The fictionalist view that I outlined in my earlier paper – ‘Can an atheist believe in God?’Footnote ¹ – combines aspects of instrumentalist and expressivist forms of religious non-realism. In brief, I argued that it can be reasonable for an atheist to commit herself to a form of life shaped by a theistic community's language and ritual practice because she recognizes the way in which participating in such a community: (1) allows her to express commitment to a distinctively religious ideal (the expressivist component); and (2) powerfully helps foster personal and communal progress toward that ideal (the instrumentalist component). The view is fictionalist insofar as the atheist in question believes that since the language in the relevant religious narratives and doctrines fails to refer successfully to an existing transcendent personal being as intended by the traditional theistic ‘believer’, the language should be reinterpreted instead as a part of a powerfully expressive and useful fiction.

Benjamin Cordry raises a series of objections to my view.Footnote ² They fall into two main categories: those alleging that an atheist would be unjustified in adopting fictionalism and those alleging that fictionalism could not be successfully implemented, or practised communally. I think that these objections can be met, and welcome the opportunity to respond since it allows for some helpful clarifications of and elaborations on my earlier account.

Cordry first argues that the kinds of reasons that an atheist has for being an atheist will also sometimes be reasons to reject fictionalism. He cites a number of reasons as candidates – e.g., that religion in general or theistic imagery/narratives in particular are a tool of oppression/exploitation; that theistic religion has hindered scientific progress; that it has tended to devalue human nature; and/or that the moral values associated with theistic traditions are untenable (79–80).Footnote ³ Certainly there is much with which to be concerned when one reflects on the history of religions (though one should be careful not to overlook the good things to which religions have contributed as well). However, fictionalism will be ruled out only if the connection between theistic discourse/practice and these causes for concern is judged to be a necessary one (or perhaps practically nearly so).

For example, if one concludes that the very idea of God – rather than a particular conception of God – is a tool of oppression, then one will, of course, have no reason to adopt a theistic version of non-realism. This is not the case for many atheists, as Cordry himself grants.Footnote ⁴ That is, many are not atheists because they regard theistic religion to be by its nature objectionable, but rather because they simply find no good reason to think that God exists and perhaps several good reasons for thinking that God does not exist.

If not strictly ruled out by the reasons she has for being an atheist, the issue now becomes one of weighing whether fictionalism offers something of value that is not realizable (or at least is not readily realizable) by alternative means. Here, Cordry seems optimistic: ‘there is nothing to block an atheist from appropriating piecemeal whatever wisdom, inspiration, or solace is there [in a religious tradition] without being a fictionalist’ (79). In my original paper, I expressed some scepticism about this, arguing that it is at least very difficult, if not impossible, to articulate the theistic ideal of imitating God in non-theistic language.Footnote ⁵ If I am correct, and it is a worthy ideal to which one might aspire, then an atheist has a good reason to entertain fictionalism if not strictly barred from it by the reasons she has for being an atheist.

I turn now to the second, and I think most important, group of Cordry's objections against fictionalism. Here, Cordry argues that fictionalism cannot be successful in what it aims to accomplish because the fictionalist cannot unilaterally invent a new use of theistic language within a community and because her interpretations of such language are likely to be at odds with those of others to such an extent that she is unlikely able to participate as a genuine co-member of the community. Rather than responding in detail to each of the objections in this group, I will highlight two problematic assumptions that seem to lie behind them. It will be helpful to begin by examining Cordry's claim that the fictionalist is engaged in a re-appropriation rather than reinterpretation of theistic language.

The fictionalist view I have proposed is an error theory. Again, this is to say that the fictionalist acknowledges that theistic language, as used by the traditional theist, either makes or presupposes realist truth claims about God, but regarding these as false, the fictionalist recommends that the use of such language be reinterpreted. Cordry maintains that my view is best understood as a re-appropriation of religious discourse rather than a reinterpretation as I have characterized it since the fictionalist no longer accepts that the function of religious language is to state truth (78–79).Footnote ⁶ The distinction Cordry seems to have in mind here is that something counts as a reinterpretation of an utterance only if there is significant continuity in the previous and subsequent use. Otherwise, the utterance is being re-appropriated – that is, used for some entirely different purpose. Cordry understands the fictionalist proposal to be that the traditional descriptive use of theistic language is to be replaced by entirely new expressivist and instrumentalist aims. Therefore, he contends that it is more aptly characterized as re-appropriation rather than reinterpretation.

Several of Cordry's central objections follow straightforwardly from this line of thought. For example, since the successful use of language is a public endeavour, how is it possible for the fictionalist privately to introduce an entirely new meaning to the theistic discourse (and practice) she shares with members of her religious community – both with respect to what she says and what she hears/interprets? (80–83). If they mean one thing, and she means something entirely different, how is successful communication between them and genuine joint participation in a religious community possible?

