Journal of Analytic Theology, Vol. 1, No. 1, May 2013
10.12978/jat.2014-1.091413100417a
©2014 John A. Keller • © 2014 Journal of Analytic Theology
Theological Anti-Realism
John A. Keller
Niagara University
Man is the measure of all things.
Protagoras
1. Introduction
We may provisionally think of realism in a domain as the view that truths in that domain are mind-independent: that the existence and nature of the phenomena in that domain are independent (except in trivial ways) of human mental activity. One particularly important consequence of this is that what can be known and what is true are not guaranteed to coincide: whatever Protagoras may have thought, metaphysics and epistemology are importantly distinct.
There is a strong presupposition in favor of realism within the “analytic” tradition.[1] While there are analytic anti-realists,[2] in papers not addressing the question of realism an analytic philosopher may take realism for granted without significantly limiting her audience. Interestingly, this is true of all major species of realism: scientific realism, realism about the mental, realism about truth, moral realism, etc. Analytic anti-realists are typically famous for being anti-realists—they have, in most cases, spent their careers defending their anti-realism. To be an analytic anti-realist is to be a special kind of analytic philosopher.
Things are otherwise within the “continental” traditions, where some important form of anti-realism is often taken for granted. A continental philosopher may presuppose significant forms of anti-realism—anti-realism about truth or ethics, for example—without significantly limiting her audience. This is not to say that there are no continental realists. But continental realists typically spend their careers defending their realism. They are famous for being realists. To be a continental realist is to be a special kind of continental philosopher.[3]
Not all mainstream theologians are equally influenced by continental philosophy, of course, but continental philosophy has had a considerable influence on mainstream theology as a whole. It is not, then, surprising that continental presuppositions about realism have made themselves felt in contemporary theologizing. Much theology is performed against an anti-realist backdrop, and many theologians approach theology itself from an anti-realist perspective. My goal in this essay is to evaluate some of the most important arguments for, and objections to, theological anti-realism. If what I say here is correct, theological anti-realism is on much weaker footing than it is often taken to be.
2. Realism and Anti-Realism
The words ‘realism’ and ‘anti-realism’ are used and abused in a host of different ways. A common thread linking most or all of these uses is the idea that to be a realist about Xs is to think that the existence and nature of Xs is mind-independent. For the purposes of this paper, I am going to focus on what it means to treat a certain domain of discourse (such as theological discourse) realistically. It is widely (although not universally) agreed that there are three elements of realism about a domain of discourse d:[4]
Objectivity: Truths about d are mind-independent: they do not depend in any non-trivial way on human mental activity (our beliefs, desires, hopes, etc.).
Transparency: Grammatical names and predicates in d are genuine names and predicates.
Truth-Aptness: Assertions in d are truth-apt: they aim at truth (at least when sincere) and can be evaluated for truth and falsity (unlike, say, discourse in fictional or theatrical domains).
I should immediately note that these theses are separable. Indeed, realism is often identified with Objectivity (i.e., mind-independence) alone, and one might think that Transparency in particular saddles realists with unnecessary baggage. While Transparency may need to be revised (see below), I think that an adequate account of realism must incorporate something in the neighborhood. For consider a theologian who, rejecting Transparency, thinks that all ‘God exists’ really means is that human beings are fundamentally good. Such an individual might hold that ‘God exists’ is objectively true—after all, it’s (perhaps) objectively true that human beings are fundamentally good. She could, then, accept both Objectivity and Truth-Aptness as applied to ‘God exists’. But clearly she isn’t a realist about the existence of God. We need Transparency, or something like it, to rule out such cases.
Let us turn, now, to a more careful discussion of these three tenets of realism.
2.1 Objectivity
The notion of mind-independence has traditionally been at the center of debates about realism and anti-realism.[5] Unfortunately, participants in those debates have not always been clear about what it means to say that something is mind-independent. It is natural to think that saying that Xs are mind-independent is to say that Xs could exist and have the properties they do (including truth or falsity) even if there were no minds. But this is not an adequate understanding of mind-independence. Most realists are realists about minds (and mental discourse more generally), but it does not make sense to say that ‘there are minds’ is mind-independent, or that ‘there are minds’ would be true if there were no minds. The relevant notion of independence is mental activity independence: most importantly, independence from what we believe, perceive, and want. Realists hold that minds exist, and would exist even if no one believed, perceived, or desired that they did. But obviously minds would not exist if there were no minds.
There are other trivial ways in which aspects of the world can be mind-dependent. If there were no artisans, there would be no artifacts. This is not the sort of “dependence on human mental activity” at issue in debates about realism. What is at issue in these debates is whether truths about things, or things themselves, are in some way constituted by (as opposed to caused by) human mental activity.
2.2 Transparency
A second important element of the traditional conception of realism about a domain of discourse d is the idea that language in d is transparent, in the sense articulated above: that grammatical names and predicates are genuine names and predicates, where a “genuine” name or predicate is a name or predicate that functions semantically like names and predicates outside of d. The motivating idea behind Transparency is more general: it is that the language in d should not receive a special “funny” semantics. In other words, realists accept:
Semantic Uniformity: Language does not behave in a funny way in d—the same (kind of) semantics applies to language inside and outside of d.[6]
If Semantic Uniformity is true, and discourse within d receives the same kind of semantic analysis as discourse outside of d, Transparency follows. Anti-realists about a domain d who deny Semantic Uniformity are what John Burgess and Gideon Rosen (2005) call content-hermeneuticists: they think that the sentences in d have a special “funny” semantics, different from the semantics of ordinary discourse.
Contemporary nominalists are typically content-hermeneuticists, holding that discourse about universals should not be taken at face value. They maintain that grammatical referring expressions like ‘justice’ do not really have the semantic function of referring to a universal. Instead, they hold that the claim expressed by, e.g., ‘justice is a virtue’, is more perspicuously expressed by some other sentence, such as ‘all just acts are virtuous’.[7]
2.2.1 Transparency in Ordinary Language
Note that in order for Transparency to follow from Semantic Uniformity, it must be the case that ordinary language is transparent. This is a substantive, and I think false, assumption.[8] This significantly complicates debates about realism and anti-realism.
If ordinary language were transparent, grammatical referring expressions such as ‘John’s lap’, ‘Maggie’s smile’, and ‘David’s shadow’ would have to function semantically as referring expressions. According to semantic orthodoxy, this would entail that their meaning depends on their referent. But these expressions don’t have referents: nothing comes into existence when John sits, Maggie smiles, or David stands in the sun, and it is impossible to give a coherent or plausible account of the properties that John’s lap, Maggie’s smile, or David’s shadow would have. And so we must either give these expressions a “funny” semantic treatment or else take all (or most) sentences containing them to express falsehoods, or nothing at all.
Furthermore, if ordinary language were transparent, we could regiment (formalize) sentences of ordinary language simply by following the heuristic rules found in any decent logic text. It is clear, however, that expressing all but the most elementary statements in the idiom of logic is a non-trivial task. There has been significant disagreement, for example, about how to so translate ‘Pegasus does not exist’, ‘It is raining’, and ‘Lois believes that Superman can fly’.[9] And this is only the tip of the iceberg. So Transparency seems to be false of ordinary discourse. However, the motivating idea behind Transparency—Semantic Uniformity—remains intact. Accordingly, from here on I will focus on Semantic Uniformity.
