Abstract
Truth by convention, once thought to be the foundation of a uniquely promising approach to explaining our access to the truth in nonempirical domains, is nowadays widely considered an absurdity. Its fall from grace has been due largely to the influence of an argument that can be sketched as follows: our linguistic conventions have the power to make it the case that a sentence expresses a particular proposition, but they can’t by themselves generate truth; whether a given proposition is true—and so whether the sentence that expresses it is true—is a matter of what the world is like, which means it isn’t a matter of convention alone. The consensus is that this argument is decisive against truth by convention. Strikingly, though, it has rarely been formulated with much precision. Here I provide a new rendering of the argument, one that reveals its structure and makes transparent just what assumptions it requires, and then I assess conventionalists’ prospects for resisting each of those assumptions. I conclude that the consensus is mistaken: contrary to what is almost universally thought, there remains a promising way forward for the conventionalist project. Along the way, I clarify conventionalists’ commitments by thinking about what truth by convention would need to be like in order for conventionalism to do the epistemological work it’s intended to do.
This is a preview of subscription content, log in via an institution to check access.
Similar content being viewed by others
Change history
03 May 2018
The original publication of the article contains two formatting errors, the second of which significantly inhibits readability.
Notes
Conventionalism has also been invoked to explain our knowledge of those domains, our ability to have justified beliefs about those domains, etc.—indeed, this is how it’s usually presented, as a way of developing an analytic theory of a priori knowledge or justification. But I’ve chosen to focus here on reliability, for the following reason: It may be possible to defend a thin conception according to which (e.g.) justified belief is easier to come by than we might have thought—see, e.g., the pragmatic view defended by Boghossian (2003a, b), according to which what explains the justification of certain of our logical beliefs is (roughly) that we must have them in order to reason responsibly at all. If such a conception is available, conventionalism may not be needed to account for our ability to have justified beliefs in the relevant domains, since we may be able to provide an account without explaining how we manage to get at the truth in those domains. But it’s not clear that we can account for our reliability in any similar way, since no analogous thin conception of reliability seems to be available—to explain our reliability just is to explain how we manage to get at the truth. Focusing on reliability, then, is one way of bringing out what’s especially attractive about conventionalism as compared to its alternatives. (In fact, conventionalism is the only view that has ever provided any real hope of a satisfying naturalist-friendly explanation of our reliability in nonempirical domains, though a full defense of that claim is beyond the scope of my discussion here.)
The mystery is especially obvious in the case of mathematics, but even in nonabstract domains, we appear able to reliably form true beliefs about unobserved parts of the world. For example, we haven’t seen every vixen in the world, but we still believe, correctly, that all of them are foxes—and that every vixen that will ever exist will be a fox, and even that that it’s impossible for any vixen to fail to be a fox.
This isn’t to say that there are no contemporary conventionalists—Einheuser (2006), Glock (2003, 2008), Sidelle (1989, 2009, 2010) and Warren (2015a, b, 2017), to name a few, have in recent years issued defenses of conventionalism in various domains. But it’s a distinctly minority position. (García-Carpintero and Pérez Otero (2009) also defend what they say is a version of conventionalism, but their view doesn’t count as conventionalist in the sense I’ve described here—they deny that any sentence can be true by convention alone. As a result, their view doesn’t solve the epistemological mystery conventionalism is intended to solve.)
For a thorough review of Quine’s argument here, see Warren’s (2017) recent discussion of the Quinean case against truth by convention.
Furthermore, if there’s a solution to be found, it’s going to be available to conventionalists just as much as to anyone else. Horwich (1998), for example, is no conventionalist, but he’s committed to there being a distinction between those dispositions that are meaning-constituting and those that aren’t. His account of this distinction is that the meaning-constituting dispositions for a given word are the ones that are explanatorily fundamental, in the sense that they can explain all other dispositions to use that word. If something like this is right, then Quine’s critique presents no problem, for conventionalists or anyone else.
Some theorists state explicitly that Quine’s critique doesn’t get to the heart of what’s wrong with conventionalism. Sider (2011, 100), for instance, says that “Quine’s argument does not go far enough” because it doesn’t “challenge the very idea of something’s being ‘true by convention”’—after all, Quine concedes that it is possible, by explicit stipulation, to make some sentences true. Sober (2000) and Benacerraf (1973) express similar sentiments.
