Quinean holism, analyticity, and diachronic rational norms


I argue that Quinean naturalists’ holism-based arguments against analyticity and apriority are more difficult to resist than is generally supposed, for two reasons. First, although opponents of naturalism sometimes dismiss these arguments on the grounds that the holistic premises on which they depend are unacceptably radical, it turns out that the sort of holism required by these arguments is actually quite minimal. And second, although it’s true, as Grice and Strawson pointed out long ago, that these arguments can succeed only if there isn’t any principled criterion for meaning change, such a criterion turns out to be hard to come by. David Chalmers has recently argued that such a criterion must exist, since the norms governing belief revision are subject to obvious exceptions that can be explained only by appeal to meaning change. But this, I argue, is incorrect: if choices about how to use language are themselves rationally assessable (as naturalists can and should take them to be), then there are no such exceptions to be explained. To show that this is so, I formulate a new kind of coherence norm that may be useful for reasoning formally about the relationship between meaning and evidence.

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


  1. 1.

    In the original version of “Two Dogmas”, Quine doesn’t cite Duhem in connection with this thesis, but in the reprint in From a Logical Point of View (1953), he adds a footnote explaining that a defense of the thesis appears in Duhem 1906/1954. Quine makes clear in subsequent work (see, e.g., his 1975, p. 313) that he takes his holism to be at least roughly equivalent to Duhem’s thesis.

  2. 2.

    Though naturalists deny the possibility of a priori justification grounded in pure rational insight, they aren’t hostile to every sort of defeasible apriority. Quine’s basic epistemological stance, after all, is that each of us is a “sailor adrift on Neurath’s boat” (1981, p. 72): all we can do is start wherever we are and make repairs as we go along. So the idea that we might be unjustified in starting with the beliefs we do can, for Quine, only be a confusion (cf. Harman’s conservatism, according to which “you start where you are” and “rationality or reasonableness then consists in trying to make improvements in your view” (1995, p. 189). That said, naturalists are hostile to indefeasible apriority (see, e.g., Bergström 2014), and this is the notion I’ll be discussing here. (In fact, the dispute between naturalists and their opponents is subtler than this. As Peacocke (2005, pp. 747–748) notes, those on both sides can agree that any belief can be defeated via evidence that the thinker has made some sort of reasoning mistake in identifying grounds for that belief—that is, every sentence exhibits “defeasibility of identification”. For example, I may be rationally required to give up a mathematical belief if an eminent mathematician tells me, incorrectly, that what I’ve identified as a proof of some claim isn’t a genuine proof. But naturalists, unlike many of their opponents, also take every belief to exhibit a kind of defeasibility besides defeasibility of identification).

  3. 3.

    I’m inclined to doubt that this argument is even implicit in Quine’s paper. After all, he’s an empiricist talking to other empiricists—all parties to the discussion share the presupposition that the only way for a sentence to be a priori is for it to be analytic. So if he really has shown that there are no analytic sentences, there’s no need for a separate argument against apriority. Still, regardless of whether Quine actually intends to advance this argument, his epistemological stance provides us with the resources to reconstruct it, as I explain below.

  4. 4.

    Carnap, as far back as his Logical Syntax of Language—which Quine (1970, p. xxiii) “read...page by page as it issued from Ina Carnap’s typewriter”—endorses precisely the view put forward by Grice and Strawson, embracing the Duhem thesis and noting that experience may induce us to “alter the language to such an extent that [a previously analytic sentence] is no longer analytic” (1934/1937, p. 319).

  5. 5.

    As Ebbs (2016) points out, this aspect of the argumentative structure of “Two Dogmas” sometimes goes unacknowledged. Grice and Strawson themselves, for example, treat the argument from holism as evaluable independently of the circularity argument in the first part of Quine’s paper, as do Russell (2008) and Juhl and Loomis (2010). (Chalmers, too, frames his discussion as though he takes these two arguments to be independent, though the substance of his response to the argument from holism suggests otherwise: he tries to “flesh out a principled distinction” between cases where meaning changes and cases where it doesn’t and so to “make inroads into the Quinean circle” (2012, pp. 204, 225).) But again, given Quine’s knowledge of Carnap’s own holism as presented in Logical Syntax, this way of understanding the arguments of “Two Dogmas” can’t be right: Quine can’t have thought that the Duhem thesis alone entails that there’s no analytic–synthetic distinction.