The failure of Transparency has important implications for certain debates about realism. Realists about time, for example, cannot be expected to take our temporal discourse at face value, i.e., to interpret temporal locutions transparently, since (independently of one’s views about time) it is implausible that our temporal locutions should be treated transparently.[10]
With that being said, I am not sure how much slack the retreat to Semantic Uniformity buys theological realists. While no one thinks that the Bible can be read or interpreted like a philosophical text, theological statements—works of academic theology, creedal statements, etc.—are a different matter. Of course they are not written to the standards of late-20th Century analytic philosophy. But they are much more carefully composed than typical ordinary speech and writing. While we cannot simply “read off” the logical forms of theological statements, neither can we “read off” the logical forms of the claims made by Leibniz or Aristotle. We are, nonetheless, generally able to get a pretty good idea of the logical structure of their theories. Likewise with many theological statements.
2.2.2 Semantic Uniformity and Traditional Theology
There are a couple of lingering issues related to Semantic Uniformity. First, what is the connection between the Thomistic Doctrine of Analogy and Semantic Uniformity? If the Doctrine of Analogy entails that discourse about God is non-literal, or in some other way anomalous, then it violates Semantic Uniformity. Thomists, however, appear to be paradigmatic theological realists.
One way to reconcile the two theses is to hold that it is the ordinary uses of language that are the analogous (anomalous) ones—God is literally and straightforwardly good, while we are only good in a partial or analogous sense. On this understanding, adherents of the Doctrine of Analogy would be realists about theological discourse and anti-realists about ordinary discourse.[11] Realism about both domains might be compatible with accepting the Doctrine of Analogy if the analogous uses of language (whether they are theological or ordinary) are “dead analogies”.[12] What begin as non-literal uses of language can become literal as expressions become conventionally associated with those uses. (At this point in time, ‘legs’ literally refers to the supports for chairs and tables, but that began as a metaphorical use.) Terms with two such conventional meanings would still be “analogous” (as opposed to equivocal) in the sense that their meanings are significantly and relevantly related. And indeed, the traditional understanding of the Doctrine of Analogy allows that analogous terms can be literally applied to God.[13] So there is less conflict between Semantic Uniformity and the Doctrine of Analogy than one might have thought.
Some Thomists have a more apophatic interpretation of the Doctrine of Analogy, however, and many theologians are attracted to apophaticism in and of itself. While Semantic Uniformity plausibly entails that apophatic theologians are not theological realists, it is unclear that this runs afoul of a desideratum on an account of theological realism. And there is even hope for apophatic theologians who fancy themselves realists: Jacobs (forthcoming) argues that we may understand apophaticism as the doctrine that our discourse about God does not express fundamental truths—that all theological truths are derivative.[14] Such a view does not run afoul of Semantic Uniformity since most ordinary truths are derivative, and so derivative theological truths may well stand in the very same sort of messy relation to fundamental reality in which derivative non-theological truths stand.
What about the connection between Semantic Uniformity and the Doctrine of Divine Simplicity? Again, there is a superficial incompatibility: if God does not have properties (in the way that other substances do), then predicating, say, personhood of God must involve something very different that predicating it of me. But it is not clear that we should count adherents of the Doctrine of Divine Simplicity as anti-realists either. The two doctrines can perhaps be reconciled by denying that ordinary predications require there to be an object/property pair such that the object is the subject, the property is expressed by the predicate, and the subject instantiates the property. Such a move would not be as ad hoc as it might first appear, once we notice that we can make grammatical predications of things that do not exist (‘Santa Claus is fat’, ‘that shadow is moving quickly’) and we can use grammatical predicates (or predicative expressions) that do not express properties. For example, pace Descartes, ‘can be conceived not to exist’ does not express a property that my body but not I possess, and pace Newton, ‘is moving’, while a grammatical predicate, must ultimately be analyzed as a relation.
Finally, does Semantic Uniformity embroil metaphysical realists in some sort of pernicious “onto-theology”? Many theologians think that realism about theology amounts to a form of idolatry: that saying, e.g., that ‘exists’ applies to God commits one to thinking that God is an “empirical object”, in the world and hence non-transcendent.[15] While there are issues worth discussing here, as stated this charge is without merit. Within the analytic tradition at least, one may happily apply ‘exists’ to non-empirical objects such as numbers and properties. To say that God exists is just to say that God is there—that God is. We can apply ‘exist’ to both God and Adam without presupposing that there is not a vast difference between them, just as we can apply ‘exist’ to the number two and Adam without presupposing that there is not a vast difference between them.
Of course, much more could be said about these reconciliation projects—and about whether these examples suggest that Semantic Uniformity shouldn’t be considered a tenet of realism after all. Limitations of space, however, prevent me from saying anything more here.
2.3 Truth-Aptness and Aiming at Truth
If Truth-Aptness for a domain d was just a matter of the truth-aptness of the sentences uttered in d, then Truth-Aptness would follow from Semantic Uniformity, at least on the assumption that ordinary language has a truth-conditional semantics. But Truth-Aptness says more than this: it is also a claim about the cognitive attitudes we bear towards the sentences we utter. It is widely held that in some contexts—perhaps including fictional writing and play-acting—we utter sentences with truth-conditions without thereby expressing belief in their truth. Some anti-realists—what Burgess and Rosen call attitude-hermeneutic anti-realists—claim that this phenomenon is more widespread that one might have thought. Van Fraassen, for example, thinks that scientific discourse aims at empirical adequacy, not truth.[16] He doesn’t claim that scientific discourse has a funny “empirical adequacy-conditional” semantics, however. Rather, he claims that typical scientific utterances of a sentence s are not intended to express belief in s, but merely to express one’s commitment to the empirical adequacy of s. The truth of s is, on this view, beside the point: in scientific contexts, the norm of assertion is empirical adequacy. It is this sort of maneuver that Truth-Aptness is supposed to rule out. To be a realist, you have to believe what you say.
2.3.1 Believing In
Before moving on I should address a source of confusion concerning debates about theological realism that arises from an ambiguity in our talk of “believing in God”.[17] James 2:19 says that “even the demons believe—and tremble”. But the faith of demons remains lacking: they do not “believe in God” in the relevant sense.
There are, then, at least two importantly different senses of ‘believe in’: an “existential” sense and a “pro-attitude” sense. On the pro-attitude sense of “believing in x”, one believes in x just in case one has a pro-attitude towards x. This is the sense in which I “believe in” honest politicians. I admit that there are not any honest politicians, of course—I am not that naive! But I am for them: I believe in them in the pro-attitude but not the existential sense.
Just the opposite is true in the case of the demons and God: the demons believe in God in the existential but not the pro-attitude sense. Traditional theism (realistically construed) involves “believing in God” in both senses. Theological anti-realists sometimes stress the relative importance of believing in God in the pro-attitude sense, given that even demons believe in God in the existential sense. This emphasis is correct—it is better to have a pro-attitude towards God and doubt She exists than to be certain She exists and doubt Her goodness.[18] But this does not show that the existential belief is unimportant. For there are a variety of different pro-attitudes that one can have, and we can have only the shallowest of them if our pro-attitude is not accompanied by an existential belief. I am “for” honest politicians, for example, but I cannot trust them, since there aren’t any. And a proper relationship with God involves much more than being “for” such a being. After all, many professed atheists would be happy to admit that they have, in the abstract, a pro-attitude towards perfect beings: they just do not think there are any. Believing (or accepting) that God exists is not everything. But it is an essential part of traditional theism.[19]
2.4 Realism, Anti-Realism, and Skepticism
A significant complication in discussions of realism and anti-realism is that whether one counts a view as “realist about Fs” often depends on one’s substantive views about the nature of Fs. Some anti-realists are proud of having that status, whilst others insist that their view is a form of “enlightened realism”.[20] Many of the things about which people are anti-realists—ethics, reason, and truth, for example—are things that “no one would admit to being against, [since] that’s like being against Motherhood and Apple Pie”, as Alan Sokol colorfully puts it.[21] When it comes to other cases, such as creatures of fiction and state governments, anti-realism is much more attractive.