The objection (or something very much like it) has been endorsed in print by Sider (2011), Russell (2008), Williamson (2007), Hale (2002), Sober (2000), BonJour (1998), Horwich (1998), Boghossian (1996), Yablo (1992), Lewy (1976), Benacerraf (1973), David Lewis (1969), Pap (1958), Kneale (1947), C. I. Lewis (1946) and Ewing (1940), among others. Note, too, that neo-Fregeanism about arithmetic (see, e.g., Wright 1983), Schiffer’s (2003) theory of pleonastic entities, and other flavors of what Thomasson (2015) calls “easy ontology” are often thought to be vulnerable to an objection of the same sort: that such approaches are workable only on the absurd view that we have the power to conjure objects into existence via our choice of language. Versions of this latter objection have been endorsed by van Inwagen (2016), Bennett (2009), Chalmers (2009), Boolos (1997) and Field (1984), among others.
I discuss this matter in some detail in Sect. 5 below.
He later decides that this view is untenable; see his (1946, 17).
See also, e.g., Carnap (1937, Sect. 74). Note, too, that Ayer later disavows any view on which this claim is taken literally, saying that, although he “sometimes seem[s] to imply that” analytic sentences “describe the way in which certain symbols are used”, he neither “wish[es] to hold” that position nor “think[s] that [he is] committed to it” (1946, 16).
For more on contemporary theorists’ failure to purge their thinking of conventionalist metaphor, see Sider (2011, Sect. 6.5).
There are, of course, questions to be answered here about just how it is that we have access to facts about our conventions. But there’s no bar in principle to a naturalist-friendly way of answering those questions.
Though this, like most claims about causation, is controversial.
In what follows, I’ll often point out where my claims about the logic of conventionalists’ explanatory relations are structurally analogous or disanalogous to Fine’s claims about the logic of ground.
For what it’s worth, I take this claim to be incorrect. What needs explanation, again, is our near-perfect reliability in the relevant domains, and in order to explain that, we need to explain our ability to be nearly perfectly reliable about when a meaning has successfully been given to E. And it’s not clear, on a view like this, that we have any way of explaining the latter ability. But we can set this worry aside for the sake of argument.
Hale and Wright (2000, 294) call this “the understanding problem” and take it to be a conclusive objection to the sort of view under discussion here.
I want to note that these in-virtue-of relations, whether identical to grounding relations or not, are at the very least formally similar to grounding relations. Their logic, on my view, resembles (but isn’t identical to) the logic of the relations of strict partial ground and strict full ground, respectively, in Fine’s (2012a, b) pure logic of ground. (For more details, see my manuscript, “Metaphysical Analyticity and the Logic of ‘In Virtue Of”’.)
Here I’m casting the objection in terms of the fact that all vixens are foxes, but it could just as easily be cast in terms of (the truth of) the proposition that all vixens are foxes. (Indeed, it usually is cast in terms of the proposition.) I’ve chosen to appeal to the fact rather than to the proposition simply because talk of facts, unlike talk of propositions, is common in nonphilosophical contexts and so can’t be dismissed as mere philosophical extravagance. As I suggested in Sect. 2, conventionalism would be an easy doctrine to dismiss if it required us to deny that it’s a fact that all vixens are foxes.
All that said, the objection, I take it, is in the end just as strong when cast in terms of the proposition as it is when cast in terms of the fact. After all, skepticism about the existence of abstracta such as propositions seems unwarranted from a conventionalist perspective—one of the points of conventionalism, as I suggested in Sect. 1, is to deflate claims about abstracta and so to give naturalists a way of vindicating our claims to knowledge in abstract domains such as mathematics. (Recall that Carnap’s (1956) attempt to allay skeptical worries about abstracta includes discussion of propositions as well as of numbers and classes.) Conventionalists shouldn’t be in the business of rejecting out of hand abstracta that may prove theoretically useful.
