This paper is about how epistemic and practical reasons for belief can be compared against one another when they conflict. It provides a model for determining what one ought to believe, all-things-considered, when there are conflicting epistemic and practical reasons. The model is meant to supplement a form of pluralism about doxastic normativity that I call ‘Inclusivism’. According to Inclusivism, both epistemic and practical considerations can provide genuine normative reasons for belief, and both types of consideration can contribute to (metaphysically) determining what beliefs one ought, all-things-considered, to have.
This is a preview of subscription content, log in to check access.
Buy single article
Instant access to the full article PDF.
Price includes VAT for USA
Subscribe to journal
Immediate online access to all issues from 2019. Subscription will auto renew annually.
This is the net price. Taxes to be calculated in checkout.
I borrow the name ‘Alethism’ and this way of characterizing the view from Leary (2017). Many Alethists are Evidentialists—they hold that all reasons for belief are epistemic, and also that all epistemic reasons are evidence-based. But Alethism doesn’t require this commitment.
I borrow the name ‘Robust Pragmatism’ from Rinard (2015). Rinard counts moral considerations among practical reasons for belief, since the benefit indicated or (partly) grounded by a practical consideration might be a benefit to another, as opposed to oneself. I follow Rinard in this here.
With the exception of Pascal (1670) and James (1897), and perhaps Foley (1993, 2003), this view is rarely, if ever, explicitly endorsed in print. On one reading, Marušić (2016) can be interpreted as accepting the view, though he denies this in conversation. Reisner (2008) goes some way toward defending Inclusivism (what he calls “The Mixed View”) but never officially endorses it. Markovits (2010, 2014) also makes some apparently sympathetic remarks, but, again, never makes it official.
All but one of these authors provide no argument that epistemic and practical reasons can’t be compared against one another—they merely report the intuitive difficulty of seeing how such comparisons could be possible (one tends to encounter this in informal discussion of the matter, too). Berker (2018) is the exception. He surveys some existing and possible models and argues that none are satisfactory. Berker’s work will feature prominently in what follows.
In the historical literature, a possible model for comparing epistemic and practical reasons is suggested by some remarks made by William James (1897). However, as I explain below, there are good (epistemic) reasons to think that James’s model is unsatisfactory (see note 14). Reisner (MS) is currently developing a new model as part of a larger book project; I hope to discuss the final product in future work.
Strictly speaking, I think it’s a (fairly common) mistake to talk about comparing oughts, as opposed to comparing the reasons underlying each respective ought. Here, I’ve stated the position under consideration in the former way, since this is how those who endorse it tend to state it.
I’ll assume throughout that there are three alternative positive doxastic attitudes that one might take toward a given proposition: belief, disbelief, and suspension of judgment. I realize that it’s still somewhat (though not very) controversial that suspension of judgment constitutes a positive doxastic attitude. But for a powerful and rigorous defense of this claim, see Friedman (2013).
This difference in the weighing behavior of epistemic and practical reasons for belief mirrors the well-known asymmetry in the weighing behavior of epistemic reasons for belief and practical reasons for action (see, e.g., Feldman 2000; Harman 2004; Cohen 2016). If one has equally strong practical reasons for two or more alternative actions, and there are no other reasons for action in play, then one has sufficient reason to perform any of those actions, and lacks sufficient reason to perform any other. Likewise for practical reasons for belief. But not so with epistemic reasons for belief. As far as I’m aware, Berker (2018) was the first to point this out; see also, for discussion, Leary (forthcoming).
Berker (2018) raises a similar objection to the idea that what I’m calling a “simple weighing procedure” could be used to make comparisons among epistemic and practical reasons.
This point is due to Reisner (2008). As he puts it: in situations like the above, “it seems that no collection of [epistemic] reasons would be relevant in the face of the overwhelmingly strong pragmatic ones. The [epistemic] reasons, if we feel that they are not the only thing that count, seem […] to lose their force altogether in determining what [we] ought to believe” (p. 22). More on this below.