  6. 6.

    Chalmers’s argument appears both in Chap. 5 of Constructing the World and in his 2011a. I’ll be referring to the former work.

  7. 7.

    Ebbs (2016) also notes that this is the standard interpretation, though he doesn’t endorse it. But his own interpretation is similar in certain respects to what I below call the empiricist reading, and like that reading, it can be ruled out on the grounds that it requires us to pretend that Quine doesn’t endorse Holding-true.

  8. 8.

    Those who endorse the empiricist reading tend to ignore Holding-true altogether and so to interpret Quine here as giving what I’ve called the argument from universal susceptibility: if every sentence is such that we’re rationally required to reject it in the face of some body of evidence, then no sentence is (indefeasibly) a priori. But Quine includes Holding-true as a premise in his argument—any interpretation on which it’s not part of his view is to be rejected on textual grounds. (It’s worth noting that Chalmers himself endorses the pragmatist reading as an interpretation of Quine’s own text. But the empiricist reading, he says, “has been more influential among later Quineans” (2012, p. 215fn).)

  9. 9.

    Chalmers, for his part, tries to make sense of what Quine is doing by attributing to him a general skepticism about rationality—he says that Quine in “Epistemology Naturalized” (1969) argues for “a sort of skepticism about norms of rationality” and that there’s a “deep linkage” between that skepticism and Quine’s holism (2012, pp. 221–222). If this reading of “Epistemology Naturalized” were correct, it would go some way toward making the pragmatist reading of the argument from holism plausible: it would make sense for Quine to have a radically permissive conception of epistemology if he thought there weren’t any rational norms at all. But Quine has repudiated this reading of “Epistemology Naturalized” on more than one occasion. Here’s one example:

    [My traditionalist critics] are wrong in protesting that the normative element, so characteristic of epistemology, goes by the board. Insofar as theoretical epistemology gets naturalized into a chapter of theoretical science, so normative epistemology gets naturalized into a chapter of engineering: the technology of anticipating sensory stimulation. (1992, p. 20)

    Quine’s epistemology is indeed radical in certain respects, but he’s not a skeptic about rational norms.

  10. 10.

    He continues:

    But in point of epistemological footing the physical objects and the gods differ only in degree and not in kind. Both sorts of entities enter our conception only as cultural posits. The myth of physical objects is epistemologically superior to most in that it has proved more efficacious than other myths as a device for working a manageable structure into the flux of experience. (1951, p. 41)

    This passage invites an interpretation on which Quine takes physical objects to be mere fictions, to be accepted on purely pragmatic grounds. But that interpretation is to be resisted. He clarifies his understanding of the epistemological significance of posits in “Posits and Reality” (1960/1966, p. 238):

    Having noted that man has no evidence for the existence of bodies beyond the fact that their assumption helps him organize experience, we should have done well, instead of disclaiming evidence for the existence of bodies, to conclude: such, then, at bottom, is what evidence is.

    Though physical objects, on Quine’s view, are to be accepted on pragmatic grounds, they aren’t thereby fictions—the relevant pragmatic grounds, for Quine, are epistemic grounds. For further discussion of Quine’s epistemological pragmatism, see Sect. 4.

  11. 11.

    This is confirmed in Quine and Ullian’s Web of Belief, where they offer the following clarification of Holding-true: “Just about any hypothesis...can be held unrefuted no matter what, by making enough adjustments in other beliefs—though sometimes doing so requires madness” (1978, p. 79, my emphasis). It’s clear here that Holding-true, whatever sort of claim it turns out to be, must be consistent with the claim that, in the face of some bodies of evidence, the only rational thing to do is to give up a particular hypothesis.

  12. 12.

    This is a bit of a simplification. Quine’s holism, as he well knows, is broader in scope than Duhem’s, and there are some differences of emphasis as well. See Vuillemin (1986) and Quine’s (1986) reply.

  13. 13.