In any case, views that one might regard as forms of theological anti-realism might be thought of, by their defenders, as simply being the proper (enlightened but realistic) understanding of theology. As Thomas Nagel puts it:
The usual set of moves among realism, skepticism, and reductionism occurs here as everywhere in philosophy: Reductionism. . . seems to offer a refuge from skepticism if realism. . . seems too hard to sustain. Being a realist. . . I regard these reductive “rescues” as equivalent to skepticism; that is, they are forms of skepticism about the reality of what I myself take [X] to be. Their proponents would describe them differently—as denials that my understanding of the nature of [X] is correct.[22]
Or, as Alexander Bird says:
[O]ne way of avoiding skepticism about Xs is to re-construe what Xs are or what ‘X’ refers to…Since doubt and even disbelief are clearly reasonable and perhaps well-grounded options in religion, even without recourse to sophisticated philosophical arguments, the claims of skepticism are especially pressing in the case of religion. Consequently there is a greater pressure in theology to…evade the pull of agnosticism or atheism by resorting to metaphysical or semantic antirealism.[23]
Instances of this dialectic are not hard to find. Scott Shalkowski, for example, writes that:
Critics seem to assume that revising one’s account of God is having one’s cake and eating it too. In this paper I will argue that too little attention has been given to the dialectical option of relinquishing traditional theological doctrines as part of the process of finding a more suitable theological package…typical atheological strategies, such as attempts to refute Christian theism via paradox or conflict with our experience of evil…have limited value in the light of a proper, looser, commitment to traditional theology.[24]
Note that Nagel seems to be assuming above that ‘reductionism’ and ‘anti-realism’ are equivalent: that reductionism and skepticism are the two alternatives to realism. (Pettit (1991) presents the dialectic similarly.) This is contentious, but there is some pull to the idea that ‘realism’ and ‘anti-realism’ are mere contrast terms: realists about a domain hold the intuitive or pre-theoretic view about objects or truths in that domain, while anti-realists hold some sort of deflationary, reductionistic, or eliminativistic view. On this way of thinking about realism, one can be a realist about nation states while admitting that they are mind-dependent, since that is the pre-theoretic view. Anti-realism about states would then only apply to views that are (to some substantial degree) more non-objective than that.
This seems like a merely terminological question: there don’t appear to be principled reasons for deciding between calling philosophers with the “intuitive” view of states (as mind-dependent) ‘realists’ or ‘anti-realists’.[25] One reason to define ‘realism’ as I have, such that anti-realism about states is trivially true, is that it minimizes the worry that appears at the beginning of this section. If ‘anti-realism’ were defined such that it required one to think that a domain was less objective than is intuitively (or pre-theoretically) plausible, simple classifications of positions would be held hostage to disagreement about which position is pre-theoretically correct.
In any case, if realism involves a “face value” interpretation of our theories, it will be incompatible with many forms of reductionism, according to which, say, a sentence which appears to predicate F of a (the room is hot) is really predicating G of b (the air molecules in the room are moving especially rapidly).[26] Of course, as Nagel notes, not all of those who reject realism become anti-realists, what he calls reductionists. Some happily (or reluctantly) embrace skepticism: in the case of religion, atheism.
3. Metaphysical Anti-Realism
One obvious way of being a theological anti-realist is to be an anti-realist tout court—to be a metaphysical anti-realist. Many theological anti-realists are anti-realists in full metaphysical generality: Don Cupitt, for example, rejects a “realistic ontology, the notion that there is something out there prior to and independent of our language and theories, and against which they can be checked.”[27] As Joseph Runzo puts it, Cupitt holds that “reality has now become a mere bunch of disparate and changing interpretations.”[28]
In this section I would like to look at some influential arguments for metaphysical anti-realism. I am going to ignore some of the most discussed arguments: Kant’s Transcendental Deduction, Putnam’s Model-theoretic Argument, and Dummett’s Manifestation and Acquisition arguments.[29] The literature on these arguments has grown so large that even a cursory summary would require a significant amount of space. Furthermore, the range of responses to these arguments is now well known, and it is generally agreed that the arguments are far from compelling, at least for those without other commitments that predispose them towards anti-realism. There is much less of a literature on the arguments discussed below, however, and so it is worth taking the time to point out their weaknesses.
3.1 Realism, Anti-Realism, and the Burden of Proof
Before we look at the arguments for anti-realism, I want to say something about the dialectical situation, and in particular about which side has the “burden of proof”. When it comes to assigning the burden of proof, I think the proper view is disappointingly simple: the side trying to prove something has the burden of proof.[30] With that being said, the dialectical situation facing realists and anti-realists is quite different. There are three considerations that together constitute a powerful prima facie case for realism.
3.1.1 The Argument from Appearances
The first consideration is simply that realism seems to be true. For example, most people think there is a fact about whether Caesar’s heart skipped a beat when he crossed the Rubicon, independently of our ability to know (or constitute) that fact.
While this is not a very powerful consideration in favor of realism, it shifts the burden to those who maintain that things are not as they seem, since it is irrational to believe that things are not as they seem without special reason. For example, it seems to me that it is cold in here and that Gettier cases are not knowledge, and if I have no reason to doubt these appearances, it is irrational for me to deny them. For similar although less powerful reasons, it would be irrational for me to be agnostic about the truth of those claims. Note that this does not presuppose “phenomenal conservatism”, the view that seemings provide justification, or constitute evidence. The Argument from Appearances merely maintains that it is irrational to ignore one’s seemings (without reason). Finding a good fortune in a cookie gives me no reason, justification, or evidence to think that I will soon be rich. Nonetheless, if I believe that fortune cookies are reliable predictors, and I found such a cookie, it would be irrational to continue fretting about money. It can be irrational to ignore things that are not reasons, if it seems to one that they are.[31]
3.1.2 The Argument from Science and History
The second consideration in favor of realism is that anti-realism is (apparently) inconsistent with the deliverances of science and history, since they tell us that the universe is much older than the human species. Long before humans came on the scene, it was true that 2+2=4, that massive bodies are subject to gravitational acceleration, and that the Earth is the third planet from the Sun. But if that is the case, the truth cannot be linked to human mental activity in any significant way.
This is not a knockdown argument, but it is an obvious argument: one that anyone would think of who was considering whether to accept realism or anti-realism. And so it is an argument to which anti-realists must respond. Indeed, its obviousness is so clear that it is, in a sense, included in the “baseline”: a response to it must be included in any plausible or interesting defense of anti-realism.
