This paper focuses on two contrasting approaches to the theory of linguistic meaning and asks how they color a range of issues of interest to scholars of religion. The so-called truth-conditional approach makes truth basic. It trades on the thought that we sometimes or perhaps often know what someone has said when we know what it would be for what she has said to be true. The other approach pegs meaning to how expressions and sentences are used in communicative situations. Dummett and Davidson are front and center. Davidson is of course in one sense a champion of truth-conditional semantics, but, over the issues I have in view, his case is instructively mixed. This discussion leads us toward an account of linguistic meaning which elevates over truth a family of concepts associated with use, including verification, justification, and pragmatic success.
Perhaps there are features of religion that do not depend on linguistic communication, but it is hard to think what they might be or in what their interest might consist. Belief, experience, liturgy, myth, pilgrimage, prayer, scripture, sermon, song, ritual, text, value—each of these draws, directly or indirectly, on linguistic meaning, and this list merely scratches the surface of religious expression. When we study religion, no doubt we implicate linguistic meaning. But must we implicate a theory of linguistic meaning—a semantic theory? Clearly not. Just as the political scientist studying voting behavior needn’t ask herself how words become meaningful so we may examine any item on this list without raising foundational questions about what it is for words to mean what they do. Just as we ourselves use language without knowing how we do it, so we can quite well understand what religious people are saying without knowing how they are managing to say it.
Still, with some regularity, we do encounter general questions about “the meaningfulness of religious language.” Sometimes the context is polemical: witness the mid-twentieth century collision of logical positivism with God-talk (on which more later). Sometimes the occasion is a surprising claim from the field—for example, that ritual action conceals unsuspected semantic content, or that an apparently meaningful vocalization has none.1 Or perhaps someone suggests that only an “insider” could understand this or that piece of religious discourse, or that this or that community sees the world through a radically different conceptual framework. Reflection on any of these claims might lead us to ask how we manage to do things with words.
In this paper, I focus on two contrasting approaches to the theory of linguistic meaning and ask how they color a range of issues of interest to scholars of religion. The so-called truth-conditional approach makes truth basic. It trades on the thought that we sometimes or perhaps often know what someone has said when we know what it would be for what she has said to be true. The other approach, to put it crudely, pegs meaning to how expressions and sentences are used in communicative situations. In Sections 2 and 3 I discuss a version of the truth-conditional approach. My goal there is not so much to criticize the approach as it is to bring out several ironical connections between it and the study of religion. In Section 4 I take up Donald Davidson’s work. He is of course in one sense a champion of truth-conditional semantics, but, at least over the issues I have in view, his case is instructively mixed. This discussion leads us toward an account of linguistic meaning which elevates over truth a family of concepts associated with use, including verification, justification, and pragmatic success. Section 5 takes stock, and Section 6 takes up an example.
2 Truth-Conditions and the Plenum
Let us consider the claim that the meaning of a statement is determined or explained by the conditions under which it is true. That at least many declarative sentences have truth-conditions seems clear enough. As Robert Brandom notes, “one can say of anything that has propositional content that it has truth-conditions…. [T]his characterization is just a harmless compliment paid to things whose contents can be expressed in declarative sentences” (1994: 329). Perhaps the most straightforward case involves using a declarative sentence to make a serious assertion: to say that this or that is true under certain conditions. It is sometimes said that religion is not primarily or even at all in the business of making serious assertions. I do not agree—but let us put this question aside for the moment. I hope we can at least agree that religious people put declarative sentences to all kinds of uses—for example, to give commands, to ask questions, to give praise, and to express hope. In any event, to say that declarative sentences have truth-conditions is, per Brandom, harmless.
However, the claim that interests us isn’t that declarative sentences have truth-conditions but rather that grasping the meaning of a sentence consists in or is explained by knowing the relevant truth-condition. So we have the familiar Tarskian truth-schema as portrayed by Mark Gardiner (introduction to this volume p. 327):
T: “s” is true if and only if p
The T sentence is disquotational because it uses quotation on the left of the biconditional and removes it on the right. The left-hand scare quotes set off a sentence we do not understand—an acceptable interpretation of which is given on the right. The suggestion, as Gardiner says (introduction to this volume p. 327), is that theorems of this sort can replace meaning-specifying theorems of the form:
M: “s” means that-p
This kind of replacement is said to be an advance because associating a statement we don’t understand with the condition under which it is true—or, as it is sometime put, knowing its truth-condition—is tantamount to understanding that statement and, so, to grasping its meaning.
One immediate question, much discussed in the literature, is how to read “knows” in this context. In attributing such knowledge, clearly we are going to require more of the speaker than the ability to produce, on the right side, an equivalent expression. Thus, beginning with the unsatisfactory, ‘Every diamond is hard’ is true if, and only if, every diamond is hard, we might try to make progress by replacing the right side of the bi-conditional with an apparently informative expression—perhaps, ‘Every diamond is hard’ is true if, and only if, every diamond has a certain molecular structure. But this will help only if we already grasp the truth-conditions having to do with substances having this certain molecular structure. For the same reason it will be no help to “explain” the nature of this certain molecular structure in still more basic terms, for here again we invite triviality. The difficulty we are having—a long-time theme of Michael Dummett’s—reflects the fact that truth-conditional theories of this sort explain meaning in terms of truth and not in terms of our ability to recognize or grasp it as such. Thus, Dummett remarks that a truth-conditional theory of meaning “provides us with no conception of how a knowledge of the truth-condition is delivered to the speaker” (1991: 316). This deficiency (as Dummett sees it) will be most damaging in those areas of discourse where we are not, or perhaps are not normally, in a position to recognize associated truth-conditions.
For those of us interested in religion, it is tempting to take Dummett’s criticism as a version of the standard empiricist challenge to the meaningfulness of religious discourse—familiar in its verificationist form to generations of graduate students in both philosophy and religion. It isn’t. The empiricist challenge has to do with evidential strength; it alleges the subject matter at issue is so far removed from human cognition as to hollow-out the general terms at issue. For example, it asks how closely the universe resembles a product of intelligent design; it asks how a wholly spiritual being can be said to possess a mental life, etc. By contrast, Dummett is alleging an in-principle problem. Consider Socrates’ exchange with Euthyphro, which I paraphrase thusly:
E: There is war amongst the gods.
S: What does that mean?
E: “There is war amongst the gods” means there is war amongst the gods.
S: That’s uninformative. Please try again.
E: “There is war amongst the gods” is true if and only if there is war amongst the gods.
S: That’s still uninformative. Please try again.
E: “There is war amongst the gods” is true if and only if there is fighting between Uranus and Cronos and Zeus.
S: Appealing to truth here doesn’t help. These truth-conditions are inaccessible to finite creatures like you and me.
My Socrates—I don’t think he’s very far from Plato’s Socrates—is asking Euthyphro to explain how he can be in a position to recognize or somehow take in the truth of sentences purporting to be about the gods. In Dummett’s terms, Socrates is asking Euthyphro to “manifest”—to make explicit through his linguistic and other behavior—the combination of inference and empirical cognition that gives them meaning. Euthyphro is claiming to know their meaning but can say nothing about how he would go about showing them to be true, nothing about how their truth has been delivered to him. Behind Socrates’ question is the thought that such limited creatures as ourselves are not in a position to recognize that such sentences are true. If so, then we cannot know the truth-conditions attached to them, and so, by hypothesis, we cannot know their meaning.
Dummett’s point is that the truth-conditional theorist must take truth as radically independent of our ability to recognize the right half of the biconditional as true. We cannot reply on Euthyphro’s behalf that the supporting pattern of inferences and experiences is, as is often the case with linguistic competence, implicit—that, as I noted at the outset, we often use words meaningfully without knowing how we manage it. Nor can we appeal to the standard division of linguistic labor—to the thought that Euthyphro knows a little (but enough) about the gods in the way that someone knows a little (but enough) about plantar fasciitis. You can’t have knowledge, explicit, implicit, or partial, of what cannot be known.
This is of course not a general argument against the claim that meaning is best explained by appealing to truth-conditions. It is just pointing out that the truth-conditional proposal can get no traction in spheres of discourse where we cannot legitimately attribute knowledge of truth-conditions. To argue in this fashion against truth-conditional semantics per se would require showing something much stronger, viz., that speakers must always be capable of fully manifesting their (explicit or implicit) mastery of relevant truth-conditions. Dummett has famously prosecuted this case, arguing that defenders of truth-conditional semantics must avail themselves of concepts aligned with use—verification, justification, etc.—rather than truth. I cannot evaluate that argument here. What’s more, an inability to engage religious discourse may not be peculiar to a truth-conditional approach to linguistic meaning. We are familiar with the claim that no semantic theory can get a purchase on religious discourse because religious discourse has no or little content expressible in declarative sentences—that by its nature it is not truth-evaluable. This sort of claim would be well received in many precincts of religious studies. It would be endorsed by Lars Albinus (this volume), as I understand his paper, and (as Albinus points out) by Roy Rappaport (1999). (Richard Rorty holds out this prospect as a virtue. If religious speech could be walled off from truth and falsity it would likely attract fewer fanatics (Rorty 1997).) On the other hand, perhaps other semantic theories are better positioned to make sense of the content of religious discourse. This suspicion might be motivated by the thought that we do know, well enough, what Euthyphro is saying even though we do not and could not recognize the associated truth-conditions and that a variety of religious fundamentalisms are making perfectly straightforward truth-claims. I leave this as an open question for the moment.2
Having argued that truth-conditional semantics, whatever its wider prospects, is not apt for interpreting religious discourse, I now want to suggest that it implicates what amounts to a theological metaphysics and epistemology.
We have seen that the truth-conditional theorist wants to make truth explanatorily basic, and that this makes for a split between what gives a sentence or an utterance its meaning (its truth-condition) and a speaker’s ability to recognize whether that condition obtains—to cognize or grasp it. Dummett contrasts the truth-conditional theory with one that rests on our ability to credit a speaker with knowledge, explicit or implicit, of how an assertion might be verified or given some empirical support:
The truth-conditional meaning-theorist has an entirely different conception. For him, the entire history of the universe, past and future, subsists in an eternal plenum of reality and is that reality to which our assertions relate. We pick out individual objects within this plenum by associating with the names we employ certain uniquely satisfiable conditions.1991: 310
Dummett is trading here on an idea of fundamental importance to the truth-conditional picture, namely, that the world is as it is entirely independent of us. This goes for talk about the cat on the mat (the one I see in front of me) as much as it does for putative god-talk—the question of the accessibility of the world to finite creatures is not to the point. On the contrary, reality is transcendent partly in the sense that it is indifferent to the cognitive apparatus of any sentient creature. Thus, we can speak of “the entire history of the universe.” It is this reality that makes our utterances true.
Dummett says of this claim—that truth-conditions may somehow be grasped by us even though we are not equipped to recognize them—that it “is a piece of mythology” (1991: 316).3 An ironical assessment given our present interest. I would extend Dummett’s figure to say that it is an essentially theological doctrine. Indeed, the conception of “an eternal plenum of reality” is one traditional conception of god in Judaism, Christianity, and Islam. The idea is to refer to our subject matter—God, reality—while at the same insisting that it may in principle be inaccessible to us. One wants to ask: Then whose cognition is it? Here too we are in well-worn theological territory. On Leibniz’s view, for example, God grasps an object and (what amounts to the same thing) the infinite complexity of the world in a single intuition without requiring a contribution from sensation. Here we may recall Dummett’s complaint that the truth-conditional account gives no attention to how this picking-out is to be “delivered to the speaker.” But Dummett’s complaint has no place in our current context, for the subject now under discussion does not require empirical cognition. Rather, it views the world sub specie aeternitatis. It would seem that for god such truth-conditions are tailor made.
3 Truth-Conditions and Absolute Determination
Like others in the humanities and social sciences, scholars of religion often want to understand the life-worlds of people and communities, including their attitudes, values, and practices. Often this involves figuring out what people mean by their words. In Gardiner’s terms this takes us from the ontological to the methodological side of semantics (introduction to this volume p. 323). It takes us from questions about the nature of meaning—we have so far been considering the suggestion that it consists in a sentence’s truth-condition—to questions about how, in practice, truth-conditions are to be identified. In the next section I want to turn to how this distinction plays out in Davidson’s work—work which has been deservedly influential in the religious studies community. But before turning from ontology to methodology, let us ask how the two are related. Writing approvingly of Davidson, Gardiner says
This answers the ontological question: the meaning of a sentence is to be identified by its truth-conditions, where the truth-conditions are to be thought of as instances of the T-schema (called T-sentences) (introduction to this volume p. 327) … [Methodologically,] the interpreter must begin postulating certain propositional attitudes—chiefly beliefs and desires—to the speaker that would explain how and why she is moving her body in that way and in that circumstance. The underlying premise of this maneuver is that the meaning of a sentence is to be identified with the content of the propositional attitude it expresses.this volume p. 328
In this way, Gardiner thinks the ontological and methodological projects are, as he says, “intimately related,” (introduction to this volume p. 324). Here I think we must proceed with caution.
We saw at the outset that, when employed in the service of understanding, the threat of triviality hangs over the Tarskian T-schema. It is no help to trade “water” for “H2O,” or “the gods” for “Uranus, Cronos, and Zeus” unless we already understand the replacements. However, in Tarski’s original paper, the threat of triviality does not arise. That is because Tarski confines his interest to formalized languages. When Tarski speaks of a “formalized” language, he means “an artificially constructed language in which the sense of every expression is unambiguously determined by its form” (1933 (1983): 165-6). He cautions that his results do not hold for ordinary, natural languages. He notes that, in formalized languages we can guarantee that the entities we quantify over—for example, strokes, “|||||”, rather than words—are given precise or absolute determination, something that is not possible in ordinary, natural language.
Now the distinction between, on the one hand, the unambiguous or absolute determination of logical notation and, on the other, the incomplete or partial determination of empirical objects had earlier been emphasized by C.S. Peirce. He notes that first-order quantificational logic with bivalence presupposes that all individual objects in the range of the variables are absolutely determinate (Peirce 1931-58: vol 8: 208).4 That is, first-order quantificational logic with bivalence can admit an object that is both X and not X or neither X nor not X, whatever predicate ‘X’ may represent, only on pain of inconsistency. But, as he notes elsewhere, “I had long before declared that absolute individuals were entia rationis, and not realities. A concept determinate in all respects is as fictitious as a concept definite in all respects” (vol 8: 208). Continued inquiry can always further refine our knowledge of this or that empirical object. Here one of the many Sidney Morganbesser stories is apposite. Morganbesser has seen a hit-and-run, and a policeman asks him what color was the car. Morganbesser replies, The side facing me was blue. The point is one over which Tarski and Peirce agree: the objects over which we quantify in ordinary language are not and cannot be absolutely determined. It is partly for that reason that Tarski confines himself to languages whose elements are by design entia rationis, the sorts of things whose unambiguous or absolute determination can be stipulated.
What, then, of Gardiner’s claim that the ontological and methodological families of questions in semantics are intimately related? What threatens this claim is the gap—emphasized by Peirce and Tarski—between the sharp, precise entities of formal logic and the fuzzy, partly determined objects of our everyday experience. It is not clear to me how to bridge this gap5—that is, how to accommodate one side to the other. The methodological side seems non-negotiable. Fuzzy empirical objects and such finite subjects as ourselves are a matched pair. Unlike logical symbols, empirical objects are fuzzy because they are presented to us from a limited perspective and not all at once. We are all in Morganbesser’s position. We encounter automobiles with finite, rather than with Leibnizian divine intellects. We cannot see the side facing away from us. Thus, the only prospect would seem to be fuzzing-up the logic underlying the ontological side. I don’t know that this cannot be done, but neither am I aware of any appealing proposals. For our purposes it is enough to note that, to gloss over this gap takes us back to theology, for the notion of an absolutely determined individual—one that therefore has no place among objects of experience—is just that of the god of the philosophers.
Talk of “encountering” objects or of having them “presented” suggests a different way of connecting truth and meaning. Consider a standard Peircean schema, this one from his well-known 1878 example:
“This is a diamond” is true if and only if it is not scratched by certain substances under certain temperatures and pressures.1878 (1935): 259-60
What has happened to truth in the Peircean schema? Let me just make three points about what is obviously a complex question. First, truth-conditions are still very much with us. Peirce is stating the meaning of the sentence in truth-conditional terms; so much is, as Brandom says, harmless. The difference is that, with Peirce, we are no longer explaining meaning in terms of truth-conditions; we are no longer claiming that meaning consists in truth-conditions. Truth is no longer explanatorily basic. Rather, we are now promising to explain meaning by appealing—loosely expressed—to the difference the terms at issue make in experience. Second, as in the Tarskian schema, the approach to truth is, in contemporary terms, deflationary or minimalist, and is exhausted by the simple equivalence:
‘p’ is true if and only if p.
Our truth-talk is deflationary or minimalist in the sense that it affirms our correspondence intuitions without assimilating truth to one or another pragmatic, epistemic standard. That is, it does not assimilate truth (contra Peirce himself) to what the ideal community of inquiry is fated to believe, or with whatever is good in the way of belief (James). Truth is a correspondence, as Kant puts it, between cognition and its object—but in so saying we do not move beyond what he calls its “nominal” (die Namenserklarung) sense. Third, truth is available for use as what Quine calls a device for semantic ascent (1960: §56). That is, it allows us to pass judgment on certain kinds of other sentences—e.g., “It’s true that whatever Euthyphro says is misleading.” It is, in this sense, a meta-linguistic predicate—we are “ascending” in subject-matter from talk of the world to talk of sentences. If truth has been displaced it is still indispensable.
I have portrayed the move from Tarski to Peirce as a move from explaining to describing meaning in truth-conditional terms. Now, as Gardiner will note (this volume, pp. 403-405), Davidson at an early point moved from a program of capturing the meaning of sentences in a natural language in a Tarski-style truth theory to a project in which, on the methodological side, the heavy lifting is done by a set of auxiliary constraints having nothing in particular to do with Tarski’s work in the semantics of formalized languages. These constraints include assumptions about the preponderant rationality of the speaker (the “charity” constraint), an account of general terms or concepts on which the content of any one depends on the content of many others (the “holism” constraint), and the claim that agreement about the comings and goings of middle-sized objects—roughly speaking, about the way things appear—is more important to understanding than is agreement over how things are. Since I have discussed each of these on other occasions I will not dwell on them here.6
To come right to the point: I am with Michael Williams and others who find that, when Davidson gets down to the business of real-world interpretation, he is actually working with a broadly verificationist picture of meaning. To see the element of verificationism in Davidson we must recognize within his work both a theory of meaning in a narrow sense, and a more general theory of interpretation. Williams writes that, in the
narrow sense, a Davidsonian theory of meaning is an axiomatic theory that generates the meaning of every sentence of some target language. Now, since the ‘meaning’ thus generated is simply the sentence’s truth condition, it is natural to suppose that Davidson identifies meanings with truth conditions…. [But] Davidson owes (and provides) an account of why a device for generating truth conditions should be thought to specify meanings, as well as an account of the adequacy criteria to which such a theory is subject.1999: 553
The narrow sense is the theory of meaning strictly considered. It takes the form of “a recursive device for specifying the meaning of every sentence of a given language or, more precisely, of the current idiolect of a particular speaker” (553). But, as Williams points out, in providing an account of the adequacy criteria to which the theory of meaning is subject, Davidson produces a freestanding theory of interpretation—free-standing in the sense that the adequacy criteria (charity and holism prime among them) in no way depend upon or even implicate the claim that meaning consists in or is explained by truth-conditions. Rather, as Davidson himself emphasizes, it’s the other way around: together with whatever background and local knowledge the interpreter brings to the table, charity and holism constrain interpretation, radical or every day, foreign or domestic, insider or outsider (1994: 122).
This account lines Davidson up with Peirce, who defended a brand of realism on which the process of verification may never end—but during which time what we come to believe about, say, the hardness of diamonds may very well be true. On the epistemological side we cannot claim a final determination of truth, while, on the metaphysical side, we may well have achieved it. This seems to be Davidson’s position on assigning content to mental states. The process of interpretation, radical or real, involves “optimizing” a host of variables. While this may preclude certainty it does not preclude truth. Thus, Davidson bridges the gap between the ontological and methodological families of questions in semantics by denying that there is anything to the nature of meaning over and above the result of this kind of optimizing. Such is Davidson’s “intentional realism” with respect to meaning.7 I find it a very compelling story. What role does truth play in it? The same three-fold usage as above. The nominal sense of truth as correspondence is required by the T-schema—both in that the interpreter assumes the speaker has the concept that things are thus-and-so and also in that the interpreter herself judges of adequate T-schemas that they capture how things are. Further, the interpreter must take the speaker to be right about much of the inferential and evidential purport of his utterances (the charity constraint), but truth in this context merely allows for the expression of agreement. As above, it is expressive in Quine’s sense, a device for semantic ascent.
We are now ready to appreciate the element of verificationism in Davidson’s account of interpretation. To see speakers as conveying thoughts, rather than just making word-sounding noises, we must take it that they know what would count as substantiating this or that assertion. The evidence will come, as Davidson says, in the form of “readily available behavior” (1994: 128).8 In a late essay, Davidson is explicit about this, writing that “the problem … is that of explaining what makes the difference between showing, by one’s behavior, that one merely responds to perceptual similarities, and that one has a criterion for grouping things found perceptually” (1999: 39). Now the notion of applying a criterion takes us deep into the conceptual neighborhood of justification. As good Davidsonian interpreters we are interested in what speakers take to be justifying their assertions—and that will presumably include grappling with the practical consequences of their remarks—with implications, successes, failures, borderline cases, etc.
I have been speaking of a broadly verificationist element in Davidson’s approach to interpretation, but in fact, as I have just sketched it, it also contains elements of pragmatism. Verificationism puts the emphasis on the conditions for uttering a certain sentence. As Dag Prawitz has it, this “covers the conventions that tell on what occasions it is appropriate to utter the sentence as far as it has to do with the meaning of the sentence” (2007: 457). This is a central part of the principle of charity, and is continuous with Davidson’s emphasis on detecting the speaker’s criterion for using a sentence under perceptual circumstances. By contrast, a pragmatist approach to meaning (again, Prawitz), “is concerned with what reactions are appropriate when the sentence has been uttered, both what the speaker commits himself to and what the hearer has the right to expect” (2007: 457).9 We have already met with the makings of this idea in the context of charity: competent speakers don’t merely appreciate many of the logical and evidential relationships between their sentences but they agree to allow those relationships to constrain their reactions to utterances of them; that is, they commit themselves to react, on the whole, appropriately. Thus, for Peirce, as an initial approximation, to know the meaning of “is hard” is to expect that an object will not be scratched by a certain substance under certain conditions. For someone in possession of such knowledge it would be appropriate to evince surprise at that outcome only under non-standard conditions—e.g., in the context of irony or subterfuge.10
None of this commits us to a simple identification of meaning with use, and certainly not to a systematic theory of meaning as use, whatever that might come to. Such an identification is blocked by the fact that the appropriate use of many expressions depends on, among other things, their meaning. (Davidson makes in this point in many places. It supports his well-known views on metaphor.) Indeed, it seems unlikely that mastery of such verificationist or pragmatic rules as these could give us, by themselves, a systematic theory of meaning in Williams’ narrow sense, that is, a recursive device for specifying the meaning of every sentence of a given language. But, equally, Davidson’s treatment of charity and holism suggests that grasping the appropriate conditions for uttering a sentence and undertaking to react appropriately to its utterance are central to the philosophical project of giving an account of meaning.
The emphasis I am placing on continued inquiry into the conditions and consequences of use echo Davidson’s elevation of idiolect over language, “passing theory” over “prior theory” (Davidson 1986)—more generally his sensitivity to how context-sensitive and interest-relative are the notions of a good or correct interpretation of another’s speech. It is uncontroversial that Davidson counts the principle of charity as an adequacy condition by which to measure such goodness or correctness. I have suggested in this section that a commitment to Davidsonian charity pushes the theory of meaning in the direction of verificationism and pragmatism. At the most basic level this would be to replace knowledge of truth-conditions with knowledge of relevant extra-linguistic practices and procedures—namely, conditions and consequences of use. I intend this move to track Williams’ distinction between a theory of meaning as a systematic, recursive device, and a theory of interpretation. What work is left for the theory of meaning in this narrow sense seems to me an open question.
5 Four Takeaways for Religious Studies
Let us take stock.
5.1 The Move from Explaining to Describing Meaning in Terms of Truth-Conditions is the Move from Theology to Religious Studies
I hope that, for students of religion, the contrast between the truth-conditional approach to linguistic meaning and the Peircean one has a familiar ring. On the one side, we have a picture of the world as composed of absolutely determinate individuals, as wholly indifferent to the structure of finite cognition, and as making (some of) our sentences and utterances true. With Peirce by contrast we tie meaning to human experience.
We trade the god’s eye point of view for that of continued inquiry. In so doing, we incline toward Dummett’s dictum that an adequate theory of linguistic meaning should be a theory of understanding; that it should capture what speakers know about how to use expressions in the give and take of actual communicative encounters. This is just the move from theology to religious studies.
5.2 Giving Up Explaining Meaning in Terms of Truth does not Mean Giving Up Assigning Truth and Falsity to Religious Beliefs, Utterances, and Inscriptions
On the minimalist or deflationary picture I am painting, we can quite well speak of truth as a correspondence between cognition and its object. That, again, is the merely nominal sense of truth. The mistake is to think we can give a substantive general characterization of what that correspondence consists in—as though truths share a common truth-making property. Truth-talk also allows us to comment on individual or, especially, groups or kinds of propositions. Thus, Nancy Frankenberry can indict the American left for its failure to take Christian and Islamic fundamentalisms as partly motivated by false beliefs (2005). And Gabriel Levy can point out the irony in the failure to acknowledge religious belief as false (when it is), namely, that, in the name of respect, we erase the agency and the rationality of the speaker (Levy 2012: 151).11
5.3 To Bypass Awareness for Truth is to See—to Paraphrase Davidson—Spirituality in the Mirror of Meaning
Our two approaches to linguistic meaning, the truth-conditional and the verificationist-pragmatic, give opposing accounts of reference. The truth-conditional account treats reference as byproduct of truth. Thus, Quine, Davidson, and Rorty would prefer to treat seeing a diamond as a relationship between an object and a verbalization, “There is a diamond,”12 namely, the relationship of “holding true.” Awareness is cut out altogether. By contrast, the Peircean truth schema cannot do without awareness and what Peirce calls the “secondness,” that, as he puts it, “jabs you perpetually in the ribs” (1931-58: 6.95). For Peirce, following Kant, reference is unavoidably mentalistic. We are aware of being affected by what we judge, perhaps falsely, to be a diamond. These two options in semantics track a fundamental divide in Western spirituality: on the one hand, the attempt to erase the self—often a form of spiritual submission—as contrasted to its intensification with an attendant emphasis on inwardness and subjectivity. It is hard to imagine the history Judaism, Christianity, or Islam without this internal tension—sometimes creative, often destructive.13 The attraction we feel to explaining meaning by appealing to truth rather than to experience would then be a kind of antiseptic spirituality, one characterized by the attempt to view a world of absolutely determinate objects from the point of view of no one in particular.
5.4 The Logical Positivists were Right about Religious Discourse, Almost
Gardiner will remind us of the logical positivists’ harsh attitude toward, as he says, much religious discourse (this volume p. 410). I am a bit more inclined toward charity. It is instructive to think of the positivists as students of religion of a very narrow sort. They fastened on putative discourse about a god so abstract as to resist verificationist and pragmatic tests—a god so abstract as to resist being grouped under any general concept, one with respect to whom there is no gauging whether we are reacting appropriately because, as, for example, with Tillich’s “god beyond the god of theism,” it finds no place among objects and events in the world. Finding no prospect of truth-conditions or means of verification or justification, they heard noise rather than meaningful speech.14
We might try to find the positivists guilty of what Stephen Yablo has recently called “truth-puritanism” (2014). Yablo points out that someone who says, wrongly, that it is Tuesday does nevertheless get something right. Even if it isn’t Tuesday, still, it is a day of the week (2014: 2-3). We “look past” the content of what was said to the underlying subject matter. One reaction to the positivists is to say they should have taken that route in the case of religious discourse. But is that fair? What underlying subject matter should they have kept in view? In Yablo’s example the speaker is implicitly committed to and is right about it being some day of the week—but it is hard to find a parallel in Tillich’s case. In Yablo’s terms, the positivists are saying: such god-talk isn’t even partly true. I am inclined to say that, if the positivists can be faulted, it is for a certain blindness to the wider, more humane work of Davidsonian optimization, not to be distinguished from the humane work of understanding. This is an area where the modem study of religion excels. Scholars of religion are trained, as Durkheim puts it, “to go beneath the symbol, to uncover the reality which it represents” (1912 (1995): 14-5). In cases where our specifically semantic inquiry isn’t panning out taking Durkheim’s advice may mean looking way past the surface, perhaps all the way to a social scientific subject matter. Thus, in The Elemental Forms of Religious Life, Durkheim sees a deep rationality in the apparently irrational practice of worshipping of plants and animals. He illustrates in one stroke both Levy’s point about how falsity can sometimes preserve agency and Yablo’s point about the usefulness of asserting false statements “with truth in them” (Yablo 2014: 3). Durkheim is looking past the expressed claim that a plant or animal has superhuman power to what is arguably an important truth about the power of society.15
More generally, the student of religion diverges from the positivists in looking to the distinctive uses to which the local discourse is put. She does not begin, and so end, by holding meaning hostage to locating an abstract god in the world of objects and events. Indeed, she does not assume that the talk is even about an abstract god—even in a Christian Eucharistic setting, even if the English word “god” appears in the subject position. In this she is shoulder to shoulder with deflationists of various stripes working out, for example, the meaning of value-talk in a world of objects and events, or of causal judgments in a world without necessary connection. Methodologically, the upshot seems to be a principled promiscuity: we are free to see some religious speech acts as disguised moral language, or perhaps in expressivist or functionalist or other terms.16 Promiscuous because, as contrasted with other such problematic areas of discourse as morals and modality, there seems no hope of giving religious discourse a univocal treatment.
6 A Concluding Application
Let us pull some of these threads together by looking at Peirce’s discussion of the Eucharist:
The Protestant churches generally hold that the elements of the sacrament are flesh and blood only in a tropical sense; they nourish our souls as meat and the juice of it would our bodies. But the Catholics maintain that they are literally just that; although they possess all the sensible qualities of wafer-cakes and diluted wine. But we can have no conception of wine except what may enter into a belief, either:
That this, that, or the other, is wine; or,
That wine possesses certain properties.
Such beliefs are nothing but self-notifications that we should, upon occasion, act in regard to such things as we believe to be wine according to the qualities which we believe wine to possess. The occasion of such action would be some sensible perception, the motive of it to produce some sensible result. Thus our action has exclusive reference to what affects the senses, our habit has the same bearing as our action, our belief the same as our habit, our conception the same as our belief; and we can consequently mean nothing by wine but what has certain effects, direct or indirect, upon our senses; and to talk of something as having all the sensible characters of wine, yet being in reality blood, is senseless jargon.… It is foolish for Catholics and Protestants to fancy themselves in disagreement about the elements of the sacrament, if they agree in regard to all their sensible effects, here or hereafter.1878 (1935): 257-8
This passage draws on several of our central themes. On Peirce’s broadly verificationist view, the reality represented by the concept “wine” is tied to what “sensible effects” it presents to sentient observers. Thus, the reference the observer makes to the underlying substance has a mentalistic or phenomenological component, and cannot be captured by his relationship to a sentence, true or false. Rather, just the reverse: the presentation of the object grounds what judgments he makes about it, true or false. When all goes well, or reasonably well, this directly referential component prevents the resulting concept from straying too far from the “sensible characters” in play—for example, one knows the meaning of “is wine” when one knows that a substance is wine if, and only if, under testing it reflects a certain chemical composition, or tastes of oak with traces of cement, etc. Certainly it will prevent the resulting concept from straying all the way to blood.
But suppose we turn our backs on Peirce’s verificationism (or any use-based approach to meaning) and ask about the nature of the object as it is in itself, that is, apart from any considerations having to do with sentient cognition. For example, on an Aristotelian view, meaning is fixed objectively by the essential nature of a thing rather than subjectively by what merely accidental properties may strike us in sensation (see Gardiner’s introduction, this volume, p. 324). When paired with nous, a faculty which allows us to pick out a substance’s essential properties, the Catholic view may get some traction. Thus, the Thomistic tradition holds that one knows the meaning of “is blood” when one knows that a substance is blood if, and only if, God deems it so. Might God deem a substance with the “sensible effects” of wine to be, in fact, blood? Since sensation no longer delivers the true nature of objects to us, the thought is no longer unintelligible that God might see its real, essential nature as something other than what is presented in experience. The suggestion, then, is that a truth-conditional approach to meaning—because it divorces truth from anything in the neighborhood of verification, justification, or proof—allows the relevant speech to be about flesh and blood. In which case Protestants and Catholics may, contra Peirce, disagree after all.
For those of us interested in the study of religion these two approaches to linguistic meaning bring different strengths to the table. As we have seen, the verificationist or use-based approach integrates naturally into a central mission of the humanities and social sciences: understanding. I have argued that we should see Davidson’s work in the theory of interpretation in this light. Applied to the case at hand, we may take it that both Protestant and Catholic parishioners have the right criteria for grouping objects under the concepts “flesh,” “blood,” “water,” and “wine” (the verificationist element), and that they retain this ability going forward (the pragmatic one). In inquiring into what comes next—to what use a judgment about water or wine is being put—we will tailor our assessment to concrete cases. At this point we have left the specifically semantical inquiry behind. Whether we should see this or that Protestant as using the judgment as a trope, or whether we should see this or that Catholic as believing in the real presence—these questions depend on a thick description of the facts on the ground. Perhaps this person says, “This is bread,” but intends it, with Peirce, “only in a tropical sense.” Perhaps another person is making word-like vocalizations, but, in fact, is just going through the motions in what only appears to be a language-game. Perhaps this is a Durkheimian invitation to go “beneath the symbol,” where the “reality it represents” has to do with social stability, or natural selection on the group level. At this point the entire realm of symbolic expression is open to the speaker, whereas, for the interpreter, the widest and deepest understanding may require “looking past” what is said.
But, as Peirce understands, a broadly verificationist semantics cannot portray orthodox Protestants and Catholics as in real disagreement about the objects on the communion plate. That requires taking seriously the thought that the real or essential nature of objects eludes evidence delivered through the senses. It requires a commitment to truth beyond justification and realism beyond verification. The question is whether we should take the associated absolute view of the world as a legitimate goal of our best metaphysics and epistemology, or whether we should see the appeal of realism in this absolute sense as continuous with the attraction of the kind of transcendence familiar to us in western religious traditions. To take the latter route would be to take seriously Dummett’s assessment of the truth-conditional model as “mythology,” to take the attempt to divorce meaning from experience as a phenomenon itself of interest to the student of religion. The student of religion will then want to retain both the verificationist or use-based and the truth-conditional approaches to linguistic meaning. The first she will rely on in making sense of the movements, including the vocalizations, of those she wishes to understand. The second she will take, in an ironical spirit, as continuous with her object of study.17
Brandom, Robert. (1994). Making It Explicit: Reasoning, Representing, and Discursive Commitment. Cambridge, Mass.: Harvard University Press.
Davidson, Donald. (1977 (1984)). Reality without reference. In Inquiries into Truth and Interpretation, 215-225. Oxford: Oxford University Press.
Davidson, Donald. (1986). A Nice derangement of epitaphs. In E. LePore (ed.), Truth and Interpretation: Perspectives on the Philosophy of Donald Davidson, 433-446. Oxford: Blackwell.
Davidson, Donald. (1999). “Interpretation: hard in theory, easy in practice.” In M. De Caro (ed.), Interpretations and Causes; New Perspectives on Donald Davidson’s Philosophy, 31-44. Dordrecht: Kluwer Academic Publishers.
Dummett, Michael. (2007). Reply to John Campbell. In R.E. Auxier and L.E. Hahn (eds.), The Philosophy of Michael Dummett, 301-314. Chicago: Open Court.
Frankenberry, N.K. (2005). “Sleepwalking through history”: reflections on democracy, religion, and pragmatism in a time of terror. American Journal of Theology & Philosophy 26 (1/2): 5-59.
Godlove, Terry F. (1989). Religion, Interpretation, and Diversity of Belief: The Framework Model from Kant to Durkheim to Davidson. Cambridge: Cambridge University Press.
Godlove, Terry F. (2002). Saving belief: on the new materialism in religious studies.” In N.K. Frankenberry (ed.), Radical Interpretation in Religion, 10-24. Cambridge: Cambridge University Press.
Godlove, Terry F. (2005). Teaching the critics: one route through The Elementary Forms. In T.F. Godlove (ed.), Teaching Durkheim, 107-130. Oxford: Oxford University Press.
Godlove, Terry F. (2011). Hanna, Kantian non-conceptualism, and Benacerraf’s dilemma. International Journal of Philosophical Studies 19 (3): 447-464.
Godlove, Terry F. (2016). Subjectivity, enchantment, and truth: Frankenberry among the Puritans. American Journal of Theology and Philosophy 37 (1): 21-35.
Lawson, E. Thomas, and Robert N. McCauley. (1990). Rethinking Religion: Connecting Cognition and Culture. Cambridge: Cambridge University Press.
Levy, Gabriel. (2012). False but significant: the development of falsity in religious cognition in light of the holism of the mental. Method & Theory in the Study of Religion 24 (2): 143-165.
Peirce, Charles S. (1878 (1935)). How to make our ideas clear. In C. Hartshorne and P. Weiss (eds.), Collected Papers of Charles Sanders Peirce, Volume V, 248-271. Cambridge MA: Harvard University Press.
Peirce, Charles S. (1931-1958). The Collected Papers of Charles Sanders Peirce, ed by P. Weiss (Vols. 1-6) and A. Burks (Vols. 7-8). Cambridge MA: Harvard University Press.
Penner, Hans H. (1995). Why does semantics matter to the study of religion? Method & Theory in the Study of Religion 7 (3): 221-249.
Prawitz, Dag. (2007). Pragmatist and verificationist theories of meaning. In R.E. Auxier and L.E. Hahn (eds.), The Philosophy of Michael Dummett, 455-481. Chicago, Ill.: Open Court.
Rorty, Richard. (1997). Religious faith, intellectual responsibility, and romance. In R.A. Putnam (ed.), The Cambridge Companion to William James, 84-102. Cambridge: Cambridge University Press.
Staal, Frits. (1990). Rules without Meaning: Ritual, Mantras, and the Human Sciences, Toronto Studies in Religion. New York: P. Lang.
Tarski, Alfred. (1933 (1983)). The concept of truth in formalized languages. In J. Corcoran (ed.) Logic, Semantics, Metamathematics: Papers from 1923 to 1938, 152-278. Indianapolis: Hackett Publishing Company.
Williams, Michael. (2010). Pragmatism, minimalism, expressivism. International Journal of Philosophical Studies 18 (3): 317-330.
For the first, E. Thomas Lawson and Robert N. McCauley, Rethinking Religion: Connecting Cognition and Culture (1990); for the second. Frits Staal, Rules without Meaning: Ritual, Mantras and the Human Sciences (1990).
I intend none one of this to run afoul of Gardiner’s plea for semantic continuity across all forms of discourse (this volume, p. 411), a plea which I am glad to echo. The present thought is not that religious discourse requires a somehow non-standard or custom-built semantics.
Elsewhere, Peirce writes that such an individual “must be one of which every predicate may be universally affirmed or denied. For, let A be such a term. Then, if it is neither true that all A is X nor that no A is X, it must be true that some A is X and some A is not X, which is contrary to its nature as a logical atom. Such a term can be realized neither in thought nor in sense…. A logical atom, then, like a point in space, would involve for its precise determination an endless process. We can only say, in a general way, that a term, however determinate, may be made more determinate still, but not that it can be made absolutely determinate” (1931-58: vol 3: 93). For discussion, see (Thompson 1978: 78).
“I have no objection if ‘Intentional Realism’ entails no more than that there are correct and incorrect interpretations of what a speaker means by his words … One can, in my opinion, be a realist in this sense and consistently hold that the primary and only ultimate source of meaning lies in successful interpersonal communication.” (1994: 126).
But, as for Dummett, the overall picture is emphatically not behavioristic. For example, behavioral evidence may point to implicit deductive or inductive reasoning.
Thus, Brandom writes of “the commitment undertaken by one who employs a given content to the propriety of the transition from the circumstances of appropriate application to the appropriate consequences of application of a conceptual content” (1994: 120).
Since Quine famously criticizes Peirce on holistic grounds, and since I am endorsing Davidson’s closely related holism, my appeal to Peirce may appear self-defeating. But, for Peirce, verificational meaning is just a stop on the way to the highest grade of clearness of thought. As Thompson puts it, “the ultimate meaning of ‘is hard’ lies in a network of true generalizations and not in the verificational meaning by virtue of which it qualifies for a place in the network” (1978: 74). These issues are outside the scope of this paper.
Hence, Davidson’s (1977 (1984)). I discuss the attempt undercut awareness in Quine, Davidson, and Rorty in (Godlove 2014), esp. chapters 4 and 5.