Skip to main content

Explaining Away Kripke’s Wittgenstein


The paradox of rule-following that Saul Kripke finds in Wittgenstein’s Philosophical Investigations purports to show that words and thoughts have no content—that there is no intentionality. This paper refutes the paradox with a dilemma. Intentional states are posited in rational explanations, which use propositional attitudes to explain actions and thoughts. Depending on which of the two plausible views of rational explanation is right, either: the paradox is mistaken about the a priori requirements for content; or, a fatal flaw in content ascription alleged by the paradox is no flaw at all, rather a necessary component of the proper method of propositional-attitude ascription. On either lemma, rational explanation defuses the paradox.

This is a preview of subscription content, access via your institution.


  1. The paradox in WRPL is a presentation of “Wittgenstein’s argument as it struck Kripke,” so it should not be understood to express Kripke’s non-exegetical views (ibid., 5). This paper does not assess WRPL as Wittgenstein exegesis.

  2. See note 10 regarding other options.

  3. “ + ” might be a meaningful non-referring expression. But, following Kripke, I accept ‘“ + ’ means plus” and “‘ + ” refers to the addition function” as equivalent, assuming that the latter is intelligible even if “ + ” is a meaningful nonreferring expression or there are no mathematical abstracta.

  4. Perhaps the primary interpretative controversy about the paradox centers around what sort of justification JC requires. According to the classical interpretation, a “justified” use may just be one that is right/correct in light of what the speaker means (cf. Boghossian 1989). The newer rival claims that a “justified” use is one that the speaker’s grasp, as a representation dictating the correct uses, must guide (cf. Bridges 2014; Kusch 2006). (For an attempt to plumb the textual evidence, see Green 2014, 68–90.]) Since this paper’s solution works under either interpretation of “justification,” my rendering of JC is neutral. The official presentation of the paradox remains neutral on this interpretative question, as well (which explains why it is less detailed than some other renderings).

    The newer rival may object that the solution on one lemma cannot satisfy JC, given how that interpretation understands the condition. I address this below (Sect. 4.6).

  5. I assume, as Kripke seems to, that the speaker being justified in her uses is equivalent to her uses being justified.

  6. Some authors have rejected this feature of meaning and content. (For an anti-skeptical solution that embraces “meaning finitism,” see Bloor 1997, 58–73.) They may be right, and Kripke probably only needs words to have a finite-yet-very large number of application conditions to give the skeptical doubts purchase. In any case, my solution need not question infinitary projection; if the paradox fails to annihilate meaning even when granted this stringent necessary condition, it will fail with less stringent cousins.

  7. One anti-skeptical approach makes no such appeal: the dispositionalist proposal (cf. Kripke 1982, 23–37). It may attempt to reduce all grasps to the speaker’s dispositions for use. Kripke’s main objection to this proposal is straightforward (in form)—that I am disposed to does not give me any reason or obligation to (ibid., 37). The proposal fails prima facie to specify a justificatory state.

  8. Some might think genuine rational explanation does not essentially involve propositional attitudes (e.g., Thompson 2008). For example, rational creatures may have essential teloi that explain what they do and think. Content-ascribing rational explanation may be translatable into these views’ preferred idiom.

  9. Some might argue that RE’s generalizations are not real laws, which is not relevant to the paradox.

  10. One might wonder whether Austere and Evaluative RE are the only two possible views. They are not. Certain other options have been defended in the past, e.g.., some twentieth-century views inspired by Wittgenstein’s later writings that took rational explanation to be acausal. I am assuming that RE, Austere or Evaluative, is causal. The two approaches are the ones with significant contemporary favor. Hence my dilemma for the paradox does not exhaust logical space—only the space of plausibility.

  11. Defenses of the merely causal approach to RE include Henderson (2010), Fodor (1974, 1997) (who never declares RE merely causal, but defends it as such), and Schroeder (2003). Dissidents include Bridges (2011), Haugeland (1998), Hieronymi (2011), McDowell (1996, 66–86; 1998, 25–340), and Millar (2004).

  12. Although some understand “evaluative” normativity as importantly distinct from the “prescriptive” sort, I choose “Evaluative” RE. This form of RE involves the interpreter’s evaluation of ascribees and their behavior; this is not to say that Evaluative RE does not deal in prescriptions.

