Abstract
In recent years, there has been a “value turn” in epistemology. We intuitively think of knowledge as having a value, a value that mere true belief does not have, and it has been held to be a condition of adequacy on theories of knowledge that they be able to explain why. Unfortunately, for most theories their explanations suffer from the “swamping problem” because what has to be added to turn true belief into knowledge has value only instrumentally to truth; for example, we take being justified to be valuable, but only because being justified is our way of trying to believe what is true and only what is true, and it follows from this that for a belief already granted as true, no extra value is added by the fact that it is justified as well. So, the task is to solve the value problem while avoiding the swamping problem. I will argue that, in fact, the value turn leaves epistemological theorizing much as it was. My reasoning goes briefly as follows: on the usual interpretation of the value problem, the demand it places on theories of knowledge is internally incoherent and hence implausible as a condition of their adequacy. It is a condition of adequacy nonetheless that the theory avoid the swamping problem if we really do have the intuition supposed. I will then use a thought-experiment to consider different kinds of knowledge and argue that, with the exception of a priori knowledge, there is no such intuition. With this exception, there is no swamping problem either, and hence theories cannot be ruled inadmissible on the grounds that the knowledge-making features they propose fail to have a non-instrumental value. My conclusion is that it is only whatever has to be added to true a priori belief in order to turn it into knowledge that has to have a non-instrumental value, that is to say, only a priori knowledge that has a value problem. But this problem is easily solved.
This is a preview of subscription content, log in via an institution to check access.
Similar content being viewed by others
Notes
Is the value of truth swamped by that of infallibility? I do not think so, but even if it were, that would not be a problem, because it would still mean that a priori knowledge has a value that true a priori belief lacks.
Thus, for example, it would not be inconsistent to say that what makes my belief that p a case of knowledge is simply that it is true, and our second-order belief about whether our belief is justified are the conditions of satisfaction for using the verb “know” as this kind of performative verb, or conversational maxims pace Grice governing its use in discourse.
A variety of accounts of this type are discussed under the title “epistemic attitudinalism” in chapter 7 of Kvanvig (2003).
Pagination is taken from a draft version available at http://digitalcommons.lmu.edu/phil_fac/18.
Similarly, there may be cases where some other procedure is normative in the sense of being the most cognitively efficient though less likely to reach the truth; it is the procedure that we epistemically ought to follow. Does it follow that it is normative to believe the outcome of this procedure? I do not think so. One may take some cognitive attitude towards the outcome, e.g., acceptance, but not belief. Instrumental rationality (following the epistemically best procedure) and theoretical rationality (being in the epistemically best state) may not match up, except possibly for ideal reasoners.
Obviously, the solutions that I will now canvass are responses to the value problem of knowledge as it is traditionally understood, but my alternative understanding is near enough to the traditional one for it to be largely ignored in the ensuing discussion.
Pritchard’s response to the “dying man” example in (2016, 205–206) misses the point because he seems to say simply that the reliabilist is not committed to the view that on all occasions knowledge is more valuable than true belief, the case of the dying man being presumably (though he does not actually say so) one of those occasions. This is true enough, but the dying man is carefully set up to be a case where intuitively knowledge is more valuable than truth, and the reliabilist response is not rejected just because it allows for a case where knowledge is not more valuable than truth, but because reliabilism says that knowledge is not more valuable than truth in a case where intuitively we think it is. Anyway, I think the generality constraint rules out making this distinction: the distinction between cases of knowledge where the intuition is present and those where it is not is theoretically motivated and not theoretically neutral or a result of common-sense thinking about knowledge.
I owe thanks to an anonymous reviewer of this journal for alerting me to these papers.
Pagination is taken from a version available at http://eprints.gla.ac.uk/133154/.
Perhaps we can read an explanation like this back into the account of Carter and Jarvis.
It is open to Carter and Jarvis and Stapleford to reject the generality constraint and the reasoning that motivates it, so appealing to it here may be dialectically ineffective against them.
In this circumstance (S4*) seems to be vacuously true, as the “item” (i.e., the false belief) does not have the property (of being justified). According to Carter and Jarvis, (S4*) is false because justification confers additional value when the belief is false provided that you do not know that the belief is false.