Note that an adequate response to this argument must explain away the tension between anti-realism and the above-mentioned facts: no argument for anti-realism could possibly trump, say, the argument that is was true long before there were any humans that massive bodies attract. The premises in this argument are going to be more plausible than the premises of any argument for anti-realism, given (the exceedingly plausibly assumption) that anti-realism cannot be proven from apodictic premises.[32]
3.1.3 The Argument from Linguistics
The third consideration is less obvious than the first two, but is potentially much more powerful. Realism is a presupposition of most contemporary work in logic, linguistics, and semantics, and a strong presumption in favor of realism arises from the fact that we have a relatively well worked out truth-conditional semantics for large portions of language. If anti-realists wish to make their view plausible, they will have to sketch out alternative theories of language, truth, and logic. But anti-realists have conspicuously failed to perform this task. Heyting gave an intuitionistic (anti-realist) account of the semantics of mathematical language, but—putting worries about this semantics to the side—there has been almost no work extending this (or any alternative) anti-realist semantics to an interesting fragment of natural language. As Timothy Williamson says,
The obvious and crucial challenge [for anti-Realists] was…as a first step, to develop a working assertability-conditional semantics for a toy model of some small fragment of empirical language. But that challenge was shirked. Anti-realists preferred to polish their formulations of the grand programme rather than getting down to the hard and perhaps disappointing task of trying to carry it out in practice…Anti-realists have simply failed to develop natural language semantics in that form, or even to provide serious evidence that they could so develop it if they wanted to. They proceeded as if Imre Lakatos had never developed the concept of a degenerating research programme. (Williamson (2007), p.282-4)
Since Williamson wrote that passage, this failure has been partially remedied by Mark Schroeder (2008), who develops a content-hermeneutic semantics for natural language. But Schroeder’s work ultimately seems to support the realist’s position—after making concession after concession, Schroeder concludes that, while he will “stop short of announcing that [he’s] constructed a reductio of [this form of anti-realism]” (p.177), he does “hope to have assembled significant cause to believe that [it] is false.” (p.179)
Metaphysical anti-realism, then, cannot plausibly be held to be the default or pre-theoretic position. Hence, whether we should accept the view will depend on the arguments that can be marshaled in its favor. It is to those arguments that we now turn.
3.2 The Argument from Non-Checkability
Recall Cupitt’s comment from the beginning of this section. Cupitt’s concern with realism seems to be with the idea that there is something “out there” that we can check our theories against. Realists do hold that there is something out there, prior to and independent of our language and theories. After all, the universe is 13.77 billion years old, and humans have only existed for the barest fraction of that time. Any theory that denies this is refuted by contemporary cosmology. (C.f. §3.1.2) So it is natural to think that it is the “checkability” claim that is really bothering Cupitt. But are realists really committed to this claim? And if they are, is that a problem?
First, note that only non-skeptical realists hold that we can check our theories against an independently existing reality. Realism is compatible with skepticism, and indeed, most skeptics are realists. Should most or all realists be skeptics? The claim that we can check our theories against reality may sound hubristic or implausible, but it contains a crucial ambiguity: what do we mean by “check”? On the one had, of course we can check our theories against reality: that’s what experiments and other observations do. (I think it’s raining out, but I want to make sure. I look out the window and check.) In another well-known sense, however, we cannot “check” our theories—we cannot get outside of our minds to somehow compare our representations with reality as it “is in itself”. But the realist is not committed to this kind of ability to check our theories. Indeed, typical realists will deny that the demand that we check our theories in this sense is even coherent. Rather, realists are committed to the much more banal thesis that many of our theories have empirical consequences which can be confirmed or disconfirmed. Of course not all of our theories have empirical consequences. Many philosophical, theological, mathematical, political, and logical claims have no empirical consequences of which we are aware, and hence cannot be directly checked against reality in any intuitive sense. But this claim about uncheckability is a paradigmatically realist claim: realism is sometimes defined as being nothing more than a metaphysical conception of truth.[33] Realists hold that there are truths that cannot be checked (or verified)—that truth is radically non-epistemic. It is the anti-realist who (typically) wants to construe truth as epistemic, to say there cannot be truths that are unknowable. So it is difficult to see how doubts about checkability could be used to argue against realism.[34]
3.3 Social Construction Arguments
The next argument for anti-realism I would like to discuss is
The Social Construction Argument: For any sentence s, the truth of s will depend on what the words in s mean. But the meanings of words are socially constructed—they are the result of our choices and behavior. Hence, for any sentence s, whether s is true depends on human choices and behavior.
For example, ‘There is more than one thing’ expresses a truth. (I assume here and throughout that Monism is false.) But we—collectively and over time at least—are language’s masters: meaning is “socially constructed”. So we could alter the meaning of the word ‘thing’ so that it refers to unicorns. But in so doing we will alter the truth-value of the sentence ‘There is more than one thing’ (I assume here and throughout that there are no unicorns). Hence, human activity determines the truth of that sentence. Over the long run, and collectively, we determine what is true and what is false. So the social constructedness of meaning entails the social constructedness of truth.
To see what’s wrong with this argument, consider the sentences ‘There is more than one thing’ and ‘Es gibt mehr als ein Ding’. These sentences are synonymous: they express the same claim. The claim they express is not itself a linguistic item; it was true long before there were languages. And it is the truth of this claim that we care about when we ask whether truth is socially constructed, or dependent on human activity in some other way. Consider the diagram below.
The claim expressed by the two sentences is represented in the upper middle by a formula of first order logic, reminding us that claims are not linguistic. There are three different kinds of arrow on the diagram. The horizontal arrows, signifying the expressed by relation, represent facts that are socially constructed, as linguistic meaning is a social construct. Because of this the diagonal arrows, signifying what the derivatively makes true relation, also signify socially constructed facts. The derivatively makes true relation holds in virtue of the (non-derivative) makes true relation (symbolized by the vertical arrow) and the expressed by relation. But since the expressed by relation is socially constructed, it follows that the derivatively makes true relation is socially constructed as well: whether it holds between the world and a sentence depends on human mental activity.
But does this show that “truth is socially constructed”? Not in any interesting sense—and certainly not in any sense in which a realist would deny. The whole social construction debate is about the relation represented by the vertical arrow: the non-derivative relation that holds between (some) claims and the world. Of course the relation symbolized by the diagonal arrows is socially constructed: it is the product of the relation symbolized by the horizontal arrows and the relation symbolized by the vertical arrow, and the relation symbolized by the horizontal arrows is socially constructed. No one denies this.
3.3.1 Epistemological Social Constructivism
A related version of social constructivism has been inspired by the sociology of science. Many sociologists of science are anti-realists, largely because they think that the acceptance of scientific theories is primarily explained by social, as opposed to truth-related, reasons. In effect, this is a version of the genealogical argument discussed in §4.1. The problem with this argument is that the existence of social reasons for accepting a theory is fully compatible with having justification for thinking that the theory is true. Society is ubiquitous, and very little could be accomplished outside of its sphere of influence. But it doesn’t follow that the truth or justification of our theories is thereby cast into doubt. If one wants to argue that we are not justified in accepting a theory t, or that t is not objectively true, one must get down to the business of providing arguments for these conclusions. Social constructivists put forward such arguments with surprising rarity.
4. Theological Anti-Realism
Even if the arguments for metaphysical anti-realism are less than compelling, this does not undermine localized anti-realism, at least in certain domains: even the most diehard metaphysical realists are often anti-realists about something. In this section I evaluate some arguments for localized theological anti-realism. This is done from the perspective of someone sympathetic with theism, as most theological anti-realists seem to be. Since going anti-realist about a domain d typically results in a lowering of the bar for accepting the claims in d, those sympathetic with theism will be more inclined towards theological anti-realism than those sympathetic with atheism.