    Duhem’s own view is that experimental results can confirm and disconfirm single hypotheses. He says that, when we’re deciding how to revise a system of hypotheses in the face of experimental contradiction, “Pure logic is not the only rule for our judgments; certain opinions which do not fall under the hammer of the principle of contradiction are in any case perfectly unreasonable” (1906/1954, Sect. II.VI.10). That is, although a body of evidence can never be inconsistent with a given hypothesis, it is possible for a body of evidence to render the hypothesis unreasonable.

  14. 14.

    Lakatos (1978, p. 97) distinguishes between a weaker version of the Duhem thesis, which “only denies the possibility of a disproof of any separate component of a theoretical system”, and a stronger one, which “excludes any rational selection rule among the alternatives”, and he attributes the weaker thesis to Duhem and the stronger one to Quine. What I’m arguing here is that this is a misinterpretation of Quine: as far as his argument from holism is concerned, he, like Duhem, is committed only to the weaker thesis.

  15. 15.

    For a fuller presentation of Quine’s mature holistic doctrine, see Chap. 1 of his Pursuit of Truth (1992).

  16. 16.

    Duhem’s logical point—that a single hypothesis has empirical consequences only when conjoined with a system of auxiliary hypotheses—is widely taken to be obviously correct by contemporary philosophers of science.

  17. 17.

    For this reason, Grünbaum (1962, p. 20) considers roughly the interpretation I’m defending and dismisses it on the grounds that it would turn the Duhem thesis into a “thoroughly unenlightening truism”, and certain of my teachers and colleagues have expressed similar reservations in conversation with me. But Quine’s response to Grünbaum is telling:

    I would say that the thesis as I have used it is probably trivial. I haven’t advanced it as an interesting thesis as such. ...I am not concerned even to avoid the trivial extreme of sustaining a law by changing a meaning; for the cleavage between meaning and fact is part of what, in such contexts, I am questioning. (1976, p. 132)

    It appears, then, that the Duhem thesis, as employed by Quine, should be understood as a truism. (See Becker 2001 for further discussion of this point.)

    One further point: textual analysis aside, Holding-true and Revisability are premises of the argument from holism, which means their truistic character is not a cost but a benefit. So I’m not sure what to make of the impulse to reject the proposed interpretation on the grounds that it makes these claims trivial—to do so, after all, is to reject it on the grounds that it makes Quine’s argument too strong.

  18. 18.

    Note that Revisability, on the empiricist reading, is equivalent to this principle. The empiricist reading, though inaccurate as an interpretation of Quine’s argument from holism in “Two Dogmas”, is perfectly acceptable as a distillation of the naturalist case against the a priori.

  19. 19.

    There need not be a unique best system of beliefs here. Since the different desiderata will often pull in different directions, a balance must be struck, and how to strike this balance is (to some degree) up to the individual thinker.

  20. 20.

    Or at least for declining to accept its negation. And for the purposes of the argument from universal susceptibility, this is all that’s really necessary. Quineans’ primary objection to apriority, after all, is methodological: since we can’t guarantee that our justification for believing any particular sentence will remain undisturbed in the face of new evidence, there aren’t any sentences such that we can be sure now that we won’t be required to reject them later. So, on the Quinean view, we have no reason for taking any of our beliefs to be wholly secure; we should be open to considering rejection of any sentence whatsoever. And we can be open in this way as long as we don’t deny Defeasibility. Hill (2013) and Ebbs (2016) each make essentially this point.

  21. 21.

    He could also have used Jeffrey’s (1965) generalization of the conditionalization principle, which allows for updating even when one is less than certain that one has acquired some body of evidence. But the standard Bayesian norm is easier to work with.

  22. 22.

    Some naturalist responses to Chalmers proceed by denying that there can be any generally applicable norm of this kind, on the grounds that rational thinkers can’t in general be expected to know, in advance of actually undergoing a particular course of experience, what the right response to that experience will be. Schroeter’s (2014, building on work in, e.g., her 2006) response is of roughly this sort, as are Neta’s (2014) and Rupert’s (2016). I don’t have the space here to do justice to these responses, but I do want to say something about why I think Chalmers can resist them.