Yablo (1992, 878), for instance, says that whether a given proposition is true is a question “to which the rules of usage are quite irrelevant”, and Sober (2000, 247) says, similarly, that “the proposition expressed by [an English-language analytic sentence] does not depend for its truth on how English works”.
But stay tuned.
To take one example, the following rhetorical question, which Boghossian (1996, 364) asks in the course of sketching his version of the objection from worldly fact, presupposes that a sentence has its truth value at least partly in virtue of a corresponding fact: “How could the mere fact that \({\mathbf {S}}\) means that \({\mathbf {p}}\) make it the case that \({\mathbf {S}}\) is true? Doesn’t it also have to be the case that \({\mathbf {p}}\)?”
Indeed, in developing his notions of full ground and partial ground, Fine (2012a, 50) defines partial ground in terms of full ground using just such a biconditional. I don’t define the partly-in-virtue-of relation in terms of the purely-in-virtue-of relation, but the biconditional does turn out to hold in the logical system I’ve developed.
One point of clarification about the notion of full explanation that’s in play here: for an explanation to count as full in the relevant sense, it’s not required that the explanatory story be traced all the way back to fundamental facts (if there are such things). On conventionalism, remember, the truth of certain sentences is fully explained by facts about what conventions are in place. But conventionalists certainly don’t endorse the (absurd) claim that facts about convention are fundamental—it’s abundantly clear that those facts are to be explained by facts about the linguistic dispositions, brain states, etc., of members of the linguistic community.
In fact, given that the purely-in-virtue-of relation is transitive, it turns out to be certainly true that the fact that the sky is cerulean fully explains the fact that it’s colored.
This, I want to note, is one place where the logic of the in-virtue-of relations under discussion here differs from Fine’s pure logic of ground.
The weaker premise can be stated as follows: there’s no fact about convention that v (i.e., the fact that all vixens are foxes) plays a role in explaining and that itself plays a role in explaining the truth of S. Or, more formally:
- \((4^-)\) :
-
\(\forall f(f \in {\mathcal {C}} \rightarrow \lnot \) (S is true partly in virtue of \(f \wedge f\) obtains partly in virtue of all vixens’ being foxes))
And we can see why conventionalists are committed to this premise by thinking again about the Naturalist-Friendliness constraint. The reason conventionalism has any hope of meeting that constraint in the first place, after all, is that facts about convention are supposed to be facts our reliability about which isn’t mysterious by naturalist lights. But v is a fact about how things are with vixens, a fact our reliability about which is mysterious. So any facts that obtain even partly in virtue of v are also going to be facts our reliability about which is mysterious. And that means that, if conventionalism is to do the epistemological work it’s intended to do, v can’t have any role to play in explaining facts about convention. That is, conventionalists are committed to the following claim:
- \((4^{\prime })\) :
-
\(\forall f(f \in {\mathcal {C}} \rightarrow \lnot \) (f obtains partly in virtue of all vixens’ being foxes))
So, since \((4^-)\) is strictly weaker than \((4^{\prime })\), conventionalists are committed to \((4^-)\) as well.
As for why the argument remains valid when (4) is replaced with \((4^-)\): Recall that the point of adding (4) as a premise was to ensure that there’s some fact not about convention in \(\Gamma \). (4) does this by ruling out the possibility that \(\Gamma \) contains some fact f such that v plays a role in explaining f and f plays a role in explaining the truth of S. But even if this possibility can’t be ruled out, there’s no problem for the objection from worldly fact as long as no such f is a fact about convention—i.e., as long as \((4^-)\) is true—for the simple reason that, if \(\Gamma \) does contain any such f, it thereby contains a fact not about convention.
Assuming, again, that the purely-in-virtue-of relation is transitive.
Unless, of course, v is fundamental. But if that’s so, S’s being true by convention certainly can’t guarantee that v obtains—whether v obtains, in that case, has nothing whatsoever to do with what conventions are in place.