If the fictionalist view required the invention of completely new and unfamiliar uses for theistic language, then these objections would have considerable force.Footnote ⁷ Fortunately, it does not. The first mistaken assumption that seems to underlie these objections is that theistic language prior to reform by the fictionalist has but one function, namely, to assert theological truths – e.g., to affirm God's existence and to describe God's nature and activity. Much of it does have (or more often presupposes) such assertions, but this is not its only function, even for the traditional realist. Though I noted this in my original paper,Footnote ⁸ I did not elaborate on its significance. Cordry's objections provide a helpful opportunity to do so.

One of the strengths of fictionalism, as I conceive it, is its ability to capitalize on those already existing functions of theistic language that are distinct from its truth-asserting aim. In other words, my claim is that the expressivist and instrumentalist elements central to the fictionalist view are already present in the language being reinterpreted. The fictionalist's reinterpretation consists in subtracting the truth-asserting function and raising the prominence of the expressivist and instrumentalist functions. Thus, when I earlier gave the following example:

I was not suggesting that the fictionalist need invent these expressivist and instrumentalist uses of the biblical narrative. They are already implicit in the text as well as in the way the narrative is typically used within theistic communities.

Granted, a theologically conservative realist will also believe that the biblical narrative recounts the geopolitical interventions of a supernatural being into the history of ancient Mesopotamia and the Near East. This, the fictionalist (along with many realists of more liberal theological persuasion) rejects. Nevertheless, the present point is that the way the fictionalist uses theistic language is not as discontinuous as Cordry's objections presuppose. Therefore, we need not accept that the realist means one thing and the fictionalist something entirely different (and the latter being something she must invent or mean privately). There may be considerable overlap in the way the realist and fictionalist use religious language and thus a shared foundation for successful communication, interpretation, and thereby joint participation in the community.

In noting this, I do not mean to dismiss entirely the force of Cordry's concerns. Much depends upon the nature of the particular religious community within which we imagine the fictionalist taking part. This leads us to the second assumption that seems to underlie Cordry's objections to the successful implementation of fictionalism. Cordry seems to assume that all religious (or, at least theistic) communities are principally defined by their concern with right belief. Thus, in describing what being an active participant in a religious community would entail, he says, ‘In order to be an active participant, the fictionalist must be able to make normative collective belief statements of the form ‘We should believe X’ …. The fictionalist will thus have to be able to make and justify collective belief statements. This is problematic' (84–85).

Cordry is certainly correct that a religious community's identity presupposes some considerable shared understanding, but the issue is whether that shared understanding must be understood in terms of convergence on collective beliefs. Theistic communities, like other religious communities, exist on a continuum with respect to the emphasis they place on a shared realist commitment and the importance of working toward convergence on right belief, or orthodoxy. Cordry's concerns over whether the fictionalist can successfully communicate and participate as a genuine co-member of theistic community have their greatest purchase if one imagines a community wherein members not only share a realist commitment but also conceive of their shared project as tightly tied to maintaining consensus over correct theological belief. Even in these contexts, the descriptive use of theistic language is often accompanied by or subordinate to its expressive and instrumental functions, so the realist and fictionalist will often continue to aim at accomplishing much of the same thing in their speech and practice.

My own observations in many different types of theistic communities suggest that assenting to theological truths is rarely the sole or even primary function of theistic language. Yet I grant that this is sometimes the case in theologically conservative communities. Earlier I had focused only on a moral concern about practising fictionalism in a setting where realism was presupposed (namely, that if one kept one's fictionalism to oneself, one would be allowing others to presume falsely that one shared their realist commitment).Footnote ¹⁰ Cordry has helpfully highlighted that if the use of one's theistic language is on occasion sufficiently discontinuous with that of others in the community, then the fictionalist's ability to communicate successfully and thus participate fully as a co-member may itself be hampered.

However, I think these instances will, for the reasons cited above, be rare. Moreover, not all theistic religious communities are defined by a shared commitment to religious realism.Footnote ¹¹ In many theologically liberal communities – the sort wherein fictionalists would be expected to be most common – members would not at all be surprised that some among them were atheists. The defining self-understanding of such communities very often has little to do with correct theological belief. Instead, the relevant convergence concerns what members regard the community to be about – that is, what ideals they are seeking to realize in themselves and the world through their shared practice. In this case, a religious community is defined more by its sense that ‘We are the people who are committed to doing X and aspire to Y’, rather than ‘We are the people who believe these propositions about God.’Footnote ¹²

Given this source of self-identity, even the realists of such a community will place greater emphasis on the expressive and instrumentalist functions of theistic language.Footnote ¹³ When we imagine the fictionalist in this setting, Cordry's concerns about whether a fictionalist can successfully communicate and jointly practise with realists seem largely unmotivated. What the fictionalist and realist in such a community are most often saying, hearing, and interpreting, will be very much alike, if not the very same thing, for they both regard their community to be defined by the way its members pursue an ideal embedded in their shared aspiration to imitate God.Footnote ¹⁴