An instrumentalist unification of zetetic and epistemic reasons

ABSTRACT

Inquiry is an aim-directed activity, and as such governed by instrumental normativity. If you have reason to figure out a question, you have reason to take means to figuring it out. Beliefs are governed by epistemic normativity. On a certain pervasive understanding, this means that you are permitted – maybe required – to believe what you have sufficient evidence for. The norms of inquiry and epistemic norms both govern us as agents in pursuit of knowledge and understanding, and, on the surface, they do so in harmony. Recently, however, Jane Friedman (2020, “The Epistemic and the Zetetic.” The Philosophical Review 129: 501–536) has pointed out that they are in tension with each other. In this paper, I aim to resolve this tension by showing that reasons for acts of inquiry – zetetic reasons – and epistemic reasons for belief can both be understood as flowing from the same general normative principle: the transmission principle for instrumental reasons. The resulting account is a version of epistemic instrumentalism that offers an attractive unity between zetetic and epistemic normativity.

KEYWORDS:

• Inquiry
• zetetic reasons
• epistemic reasons
• instrumental reasons
• epistemic instrumentalism
• normativity

Disclosure statement

No potential conflict of interest was reported by the author(s).

Notes

1 As Friedman (2020, fn 1) explains, ‘zetetic’ derives from the Greek verb ‘ζητϵω’, which translates to ‘seek for’ or ‘inquire after’.

2 For a compelling account, see Maddox (2003).

3 For exceptions to this tendency, see e.g. Friedman (Forthcoming; 2020; see also her 2018 and 2019), Kelp (2021a, 2021b), Hookway (2006, 2008), and Thorstad (2021).

4 The example has been modified slightly for my purposes.

5 Exactly how the zetetic norms might have this implication is a question that I will take up in the next section.

6 For discussion, see e.g. Reisner (2008, 2009), Berker (2018), and Steglich-Petersen & Skipper (2019).

7 For a helpful overview and discussion of the main prevailing approaches to epistemic normativity, see Hazlett (2013, Part II).

8 This case is inspired by Rawls’ famous counterexample to desire-based theories of well-being (Rawls 1971, 432).

9 For a recent argument that the tension between epistemic and zetetic norms should not lead us to revise either set of norms, see Thorstad (2021).

10 For defenses of knowledge as the aim of inquiry, see Kelp (2021b, 2018, 2021a, 2014).

11 This argument against narrow-scope principles of instrumental rationality is known as the ‘bootstrapping problem’ (Bratman 1981). This has often been taken to support a wide-scope interpretation of instrumental rationality, e.g. by Darwall (1983), Broome (1999), and Wallace (2001). For criticism of the wide-scope interpretation, see e.g. Setiya (2007), Bedke (2009), Schroeder (2009), and Kiesewetter (2017).

12 For versions of conservative transmission principles, see e.g. Darwall (1983, 16), Setiya (2007, 660), Bratman (2009, 424), Schroeder (2009, 234), Way (2010, 225), Scanlon (2014, 85), and Kiesewetter (2015).

13 Defenses of or reliance upon liberal transmission principles can be found in Raz (2005), Bedke (2009), Way (2012), Schroeder (2007, 2009), and Kolodny (2018).

14 Although popular, liberal transmission principles are not without detractors. The most common objection is that such principles would allow the transmission of reasons to means that are highly inefficient or very objectionable in other respects. For example, it would allow my reason to become rich to transmit to the highly inefficient means of taking a degree in philosophy; and, to repeat a famous example, it would allow my reason to avoid a feeling of hunger to transmit to the very objectionable means of killing myself (Broome 2005, 7). Some authors have a strong intuition that no reason whatsoever can transmit to such means, and that defenders of the liberal transmission principle must at the very least explain away this intuition. A common strategy here appeals pragmatic considerations: when calling something a ‘reason’, it is pragmatically implied that it is not massively outweighed by other considerations. This strategy is taken by Schroeder (2005, 2007), Raz (2005), Bedke (2009), Way (2012), and Kolodny (2018). Speaking for myself, I don’t share the intuition that there is no transmitted reason to take the means in the mentioned examples, and hence I don’t think that there is an intuition to be explained away. Motivating this, however, would take us too far afield. For a recent critical discussion of liberal transmission principles, see Kiesewetter and Gertken (2021).

15 Why should the relevant probability not be identified with objective chance? Partly, this choice reflects a general commitment to a kind of ‘perspectivism’ about one’s possessed or available reasons (for both actions and beliefs) as being sensitive to one’s evidence. It may make sense to also operate with an objective sense of reasons that exist independently of one’s evidence. Such reasons, however, would not have direct implications for what one ought to do. For recent defenses of perspectivism, see e.g. Kiesewetter (2017, 2018) and Lord (2018).