    As Chalmers (2014) points out in his reply to Schroeter and Neta, each of them discusses several purportedly problematic cases, but never in these discussions do they give any reason to deny the following (overwhelmingly plausible) claim: that a thinker who’s merely supposing that she has some evidence can in principle engage in the same sort of reasoning that a thinker who actually has the evidence can, and with the same justification. And if that’s right, then it’s unclear why these cases are supposed to be problematic in the first place. (Schroeter thinks this sort of hypothetical reasoning doesn’t in general issue in judgments of the right kind—she claims that supposition is a kind of fictional role-playing and that what it justifies, in the first instance, are just metalinguistic judgments about the language of a hypothetical thinker. But it seems clear that this is false, at least if what’s in question is the kind of suppositional reasoning that’s associated with conditional credences.)

    Naturalists can respond that what’s really problematic here is the claim that thinkers are always rationally required to have the relevant suppositional beliefs. This is Rupert’s strategy: he points out that, as a matter of fact, we humans aren’t in general able to predict the effects of experience—we just don’t have the imaginative capacity. So we can be rationally required to make such predictions only if rationality is highly idealized. (Chalmers’s use of a Bayesian framework is another clue that the notion of rationality he’s working with is an idealized one.) And this idealized notion of rationality, Rupert says, isn’t one that naturalists will be inclined to endorse. But even if this is right, the idealized notion of rationality is at least coherent, and Chalmers’s diachronic rationality arguments require only that a coherent notion of this kind is available. (Here naturalists may respond that a notion of apriority associated with this sort of idealized rationality is of little theoretical interest. I’m inclined to disagree, but idealization in epistemology is a huge topic a full discussion of which would take us far outside the scope of this paper.)

  23. 23.

    On certain views, there’s another possible explanation: a thinker can rationally violate conditionalization by resetting her priors. If more than one set of priors is rationally permissible, and if there’s no ban on switching from one set to another, then this is a possibility that needs to be taken into account. And it’s relatively clear, given naturalists’ epistemological commitments, that they should think more than one set of priors is rationally permissible. (Ebbs’s response to Chalmers, for instance, appears to rely on the idea of resetting priors: on his view, rational violations of sentential conditionalization are possible simply because “changing our confirmational commitments whenever we judge it useful to do so” is not irrational (2014, 702).)

    Chalmers does discuss the possibility of resetting priors, claiming that “as long as we have a conceptual distinction between cases in which beliefs are revised by this process and cases in which they are not”, there’s no problem for his argument—we can just stipulate that violations of conditionalization we’re interested in are those that don’t involve resetting priors (2012, p. 223). But I think a stronger response is available: there are powerful Quinean reasons to avoid resetting priors. After all, even if more than one set of priors is rationally permissible, switching from one set to another amounts to arbitrarily engaging in wholesale revision of one’s system of beliefs, and this sort of arbitrary revision is exactly the sort of thing that the desideratum of familiarity is intended to rule out. So Quineans should deny that thinkers can rationally violate conditionalization by resetting priors. (Ebbs suggests that evidence can give us pragmatic reason to reset our priors, but it’s hard to see how to square this claim with Quine’s epistemological pragmatism. After all, if evidence can provide pragmatic grounds for changing our beliefs, and if these pragmatic grounds aren’t separable from epistemic grounds, then a set of priors, if it’s rational, will build in proper responses to these pragmatic grounds. So we won’t have to reset our priors in order to do what we have pragmatic reason to do.)

  24. 24.

    Here’s a simple example: if the word equilateral has its usual meaning at \(t_1\) but undergoes meaning change and so, at \(t_2\), is synonymous with equiangular, and if S is the sentence “In Euclidean geometry, all equilateral triangles are equiangular”, then \(cr_1(S \mid E) = 1\) (where E specifies the evidence acquired between \(t_1\) and \(t_2\)), and \(cr_2(S) = 1\). In this case, then, \(cr_2(S) = cr_1(S \mid E)\) despite the fact that the meaning of S has changed.

  25. 25.

    Naturalists might insist here that all truths are synthetic, but that’s not what Quine himself claims. The arguments in “Two Dogmas” are intended to establish, not merely that no sentences are analytic, but that the notion of analyticity is unprincipled.