In 2003 (xiv & 114), Jonathan Kvanvig concedes that infallibility is a value when we have it, but rejects this as a general solution only because he believes that all knowledge is always more valuable than true belief, and most knowledge is not infallible. But it should be noted that in saying that it is only a priori knowledge that has a value problem I am not appealing to any differences between types of knowledge not available at the conceptual level. This in itself does not imply that with regard to a priori knowledge our intuitions actually do accord with what I am suggesting here, however, but only shows that it does not violate any formal constraint to say that they do with regard to a priori knowledge and not with regard to knowledge of other kinds.
Or at least, are not synonymous with inferentially justified beliefs. Arguably, the way I am going to define inferred beliefs is co-extensive with inferentially justified beliefs (at least in our world), so all inferred beliefs are token-identical with inferentially justified beliefs. But they are type-distinct, so it is possible to believe that a belief is inferred but not inferentially justified and vice versa. See next footnote for much the same point.
I make the “believed by the subject to be” qualification because I am not sure that there can be cases of beliefs counter-factually dependent on other beliefs without being inferentially justified by them, as I noted in the previous footnote. But justification and counter-factual dependence are at least conceptually distinct and it is possible for a subject to have beliefs that are counter-factually dependent on others, and even that the subject believes to be so dependent, while not being committed to there being an implicational relationship between them. I will, however, omit the qualification in what follows.
Perhaps these distinctions can be made by reference to the distinction between propositionally justified and doxastically justified beliefs. What I have called inferentially justified beliefs are propositionally justified, while what I have called 'inferred' beliefs are doxastically justified, which can be read as being propositionally justified plus (at minimum) a basing condition which says that your belief in what is propositionally justified must be based on your beliefs in those propositionally justifying propositions. The contrast class consists of those propositions that comply with the basing condition, and that are true and based on other true beliefs, but without any commitment to whether the beliefs they are based on are propositionally justifying, i.e., the basing condition on its own or doxastic justification minus propositional justification. I will call these inferred beliefs, reserving the scare-quoted ‘inferred’ belief to mean what has actually been inferred, i.e., the doxastically justified. On this construal, there need not be any logical or inferential relationship between the true inferred belief and the (true) beliefs it is based on: the world might co-operate by making it a brute fact about the world that these truth-values always covary. Of course, the actual world is not like this, so it is reasonable to expect inferred beliefs and propositionally justified beliefs to be co-extensive in the actual world.
Following from the previous footnotes, I believe it can be asked on the grounds that these are conceptually distinct, even if we conceded that they are co-extensive.
I will continue with this talk of validity, but there is an interesting issue about whether we really need to know any more than that the rule of inference used is truth-preserving in the particular case in which we are using it rather than—as its validity implies—in all cases. Although he does not refer to the value problem of knowledge at all in his paper, Dogramaci (2017) has the intuition that we have a value intuition here similar to that in the case of knowledge, namely, that we value truth-preservation in all cases over and above truth-preservation in the particular case. Canvassing different accounts of validity, he decides that the substitutional account of validity provides the best explanation of this intuition because in this account alone are the other cases ones that we might be required to make in the future, whereas in other accounts the cases represent situations that are known not to obtain (and why should we care about them? Why should they have any value?). This resembles the reliabilist’s response to the value problem of knowledge (i.e., validity’s value obtains by virtue of further potential items of knowledge), and I am inclined to think it fails for much the same reasons. The same moral applies here as there: if there really is such an intuition, then it must be neutral between the different accounts and not the theoretical result of one of them. So I am not sure that the intuition that truth-preservation in all cases has value over and above the value of truth-preservation in the particular case can be correct. If this is so, continuing to talk of validity may not be quite right; knowledge of validity may be more than is strictly needed.
Swinburne (2001, 163) says: “We value the scientist for having a belief that is not merely true, but results from consciously responding to all the evidence of which he is currently aware . . . because he shows a grasp on those a priori truths . . . and is consciously guided by them in his belief formation.” Kvanvig (2003, 54–55) criticizes this because unless the value is independent of the belief’s truth, the value will be merely instrumental and hence “swamped” by the value of the belief, and Swinburne’s proposal does not give justification an independent value. However, Swinburne seems to inch towards a virtue theory here, because he says immediately afterwards: “If true belief is intrinsically important—as surely it is—conscious truth seeking is a virtue, and doing it in the right way is exemplifying that virtue paradigmatically” [my italics]. This is despite the fact that he later argues against virtue theories (2001, 188–90).