I want to note that there’s a gap in this argument. We haven’t ruled out the possibility that the relevant f plays some role in explaining v and also, independently, plays some additional role in explaining some fact about convention c (i.e., that f’s role in explaining v isn’t merely derivative of its role in explaining c). If this were true, it might be possible to tell some story of the following form: Since c is a fact about what conventions are in place, it’s no mystery how we have access to c. But facts about convention aren’t fundamental—they obtain in virtue of certain other facts about us, facts about our linguistic dispositions, brain states, etc. Our access to c, then, puts us in a position to be sure also that some other facts obtain (namely, the facts, whatever they are, in virtue of which c obtains). So, if we’re somehow in a position to be sure that, whenever c obtains, one of the facts in virtue of which it obtains will also, independently, play a role in explaining v, then we may be able to be sure, just by knowing what conventions are in place, that some suitable f will be available.
But there at least two problems with this strategy. First, no plausible story of this sort has ever been given, and it’s not at all clear what a plausible story might look like. And second (and more importantly), any motivation for endorsing (2) is equally a motivation for ruling out the possibility that f has independent roles to play in explaining v and c. The central reason for endorsing (2), after all, is to respect the thought that facts about how things are with us can’t fully explain facts about how things are with the world outside of us, such as the fact that all vixens are foxes. But if f plays a role in explaining c, then f is a fact about how things are with us, despite the fact that it’s not itself a fact about convention. So, if conventionalists embrace this strategy, they’re still forced to say that v obtains purely in virtue of facts about us. And since the motivation for endorsing (2) in the first place is to avoid that result, conventionalists who remain committed to (2) shouldn’t embrace this strategy.
One theorist who takes conventionalists to be committed to claims of this sort is Blackburn (1986, 121)—he argues that necessities can’t be explained by contingent facts about convention, since “the explanation, if good, would undermine the original modal status: if that’s all there is to it, then [for example] twice two does not have to be four”.
Boghossian (1996) and Stroud (1984) both cite conventionalists’ purported commitment to counterfactuals like this one as a serious problem for conventionalism, as does C. I. Lewis (1946, 148), whose argument against truth by convention goes via the following claim: “If the conventions were otherwise, the manner of telling would be different, but what is to be told, and the truth or falsity of it, would remain the same”.
As Wright (1985, 190) puts it, our talk is governed by the general convention that “what it is true to say of a hypothetical state of affairs...is to be determined by reference to our actual linguistic conventions, even if those are not the conventions that would then obtain”. And Sidelle (1989, 7) explains, similarly, that “if it is a convention of ours that nothing in any possible situation counts as water if it is not composed of H2O, then this very convention tells us that in the subclass of possible situations in which we have different conventions, still, nothing counts as water that is not H2O: that is, that it is necessary that water is H2O”.
The most explicit statement of this point comes from Wright (1985, 191).
Sider (2011, 102) expresses a version of this worry, saying that it’s “unclear just what sort of dependence of truth upon conventions is supposed to be distinctive of conventionalism” given that “the conventionalist will surely deny counterfactual or temporal dependence”. The worry also appears to be the basis of Elder’s case against conventionalism—he suggests that denying that counterfactual dependence follows from a fact’s obtaining by convention would be “a desperate move indeed” (2006, 14).
This is a version of a point made by Sidelle (2009, 234–235).
Explicit expressions of this worry in the literature are rare. (Stroud’s (1984, Chap. 5) discussion of Carnap’s conventionalism is an exception, although Stroud seems to take the worry about idealism to be equivalent to the worry about modal claims.) Still, I’ve heard it expressed fairly often in conversation, and so I think it’s worth saying something about.
Einheuser uses this idea to develop a conventionalist-friendly generalization of possible world semantics in which a world is understood as a pair containing a substratum and a “carving” (roughly, a function from substrata to arrangements of whatever features are taken to be conventional), and then she uses this formal framework to give a rigorous argument showing that convention-sensitivity doesn’t entail counterfactual dependence.
References
Audi, P. (2012). Grounding: Toward a theory of the in-virtue-of relation. Journal of Philosophy, 109(12), 685–711.
Ayer, A. J. (1936a). Language, truth and logic. London: Victor Gollancz Ltd.
Ayer, A. J. (1936b). Truth by convention. Analysis, 4(2/3), 17–22.
Ayer, A. J. (1946). Language, truth and logic (2nd ed.). London: Victor Gollancz Ltd.
Benacerraf, P. (1973). Mathematical truth. Journal of Philosophy, 70(19), 661–679.