  13. Usually, those who like Austere RE call rational explanation, “folk psychology.”

  14. In fact, a subject is sometimes more likely to act irrationally than rationally. Cognitive science reveals many cases where experimental subjects are more likely to reach an incorrect conclusion, even though they have everything needed to reach the right one. Ceteris paribus clauses in Austere RE might exclude conditions of unusually high rationality.

  15. There is controversy about why (and whether) simplicity is virtuous. Sober (2015) provides a nice treatment.

  16. Only one competing explanation, technically. Lower- and higher-level sciences could offer different true explanations of the same phenomenon.

    Also, adequacy is not really “all or nothing.” Even the best theories at a given time may not be perfectly adequate. They are often less adequate than clearly worse competitors that achieve perfect adequacy through ad hoc adjustments.

    I leave aside these (and other) qualifications to this rough-and-ready sketch. They do not affect Austere RE’s anti-skeptical consequences. (Thanks to Joyce C. Havstad for spurring these qualifications.)

  17. Gibbard (2012) defends an anti-skeptical solution that appeals to the relative simplicity of the plus- over the quus-hypothesis. Feldman (1986) suggests something similar. Neither answer relies on the purely empirical nature of content posits. Ludlow pursues simplicity as an answer to the paradox as adapted to Chomskyan grammar-processing rules, but does not provide a general anti-skeptical solution (2011, 105–117).

  18. If false, a subject could possess intrinsically contentful categorical states even if not at all disposed to think or act in ways characteristic of possessing them. This would be a return to the picture of content states as “mental mirrors” composed of intrinsically representational stuff. (See Rorty 1979 for the metaphor.) Although the picture has its defenders (e.g. Searle 1980), suffice it to say: its truth would be surprising.

  19. This is not intended to exclude a stronger grounding relation with the nonintentional. Admittedly, supervenience may be the strongest plausible relation if Evaluative RE is right (cf. Sect. 4.5).

  20. I am indebted to Soames (1998) and Van Cleve (1992) for the inspiration to rely on supervenience here.

    Van Cleve thinks the paradox inveighs against the possibility of intentional supervenience on the nonintentional. But his interpretation is tendentious (Green 2014, 76–89).

  21. Assume that we do not have to say exactly what it is about the speaker and her environment that add up to intentional states. Why not just then say straight off that the paradox never touches a merely supervenient conception of content, no digression into RE necessary?

    That might be ad hoc. Appeals to supervenience can “save” faux entities. Take concrete entity type E. E-entities seem to exist, though they do not. Assume that it is unclear how causally basic stuff “adds up” to E-entities. But we do know that if there are E-entities, no two worlds identical in facts about causally basic reality differ in facts about E-entities. So we can sit satisfied that E-entities supervene on the causally basic without explaining how. We would be misled—by stipulation, there are no E-entities.

    Rather, E-entities need to do something, to explain something, to merit confidence. Intentional content explains quite a bit, whatever its eccentricities. That is one reason why the paradox is “absolutely wild” (Kripke 1982, 9). Even if no one comes to believe in intentional contents through empirical confirmation, these contents evidently explain so much behavior. The paradox abnegates something with such evident explanatory power. So the thought that intentional contents supervene on the nonintentional is well-motivated.

  22. If applying “ought” to evaluative normativity seems odd, see note 12.

  23. Evaluative RE might look very similar to Davidsonian interpretation. (See Davidson 2001a, 123–179; Lepore and Ludwig 2005, 147–300). Evaluative RE uses charity basically as Davidson understands it. For a paragon evaluative theory of RE, see Schueler (2009).

    Davidson may not have held that RE is essentially evaluative. Lepore and Ludwig (2005) and Schroeder (2003) deny he did. Others disagree (e.g., Fennell 2015). To the extent that he did not, he would endorse Austere RE.

  24. The rationale for the principle is well-known, and the varied readings do not affect Evaluative RE’s approach to the paradox.

  25. Charity can play a role in Austere RE. It could be an inductively justified heuristic.

  26. This does not imply that the interpreter can only ascribe contents that she possessed before beginning the interpretation. Interpreters can obtain contents while interpreting. They may realize that they do not possess a content that “maps on” well to what an ascribee likely grasps, but eventually come to possess (thence ascribe) that content.