16 Why not simply let the strength of the transmitted reason depend on the probability of the aim being realised conditional on taking the means, rather than letting it depend on the probability of the means helping to bring about the aim conditional on taking the means? Kolodny’s rationale for this rests on cases of actions that raise the probability of an aim being realised without thereby being means to them (see Kolodny 2018, §1). For example, a boxer who aims to land a punch might grit his teeth every time he throws a punch, in which case the probability of landing a punch conditional on his gritting his teeth will be higher than otherwise. Note, however, that even on Kolodny’s account, transmitted reasons for means will be stronger, the more likely the means are to achieve the relevant aims, since, trivially, the more likely it is that a means will achieve an aim, the more likely it will be that the means help bring about the aim. In what follows, I will thus largely ignore this complication, and simply consider whether the relevant means make achieving the aim more likely. For discussion and an alternative probabilistic transmission principle, see Bedke (2017).

17 Some readers might worry about potential counterexamples where there is reason to figure out a question, and positive probability that some action will help figuring out the question, but where there intuitively is no transmitted reason to perform that action. For example, I might have reason to figure out whether I can complete a strenuous race, and be aware that injuring my foot would ensure that I will not be able to complete the race. In that case, injuring my foot would help me figuring out if I can complete the race. But do I thereby have reason to injure my foot? It seems not. What is peculiar about this case, I take it, is that I bring myself into a position of being able to correctly answer the question by causally affecting what the correct answer is. If we wish to exclude the possibility of transmission of reasons to such acts, we can include an amendment to the principle to that effect, but I will not attempt to do so here. I am grateful to an anonymous reviewer for bringing up this potential counterexample.

18 The idea of constitutive means harks back at least as far as Aristotle’s discussion of activities that are constitutive rather than causal means for eudaimonic well-being (Aristotle 1984). In a more recent context, the idea is endorsed by e.g. Kolodny (2018) and Kiesewetter & Gertken (2021). For critical exchange, see Audi (2001) and Marras (2003).

19 This idea mirrors (but is not equivalent to) the popular idea that knowing-wh (e.g. knowing what, knowing whether, knowing who, etc.) is a matter of knowing that p, where p is the true answer to the indirect question embedded in the wh-clause. For example, on this account, if the game begins at 4pm, knowing what time the game begins is a matter of knowing that the game begins at 4pm. For discussions of this account, see Hintikka (1975), Lewis (1982), Higginbotham (1996), Schaffer (2007), and Steglich-Petersen (2014).

20 This reflects the common idea in formal semantic theories of questions that the meaning of a question is determined by the set of propositions that could constitute possible answers. So, the meaning of the question ‘How many windows are there in the Chrysler Building’ is determined by the set of propositions [There is one window in the Chrysler Building; There are two windows … etc.], and an interest in the question thus implies an interest in each of the propositions constituting its meaning. For this theory, see Hamblin (1973) and Karttunen (1977).

21 Steglich-Petersen (2018) and Steglich-Petersen & Skipper (2019; 2020). See also Steglich-Petersen (2006, 2009, 2011, 2013) for earlier versions.

22 For versions of this objection, see e.g. Kelly (2003), Schroeder (2007), and Côté-Bouchard (2015).

23 For traditional forms of instrumentalism, that take the aims of believers rather than reasons to pursue aims as the normative source of epistemic reasons, see e.g. Quine (1969), Foley (1987), Laudan (1990), Kornblith (1993), Nozick (1993), and Papineau (1999). Some instrumentalists appeal, like I do, to aims that one has reason to have, but do not spell out this idea in terms of the transmission principle, e.g. Grimm (2009) and Cowie (2014).

24 What is the relationship between the proposed account of epistemic reasons and evidentialism about epistemic reasons? If we take evidentialism to be the strict view that epistemic reasons consist of evidence and nothing else, then the proposed account is not a version of evidentialism, since it makes epistemic reasons depend partly on reasons for pursuing true beliefs, and these reasons are not evidential. The account seems, however, to retain some of the spirit of evidentialism, by making evidence for p necessary for having epistemic reason to believe that p. Furthermore, although I don’t myself hold that view, the account is compatible with thinking that we have reason to pursue true beliefs about all propositions, in which case the account becomes extensionally equivalent to the stricter version of evidentialism, in the sense of implying the same epistemic reasons.

25 Among the objections: we tend to talk as if evidence in itself constitute reasons for belief, and transmitted instrumental reasons have various properties that epistemic reasons don’t seem to share. Among the virtues: this version of epistemic instrumentalism explains the distinctiveness of epistemic reasons – something that other brands of instrumentalism have trouble with. Epistemic reasons are reasons that speak in favor of beliefs in virtue of the beliefs being likely to constitute the aim of inquiry. See e.g. Steglich-Petersen (2018) and Steglich-Petersen & Skipper (2019, 2020) for discussion.

26 For more general reasons to seek an account of epistemic normativity that is unified with practical normativity, see e.g. Steglich-Petersen (2018) and Rinard (2019).

27 When is there sufficient reason for belief, on this understanding? That question is discussed in Steglich-Petersen & Skipper (2020).

28 For a more comprehensive account of how to weigh instrumental reasons, see Steglich-Petersen & Skipper (2019).

Additional information

Funding

This work was supported by Carlsbergfondet: [Grant Number CF19-0160].