Following on from footnote 1, not being based on or depending counter-factually on any other beliefs may not actually amount to infallibility and one may be induced to retract what one knows a priori after all. If so, then I suggest that it is precisely this lack of dependency, rather than infallibility, that has non-instrumental value.
For myself, I am not so convinced. Worsnip (2018) gives several examples with the aim of arguing that there are cases of rational epistemic akrasia, that is to say, cases where you doubt your a priori justification for a proposition and suspend belief in that proposition but in which it is still rational to believe the proposition inferred from it. If anything, I think that the reverse is the case: I am not sure that your doubts about a priori justification can actually rationalize suspending judgment in the a priori belief, or even cause such a suspension as a matter of psychological fact, but I think it might rationalize the choice not to draw any inference from that belief, and hence not to have the inferred belief. In other words, you believe p and that p is evidence for q, but do not believe q. It, too, would be a kind of akrasia, as you would refuse to have a belief that it appears to you that you have good reasons to have, but I think it could be justified by the higher-level thought that you would be knowingly putting yourself into a worse epistemic position by drawing inferences from a priori justifications whose rational credentials you now doubt. It seems to me that this is the course of least irrationality in this case. (This might have interesting applications in the burgeoning literature on disagreement.) For discussion of similar cases where you are psychologically unable to withdraw a belief that you may not be fully convinced is rational (though it does not focus on a priori beliefs), see Botting (2016).
This implies that young children cannot have inferred knowledge, though they may have knowledge of other kinds.
Incidentally, along with Swinburne I think that having the a priori belief adds value even if the belief is not true and hence not knowledge. If this is so, it implies that having the a priori belief has a value independently of its truth. Infallible a priori belief has even more value, as Swinburne also says.
References
Baehr, J. (2009). Is there a value problem? In A. Haddock, A. Millar, & D. Pritchard (Eds.), Epistemic value (pp. 42–59). Oxford: Oxford University Press.
Botting, D. (2016). The narrowness of wide-scope principles. Organon F, 23(2), 177–203.
Carter, J. A., & Jarvis, B. (2012). Against swamping. Analysis, 72(4), 690–699.
Dogramaci, S. (2017). Why is a valid inference a good inference? Philosophy and Phenomenological Research, 94(1), 61–96.
Kvanvig, J. (2003). The value of knowledge and the pursuit of understanding. Cambridge: Cambridge University Press.
Meylan, A. (2007). The value problem of knowledge. Against a reliabilist solution. In Latin Meeting in Analytic Philosophy Geneva Proceedings (pp. 85–92).
Pritchard, D. (2016). Veritism and epistemic value. In H. Kornblith & B. McLaughlin (Eds.), Alvin Goldman and his Critics (pp. 200–215). Oxford: Blackwell.
Stapleford, S. (2016). Epistemic value monism and the swamping problem. Ratio, 29(3), 283–297.
Swinburne, R. (2001). Epistemic justification. Oxford: Clarendon Press.
Worsnip, A. (2018). The conflict of evidence and coherence. Philosophy and Phenomenological Research, 96(1), 3–44.
Acknowledgements
An ancestor of this paper was presented at the GWP 2019 Conference held in Cologne, Germany, on 25th to 27th February 2019. Thanks is owed to the organizers of the conference and all those attending the presentation. Thanks also to anonymous reviewers whose suggestions led to improvements in the paper, and to Scott Stapleford.
Author information
Authors and Affiliations
Corresponding author
Ethics declarations
Conflict of Interest
The author declares that he has no conflict of interest.
Additional information
Publisher’s Note
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
Rights and permissions
About this article
Cite this article
Botting, D. The Value Problem of A Priori Knowledge. Acta Anal 35, 229–252 (2020). https://doi.org/10.1007/s12136-019-00399-9
Received:
Accepted:
Published:
Issue Date:
DOI: https://doi.org/10.1007/s12136-019-00399-9