  27. I am discussing ascriptions’ justification here, not etiology. Clearly, we usually just understand what others are saying without any occurrent inferences when we share fluency.

  28. Empirical knowledge certainly informs our knowledge of what the application conditions of particular contents are, and thus when they are properly ascribed. (The extensions of natural-kind concepts are probably discernible only a posteriori.) The point, rather, is that Evaluative RE content ascription involves an irreducibly nonempirical element.

  29. One might have a “shaky grasp” of a content and still be able to ascribe it to others. Remember, however, that the ascriptions that overcome the skeptic are made from the God’s Eye View (cf. Sects. 1.1, 4.4). I am not relying on the everyday cogency of the interpreter’s contents to herself as a direct answer to the paradox. (I thank Raff Donelson for pressing me on this.).

  30. The “global doubt” can take one other form, depending on one’s interpretation of the paradox (cf. Sect. 4.6).

  31. For content, existence is possession (Platonism notwithstanding).

  32. This is not an empty appeal to supervenience to save something that does not exist. (See note 21.) Austere RE incurred the risk that the appeal to supervenience covered for some shortcoming in the causal explanation of behavior. This risk had to be dispelled. But the appeal to supervenience in Evaluative RE’s case is not “covering for” an explanatory shortcoming. It explains how intentionality, an evident feature of cognition, coexists with causally basic properties. So there is nothing ad hoc about supplementing Evaluative RE with an appeal to supervenience.

  33. Davidson argues, from the perspective of radical interpretation, that justified, meaningful uses of expressions rarely issue from consciously accessible rules that speakers consult (2001b, 113–115).

  34. For a diagnosis Kripke’s confidence (assuming he really thinks JC mandates such strong justification), see Bridges 2014.

  35. I argue elsewhere, without reference to Evaluative RE, that the paradox is uncompelling if “justification” in JC is equivalent to a requirement for the strong form of justification just addressed (Green 2018, 193–195). My thanks to an anonymous reviewer for persuading me that more was needed to address the skeptic’s possible demand for such guidance.


  • Bloor, D. (1997). Wittgenstein, rules and institutions. London: Routledge.

    Google Scholar 

  • Boghossian, P. (1989). The rule-following considerations. Mind, 98(392), 507–549.

    Article  Google Scholar 

  • Bridges, J. (2011). Dispositions and rational explanation. In J. Bridges, N. Kolodny, & W.-H. Hong (Eds.), The possibility of philosophical understanding (pp. 182–215). Oxford.

  • Bridges, J. (2014). Rule-following skepticism, properly so called. In J. Conant & A. Kern (Eds.), Varieties of skepticism (pp. 249–88). Berlin: deGruyter.

    Google Scholar 

  • Carnap, R. (1956). Meaning and necessity (Expanded). Chicago: University of Chicago.

    Google Scholar 

  • Davidson, D. (2001a). Inquiries into truth and interpretation (2nd ed.). Oxford: Clarendon.

    Book  Google Scholar 

  • Davidson, D. (2001b). Subjective, intersubjective, objective. Oxford: Clarendon.

    Book  Google Scholar 

  • Feldman, F. (1986). Critical notice of Kripke, Wittgenstein on rules and private language. Philosophy and Phenomenological Research, 46(4), 683–687.

    Article  Google Scholar 

  • Fennell, J. (2015). Davidson: normativist or anti-normativist? Acta Analytica, 30(1), 67–86.

    Article  Google Scholar 

  • Fine, A. (1984). The natural ontological attitude. In J. Leplin (Ed.), Scientific realism (pp. 83–107). Berkeley: University of California Press.

    Google Scholar 

  • Fodor, J. (1974). Special sciences (or: the disunity of science as a working hypothesis). Synthese, 28(2), 97–115.

    Article  Google Scholar 

  • Fodor, J. (1997). Special sciences: still autonomous after all these years. Philosophical Perspectives, 11, 149–163.

    Google Scholar 

  • Gibbard, A. (2012). Meaning and normativity. Oxford.

  • Green, D. (2014). Intentionality, determinacy, and rules: Kripke's paradox and the normativity of intentional content. Dissertation, Northwestern University (USA).