It’s worth noting a third potential difficulty for the idea that epistemic and practical reasons could be compared by a simple weighing procedure. A simple weighing procedure requires the assumption that the weights of the reasons being compared are commensurable, i.e., that there’s some common unit in terms of which the weights of the relevant reasons can be measured (Chang 1997). This assumption is needed for the first step of the procedure in which the weights of the reasons for and against a given alternative are combined in order to determine the total reason (simpliciter) in favor of that alternative. But the assumption that the weights of epistemic and practical reasons are commensurable is on its face implausible. For example, suppose we think the weight of an epistemic reason for believing P is determined by the degree to which the relevant reason indicates that believing P would promote (or somehow show respect for) the Jamesian goals of getting at the truth and avoiding error. And suppose we think the weight of a practical reason for believing P is determined by the degree of benefit the relevant reason indicates that believing P would bring about. Then, plausibly, there will be no common unit in terms of which the weights of the two types of reasons can be measured. And if this is so, then simple weighing is a non-starter. I omit discussion of this issue in the main text since there are several ways one might skirt it. For example, one might try to give a unified account of the weights of epistemic and practical reasons (though this seems a tall order), or take facts about the weights of reasons to be absolutely basic (and simply assume that the weights of epistemic and practical reasons are commensurable). No matter: either sort of view would face the difficulties raised in the main text.
In the historical literature, a model for comparing epistemic and practical reasons is suggested by some remarks made by James (1897). James is often interpreted as having held that practical reasons can be relevant to determining what one ought all-things-considered to believe only when the epistemic reasons are equally (or roughly equally) balanced in favor of belief and disbelief. In such cases, one’s practical reasons are supposed to “break the tie.” But while James’s model might satisfy the first of our desiderata, it certainly doesn’t satisfy the second. For example, in a case where my believing P is necessary to save the world, but the balance of epistemic reasons favors disbelieving P, James’s model yields the (implausible) result that I ought, all-things-considered, to disbelieve P. So James’s model is too sensitive in high-stakes cases to the total weight of epistemic reason. So the model doesn’t satisfy our second desideratum. And see Berker (2018) for some further worries.
Reisner’s suggested account of the weights of practical reasons for belief thus resembles a value-based account of reasons for action. See Maguire (2016) for discussion and defense of the latter.
I say that this is contentious only because one might think that there are some practical reasons for belief (e.g., certain moral ones) that aren’t benefit related. Were this so, the above assumption concerning the weights of practical reasons would need to be rejected. However, for the purposes of this paper, we can assume that all such apparently non-benefit-related practical reasons can be “consequentialized”. This is only for simplicity’s sake and won’t affect the arguments to follow.
Note also that Reisner’s proposal doesn’t require the assumption that the weights of epistemic and practical reasons are commensurable, and so it avoids the difficulty mentioned in note 13.
I borrow this case from Kelly (2002).
This discussion reveals an important advantage of Inclusivism over Robust Pragmatism. On the most natural way of developing the latter view, it will deliver the result that, in every case, you ought to hold the attitude the having of which would bring about the most benefit. But, as we’ve just seen, this is intuitively implausible. And even if we supplement Robust Pragmatism with the further (admittedly plausible) thesis that it’s in general beneficial to hold epistemically well-supported beliefs, there will still be possible low-stakes cases in which it would be slightly more beneficial overall to hold an epistemically unsupported belief, but we think that, despite this, you ought to have the attitude that’s best supported by the balance of your epistemic reasons. And at least without some fancy footwork, it’s hard to see how Robust Pragmatism could accommodate judgments of this kind. But, when paired with the right kind of model for comparing epistemic and practical reasons, Inclusivism clearly can accommodate such judgments, as the above discussion demonstrates (cf. Maguire and Woods forthcoming).
It’s worth noting, as Berker does (2018, p. 446), that shifting to this proposal would require Reisner to jettison the idea that epistemic reasons are silenced when practical reasons are relevant to determining all-things-considered verdicts (and vice versa) since, in cases of the kind that generate Berker’s problem, both practical reasons and the reason generated by the first (epistemic) weighing process are needed to determine the all-things-considered verdicts ultimately generated by the second weighing process.
Berker raises a different worry for the double-weighing view, which I don’t discuss in the main text, since I’m unconvinced that it poses a genuine problem. Berker (2018, p. 446) asks us to consider a (temporal or modal) sequence of cases in which our evidence for P steadily improves as we move through the sequence, while all other facts about our reasons remain fixed. We begin in a case where we have an “extremely strong” practical reason against believing P and a “fairly strong” epistemic reason in favor of disbelieving P (where this latter reason is evidenced-based). As our evidence for P improves, we transition to a case where our epistemic reasons to believe P are equally balanced with our epistemic reasons to disbelieve P, and then finally to a case where our epistemic reasons to believe P are much stronger than our epistemic reasons to disbelieve P. Berker finds objectionable the double-weighing view’s predictions concerning the all-things-considered status of disbelief in P during this process as we move through the sequence. At first, the double-weighing view predicts that we have decisive all-things-considered reason to disbelieve P; then, as we gain epistemic reasons to believe P, we come to lack sufficient all-things-considered reason to disbelieve P; but then, as we gain yet further epistemic reasons to believe P, we go back to having sufficient all-things-considered reason to disbelieve P. Berker finds this bizarre, particularly the last transition: “How can disbelief in P go from being forbidden (all things considered) to permitted (all things considered) merely in virtue of [our] gaining [epistemic] reasons to believe P? Such a result seems intolerable” (ibid.). But contrary to Berker, I don’t find this result particularly problematic. After all, it’s not the case that, in the final transition, disbelief in P goes from being forbidden to permitted (all things considered) merely in virtue of our gaining epistemic reasons to believe P. For when we gain epistemic reasons to believe P, in the final transition, two things happen: First, we find ourselves in a situation in which the attitude favored by the balance of our epistemic reasons (belief in P) is ruled out by our “extremely strong” practical reason against believing P. Second, we come to lack sufficient epistemic reason for either of that attitude’s alternatives (disbelief in P and suspension of judgment). In virtue of these facts, the double-weighing view predicts that we lack sufficient all-things-considered reason to believe P, and have sufficient all-things considered reason either to disbelieve P or to suspend judgment. And these verdicts seem to me sensible.
Like Reisner, I’m agnostic about whether the threshold that figures in my model is best understood as being fixed or contextually variable, vague or sharp—but some discussion to follow (in Sect. 5).
The discussion in this paragraph assumes that epistemic permissivism is false. If epistemic permissivism is true, then there are possible cases in which the balance of epistemic reasons sufficiently favors multiple doxastic alternatives. And if such cases are possible, then the verdicts yielded by Reisner’s model and mine will diverge in one more kind of case (other than that which I’ve already indicated), viz., cases in which the weight of the practical reasons for (or against) any of the relevant doxastic alternatives isn’t above the threshold, the balance of epistemic reasons sufficiently favors multiple alternatives, and the practical reasons favor one of those alternatives over the other(s). In such cases, my model predicts that you have decisive all-things-considered reason to hold the attitude that’s best supported by the practical reasons among those that are compatible with the prescription of the epistemic reasons (i.e., the most beneficial attitude among those that are sufficiently favored by the balance of epistemic reasons). Reisner’s model, on the other hand, predicts that you have sufficient all-things-considered reason to hold either of the attitudes for which there’s sufficient epistemic reason. Unsurprisingly, I’m inclined to think that the verdict that my model delivers is preferable. Although I want to remain officially neutral about whether epistemic permissivism is true, the rest of the paper is written as if the view false. Beyond what I’ve said in this note, however, nothing important turns on this.
Suppose that the weight of the practical reasons is above the threshold, the preponderance of practical reason favors suspending judgment on P, but disbelieving P is a close practical second (perhaps suspending judgment on P comes with a bonus donut). And suppose also that I have decisive epistemic reason to disbelieve P. In this situation, Reisner’s model and mine (as I’ve stated it above) yield the verdict that I have decisive all-things-considered reason to suspend judgment on P. I find this verdict acceptable, but some people report finding it intuitive that in this case I have at least sufficient all-things-considered reason to disbelieve P. To accommodate this judgment, one could simply specify that practical reasons for belief are to be compared in a satisficing way, such that one has sufficient practical reason for a given doxastic attitude just in case the total weight of practical reason in favor of the attitude is above a certain threshold. If disbelieving P is a close practical second to suspending judgment on P, then if the total weight of practical reason in favor of suspending judgment on P meets the relevant threshold, then, presumably, so too will the total weight of practical reason in favor of disbelieving P. And if so, then my model will yield the verdict that I have sufficient all-things-considered reason to hold either of these attitudes, and lack sufficient all-things-considered reason to believe P. Many thanks to an anonymous referee at this journal for pushing me on this kind of case.
