Abstract
During the past decade a new twist in the debate regarding the a priori has unfolded. A number of prominent epistemologists have challenged the coherence or importance of the a priori—a posteriori distinction or, alternatively, of the concept of a priori knowledge. My focus in this paper is on these new challenges to the a priori. My goals are to (1) provide a framework for organizing the challenges, (2) articulate and assess a range of the challenges, and (3) present two challenges of my own.
This is a preview of subscription content, log in via an institution to check access.
Similar content being viewed by others
Notes
See Casullo (2003, Chap. 5).
An anonymous referee maintains that if Quine and the target of his criticism took it as common ground that (LE) is constitutive of the a priori, then Quine’s contention that the analytic—synthetic distinction is incoherent, if cogent, does indeed establish that the a priori—a posteriori distinction is also incoherent. This suggestion, however, is incorrect. The question at hand is whether Quine’s contention, if cogent, establishes that the traditional Kantian concept of the a priori—i.e., the conjunction of (APK) and (APJ)—is incoherent. From the fact that Quine’s contention, if cogent, establishes that the concept of the a priori endorsed by his target is incoherent, it does not follow that the traditional concept is incoherent unless the two concepts of the a priori are the same. But clearly they are not since (LE) is not constitutive of the traditional concept of the a priori. If it were, Kant would not have been in a position to raise (nontrivially), let alone affirmatively answer, the question: Is there synthetic a priori knowledge? Moreover, neither (APK) nor (APJ) explicitly involves the concept of analyticity and, as I (2003, Chap. 8) have argued elsewhere, the arguments that purport to show that either (APK) or (APJ) implicitly involves the concept of analyticity fail.
It is important to distinguish the claim that (LE) is constitutive of the concept of a priori knowledge from another claim regarding (LE) suggested by (Friedman (2006), p. 37):
Indeed, this rejection of pure intuition and the synthetic a priori in favor of the view that all logico-mathematical truth is analytic and has no factual content quickly became definitive of what Carnap and the Vienna Circle meant by their empiricism.
Friedman suggests that (LE) is constitutive of the concept of logical empiricism—i.e., constitutive of the epistemological view endorsed by Carnap and the Vienna Circle. The fact that logical empiricists endorsed (LE) as constitutive of their epistemological view does not commit them to the view that (LE) is constitutive of the concept of a priori (or empirical) knowledge.
Finally, it cannot be taken for granted that Quine and his target took it as common ground that (LE) is constitutive of the a priori if, as the referee suggests, his target is Carnap. For example, (Friedman 2006, pp. 51–52) argues:
Quine, as we have seen, never fully appreciated the deeply original character of the Carnapian logic of science. He assimilated it, instead, to a program in traditional epistemology, one which begins with the Kantian question how synthetic a priori knowledge is possible, replaces it with the question “How is logical certainty possible?’ and concludes with “the linguistic doctrine of logical truth” as the supposed answer to this epistemological question...From Carnap’s point of view, however, Quine’s mature epistemology represents just as much of an externally motivated, purely philosophical intrusion into the ongoing progress of empirical science and the logic of science as Quine’s earlier defense of nominalism.
Ebbs (2011, p. 213) offers a similar reading of Carnap’s project:
Carnap’s account of ‘analytic’ or ‘L-true’ is language-system-relative. Carnap did not aim to analyse or explain the traditional notion of an a priori truth in terms of his notion of ‘L-true.’ Instead he eschewed the traditional phrase ‘a priori truth’, which he regarded as vague and confused.
I am not endorsing either Friedman’s or Ebbs’s position. I am only pointing out that it cannot be assumed uncritically that Quine and Carnap took it as common ground that (LE) is constitutive of the a priori.
BonJour (1998) is an exception.
An anonymous referee suggests that if there is general agreement about the classification of some but not all sources, then the distinction is at least minimally coherent. The referee, however, also acknowledges that such a distinction will have little explanatory value.
One might attempt to circumvent the heterogeneity of experiential sources by contending that there is a single source of nonexperiential justification: a cognitive state such as rational insight with a characteristic phenomenology. This contention, even if correct, is not enough to mark the difference between experiential and nonexperiential sources of justification. Given that there are a number of different sources of experiential justification, each associated with a cognitive state that has a characteristic phenomenology, there is no basis for dividing sources of justification into two categories based on phenomenological differences unless there is some phenomenological feature common to all and only the cognitive states associated with the experiential sources.
The proposal does not accommodate theories of the a priori that deny that nonexperiential sources involve a relation to some abstract object. To accommodate such theories, the proposal might be revised to maintain that nonexperiential sources are those that do not involve a relation to some concrete object. The modified proposal is open to the same problems as the original.
This explanation of cognitive access also has the consequence that sense experience provides cognitive access to the properties themselves via causal interaction with property-instantiations, thus undermining the need for nonexperiential access to the properties.