Bennett, K. (2009). Composition, colocation, and metaontology. In D. Chalmers, D. Manley, & R. Wasserman (Eds.), Metametaphysics: New essays on the foundations of ontology (pp. 38–76). Oxford: Oxford University Press.
Blackburn, S. (1986). Morals and modals. In G. Macdonald & C. Wright (Eds.), Fact, science and morality: Essays on A. J. Ayer’s Language, truth and logic (pp. 119–141). Oxford: Basil Blackwell.
Boghossian, P. A. (1996). Analyticity reconsidered. Noûs, 30(3), 360–391.
Boghossian, P. A. (2003a). Blind reasoning. Proceedings of the Aristotelian Society, Supplementary Volume, 77, 225–248.
Boghossian, P. A. (2003b). Epistemic analyticity: A defense. Grazer Philosophische Studien, 66, 15–35.
BonJour, L. (1998). In defense of pure reason: A rationalist account of a priori justification. Cambridge: Cambridge University Press.
Boolos, G. (1997). Is Hume’s principle analytic? In R. G. Heck Jr. (Ed.), Language, thought, and logic: Essays in honour of Michael Dummett (pp. 245–262). Oxford: Oxford University Press.
Carnap, R. (1937). Logical syntax of language (A. Smeaton, Trans.). London: Kegan Paul, Trench, Trubner & Co. (Original work published 1934)
Carnap, R. (1956). Empiricism, semantics, and ontology. In Meaning and necessity: A study in semantics and modal logic (2nd ed., pp. 205–221). Chicago: University of Chicago Press. (Reprinted from Revue Internationale de Philosophie, 1950,4[11], 20–40)
Chalmers, D. J. (2009). Ontological anti-realism. In D. Chalmers, D. Manley, & R. Wasserman (Eds.), Metametaphysics: New essays on the foundations of ontology (pp. 77–129). Oxford: Oxford University Press.
Einheuser, I. (2006). Counterconventional conditionals. Philosophical Studies, 127(3), 459–482.
Elder, C. L. (2006). Conventionalism and realism-imitating counterfactuals. Philosophical Quarterly, 56(222), 1–15.
Ewing, A. C. (1940). The linguistic theory of a priori propositions. Proceedings of the Aristotelian Society, 40, 207–244.
Field, H. H. (1984). Review of Frege’s conception of numbers as objects. Canadian Journal of Philosophy, 14(4), 637–662.
Fine, K. (2001). The question of realism. Philosophers’ Imprint, 1(1), 1–30.
Fine, K. (2012a). Guide to ground. In F. Correia & B. Schnieder (Eds.), Metaphysical grounding: Understanding the structure of reality (pp. 37–80). Cambridge: Cambridge University Press.
Fine, K. (2012b). The pure logic of ground. Review of Symbolic Logic, 5(1), 1–25.
Garcia-Carpintero, M., & Perez Otero, M. (2009). The conventionaland the analytic. Philosophy and Phenomenological Research, 78(2), 239–274.
Glock, H. (2003). The linguistic doctrine revisited. Grazer Philosophische Studien, 66, 143–170.
Glock, H. (2008). Necessity and language: In defence of conventionalism. Philosophical lnvestigations, 31(1), 24–47.
Gluer, K. (2003). Analyticity and implicit definition. Grazer Philosophische Studien, 66, 37–60.
Hale, B. (2002). The source of necessity. Philosophical Perspectives, 16, 299–319.
Hale, B., & Wright, C. (2000). Implicit definition and the a priori. In P. Boghossian & C. Peacocke (Eds.), New essays on the a priori (pp. 192–205). Oxford: Oxford University Press.
Horwich, P. (1997). Implicit definition, analytic truth, and apriori knowledge. Noûs, 31(4), 423–440.
Horwich, P. (1998). Meaning. Oxford: Oxford University Press.
Jenkins, C. S. (2008). Boghossian and epistemic analyticity. Croatian Journal of Philosophy, 8(22), 113–127.
Kneale, W. C. (1947). Are necessary truths true by convention? Proceedings of the Aristotelian Society, Supplementary Volume, 21, 118–133.