My model achieves this result in a different way than Reisner’s. The “seems” qualifier is important: the practical reasons aren’t irrelevant, but only lexically posterior in the relevant cases.
Much of my thinking about this question, and the charge of arbitrariness that any specific answer to it might invite, is inspired and informed by Kagan’s (1998) commentary concerning thresholds in his discussion of the prospects for moderate (threshold) deontology (see especially pp. 78–84).
This point shouldn’t be overstated. Both Alethists and Robust Pragmatists need to specify plausible comparison functions for determining all-things-considered doxastic verdicts. And at least for Alethists, worries about arbitrariness loom. This is because many Alethists favor the view that whether you ought to believe something isn’t a matter of whether there is more or most epistemic reason for you to believe it, but rather of whether there is sufficient epistemic reason for you to believe it (i.e., these theorists opt for a satisficing way of comparing the total epistemic reason in favor of each of the relevant doxastic alternatives in order to determine all-things-considered verdicts). But precisely how much epistemic reason must there be in favor of a given doxastic option in order for the epistemic reason in its favor to be sufficient? Worries about arbitrariness seem as potent here as they do for my own proposal. (Of course, Alethists who aren’t also satisficers won’t face this worry; see Schroeder 2015 for discussion).
One suggestion that doesn’t particularly help us here, but which is nonetheless worthy of note, is that the location of the threshold might vary according to the “epistemic importance” of the proposition under consideration. Roughly: the more epistemically important the relevant proposition, the higher the threshold. This is analogous to the common view among moderate (threshold) deontologists that the location of the threshold that their theory posits is in part a function of the nature of the particular deontological constraint to be infringed upon. In principle, I’m highly sympathetic to this suggestion. My only apprehension is that I find it difficult to grasp a notion of “epistemic importance” that doesn’t collapse into practical importance—in particular, the practical importance of having an epistemically well-supported attitude with respect to the relevant proposition. For a helpful discussion of this puzzle, and of what’s needed to answer it, see Treanor (2014).
What is it for a normative factor to be intrinsically normatively significant? I suggest the following (first pass) account: For a normative factor, F, to be intrinsically normatively significant is for F to be normatively significant just in virtue being the kind of factor that it is, and not in virtue of any relation it might bear to a normative factor of any other kind. Epistemic considerations are thus intrinsically normatively significant for belief, inasmuch as they’re normatively significant just in virtue of being considerations of epistemic relevance, and not in virtue of, e.g., their being of practical import (if they are). This notion—that of intrinsic normative significance—provides a straightforward way of distinguishing Inclusivism from its competitors. For example, Robust Pragmatists (e.g., Rinard 2015) claim that practical considerations are intrinsically normatively significant for belief, but deny that epistemic considerations are. Instead, these theorists claim that epistemic considerations are normatively significant for belief only when (and because) believing in accordance with them would provide some benefit (to oneself or another). But the Inclusivist denies this. Her core commitment is that both epistemic and practical considerations are of intrinsic normative significance for belief.
Baker, D. (2018). Skepticism about ought simpliciter. Oxford Studies in Metaethics,13, 230–252.
Berker, S. (2018). A combinatorial argument against practical reasons for belief. Analytic Philosophy,59, 427–470.
Booth, A. (2008). A new argument for pragmatism? Philosophia,36, 227–231.
Chang, R. (1997). Introduction. In Ruth Chang (Ed.), Incommensurability, incomparability and practical reason (pp. 1–34). Cambridge: Harvard University Press.
Chappell, R. (2012). Fittingness: The sole normative primitive. Philosophical Quarterly,62, 684–704.
Cohen, S. (2016). Reasons to believe and reasons to act. Episteme,13, 427–438.
Copp, D. (1997). The ring of gyges: overridingness and the unity of reason. Social Philosophy and Policy,14, 86–101.
Feldman, R. (2000). The ethics of belief. Philosophy and Phenomenological Research,60, 667–695.