4.1 The Political Argument[35]
One very influential argument for theological anti-realism is:
The Political Argument:
(1) We should decide what to believe (or accept) on the basis of political and other practical considerations: specifically, considerations that support the liberation of the oppressed.
(2) The liberation of the oppressed is best supported by theological anti-realism.[36]
(3) Hence, we should be theological anti-realists.
4.1.1Evaluating the Political Argument
A full response to this argument would require an essay unto itself, but let me briefly indicate what I take to be its main weaknesses. Given what was said above about metaphysical anti-realism, we have reason to doubt the argument’s first premise. But there are serious worries about the second premise as well. The most obvious and important problem is that theological realism is a view about the status, not the content, of theology. Since whether theology is liberating depends on its content, it is unclear how anti-realism is supposed to help. A second problem is that anti-realist theology seems less politically potent than realist theology, since people are less inclined to fight, suffer, and perhaps die, for causes that they do not really “believe in”. (Although the fervor with which sports fans support their teams casts some doubt upon this claim.)
One response to the first problem would be to hold that anti-realism makes theology more liberating by making “revolutionary” conceptions of theology possible, maintaining that we will be stuck with “traditional” theology unless we go anti-realist.[37] There are three problems with this response. First, it doesn’t appear to be true: as indicated in the previous footnote, all manner of revolutionary theologizing has been conducted in the realist mode. Second, if it were true, it would appear to be an indictment of revolutionary theology—why is revolutionary theology in particular not truth-apt? Finally, even if going anti-realist were the only way to get a revolutionary theology up and running, it is unclear that this strategy would be politically effective. If we compare the realist/traditional theology package with the anti-realist/revolutionary theology package, it is far from clear that the latter package is more politically effective than the former.
Consider a parallel case from US politics. Assume that the platform of the Democratic Party is more liberating than the platform of the Republican Party, but that that platform of the Green Party is much more liberating than both. Does that mean that the Green Party actually frees more people from oppression than the Democratic Party, and that someone whose sole goal was freeing the oppressed should throw her support behind the Green Party? Not obviously, for the Democratic Party, imperfect as it is, is much more influential than the Green party. Even though, e.g., Nader was much more liberal than Gore, supporting Nader was counterproductive for liberals.
Just as the Democratic Party, flawed thought it may be, still offers significant support for the poor and oppressed, so does traditional theology, flawed though it may be. Of course, traditional theology can be, and often has been, twisted to the benefit of the powerful. (As, of course, has the Democratic Party.) But it is clear that such twisting is a deformation of traditional theism: attempts to theologically justify oppression face devastating criticisms from within theology itself. So it is not as if traditional theology offers nothing to the oppressed. Theology could perhaps be made more liberating in theory by accepting the anti-realist/revolutionary package, but it is unclear that it would be made more liberating in practice. The first reason for this is the same as in our political analogy: traditional theism has much more influence than its more politically radical offshoots. The second reason is that, once we go anti-realist, and attempt to reshape theism for political reasons, who is to say what form of “theism” will gain influence, and with whose interests that form of theism will align? The righteous will advocate forms of theism congenial to the causes of the oppressed. But other people—more powerful people—will presumably advocate forms of theism congenial to cementing the existing oppressive order. Why would an advocate for the oppressed want to play this game? Since the oppressed are less powerful than their oppressors, there is good reason to think that any significant modification of theism will hurt rather than help their cause: the powerful will have more influence on the creation of new forms of theism than the oppressed. Traditional theism has the power of tradition, if nothing else: it can use the power of tradition to counter the power of oppressors. Newly created forms of theism lack such traditional authority, and so it is likely that they will be disregarded by those they threaten. But since they threaten the established order, such forms of theism seem destined to be marginalized. If you are oppressed, the truth is one your side.[38] If you are an oppressor, going anti-realist is to your advantage—the truth is your enemy. Better to make ‘pious’ apply to whomever we think it applies to. With a little advertising, oppressors can insure that it applies to them.[39]
The Political Argument, then, fails to convince. But note that the problem is not (claimed to be) that realist forms of theology support the oppressed better than any alternative. I am not given a political argument for theological realism. Rather, I am objecting to the Political Argument for theological anti-realism, by arguing, first, that simply going anti-realist does not affect the content of theology at all, and so does not affect it in a liberating way; second, that it seems likely to sap the liberating power of theism in practice; and third, that insofar as it is claimed that anti-realism is needed to make revolutionary forms of theism plausible, this is a) contradicted by the existence of revolutionary realist theology, b) reflects poorly on such revolutionary forms of theism, and c) seems to ignore two crucial facts: that revolutions in theology will be overseen by the powerful; and that traditional theism has much more influence than any revolutionary form of theism is likely to have.
4.2 Shafer-Landau’s Arguments
Let us turn now to three arguments that have been widely used to attack both theological and moral realism. They are:
The Genealogical Argument: Since most people “inherit” their moral and religious beliefs from the people around them (especially their parents), most people’s moral and religious beliefs are unjustified, and so unknown. If they had been around different people, they would have ended up with different moral and religious beliefs. Even if there are correct moral and theological beliefs, and even if one happens to have inherited the correct moral and theological beliefs, this is just a matter of luck. But luck is incompatible with knowledge, and so there is no moral or theological knowledge.
The Argument from Disagreement: Disagreement about morality and religion is widespread and persistent, even among people that are highly intelligent and well educated. Some conclude that this is best explained by the hypothesis that there are no moral or religious facts: that there is no moral or religious reality that people are failing to come to terms about. Call this conclusion NO FACTS. Others conclude merely that such disagreement acts as a defeater to any justification one might have had for one’s moral or religious beliefs, and hence that moral and religious beliefs are unjustified, and so unlikely to be true. Call this conclusion NO JUSTIFICATION.[40]
The Explanatory Argument: We should only believe things that are essential parts of the best explanation of our experience. But neither moral nor religious claims play an essential part in the best explanation of our experience. Hence, we should not believe moral or religious claims.
Russ Shafer-Landau has recently tried to show that these three arguments are more damaging to theological realism than they are to moral realism.[41] While he admits that neither moral nor theological realists should be impressed by the first two arguments, he takes he explanatory argument to be much more potent, and to be especially damaging for theological realism. In what follows, I will argue, pace Shafer-Landau, that theological realists are better situated to respond to the explanatory argument than moral realists.[42] Given the widespread acceptance of moral realism, this leaves theological realists on relatively stable ground. But first, let us rehearse the problems with the first two arguments.
4.2.1 Evaluating the Genealogical Argument
Despite its ubiquity, there is a fundamental problem with the Genealogical Argument: it overgeneralizes. If it works against moral and religious beliefs, it will work against almost any belief whatsoever. There is, of course, some sense in which knowledge is incompatible with luck, but it cannot just be the sense in which I am luckily to live in a society with correct moral and religious beliefs (if I am). I was lucky to be born in a society that knows about quantum physics and advanced medicine, but that obviously doesn’t impugn my knowledge of those subjects. Similarly, I was lucky not to be born blind or deaf, but that doesn’t impugn the knowledge I receive from my senses. And I was lucky to have been born in a society that recognizes the moral equality of all humans, but that doesn’t impugn my knowledge of that either. As Wittgenstein says, “It is always by favor of Nature that one knows something” (1969). While we know that some kinds of luck are incompatible with knowledge, if we knew what kind of luck was incompatible with knowledge, we could analyze knowledge. But we don’t, so we can’t.[43]
So while it must be granted that luck can thwart knowledge, it does not always do so. The relationship between luck and knowledge is complicated, but unless we are willing to be fairly global skeptics we cannot think that luck’s presence in any capacity is sufficient for taking away knowledge. To make the case that one’s luck in living in a society with correct moral or religious beliefs was incompatible with knowledge about matters moral or religious, one would have to show that the society in question stumbled upon those beliefs by luck, rather than on the basis of some argument or evidence. (Even that might not be enough: I am lucky that I am not in fake barn country, but that does not impugn my knowledge that there is a barn in front of me.) But to show that the moral or religious beliefs of a society are unknown (because unjustified) one has to engage in the dirty work of criticizing the society’s reasons for accepting the claims in question. Genealogical considerations provide no shortcut around this task.[44]
4.2.2 Evaluating the Argument from Disagreement
Similar problems plague the Argument from Disagreement. While it is true that there is deep and persistent disagreement about moral and religious claims amongst people that are highly intelligent and well informed,[45] there is deep and persistent disagreement about a host of philosophical, political, counterfactual, historical, cosmological, and more generally scientific claims as well. In many cases of deep and persistent disagreement—historical questions about Homer, Shakespeare, the origins of life on Earth, etc.—it is clear that there is a fact of the matter: one side has simply got things wrong.[46] So NO FACTS simply does not follow from the existence of deep and persistent disagreement. But what about NO JUSTIFICATION?
It is more plausible that deep and persistent disagreement about some claim can take away our justification for believing the claim or its denial: that disagreement breeds agnosticism, or at least that it should. The problems with the argument for NO JUSTIFICATION are more subtle than those with the argument for NO FACTS, and while I cannot do them full justice here, I will note that one reason for being suspicious of the argument is that it would entail that it is irrational to have beliefs about many controversial philosophical, political, counterfactual, historical, cosmological, and other scientific claims. This seems hard to swallow. A second problem arises from the fact that we should believe whatever it is that our total evidence supports. But if my opponent correctly takes her total evidence to support p, and subsequently becomes aware that I disagree with her about the truth of p (and about whether the evidence supports p), it is far from clear that this new fact she has learned (about my lack of agreement) will change her evidence base such that it will, in toto, invariably fail to support p. To think this would be to think that deep and persistent disagreement always trumps or defeats proper assessments of the direct evidence for or against a proposition. And why should we think that?[47] Until this question has been satisfactorily answered—and it is unclear that it can be—the argument from disagreement remains inconclusive.
A final point about the argument from disagreement is that there is just as much disagreement about whether our (total) evidence supports agnosticism as there is about whether it supports theism or atheism. So if the Argument from Disagreement were cogent, we would seemingly be forced into a kind of “meta-agnosticism” where we remain neutral between theism, atheism, and agnosticism. It is not clear, however, that it is possible to do anything but accept, deny, or withhold on the existence of God. And even if it were, it is not clear that it is coherent to be a non-agnostic meta-agnostic. Of course, neither of these worries arises if we think of belief as coming in degrees, as perhaps we should.
4.2.3 Evaluating the Explanatory Argument
What, then, of the explanatory argument? First, I should note that it is far from clear that we should only believe things that are essential parts of the best explanation of experience. But even if we should, the theological realist is significantly better positioned to answer this explanatory demand than the moral realist. Contrary to what Shafer-Landau suggests, theological realism is more explanatorily potent that moral realism. It is no coincidence that adherents of the principle of sufficient reason are typically theists.[48] And of course the fine-tuning argument, the design argument, and the cosmological argument are often presented as involving explanatory demands or inference to the best explanation in the cases they make for the existence of God.
In order to assess the plausibility of this response, consider Shafer-Landau’s version of the explanatory argument:
(1) We have reason to believe that something exists only if it is required in the best explanation of the events that we undertake or experience.
(2) Neither religious nor moral facts are thus required.
(3) Therefore, we lack a reason to believe that there are such facts.
(4) If anything that x is invoked to explain can be better explained without positing x’s existence, then we have reason to deny x’s existence.
(5) Anything that divine or moral facts are invoked to explain can be better explained without positing their existence.
(6) Therefore, we have reason to deny the existence of divine and moral facts.
(7) There is no reason to believe in moral or divine facts, and some reason to deny their existence.
(8) Therefore, there is most reason to deny the existence of divine and moral facts.
Perhaps obviously, realists will want to deny premises (2) and (5) of the argument, at least if they don’t simply reject (1). This is the route Shafer-Landau takes in defense of moral realism, noting that various observable facts counterfactually depend on moral facts: e.g., if the CEO hadn’t been wicked, there would be more money in the employee pension plan; if adultery wasn’t wrong, Bill would have cheated more often; etc. Shafer-Landau admits that such counterfactual dependence does not show that the moral facts explain the observable outcomes. But he suggests that counterfactual dependence is at least a rough guide to explanatory potency—at a minimum, it is necessary for explanation. He grants, however, that using the guide in the case of moral facts is problematic due to the supervenience of the moral on the non-moral.
According to Shafer-Landau, theists have less room than moral realists to respond to the explanatory argument. But it is unclear why he thinks that. Note, first, that theists can also invoke counterfactual dependence in support of the explanatory potency of theistic facts: e.g., if God had not commanded Her people to keep holy the Sabbath, the parking lot at this church would not have as many cars in it; if adultery wasn’t a sin, George would have gone home with Roxy; etc. Of course, such dependencies don’t establish the explanatory potency of theistic facts. Just as in many moral examples, the counterfactual test fails to distinguish between cases where the observable phenomena depend on the truth or acceptance of theological claims. But note that theists don’t face the problem for moral realism mentioned above: no one thinks that the theistic facts supervene on the non-theistic facts.
In any case, the theological realist is better positioned to claim that her view is explanatorily potent than is the moral realist.[49] Aside from the fine-tuning argument, the traditional design argument, and the cosmological argument, one could also cite the fact that a large number of people in 1st Century Palestine thought that Jesus of Nazareth had risen from the dead (many even thought they had seen him) and was the Christ, the Son of God. One explanation for this fact is that Jesus did rise from the dead, and that many people were witnesses to this and other miracles he performed. Is this the best explanation for the rapid spread of what we now call ‘Christianity’? That is a vexed question, and a difficult one to answer without begging the question.[50] What matters for our purposes, however, is just that Christian theism has significantly greater explanatory potential than moral realism. Whether the truth of Christian theism actually is the correct explanation of the apparent design in the cosmos, the rapid spread of Christianity, etc., is a question I cannot hope to answer here.
5. Genera of Theological Anti-Realism
Even if the arguments against theological realism are less than compelling, to fully evaluate the prospects of theological anti-realism we must assess it in its most plausible form. But what is that? In this section I appraise the relative merits of content and attitude-hermeneutic forms of theological anti-realism. We saw above that content-hermeneutic anti-realism about d is the view that sentences in d have a “funny” non-face-value semantics.[51] Attitude-hermeneutic anti-realism, on the other hand, maintains that utterances of sentences in d are not like utterances in other domains—in particular, they are not expressions of belief in the contents of sentences uttered. (Those sentences may, according to attitude-hermeneuticists, be given a straightforward semantics.) On this view, utterances in d are, rather, expressing some cognitive attitude compatible with the falsity of those sentences.
5.1 Content Hermeneuticism
The fundamental problem for most forms of content-hermeneuticism is that linguistic constructions about different domains can be haphazardly combined. For example, theological language can be combined with constructions about non-theological topics. After a long discussion of the fine-tuning argument, a disenchanted student of mine recently muttered, “God doesn’t exist and class is over anyway”. Mark Schroeder has argued that making statements in some restricted domain (like morality or theology) “play nice” with other statements requires that we give a uniform semantics for sentences within and without that domain. So if we wish to give a “funny” semantics for theological language, we must give a “funny” semantics for language in general. And the project of giving a “funny” (non-truth-conditional) semantics for language in general, to put it bluntly, spirals out of control rather quickly.[52]
Let’s look at a specific example of the problem as it applies in the theological case. One common form of theological content-hermeneuticism holds that ‘God’ is not a referring expression. Alexander Bird notes an obvious problem with this view: ‘Jesus’, ‘Paul’, and ‘Mary’ are names for actual historical figures.[53] Even if those names do not refer, that is a purely historical (ordinary) fact. Those names are ordinary names, with an ordinary semantics. According to semantic orthodoxy, they are singular terms. But Christians think ‘Jesus’ and ‘the son of God’ co-refer. Hence, if ‘Jesus’ is a singular term, ‘the son of God’ is as well. But now we are forced to either give ‘the son of God’ an ordinary face-value interpretation, or else abandon semantic orthodoxy, and deny that ordinary denoting expressions (like ‘Jesus’) are singular terms. If we go the former route, we have abandoned content-hermeneuticism about ‘the son of God’,[54] but if we go the latter route we end up having to construct a novel semantic theory for ordinary names like ‘Jesus’ and ‘Mary’, and do so within the constraints set by our theological anti-realism. There are good reasons to think that this project will quickly get out of hand.[55] But even if it were to succeed, it looks like we would end up with a full-blown anti-realist semantics for ordinary language. To save theological anti-realism, we will pushed into accepting metaphysical anti-realism. And this is a cost that some theological anti-realists will and should be unwilling to pay.[56]
5.2 Attitude Hermeneuticism
Given the problems facing theologically localized content-hermeneutic anti-realism (and full-blown content-hermeneutic metaphysical anti-realism), one might be tempted to give up on the idea of providing an anti-realist semantics for theological language, and adopt attitude-hermeneutic theological anti-realism instead. This form of anti-realism does not claim that theological discourse has a “different” semantics than ordinary discourse, but rather that we do or should have a “different” attitude towards theological speech than we do towards ordinary discourse.[57] The move from content-hermeneuticism to attitude-hermeneuticism might be further motivated by reflection on the development of anti-realist views in the philosophy of science.
In the early 20th Century, most scientific anti-realists were instrumentalists: they held that our theories had, in effect, “predictive-accuracy conditions” rather than truth-conditions. On such views the instrumental usefulness of a scientific theory is a matter of its being predictively accurate—its being able to predict and organize our experiences. Accordingly, terms for unobservable entities like ‘electron’ were given a non-standard semantics, according to which ‘electron’ meant something like ‘streak in a cloud chamber’. But these attempts to define “theoretical terms” in observable language ran into insuperable difficulties. This failure ushered in a shift in opinion towards realism, but also in a reorientation of the anti-realist project.
Bas van Fraassen’s constructive empiricism is now the standard bearer for scientific anti-realism.[58] This attitude-hermeneutic approach eschews the attempt to give a special semantics for theoretical terms, and so avoids the difficulties associated with traditional content-hermeneutic forms of instrumentalism. Constructive empiricists hold that the aim of scientific discourse is not truth, but empirical adequacy. Although van Fraassen takes our scientific theories to have a realist semantics, he thinks we should decline to believe those theories. When we make scientific statements, we are not expressing belief in the truth of the sentences we utter, but rather a commitment to their empirical adequacy. Statements about unobservables are meaningful, on this view, and either true or false. But we should maintain an agnostic attitude towards such statements, declining to believe or deny them.
5.2.1 Theological Instrumentalism
It would be natural to classify people who claim that theological beliefs (of some salient variety) are instrumentally useful as “theological instrumentalists”. Defined thus, most theological realists are theological instrumentalists. We could then define a mere theological instrumentalist as someone who thinks that theological beliefs are instrumentally useful but not true. I will, however, follow established usage, according to which instrumentalists about a domain d think that the claims in d are instrumentally useful but not (necessarily) true.
What reason do we have for thinking that attitude-hermeneutic theological anti-realism is correct? Van Fraassen suggests that the lightheartedness of scientists about differences that do not make an observable difference supports his position. (As do cases like Fluid Mechanics where it is clear that we “accept” and use a theory that is not true.) This lightheartedness argument has been influentially used in support of mathematical anti-realism as well: mathematicians are strikingly lighthearted about accepting new mathematical theories—theories which purport to be about vast new domains of mathematical objects—as long as they are consistent with established mathematics.[59] And indeed, if attitude-hermeneutic anti-realism is true, we should expect people to be lighthearted about the truth of the sentences they assert, since they are not asserting that those sentences are true.
Do theologians exhibit a comparable lightheartedness? To ask this question is to answer it. The vast majority of theologians have not been lighthearted, and it seems clear that they have been realists. (For example, Augustine and Aquinas both thought that they were in possession of a good number of important and objective theological truths.) The official teaching of the Catholic Church, not to mention the Council of Nicaea, seems to be clearly realist in nature. People have died over the question of whether the Spirit proceeds from the Father, or from the Father and the Son; over whether the Eucharist involves transubstantiation or consubstantiation; etc. This might give us a practical reason to support an anti-realist revolution,[60] but it is strong evidence against anti-realism as a descriptively accurate account of what theology actually is.
It should be noted that, in contrast with the mathematical case, novel theological theories are often inconsistent with previous theology. But even in cases where mathematical developments are inconsistent with previous mathematics—the development of non-Euclidian geometries, for example—mathematicians have been far more willing to “let a thousand flowers bloom” than theologians. And so I think that even if it is possible for attitude-hermeneutic theological anti-realism to be true, it is clear that it isn’t. Or at least, it is much less plausible that attitude-hermeneutic anti-realism is true of theology than it is of mathematics. And it is far from clear that it is true of mathematics.[61] It seems, then, that the most plausible form of theological anti-realism is not very plausible at all.
6. Conclusion
The central aim of this paper has been to facilitate productive discussion between analytic and non-analytic theologians by explaining why analytically oriented theorists tend not to be attracted to anti-realism, both in general and with regard to theology in particular. The nature of the enterprise precludes me from examining any one argument in depth, and I doubt that many theological anti-realists (or non-analytic theorists more generally) will be convinced by what I have said here. I do hope, however, that this paper serves as a helpful resource for analytic theologians interested in bridging the gap with their non-analytic sisters and brothers.[62]
Bibliography
Alston, William. 1996. Belief, acceptance, and religious faith. In Jeff Jordan and Daniel Howard-Snyder, editors, Faith, Freedom, and Rationality. Rowman & Littlefield.
Aquinas, Thomas. 1981. The Summa Theologica of St. Thomas Aquinas. Fathers of the English Dominican Province (tr.). Ave Maria Press.
Bird, Alexander. 2007. Scientific and theological realism. In Andrew Moore and Michael Scott, editors, Realism and Religion. Ashgate.
Bogardus, Tomas. 2013. The Problem of Contingency for Religious Belief. Faith and Philosophy, 30:371-392.
Boyd, Richard. 1991. On the current status of scientific realism. In Richard Boyd, Philip Gasper, and J.D. Trout, editors, The Philosophy of Science. MIT Press.
Braver, Lee. 2007. A Thing of This World: A History of Continental Anti-Realism. Northwestern University Press.
Brink, David. 1984. Moral realism and the sceptical arguments from disagreement and queerness. Australasian Journal of Philosophy 62:112–25.
Brogaard, Berit. 2012. Transient Truths: An Essay in the Metaphysics of Propositions. Oxford University Press.
Burgess, John P. and Gideon Rosen. 1997. A Subject with No Object: Strategies for Nominalistic Interpretation of Mathematics. Clarendon, Oxford, 1997.
_________. 2005. Nominalism reconsidered. In Stewart Shapiro, editor, Oxford Handbook of the Philosophy of Mathematics and Logic. Oxford University Press.
Byrne, Peter. 2003. God and Realism. Ashgate.
Chignell, Andrew. 2009. ‘As Kant has Shown’…analytic theology and the critical philosophy. In Oliver D. Crisp and Michael C. Rea, editors, Analytic Theology. Oxford University Press,
Cupitt, Don. 1991. What Is a Story? SCM Press.
Davidson, Donald. 1968. On Saying That. Synthese, 19:130-146.
Doyle, Andrew. 2009. Schleirmacher’s theological anti-realism. In Oliver D. Crisp and Michael C. Rea, editors, Analytic Theology. Oxford University Press.
Dummett, Michael. 1978. Truth and Other Enigmas. Harvard University Press.
Feldman, Richard and Ted A. Warfield, editors. 2010. Disagreement. Oxford University Press.
Geach, Peter. 1960. Ascriptivism. Philosophical Review, 69:221–225.
Guyer, Paul. 2006. Kant. Routledge,
Hebbeler, James C.. 2012. The principles of the first critique. Review of Metaphysics, 65:555.
Howard-Snyder, Daniel. 2013. Propositional faith: What it is and what it is not. American Philosophical Quarterly, 50(3).
Jacobs, Jonathan D. Forthcoming. The Ineffable, Inconceivable, and Incomprehensible God: Fundamentality and Apophatic Theology. Oxford Studies in Philosophy of Religion.
Jordan, Jeff. 2006. Pascal’s Wager: Pragmatic Arguments and Belief in God. Oxford University Press.
Jubien, Michael. 2001. Propositions and the Objects of Thought. Philosophical Studies, 104:47.
Kant, Immanuel. 1781. Critique of Pure Reason. St. Martins.
Keller, John A. Forthcoming. Perspicuity, Paraphrase, and the Symmetry Paradox.
_________. 2014. Paraphrase, Semantics, and Ontology. In Karen Bennett and Dean Zimmerman, editors, Oxford Studies in Metaphysics, v.9. Oxford University Press.
Kelly, Thomas. 2010. Peer disagreement and higher order evidence. In Richard Feldman and Ted Warfield, editors, Disagreement. Oxford University Press.
_________. 2011. Consensus Gentium: Twenty-first century reflections on the ‘common consent’ argument for the existence of god. In Kelly James Clark and Raymond J. VanArragon, editors, Evidence and Religious Belief. Oxford University Press.
Lewis, David. 1984. Putnam’s paradox. Australasian Journal of Philosophy, 62:221–36.
_________. 1986. On the Plurality of Worlds. Blackwell.
Loux, Michael J. 2006. Metaphysics: A Contemporary Introduction. Routledge, 3rd edition.
Mackie, J. L.. 1977. Ethics: Inventing Right and Wrong. Penguin.
Marion, Jean-Luc. 1991. God Without Being: Hors-Texte. University of Chicago Press.
Miller, Christian. 2012. Atheism and theistic belief. Oxford Studies in Philosophy of Religion, 4.
Nagel, Thomas. 1997. The Last Word. Oxford University Press.
Pettit, Phillip. 1991. Realism and response dependence. Mind, 100:587–626.
Plantinga, Alvin. 1982. How to be an Anti-Realist. Proceedings and Addresses of the American Philosophical Association, Vol. 56, No. 1.
Pritchard, Duncan. 2005. Epistemic Luck. Oxford University Press,
Putnam, Hilary. 1987. The Many Faces of Realism. Open Court.
Rea, Michael C. 2007. Realism in theology and metaphysics. In Conor Cunningham and Peter Candler, editors, Belief and Metaphysics. SCM Press.
Runzo, Joseph. 1993. Is God Real? Palgrave Macmillan.
Russell, Bertrand. 1905. On denoting. Mind, pages 479–93.
Schroeder, Mark. 2008. Being For: Evaluating the Semantic Program of Expressivism. Oxford University Press.
Shafer-Landau, Russ. 2007. Moral and theological realism: The explanatory argument. Journal of Moral Philosophy, 4(3):311–329.
Shalkowski, Scott A. 1997. Theoretical virtues and theological construction. International Journal for Philosophy of Religion, 41.
Sider, Theodore. 2009. Ontological realism. In David Chalmers, David Manley, and Ryan Wasserman, editors, Metametaphysics. Oxford University Press.
Soames, Scott. 1987. Direct reference, propositional attitudes, and semantic content. Philosophical Topics, 15:47–87.
Sokal, Alan. 1997. A plea for reason, evidence and logic. New Politics, 6(2):126–129.
Stanley, Jason. 2001. Hermeneutic fictionalism. In Peter French and Howard K. Wettstein, editors, Midwest Studies in Philosophy XXV: Figurative Language. Basil Blackwell.
Szabo, Zoltan Gendler. 2003. Believing in things. Philosophy and Phenomenological Research, 66(3).
Tanner, Kathryn. 1997. Social theory concerning ‘new social movements’ and the practice of feminist theology. In Rebecca S. Chopp and Sheila Greeve Davaney, editors, Horizons in Feminist Theology: Identity, Tradition, and Norms. Augsburg Fortress.
Tillich, Paul. 1957. Dynamics of Faith. Allen and Unwin.
Trigg, Roger. 1999. Theological realism and antirealism. In Philip L. Quinn and Charles Taliaferro, editors, A Companion to the Philosophy of Religion. Blackwell.
Tucker, Chris. 2013. Seemings and Justification: New Essays on Dogmatism and Phenomenal Conservatism. Oxford University Press.
van Fraassen, Bas C. 1980. The Scientific Image. Clarendon.
_________. 2002. The Empirical Stance. Yale University Press.
van Inwagen, Peter. 2006. The Problem of Evil. Oxford University Press.
White, Stephen Ross. 1994. Don Cupitt and the Future of Christian Doctrine. SCM Press.
Williams, Delores. 1995. Sisters in the Wilderness. Orbis Books.
Williamson, Timothy. 2007. The Philosophy of Philosophy. Blackwell.
Wittgenstein, Ludwig. 1969. On Certainty. Oxford University Press.
Wood, William. 2013. Philosophical Theology in the Religious Studies Academy: Some Questions for Analytic Theologians. Journal of the American Academy of Religion, 81(3).
Yablo, Sephen. 1998. Does ontology rest on a mistake? Proceedings of the Aristotelian Society, Supp. 72:229–261.
_________. 2000. A paradox of existence. In Anthony Everett and Thomas Hofweber, editors, Empty Names, Fiction and the Puzzles of Non-Existence. CSLI Press.
Zimmerman, Dean W. 2005. The a-theory of time, the b-theory of time, and ‘taking tense seriously’. Dialectica, 59:401–57.