  26. 26.

    Strictly speaking, this is guaranteed to be true only if the language in which the thinker’s evidence sentences are stated is rich enough that, for any body of evidence, there’s an evidence sentence that specifies it. Otherwise, there may be a body of evidence such that no sentence specifies it and such that the thinker is required to reject S in the face of it, in which case S is defeasible despite the fact that \(cr(S \mid E)\) is high for every evidence sentence E. But we can grant, at least for the sake of argument, that the language here is rich enough to make the necessary evidence sentences available. After all, even if a body of evidence isn’t specified by any sentence, thinkers, in order to take it on board, must be able to take some attitude toward it. In particular, they must be able to accept that it obtains. And so, again, as long as we allow that they can also suppose that it obtains—though this supposition won’t take the form of a supposition that any particular sentence is true—we can make sense of a norm requiring that thinkers’ beliefs on accepting that it obtains match their beliefs on supposing it obtains. And we can, if we like, state the diachronic rationality arguments in terms of that norm rather than in terms of the sentential conditionalization principle. So the richness of the language turns out to be immaterial—appeal to evidence sentences, though convenient, isn’t strictly necessary for our purposes here.

  27. 27.

    This description of the situation may not be quite apt. Strictly speaking, if the language is communal, then it’s not even possible for Fred to have B mean what he likes. The meaning of B is just determined by patterns of use in Fred’s linguistic community, and so B means what it means regardless of how Fred chooses to use it. Still, though, how he uses it is going to be determined in part by his beliefs about what it means, and there are, of course, rational constraints on those beliefs. So changes in how he uses B are rationally constrained, which is what’s important for our purposes here.

  28. 28.

    Incidentally, Chalmers agrees that there are facts of the matter about what sentences ought to mean in particular situations. He suggests in Constructing the World that “conceptual evolution...is constant and ongoing, driven by various practical purposes” (2012, p. 231), and he claims in his “Verbal Disputes” that “there are important normative questions about what expressions ought to mean,” questions whose answers “depend on our purposes and values” (2011b, p. 542). Chalmers, though, isn’t an epistemological pragmatist: for him, the practical considerations governing language choice aren’t epistemic.

  29. 29.

    Here I’m relying on a few simplifying assumptions, each of which could be relaxed at the cost of significantly complicating the presentation of my argument. First, I’m assuming that the thinker, in taking an attitude toward S, ought to take a definite stand about what proposition S expresses. If that’s not right—if the thinker instead ought to let uncertainty about S’s meaning have an effect on the attitude she takes toward S —then our constraint will be a bit different:

    $$\begin{aligned} cr(S) = \sum _{i=1}^{n} cr({\mathsf {Exp}}(S,p_i)) \times cr(p_i) \end{aligned}$$

    where \({\mathsf {Exp}}(S,p)\) is the proposition that S expresses p.

    Second, I’m assuming that, for any given body of evidence, there will be a unique best assignment of propositions to sentences—it’s only on this assumption that it makes sense to talk about the proposition that ought to be expressed by S. If this assumption is false—if it’s possible for a body of evidence to make permissible more than one meaning assignment—then, again, our constraint will be a bit different. Let \({ asn}(A,S,p)\) be the following function:

    $$\begin{aligned} { asn}(A,S,p) = {\left\{ \begin{array}{ll} 1, &{} \quad {\text {if assignment }} A {\text { assigns proposition }} p {\text { to sentence }} S \\ 0, &{}\quad {\text {otherwise}} \end{array}\right. } \end{aligned}$$

    Then \(cr(*)\), if our thinker is fully rational, will satisfy the following constraint:

    $$\begin{aligned} {\text {For some }} A {\text { permissible on evidence }} E_t{\text {, }}cr(S) = \sum _{i=1}^{n} { asn}(A,S,p_i) \times cr(p_i) \end{aligned}$$

    Notice, though, that, in general, the thinker’s evidence will include evidence about what meaning assignments she has used in the past. Arbitrary changes in meaning assignment, like the arbitrary resetting of priors discussed in footnote 23, can be ruled out due to the familiarity desideratum. So, even on a permissive view of the relationship between evidence and meaning assignment, most bodies of evidence will place strict limits on what meaning assignments are permissible.