  • Green, D. (2018). Semantic knowledge, semantic guidance, and Kripke’s Wittgenstein. Pacific Philosophical Quarterly, 99(2), 186–206.

    Article  Google Scholar 

  • Hattiangadi, A. (2007). Oughts and thoughts. Oxford University Press.

  • Haugeland, J. (1998). Having thought. Cambridge, MA: Harvard University Press.

    Google Scholar 

  • Henderson, D. (2010). Explanation and rationality naturalized. Philosophy of the Social Sciences, 40(1), 30–58.

    Article  Google Scholar 

  • Hieronymi, P. (2011). Reasons for action. Proceedings of the Aristotelian Society, 111(3), 407–427.

    Article  Google Scholar 

  • Kripke, S. A. (1982). Wittgenstein on rules and private language. Cambridge, MA: Harvard.

    Google Scholar 

  • Kusch, M. (2006). A sceptical guide to meaning and rules. Montreal: McGill-Queen’s University Press.

    Google Scholar 

  • Lepore, E., & Ludwig, K. (2005). Donald Davidson. Oxford: Oxford University Press.

    Book  Google Scholar 

  • Ludlow, P. J. (2011). Philosophical issues in generative linguistics. Oxford: Oxford University Press.

    Book  Google Scholar 

  • McDowell, J. (1998). Mind, value, and reality. Cambridge, MA: Harvard.

    Google Scholar 

  • McDowell, J. (1996). Mind and world. Cambridge, MA: Harvard.

    Book  Google Scholar 

  • Millar, A. (2004). Understanding people. Oxford.

  • Rorty, R. (1979). Philosophy and the mirror of nature. Princeton.

  • Schroeder, T. (2003). Donald Davidson’s theory of mind is not normative. Philosopher’s Imprint, 3(1), 1–14.

    Google Scholar 

  • Schueler, G. F. (2009). Interpretative explanations. In C. Sandis (Ed.), New essays on the explanation of action (pp. 112–131). London: Palgrave Macmillan.

    Chapter  Google Scholar 

  • Searle, J. C. (1980). Minds, brains, and programs. Behavioral and Brain Sciences, 3(3), 417–457.

    Article  Google Scholar 

  • Soames, S. (1998). Skepticism about meaning: indeterminacy, normativity, and the rule-following paradox. Canadian Journal of Philosophy, 23(supp), 211–249.

    Google Scholar 

  • Sober, E. (2015). Ockham’s Razors. Cambridge, UK: Cambridge University Press.

  • Thompson, M. (2008). Life and action. Cambridge, MA: Harvard.

    Book  Google Scholar 

  • Van Cleve, J. (1992). Semantic supervenience and referential indeterminacy. Journal of Philosophy, 89(7), 344–361.

    Article  Google Scholar 

  • Van Fraassen, B. (1980). The scientific image. Oxford: Oxford University Press..

  • Wittgenstein, L. (2001 [1950]). Philosophical investigations. G. E. M. Anscombe, M. S. Hacker, & J. Schulte (Eds. & Trans.). 4th edition. Cambridge: Wiley-Blackwell.

Download references


Previous versions of this paper were presented to audiences at the Philosophy and Linguistics Colloquium Group at Northwestern University (Evanston, IL, USA), the 3rd Annual Philosopher’s Cocoon Philosophy Conference at the University of Tampa (FL, USA), and PhiLang: The Fifth International Conference on Philosophy of Language and Linguistics at the University of Łódź (Poland). All were very helpful. Special thanks to Sanford Goldberg, Mark Alznauer, Michael Glanzberg, Axel Müller, Pengbo Liu, Raff Donelson, Amy Flowerree, John Halpin, Joyce C. Havstad, Nick Leonard, Rebecca Mason, Jared Peterson, and Adam Streed for their valuable commentary.

Author information

Authors and Affiliations


Corresponding author

Correspondence to Derek Green.

Additional information

Publisher's Note

Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.

Rights and permissions

Reprints and Permissions

About this article

Verify currency and authenticity via CrossMark

Cite this article

Green, D. Explaining Away Kripke’s Wittgenstein. Erkenn (2021).

Download citation

  • Received:

  • Accepted:

  • Published:

  • DOI: