Abstract
Earl Conee and Richard Feldman have argued that epistemic support should be understood in terms of explanatory considerations. Very roughly, they hold that one’s evidence supports a given proposition when that proposition is part of the best explanation of one’s evidence. This proposal is attractive, but T. Ryan Byerly has recently argued that it is false. Byerly claims that such explanationist accounts of epistemic support cannot account for the fact that one’s evidence can support propositions about the future. Although Byerly presents an interesting challenge, his argument is problematic and, ultimately, unconvincing.
This is a preview of subscription content, log in via an institution to check access.
Similar content being viewed by others
Notes
Conee and Feldman (2008) p. 83 In their earlier work (1985) they describe Evidentialism as: “EJ Doxastic attitude D toward proposition p is epistemically justified for S at t if and only if having D toward p fits the evidence S has at t” (15). There are two notable differences between EJ and Conee and Feldman’s most recent formulation of Evidentialism. First, the most recent formulation is in terms of belief instead of doxastic attitudes more generally. However, Conee and Feldman (2008) make it clear that E can be formulated in terms of other doxastic attitudes as well. Second, EJ references “fit” with evidence whereas E is presented in terms of “support”. This is merely a terminological difference because belief that p “fits” the evidence just in case p is “supported” by the evidence.
Some might disagree with this assessment on the grounds that epistemic support is basic. However, even if epistemic support is basic and cannot be given a reductive analysis, before E provides much of a theory one would still need to provide conditions under which a given body of evidence on balance supports a proposition. Without providing such conditions E would not give us much guidance in determining whether particular beliefs are justified. Thanks to an anonymous reviewer for suggesting that some might think that epistemic support is basic.
All references to Byerly’s argument are from this source. Page numbers refer to typescript of the online version.
See their (2008).
Earl Conee (private conversation) has suggested that he and Feldman in fact only intended to provide a sufficient condition for epistemic support. If Conee and Feldman only mean to provide sufficient conditions, Byerly’s objection has no hope of being effective because he only argues against the necessity of Conee and Feldman’s conditions. Despite these facts, it is worth considering whether the objection that Byerly raises would succeed against a stronger form of Conee and Feldman’s account which provides necessary and sufficient conditions for epistemic support. So, for the purposes of exploring Byerly’s objection I am willing to consider Conee and Feldman’s account as providing necessary and sufficient conditions for epistemic support.
I think that it is in fact quite probable that Conee and Feldman have something else in mind in light of their claim that “inferences justify by identifying to one further propositions that either require inclusion in one’s best explanation for it to retain its quality or enhance the explanation to some extent by their inclusion” (98). I will set this worry aside for now, but I will return to it when I explain why Byerly’s argument is problematic.
Saying precisely what circumstances C are is difficult, but the general idea is straightforward. Presumably, they are the circumstances described in Byerly’s case that are relevant to the golf ball going into the cup. They include things like the weather conditions, the speed of the ball, the line of the ball relative to the cup, etc.
One might worry that the fact that all golf balls rolling toward a cup in circumstances C go into the cup does not really explain why the particular golf balls that Byerly observes behave as they do. This is true. The general fact about golf balls in circumstances C does not explain what it is about the golf balls that Byerly observes that makes them go into the cup. However, this does not mean that <all golf balls rolling toward a cup in circumstances C go into the cup> is not part of the best explanation of Byerly’s evidence that <all observed golf balls rolling toward a cup in circumstances C have gone into the cup>. The explanadum here is the fact that all observed golf balls in C have went into the cup, which is very different from the fact that the particular golf balls Byerly observes go into the cup. Since our explanadum concerns the behavior of observed golf balls in C, the behavior of all golf balls in C (observed and unobserved) is relevant. As White (2005:9) explains, when we are explaining why all observed Fs are G “the properties of unobserved things is crucial” because “if the unobserved Fs are G, then it is to be expected that we only observe Fs which are G.” So, it is plausible that in this case part of the best available explanation of Byerly’s evidence is that <all golf balls rolling toward a cup in circumstances C go into the cup>. After all, this proposition helps account for the fact that all observed golf balls in circumstances C have went into the cup and it does so in an intuitively simple manner. Thanks to an anonymous reviewer for pushing me for further explanation of this point.
One might insist that such generalizations are themselves laws of nature. It is not clear that such a view is correct, but even if it is, there is no problem for the modified version of EXP + E. First, it is plausible that ordinary thinkers do have these sorts of generalizations available as parts of explanations. Second, these sorts of generalizations do not rule out indeterministic laws a priori. So, whether these generalizations are laws or not they do not pose a problem for the modified version of EXP + E.
This view has been defended by Harman (1973). But, this position is controversial. Salmon (1989) has argued that many entailment relations are explanatorily irrelevant. Most notably, in his classic counterexample to the D-N model of explanation Salmon argues that the although the length of a flagpole’s shadow and the position of the sun entail that the flagpole is a particular height, the shadow and the position of the sun do not explain the flagpole’s height. So, he claims that entailment does not always amount to explanation. It is worth noting that Salmon’s arguments do not themselves establish that the fact that a proposition is entailed by an explanation is not itself sufficient for that proposition to be part of the explanation. Rather, Salmon’s arguments show that the fact that some propositions entail another proposition is not sufficient for the former to explain the latter. These are very different claims. That being said, there are some difficult issues that require working through before this sort of response to Byerly’s challenge could be deemed successful. However, this response is worth mentioning because until these issues have been settled it remains an option that deserves exploring.
It may be that this is the sort of case that Byerly had in mind all along. Byerly does say that his putting stroke has been working “most of the day”. Of course, this is a bit ambiguous between Byerly’s only putting most of the day, but with perfect accuracy and his putting with very good, but still imperfect accuracy. Regardless, both versions of the case are worth considering because either could be thought to pose a problem for explanationist accounts of epistemic support.
Thanks to an anonymous reviewer for raising this objection.
Another way that one might be tempted to respond to this sort of objection is to modify EXP + E by making use of probabilistic relations instead of entailment. So, one might be tempted to modify EXP + E to something like the following: S is justified in believing that p at t iff p is part of the best explanation for e available to S at t or p is available to S at t as probable consequence of the best available explanation of e. It might seem that such an account would offer a more straightforward response to this objection than the two ways of responding outlined here. There are, however, some significant drawbacks to such a modification. First, it is not clear what sort of probability would be operative here. Objective probability may sacrifice the internalist motivation of EXP + E. Epistemic probability is typically understood to assume epistemic support (i.e. the epistemic probability of p given e depends on the degree to which e supports p), so it is not clear how it could be used as part of an analysis of such support. Second, if probabilistic relations could do the necessary work, it is not clear what the motivation for the explanatory component of EXP + E would be. There are further issues with probabilistic accounts of epistemic support, but for present purposes it is clear enough why a probabilistic modification of EXP + E is not an appealing option. Thanks to an anonymous reviewer for pointing out that one might be inclined to make this sort of modification to EXP + E and for noting some of the difficulties with such a modification.
Thanks to an anonymous reviewer for helping me see that this understanding of epistemic probability yields this result.
See Hawthorne and Stanley (2008) for such a view. Of course, there are other reasons that one might accept such a view. One might simply be skeptical of being able to know more than probabilistic claims about the future, for example. So, one might not accept that Byerly should believe that the ball will go into the cup, but rather, she might think that Byerly should believe that it will probably go into the cup.
‘This’ here refers to the situation that Byerly is currently observing in the case.
One might worry about whether this will be the case in other situations—particularly, situations where things that occasionally interfere with golf balls rolling into cups are imperceptible. In this sort of situation Byerly has observed that most golf balls rolling toward a cup in circumstances C go into the cup, but sometimes something imperceptible interferes and the golf ball does not go into the cup in C. It seems that what one should say about this situation depends on what is part of the best explanation of Byerly’s evidence. If it is part of the best explanation of his evidence that there are not any imperceptible interferers in this case, then <the ball will go into the cup> is entailed by the best explanation of his evidence and he is justified in believing that the ball will go into the cup. If it is not part of the best explanation of his evidence that there are not any imperceptible interferers in this case (presumably, because he cannot tell whether they are there or not), then the best explanation of his evidence will not entail that the ball will go into the cup. So, EXP + E would yield the result that Byerly is not justified in believing that the ball will go into the cup. This result is not at all unreasonable though. After all, in this situation Byerly knows that imperceptible interferers have been active in situations that are completely indistinguishable from his current situation and when they are active they prevent the ball from going into the cup. Thus, Byerly knows that if such an interferer is present it will stop the ball from going into the cup and he knows that he cannot tell whether such an interferer is present or not. In this sort of situation it is not implausible that Byerly is not justified in thinking the ball will go into the cup. Thanks to an anonymous reviewer for pointing out this worry.
This response is suggested by Lycan’s (1988) explanationist account of statistical inference.
I am grateful to anonymous reviewers for extensive comments on earlier drafts. I am also grateful to the following individuals for helpful comments and discussion: Marshall Abrams, John Bennett, Earl Conee, Trent Dougherty, Matt Frise, David Grober-Morrow, Ted Poston, and Bill Rowley.
References
Byerly, T. R. (forthcoming). Explanationism and justified beliefs about the future. Erkenntnis. doi:10.1007/s10670-012-9374-7.
Conee, E., & Feldman, R. (2004). Evidentialism. New York: Oxford University Press.
Conee, E., & Feldman, R. (2008). Evidence. In Q. Smith (Ed.), Epistemology: New essays (pp. 83–104). Oxford: Oxford University Press.
Conee, E., & Feldman, R. (2011). Replies. In T. Dougherty (Ed.), Evidentialism and its discontents (pp. 428–501). New York: Oxford University Press.
Feldman, R., & Conee, E. (1985). Evidentialism. Philosophical Studies, 48, 15–34.
Harman, G. (1973). Thought. Princeton: Princeton University Press.
Hawthorne, J., & Stanley, J. (2008). Knowledge and action. Journal of Philosophy, 105, 571–590.
Lycan, W. (1988). Judgement and justification. New York: Cambridge University Press.
Salmon, W. (1989). Four decades of scientific explanation. Minneapolis: University of Minnesota Press.
White, R. (2005). Explanation as a guide to induction. Philosopher’s Imprint, 5. www.philosophersimprint.org/005002/.
Author information
Authors and Affiliations
Corresponding author
Rights and permissions
About this article
Cite this article
McCain, K. Evidentialism, Explanationism, and Beliefs About the Future. Erkenn 79, 99–109 (2014). https://doi.org/10.1007/s10670-013-9470-3
Received:
Accepted:
Published:
Issue Date:
DOI: https://doi.org/10.1007/s10670-013-9470-3