    At any rate, for my purposes here it doesn’t matter which of these constraints we use, so I’m using the simplest one.

  30. 30.

    Note the restriction to eternal sentences. The interaction between conditionalization and indexicality gives rise to lots of problems, none of which is relevant here.

  31. 31.

    As above, if either of my simplifying assumptions is false, things are a bit different. If uncertainty about S’s meaning ought to have an effect on the thinker’s attitude toward S, the candidate constraints are the following:

    $$\begin{aligned} cr(S \mid E)= & {} \sum _{i=1}^{n} cr({\mathsf {Exp}}(S,p_i)) \times cr(p_i \mid E) \\ cr(S \mid E)= & {} \sum _{i=1}^{n} cr({\mathsf {Exp}}(S,p_i) \mid E) \times cr(p_i \mid E) \end{aligned}$$

    If bodies of evidence don’t always pick out unique best meaning assignments, things get more complicated. In that case, these are the candidates:

    $$\begin{aligned}&\displaystyle {\text {For some }} A {\text { permissible on evidence }} E_t{\text {, }}cr(S \mid E) = \sum _{i=1}^{n} { asn}(A,S,p_i) \times cr(p_i \mid E) \\&\displaystyle {\text {For some }} A {\text { permissible on evidence }} E_t \wedge E{\text {, }}cr(S \mid E) = \sum _{i=1}^{n} { asn}(A,S,p_i) \times cr(p_i \mid E) \end{aligned}$$

    But we must remember that the thinker’s future total evidence will include evidence about her present suppositional judgments. So, if the thinker actually goes on to acquire total new evidence specified by E, we can be sure that E includes evidence about what meaning assignment she used in arriving at these suppositional judgments. The question, then, is whether she’s required, on actually acquiring the evidence, to abide by that suppositional meaning assignment.

    If the first candidate constraint is correct, it’s relatively clear that she isn’t so required—after all, the evidence she used in choosing that meaning assignment is not the same as the evidence she now has. But if the second candidate constraint is correct, it’s plausible that she is so required, since she has strong pragmatic reasons to remain faithful to her previous suppositional assignments. Some of these reasons arise from the familiarity desideratum, and others arise from standard Bayesian concerns such as the avoidance of Dutch books.

  32. 32.

    Chalmers himself endorses roughly this conception of suppositional reasoning: “In cases of supposition, we take [a sentence] to be true and we reason just as if it were true” (2014, p. 685).

  33. 33.

    Chalmers (2012, p. 216fn), it’s worth noting, states explicitly that his argument relies on an understanding of entailment according to which, if A entails B, then rationality requires that \(cr(B \mid A) = 1\). This, I take it, is equivalent to (ii): for the sentences in some set \(\Gamma \) to entail S is just for their conjunction to entail S, and a supposition that the sentences in \(\Gamma \) are true just amounts to a supposition that their conjunction is true. The only problem is that, since Chalmers specifies the content of the thinker’s supposition via a single sentence, we need to employ the logical device of the empty conjunction in order for Chalmers’s formulation to be able to deal with the degenerate case where S is entailed by the sentences in the empty set. But this is just an artifact of the decision to use conditional-on-a-sentence credences to formally represent what’s going on in cases of suppositional reasoning. So, to avoid empty conjunctions, I’m using (ii) rather than Chalmers’s version.

  34. 34.

    And the examples I’ve mentioned aren’t the only places where (i) and (ii) are presupposed. Examination reveals that it happens throughout the argument.


  1. Becker, K. (2001). Understanding Quine’s famous ‘statement’. Erkenntnis, 55(1), 73–84.

    Article  Google Scholar 

  2. Bergström, L. (2014). Quine and the a priori. In G. Harman & E. Lepore (Eds.), A companion to W. V. O. Quine (pp. 38–53). Hoboken: Wiley.

    Google Scholar 

  3. BonJour, L. (1998). In defense of pure reason: A rationalist account of a priori justification. Cambridge: Cambridge University Press.

    Google Scholar 

  4. Carnap, R. (1937). Logical syntax of language (A. Smeaton, Trans.). London: Kegan Paul, Trench, Trubner & Co.. (Original work published 1934)

  5. Carnap, R. (1950). Empiricism, semantics, and ontology. Revue Internationale de Philosophie, 4(11), 20–40.

    Google Scholar 

  6. Chalmers, D. J. (2011a). Revisability and conceptual change in “Two dogmas of empiricism”. Journal of Philosophy, 108(8), 387–415.

    Article  Google Scholar 

  7. Chalmers, D. J. (2011b). Verbal disputes. Philosophical Review, 120(4), 515–566.

    Article  Google Scholar 

  8. Chalmers, D. J. (2012). Constructing the world. Oxford: Oxford University Press.

    Google Scholar 

  9. Chalmers, D. J. (2014). Frontloading and Fregean sense: Reply to Neta, Schroeter and Stanley. Analysis, 74(4), 676–697.

    Article  Google Scholar 

  10. Devitt, M. (2005). There is no a priori. In M. Steup & E. Sosa (Eds.), Contemporary debates in epistemology (pp. 105–115). Oxford: Blackwell.

    Google Scholar 

  11. Duhem, P. (1954). The aim and structure of physical theory (P. P. Wiener, Trans.). Princeton: Princeton University Press. (Original work published 1906)

  12. Ebbs, G. (2014). Conditionalization and conceptual change: Chalmers in defense of a dogma. Journal of Philosophy, 111(12), 689–703.

    Article  Google Scholar 

  13. Ebbs, G. (2016). Reading Quine’s claim that no statement is immune to revision. In F. Janssen-Lauret & G. Kemp (Eds.), Quine and his place in history (pp. 123–145). London: Palgrave Macmillan.

    Google Scholar 

  14. Grice, H. P., & Strawson, P. F. (1956). In defense of a dogma. Philosophical Review, 65(2), 141–158.

    Article  Google Scholar 

  15. Grünbaum, A. (1962). The falsifiability of theories: Total or partial? A contemporary evaluation of the Duhem-Quine thesis. Synthese, 14(1), 17–34.

    Article  Google Scholar 

  16. Haack, S. (1974). Deviant logic: Some philosophical issues. Cambridge: Cambridge University Press.

    Google Scholar 

  17. Harman, G. (1995). Rationality. In E. E. Smith & D. N. Osherson (Eds.), An invitation to cognitive science, volume 3: Thinking (2nd ed., pp. 175–211). Cambridge: MIT Press.

    Google Scholar 

  18. Harman, G. (1996). Analyticity regained? Noûs, 30(3), 392–400.

    Article  Google Scholar 

  19. Harman, G. (2001). General foundations versus rational insight. Philosophy and Phenomenological Research, 63(3), 657–663.

    Article  Google Scholar 

  20. Hill, C. S. (2013). Concepts, teleology, and rational revision. In A. Casullo & J. C. Thurow (Eds.), The a priori in philosophy (pp. 134–157). Oxford: Oxford University Press.

    Chapter  Google Scholar 

  21. Jeffrey, R. C. (1965). The logic of decision. Chicago: University of Chicago Press.

  22. Juhl, C., & Loomis, E. (2010). Analyticity. London: Routledge.

    Google Scholar 

  23. Lakatos, I. (1978). Falsification and the methodology of scientific research programmes. The methodology of scientific research programmes: Philosophical papers , volume 1 (pp. 8–101). Cambridge: Cambridge University Press.

  24. Neta, R. (2014). Chalmers’s frontloading argument for a priori scrutability. Analysis, 74(4), 651–661.

    Article  Google Scholar 

  25. O’Grady, P. (1999). Carnap and two dogmas of empiricism. Philosophy and Phenomenological Research, 59(4), 1015–1027.

    Article  Google Scholar 

  26. Peacocke, C. (2005). The a priori. In F. Jackson & M. Smith (Eds.), The Oxford handbook of contemporary philosophy (pp. 739–763). Oxford: Oxford University Press.