One might propose that the problem can be circumvented by rejecting (APJ) in favor of the following positive characterization of a priori justification:
-
(APJ*) S’s belief that p is justified a priori iff S’s belief that p is justified by \(\Phi \),
where ‘\(\Phi \)’ designates some specific source of justification such as logical intuition or rational insight. Since (APJ*) does not contain the concept of experience, the proposal continues, its coherence does not depend on the coherence of the concept of experience.
Although the problem can be circumvented by embracing (APJ*), there is a price to pay. Embracing (APJ*) has the consequence of divorcing the question of whether there is a priori knowledge from the question of whether radical empiricism is true. Given (APJ*), the claim that some knowledge is a priori does not entail that some knowledge is nonexperientially justified. Consequently, one cannot draw the conclusion that radical empiricism is false from the premise that some knowledge is a priori. In order to draw that conclusion, the further premise that \(\Phi \) is a nonexperiential source of justification is necessary, which reintroduces the original challenge.
-
There is an important difference between the two types of conceptual argument. Those in the first category presuppose that there is a coherent concept of a priori knowledge. Those in the second do not.
Some theorists, such as Field (1989) and Maddy (1990), have attempted to revive a version of Benacerraf’s argument that does not rest on the causal theory of knowledge. They maintain that any acceptable account of mathematical knowledge must explain the reliability of the mathematical beliefs of mathematicians but the fact that the truth conditions of mathematical statements refer to abstract entities precludes such an explanation. See Casullo (2003) for further discussion.
See Casullo (2003) for a defense of this claim. An anonymous referee worries that there is a tension between my rejection of the causal condition on knowledge and my claim in Sect. 2 that explaining how cognitive states such as rational insight can provide access to abstract entities is challenging since such an explanation would presumably require causal interaction between those cognitive states and abstract entities. The tension, however, is merely apparent. One must distinguish between two different claims: (1) The concept of knowledge entails that S knows that p only if S stands in a causal relation to the entities referred to by the truth conditions of p, and (2) A satisfactory explanation of how cognitive state S provides access to entities of type E requires that S stands in some causal relation to entities of type E. The claims are logically independent: the denial of (1) does not entail the denial of (2).
An anonymous referee finds Hawthorne’s reference to “epistemological joints” hard to understand and, moreover, suggests that a distinction can be important but fail to carve at the epistemological joints. I agree with these remarks. My goal in Sect. 4 is to offer an articulation of “epistemological joints” in terms of the features of one’s general theory of knowledge and to maintain that the best way to understand the criticism that the a priori—a posteriori distinction is unimportant is that the distinction does not mesh well with the features of the critic’s general theory of knowledge.
An anonymous referee offers a different challenge that falls into the third category: There is a distinction between a priori and a posteriori knowledge that cuts at the epistemic joints but all significant knowledge is a mixture of both a priori and a posteriori knowledge. I don’t regard this challenge as significant. To make matters more concrete, let us suppose that all significant mathematical, scientific and philosophical knowledge is a mixture of both a priori and a posteriori knowledge. The question at issue is whether the concept of a priori knowledge remains significant for the theory of knowledge. Since one of the main goals of the theory of knowledge is to identify basic sources of knowledge and explain how beliefs derived from those sources are known, the concept remains significant for the theory of knowledge. Given that the challenge concedes that the concept cuts at the epistemological joints, any theory of knowledge that does not employ it will be incomplete. Its taxonomy of basic sources will be incomplete.
Goldman (1999, p. 23), offers a version of this challenge:
A significant number of people’s beliefs have a warranting history that includes both perceptual and ratiocinative processes. By calling such beliefs “empirical,” the classificational system automatically gives pride of place to the first of these components. To my mind, this is misleading.... We need an epistemology that puts the two sources of warrant on a more balanced footing.... What should we call an epistemology that gives roughly equal credit to perceptual and ratiocinative sources of warrant: empirico-rationalism? Unfortunately, the label “empirico-rationalism” tends to suggest that all warranted beliefs are warranted by perception, ratiocination, or a combination of the two. That, as we have seen, is false.... So it is best to reject not only the traditional options of empiricism and rationalism but even the appealing but simplistic synthesis of empirico-rationalism. Warrant is just a complex and multi-dimensional affair. Why try to force it into some neat little container or pair of containers that simply disguise its true contours?
The problem that Goldman raises is terminological. Cases of knowledge that derive from both a priori and a posteriori sources are classified as a posteriori as a matter of definitional stipulation. This problem, however, is easily resolved by introducing a third category of knowledge: knowledge that derives from both a priori and a posteriori sources, where neither source alone suffices for knowledge. (The final qualification is necessary to distinguish this category from cases of epistemic overdetermination—i.e., cases of knowledge that have two independent sources, one a priori and one a posteriori, where each source alone suffices for knowledge). As Goldman suggests, coming up with a suitable label for this category—i.e., one that is not misleading—may be challenging. Nevertheless, the categories themselves are quite clear. Setting aside cases of epistemic overdetermination, all knowledge falls into three categories: pure a priori, pure a posteriori and impure (or involving a mixture of both). If Goldman is right, then all three categories are nonvacuous. If the referee is right, then all significant knowledge is impure.
I have argued elsewhere (Casullo 2012) that the complaints of Hawthorne, Jenkins and Williamson can be addressed in this fashion.
According to her account, (1) S’s (basic) arithmetical knowledge that p depends epistemically on experience in just one respect: the concepts constitutive of S’s belief that p must be grounded by the senses, but (2) the sensory input that grounds those concepts does not constitute evidence for S’s belief that p. The conjunction of (A) and (2) entails that basic arithmetical knowledge is not a posteriori. The conjunction of (B) and (1) entails that basic arithmetical knowledge is not a priori. Hence (C) is false.
Jenkins’s analysis of the concept of a priori knowledge is presented in (A) in terms of a posteriori knowledge: S knows a priori that p iff S knows that p and S’s knowledge that p is independent of empirical evidence.
See Casullo (2012) for further discussion.
Wright’s concept of entitlement is different from Burge’s.
Two anonymous referees question the significance of the choice. One referee wonders why it matters if we have to give up (JB). It matters for both theoretical and practical reasons. On the theoretical side, a comprehensive theory of knowledge must acknowledge that there is knowledge that is neither a priori nor a posteriori, articulate its nature and scope, and coherently accommodate it within the framework of its more general commitments. On the practical side, philosophical arguments that employ (JB) or analogues of it (either explicitly or tacitly) must be reevaluated. For example, as Ayer (1952) recounts, the primary motivation behind logical empiricism is the rejection of synthetic a priori knowledge. The logical empiricists maintained that there are only two options for doing so: embrace radical empiricism or defend (LE). Since they found radical empiricism’s account of mathematical and logical knowledge to be deficient, they opted to defend (LE). Their argument tacitly presupposes (JB); it overlooks the possibility of rejecting synthetic a priori knowledge by maintaining that the truths of mathematics and logic are synthetic but knowledge of them is neither a priori nor a posteriori. The other referee wonders why we need to respect the traditional concept of the a priori. The primary reason is that there is a cost to abandoning it. Abandoning it in favor of the negative concept of the a priori results in a less natural classification of warranted acceptances (or so I argue). The fact that abandoning the traditional concept comes at a cost does not establish that it cannot be abandoned. It establishes only that, prior to abandoning it, one must weigh the cost and benefit of doing so.
An anonymous referee suggests that if all a priori justified beliefs are either default reasonable or derived from default reasonable beliefs via default reasonable rules, then the purely negative characterization of the a priori would suffice for purposes of epistemology. I disagree. If there are no nonexperientially justified beliefs, then the purely negative characterization of the a priori is adequate in the sense that the beliefs that satisfy it do not, as a matter of fact, include both beliefs justified nonexperientially and beliefs whose justification does not derive from any source. The resulting class of a priori justified beliefs is not unnatural. But the purely negative characterization remains theoretically inadequate in the sense that it masks the important distinction between beliefs justified nonexperientially and beliefs whose justification does not derive from any source. Consequently, even if all a priori justified beliefs are either default reasonable or derived from default reasonable beliefs via default reasonable rules, the negative characterization fails to highlight the unique feature of such justified beliefs.
Earlier versions of this paper were presented at the Basic Knowledge: A Priori Conference, Northern Institute of Philosophy, University of Aberdeen, June 16–17, 2012; The Roles of Experience in A Priori Knowledge Conference, University of Cologne, October 13–14, 2012; and the Kline Workshop on A Priori Knowledge, University of Missouri, March 8–9, 2013. I would like to thank the organizers for inviting me and the audiences for their challenging questions, comments and criticisms. Special thanks are due to Mikkel Gerken, my commentator at the Basic Knowledge Conference, and to Andrew Melnyk, my commentator at the Kline Workshop. I would also like to thank three anonymous referees for this journal for their criticisms and suggestions, which have led to significant improvements in the paper.
References
Ayer, A. J. (1952). Language, truth and logic. New York: Dover Publications.
Benacerraf, P. (1973). Mathematical truth. Journal of Philosophy, 70, 661–679.
Boghossian, P. (1996). Analyticity reconsidered. Nous, 30, 360–391.
BonJour, L. (1998). In defense of pure reason. Cambridge: Cambridge University Press.
Brueckner, A. (2011). Albert Casullo’s a priori justification. In M. J. Shaffer & M. L. Veber (Eds.), What place for the a priori?. Chicago: Open Court.
Burge, T. (1993). Content preservation. Philosophical Review, 102, 457–488.
Casullo, A. (1988). Revisability, reliabilism, and a priori knowledge. Philosophy and Phenomenological Research, 49, 187–213.
Casullo, A. (2003). A priori justification. New York: Oxford University Press.
Casullo, A. (2009). Analyzing a priori knowledge. Philosophical Studies, 142, 77–90.
Casullo, A. (2011). Reply to my critics: Anthony Brueckner and Robin Jeshion. In M. J. Shaffer & M. L. Veber (Eds.), What place for the a priori?. Chicago: Open Court.
Casullo, A. (2012). Articulating the a priori–a posteriori distinction. In essays on a priori knowledge and justification. New York: Oxford University Press.
Ebbs, G. (2011). Carnap and Quine on truth by convention. Mind, 120, 193–237.
Field, H. (1989). Realism, mathematics and modality. Oxford: Blackwell Publishers.
Field, H. (2000). Apriority as an evaluative notion. In P. Boghossian & C. Peacocke (Eds.), New essays on the a priori. Oxford: Oxford University Press.
Friedman, M. (2006). Carnap and Quine: Twentieth-century echoes of Kant and Hume. Philosophical Topics, 34, 35–58.
Fumerton, R. (2006). Epistemology. Malden: Blackwell.
Goldman, A. (1999). A priori warrant and naturalistic epistemology. Philosophical Perspectives, 13, 1–28.
Goldman, A. (2013). Philosophical naturalism and intuitional methodology. In A. Casullo & J. Thurow (Eds.), The a priori in philosophy. Oxford: Oxford University Press.
Hawthorne, J. (2007). A priority and externalism. In Sanford Goldberg (Ed.), Internalism and externalism in semantics and epistemology. Oxford: Oxford University Press.
Hempel, C. (1972). On the nature of mathematical truth. In R. C. Sleigh (Ed.), Necessary truth. Englewood Cliffs: Prentice-Hall.
Jenkins, C. S. (2008). Grounding concepts. Oxford: Oxford University Press.
Jeshion, R. (2011). Experience as a natural kind: Reflections on Albert Casullo’s a priori justification. In M. J. Shaffer & M. L. Veber (Eds.), What place for the a priori?. Chicago: Open Court.
Kitcher, P. (1983). The nature of mathematical knowledge. New York: Oxford University Press.
Kitcher, P. (2000). A priori knowledge revisited. In P. Boghossian & C. Peacocke (Eds.), New essays on the a priori. Oxford: Oxford University Press.
Kripke, S. (1971). Identity and necessity. In M. K. Munitz (Ed.), Identity and individuation. New York: New York University Press.
Kripke, S. (1972). Naming and necessity. In D. Davidson & G. Harman (Eds.), Semantics of natural language. Dordrecht: D. Reidel Publishing Company.
Maddy, P. (1990). Realism in mathematics. Oxford: Oxford University Press.
Malmgren, A.-S. (2006). Is there a priori knowledge by testimony? Philosophical Review, 115, 199–241.
Putnam, H. (1975). The analytic and the synthetic. In Language mind and reality: Philosophical papers, vol. 2. Cambridge: Cambridge University Press.
Putnam, H. (1983). Two Dogmas’ revisited. In realism and reason: Philosophical papers, Vol. 3. Cambridge: Cambridge University Press.
Quine, W.V. (1963). Two Dogmas of Empiricism. In From a Logical Point of View, (2nd ed. revised). New York: Harper and Row.
Sosa, E. (2007). A virtue epistemology, Vol. 1. Oxford: Oxford University Press.
Sosa, E. (2013). Intuitions and foundations: The relevance of Moore and Wittgenstein. In A. Casullo & J. Thurow (Eds.), The a priori in philosophy. Oxford: Oxford University Press.
Williamson, T. (2007). The philosophy of philosophy. Oxford: Blackwell.
Williamson, T. (2013). How deep is the distinction between a priori and a posteriori knowledge? In A. Casullo & J. Thurow (Eds.), The a priori in philosophy. Oxford: Oxford University Press.
Wright, C. (2004a). Intuition, entitlement and the epistemology of logical laws. Dialectica, 58, 155–175.
Wright, C. (2004b). Warrant for nothing (and foundations for free)? Proceedings of the Aristotelian Society, 78, 167–212.
Author information
Authors and Affiliations
Corresponding author
Rights and permissions
About this article
Cite this article
Casullo, A. Four challenges to the a priori—a posteriori distinction. Synthese 192, 2701–2724 (2015). https://doi.org/10.1007/s11229-013-0341-x
Received:
Accepted:
Published:
Issue Date:
DOI: https://doi.org/10.1007/s11229-013-0341-x