990] may not have explicitly mentioned conceptual priority, but have defended the priority of propositional justification as a result of the consideration that doxastic justification equals propositional justification plus basing. See Turri [Citation2010: 313–14] for a more comprehensive list of philosophers who have argued along similar lines.
10 Williamson [2000: 3] makes a similar point in a different discussion.
11 Similar claims had already been advanced by Goldman [Citation1979: 345] and Alston [Citation1985: 104], and considered by Firth [Citation1978: 220]. Zardini [Citation2014: 34] introduces the distinction between propositional and doxastic justification in the same way.
12 As recalled in note 6, Turri argues that propositional justification plus the obtaining of the basing relation is not sufficient for doxastic justification. Through the discussion of some examples, he goes on to draw the lesson that it is a mistake to endorse the orthodox explanation of doxastic justification in terms of propositional justification. Silva [Citation2015] argues that the problem for the orthodox account highlighted by Turri is amenable to a rather straightforward solution. While this paper does not engage with Turri's or Silva's examples, the present section may be described as aiming to vindicate what's appealing in Turri's suggestion, and section 3 may be broadly portrayed as an attempt to defend the orthodoxy attacked by Turri. See Vahid [Citation2016: sec. 1] for an overview of the debate on Turri's paper.
13 Let me emphasize again that I intend to remain neutral on what is the best characterization of propositional justification. The only reason why I am choosing to discuss the characterization in terms of evidential likelihood in this section is that, on the face of it, it is rather promising with respect to the task at hand—namely, describing propositional justification without appealing, at some level, to doxastic justification.
14 A worry along these lines is raised by Turri [Citation2010: 321–2] and Boghossian [Citation2014: sec. 5].
15 One concern here is that such a move over-intellectualizes the situation. Since some agents seem to form justified beliefs without any grasp of the relevant relation of epistemic support—think about justified beliefs of very young children, for example—we might wonder whether a notion of propositional justification that imposes the requirement that the agent needs to be able to grasp the support relation between e and p can apply to all intended agents. For an early complaint along these lines, see Alston [Citation1983: 84ff].
16 More on this point in a minute: see the upcoming discussion of the thought that evidence can be evidence only for some epistemic agent.
17 See also Conee and Feldman [Citation1985: 86–7].
18 The two notions of propositional justification have also been distinguished by Ichikawa and Jarvis [Citation2013: 163] and Coliva [Citation2014: 254]. However, the terms used to describe the properties differ. What I call ‘objective propositional justification’ is called simply ‘propositional justification’ by Ichikawa and Jarvis, and simply ‘propositional warrant’ by Coliva. What I call ‘ordinary propositional justification’ is called ‘ex-ante justification’ by Ichikawa and Jarvis, and ‘warrant that is available to the subject’ by Coliva. In general, there seems to be some terminological confusion in the literature: contrary to Ichikawa and Jarvis and Coliva, some philosophers, like Turri [Citation2010] and Boghossian [Citation2014], use the expression ‘propositional justification’ to refer to what I call ‘ordinary propositional justification’.
19 Here's an example. Suppose that there is a strong statistical correlation between ‘S speaks good English’ and ‘S is British.’ For an agent who, through a quick chat, is trying to identify the British citizens in a group of mixed nationalities, the first proposition may be good evidence for the second; for an agent who is trying to figure out who might speak English by reading details in some document about the members of the mixed group, the second proposition may be good evidence for the first.
20 That is why I did not have to engage with any specific probabilistic theory of justification in order to make the point.
21 For example, in a theory of car driving, the concept of a steering wheel has theoretical priority over the notion of turning (the theory will say that one has to operate the steering wheel in order to turn the car), but it is not conceptually prior to it (one can grasp the notion of turning without grasping the notion of steering wheel). Similarly, as we saw in the previous section, a Bayesian theory might assign theoretical priority to propositional justification, but that doesn't show that propositional justification also enjoys conceptual priority.
22 At the risk of coming across as pedantic, let me recall that the relevant notion of reasons is very broad. I do not wish to suggest that in this section propositional justification is still to be understood in terms of evidential likelihood. Rather, at this stage it should be understood in the most theoretically neutral way: namely, as the justification that does not entail belief (regardless of what general theory of epistemic justification one might favour). The characterization of justification in terms of evidential likelihood was only adopted for the purposes of section 2 and, as will soon be clear, for the purposes of section 3 it will be helpful to adopt a different view of justification.
23 Note, however, that the focus on the evidentialist account in the previous section was warranted: had we found that the broad evidentialist view considered could vindicate the conceptual priority of propositional justification, we would have falsified the claim that all general characterizations of justification assign conceptual priority to doxastic justification. That was the claim under investigation in the previous section, and the focus on evidentialism was due to the fact that it offered a prima facie good candidate for delivering that result. By contrast, in this section the claim that we will be trying to falsify is that all general characterization of justification assign theoretical priority to propositional justification, and we will work with a general account of justification—namely, reliabilism—that has at least some prima facie credibility to successfully carry out the task.
24 As suggested in section 1, it is not obvious that the way in which we think or talk about reality (or about a specific subject matter) should mirror the hierarchical relations that hold in reality itself. If so, metaphysical priority, in and of itself, does not guarantee representational priority (see note 2). And, since I have characterized theoretical priority as a species of representational priority, the theoretical priority of propositional justification over doxastic justification would not follow merely from establishing that propositional justification has metaphysical priority. While a discussion of metaphysical priority goes beyond the scope of this paper, it is worthwhile noting that the view recently explored in Vahid [Citation2016]—according to which propositional justification is a dispositional property and doxastic justification is its manifestation—may be seen as ascribing metaphysical priority to propositional justification (to the extent that the relation between a disposition and the occurrence of the relevant manifestation is characterized as distinctively metaphysical in nature). See Ichikawa and Jenkins [forthcoming: secs 3–4] for some discussion of the relationship between the two main families of priority.
25 This is very similar to what an advocate of the conceptual priority of doxastic justification says propositional justification is. Compare Bergmann's and Turri's suggestion, mentioned in section 2.
26 One might wonder what it is for a belief-forming process to be available to the subject. Goldman himself admits that the issue is somewhat vague, but there seem to be some good examples. Suppose that, whilst having a conversation with someone, you are distractedly watching a dog playing with a ball in the park. As it happens, you do not form any belief about the dog, but, since you are looking at it, there is a clear sense in which you have propositional justification to believe that there is a dog playing in the park: the proposition ‘there is a dog playing in the park’ is within reach of a reliable belief-forming process of yours (which involves vision).
27 Indeed, Goldman himself [Citation2008: 64] draws the distinction between propositional and doxastic justification in these terms.
28 Such as, for example, that propositional justification needs to be understood as an abstract relation between propositions.
29 See Turri [Citation2010] and the references offered in note 11 for very similar characterizations of propositional justification. See also Becker [sec. 2.a] for the attribution to Goldman of the propositional/doxastic distinction.
30 I would like to thank Carrie I. Jenkins, Crispin Wright, Jonathan J. Ichikawa, and Jim Pryor for discussion. The article has also benefitted from the feedback provided by anonymous referees and by audiences at the Autonomous University of Madrid, the University of Aberdeen, and the University of Manchester.