    Google Scholar 

  27. Putnam, H. (1983). “Two dogmas” revisited. In Realism and reason: Philosophical papers, volume 3, (pp. 87–97). Cambridge: Cambridge University Press. (Reprinted from Contemporary aspects of philosophy, pp. 202–213, by G. Ryle, Ed., 1976, Stocksfield: Oriel Press).

  28. Quine, W. V. O. (1951). Two dogmas of empiricism. Philosophical Review, 60(1), 20–43.

    Article  Google Scholar 

  29. Quine, W. V. O. (1953). From a logical point of view: 9 logico-philosophical essays. Cambridge: Harvard University Press.

    Google Scholar 

  30. Quine, W. V. O. (1966). Posits and reality. In The ways of paradox and other essays, (pp. 233–241). New York: Random House. (Reprinted from Basis of the contemporary philosophy, pp. 391–400, by S. Uyeda, Ed., 1960, Tokyo: Waseda University Press).

  31. Quine, W. V. O. (1969). Epistemology naturalized. Ontological relativity and other essays (pp. 69–90). New York: Columbia University Press.

    Google Scholar 

  32. Quine, W. V. O. (1970). Homage to Rudolf Carnap. PSA: Proceedings of the Biennial Meeting of the Philosophy of Science Association, 1970, pp. xxii–xxv.

  33. Quine, W. V. O. (1975). On empirically equivalent systems of the world. Erkenntnis, 9(3), 313–328.

    Article  Google Scholar 

  34. Quine, W. V. O. (1976). A comment on Grünbaum’s claim. In S. G. Harding (Ed.), Can theories be refuted? Essays on the Duhem-Quine thesis (p. 132). Dordrecht: D. Reidel.

    Chapter  Google Scholar 

  35. Quine, W. V. O. (1981). Five milestones of empiricism. Theories and things (pp. 67–72). Cambridge: Harvard University Press.

    Google Scholar 

  36. Quine, W. V. O. (1986). Reply to Jules Vuillemin. In L. E. Hahn & P. A. Schilpp (Eds.), The philosophy of W. V. Quine (pp. 619–622). Chicago: Open Court.

    Google Scholar 

  37. Quine, W. V. O. (1992). Pursuit of truth (revised ed.). Cambridge: Harvard University Press.

    Google Scholar 

  38. Quine, W. V. O., & Ullian, J. S. (1978). The web of belief (2nd ed.). New York: McGraw-Hill.

    Google Scholar 

  39. Rupert, R. D. (2016). Embodied knowledge, conceptual change, and the a priori; or, justification, revision, and the ways life could go. American Philosophical Quarterly, 53(2), 169–192.

  40. Russell, G. (2008). Truth in virtue of meaning. Oxford: Oxford University Press.

    Book  Google Scholar 

  41. Schroeter, L. (2006). Against a priori reductions. Philosophical Quarterly, 56(225), 562–586.

    Article  Google Scholar 

  42. Schroeter, L. (2014). Scrutability and epistemic updating: Comments on Chalmers’s Constructing the World. Analysis, 74(4), 638–651.

  43. Sober, E. (2000). Quine’s two dogmas. Proceedings of the Aristotelian Society, Supplementary Volumes, 74, 237–280.

  44. Vuillemin, J. (1986). On Duhem’s and Quine’s theses. In L. E. Hahn & P. A. Schilpp (Eds.), The philosophy of W. V. Quine (pp. 595–618). Chicago: Open Court.

    Google Scholar 

Download references


Thanks to Zachary Barnett, David Christensen, Phillip Galligan, Geoffrey Grossman, Richard Heck, Christopher Hill, Iain Laidley, Miquel Miralbés del Pino, Joshua Schechter, Richard Stillman, Leo Yan, and three anonymous referees for helpful discussion of earlier drafts of this paper.

Author information



Corresponding author

Correspondence to Brett Topey.

Rights and permissions

Reprints and Permissions

About this article

Verify currency and authenticity via CrossMark

Cite this article

Topey, B. Quinean holism, analyticity, and diachronic rational norms. Synthese 195, 3143–3171 (2018). https://doi.org/10.1007/s11229-017-1366-3

Download citation


  • A priori
  • Chalmers
  • Conditionalization
  • Epistemology
  • Naturalism