Lewis, C. I. (1946). An analysis of knowledge and valuation. La Salle: Open Court.
Lewis, D. (1969). Convention: A philosophical study. Cambridge: Harvard University Press.
Lewy, C. (1976). Meaning and modality. Cambridge: Cambridge University Press.
Malcolm, N. (1940). Are necessary propositions really verbal? Mind, 49(194), 189–203.
Pap, A. (1958). Semantics and necessary truth: An inquiry into the foundations of analytic philosophy. New Haven: Yale University Press.
Quine, W. V. O. (1936). Truth by convention. In O. H. Lee (Ed.), Philosophical essays for Alfred North Whitehead (pp. 90–124). London: Longmans, Green and Co.
Quine, W. V. O. (1960). Carnap and logical truth. Synthese, 12(4), 350–374.
Rosen, G. (2010). Metaphysical dependence: Grounding and reduction. In B. Hale & A. Hoffmann (Eds.), Modality: Metaphysics, logic, and epistemology (pp. 109–135). Oxford: Oxford University Press.
Russell, G. (2008). Truth in virtue of meaning. Oxford: Oxford University Press.
Schaffer, J. (2009). On what grounds what. In D. Chalmers, D. Manley, & R. Wasserman (Eds.), Metametaphysics: New essays on the foundations of ontology (pp. 347–383). Oxford: Oxford University Press.
Schiffer, S. (2003). The things we mean. Oxford: Oxford University Press.
Sidelle, A. (1989). Necessity, essence, and individuation: A defense of conventionalism. Ithaca: Cornell University Press.
Sidelle, A. (2009). Conventionalism and the contingency of conventions. Noûs, 43(2), 224–241.
Sidelle, A. (2010). Modality and objects. Philosophical Quarterly, 60(238), 109–125.
Sider, T. (2011). Writing the book of the world. Oxford: Oxford University Press.
Sober, E. (2000). Quine’s two dogmas. Proceedings of the Aristotelian Society, Supplementary, 74, 237–280.
Stroud, B. (1984). The significance of philosophical scepticism. Oxford: Oxford University Press.
Thomasson, A. L. (2015). Ontology made easy. Oxford: Oxford University Press.
van Inwagen, P. (2016). The neo-Carnapians. Synthese. https://doi.org/10.1007/s11229-016-1110-4.
Warren, J. (2015a). Conventionalism, consistency, and consistency sentences. Synthese, 192(5), 1351–1371.
Warren, J. (2015b). The possibility of truth by convention. Philosophical Quarterly, 65(258), 84–93.
Warren, J. (2017). Revisiting Quine on truth by convention. Journal of Philosophical Logic, 46(2), 119–139.
Williamson, T. (2007). The philosophy of philosophy. Malden: Blackwell.
Wisdom, J. (1938). Metaphysics and verification (I). Mind, 47(188), 452–498.
Wright, C. (1983). Frege’s conception of numbers as objects. Aberdeen: Aberdeen University Press.
Wright, C. (1985). In defence of the conventional Wisdom. In I. Hacking (Ed.), Exercises in analysis: Essays by students of Casimir Lewy (pp. 171–197). Cambridge: Cambridge University Press.
Yablo, S. (1992). Review of Necessity, essence, and individuation: A defense of conventionalism. Philosophical Review, 101(4), 878–881.
Acknowledgements
Thanks to Zachary Barnett, David Christensen, Nina Emery, Tobias Fuchs, Phillip Galligan, Geoffrey Grossman, Louis Gularte, Yongming Han, Richard Kimberly Heck, Christopher Hill, Paul Horwich, Iain Laidley, Rachel Leadon, Miquel Miralbés del Pino, Mary Renaud, Kirun Sankaran, Joshua Schechter, and Leo Yan for helpful discussion of the material in this paper.
Author information
Authors and Affiliations
Corresponding author
Rights and permissions
About this article
Cite this article
Topey, B. Linguistic convention and worldly fact. Philos Stud 176, 1725–1752 (2019). https://doi.org/10.1007/s11098-018-1088-5
Published:
Issue Date:
DOI: https://doi.org/10.1007/s11098-018-1088-5