Foley, R. (1993). Working without a net: A study of egocentric epistemology. New York: Oxford University Press.
Foley, R. (2003). Justified belief as responsible belief. In M. Steup & E. Sosa (Eds.), Contemporary debates in epistemology (pp. 313–326). Malden: Blackwell Publishing.
Friedman, J. (2013). Suspended judgment. Philosophical Studies,162, 165–181.
Harman, G. (2004). Practical aspects of theoretical reasoning. In A. R. Mele & P. Rawling (Eds.), The oxford handbook of rationality (pp. 45–56). Oxford: Oxford University Press.
Howard, C. (2016). In defense of the wrong kind of reason. Thought A Journal of Philosophy,5, 53–62.
Howard, C. (forthcoming). The fundamentality of fit. Oxford Studies in Metaethics 14.
James, W. (1897). The will to believe and other essays in popular philosophy. Cambridge: Harvard University Press.
Kagan, S. (1998). Normative ethics. Boulder: Westview Press.
Kelly, T. (2002). The rationality of belief and some other propositional attitudes. Philosophical Studies,110, 163–196.
Leary, S. (forthcoming). Grounding the domain of reasons. Australasian Journal of Philosophy.
Leary, S. (2017). In defense of practical reasons for belief. Australasian Journal of Philosophy,95, 529–542.
Maguire, B. (2016). The value-based theory of reasons. Ergo 8.
Maguire, B., & Woods, J. (forthcoming). The game of belief. Philosophical Review.
Markovits, J. (2010). Internal reasons and the motivating intuition. In M. Brady (Ed.), New waves in metaethics (pp. 141–165). Basingstoke: Palgrave Macmillan.
Markovits, J. (2014). Moral reason. New York: Oxford University Press.
Marušić, B. (2016). Evidence and agency: Norms of belief for promising and resolving. Oxford: Oxford University Press.
McCormick, M. (2014). Believing against the evidence: Agency and the ethics of belief. New York: Routledge.
McLeod, O. (2001). Just plain ‘ought’. Journal of Ethics,5, 269–291.
Papineau, D. (2013). There are no norms of belief. In T. Chan (Ed.), The aim of belief (pp. 64–79). Oxford: Oxford University Press.
Parfit, D. (2011). On what matters. Oxford: Oxford University Press.
Pascal, B. 1670 (1910). Pensées, tr. W.F. Trotter, London: Dent.
Reisner, A. (2008). Weighing pragmatic and evidential reasons for belief. Philosophical Studies,138, 17–27.
Reisner, A. (2009). The possibility of pragmatic reasons for belief and the wrong kind of reasons problem. Philosophical Studies,145, 257–272.
Reisner, A. (MS). The pragmatic foundations of theoretical reason.
Rinard, S. (2015). Against the new evidentialists. Philosophical Issues,25, 208–223.
Rinard, S. (2017). No exception for belief. Philosophy and Phenomenological Research,94, 121–143.
Ross, W. D. (1930). The right and the good. Oxford: Clarendon Press.
Schroeder, M. (2015). What makes reasons sufficient? American Philosophical Quarterly,52, 159–170.
Shah, N. (2006). A new argument for evidentialism. Philosophical Quarterly,56, 481–498.
Steglich-Peterson, A. (2011). How to be a teleologist about epistemic reasons. In A. Steglich-Peterson & A. Reisner (Eds.), Reasons for belief (pp. 13–33). Cambridge: Cambridge University Press.
Stich, S. (1990). The fragmentation of reason. Cambridge: MIT Press.
Thomson, J. J. (2008). Normativity. Chicago: Open Court.
Tiffany, E. (2007). Deflationary normative pluralism. Canadian Journal of Philosophy,37, 231–262.
Treanor, N. (2014). Trivial truths and the aim of inquiry. Philosophy and Phenomenological Research,89, 552–559.
Way, J. (2016). Two arguments for evidentialism. Philosophical Quarterly,66, 805–818.
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
About this article
Cite this article
Howard, C. Weighing epistemic and practical reasons for belief. Philos Stud 177, 2227–2243 (2020). https://doi.org/10.1007/s11098-019-01307-y
- Ethics of belief
- Epistemic reasons
- Practical reasons