Two significant contributions to recent discussions of religious language are offered by Janet Soskice and William P. Alston. In Metaphor and Religious Language (1985) Soskice offers as a working definition "metaphor is that figure of speech whereby we speak about one thing in terms which are seen to be suggestive of another" (p.15). The minimum unit in which a metaphor is established is semantic. A satisfactory theory of metaphors "should regard metaphors neither as a simple substitution for literal speech nor as strictly emotive. Metaphors should be treated as fully cognitive and capable of saying that which may be said in no other way. It should explain how metaphor gives us "two ideas for one," yet do so without lapsing into a comparison theory" (p.44). Noncognitive accounts of metaphor are rejected because "we cannot conceive of emotive 'import' apart from a cognitive content which elicits it" (p. 27). The "two ideas for one" feature involves a metaphor having a "unity of subject matter" that "draws upon two (or more) sets of associations, … characteristically, by involving the consideration of a model or models" (p. 49). A model is "an object or a state of affairs … viewed in terms of its resemblance, real or hypothetical, to some other object or state of affairs" (p. 100).
Models come in two types: paramorphic (the source and subject differ, as in billiard ball movement serving as a model for the properties of gases); or homeomorphic, where the subject is the source (e.g., a dummy used to teach lifesaving skills). Models, in both theology and science, are essential to theories because they carry their explanatory force. "The fertility of a theory lies in its ability to suggest possibilities of explanation which, while not inconsistent with, are more than simply the logical extensions of mathematical formulas … this suggestive capacity … constitutes the fruitfulness of a theory, and gives the theory the predictive nature which is its raison d'etre" (p. 114). We do not describe God but point to God through effects, and beyond them to him. We refer without defining. "This is the fine edge at which negative theology and positive theology meet, for the apophatic insight that we say nothing of God, but only point to Him … this separation of referring and defining is at the very heart of metaphorical speaking.…" (p. 140).
Nothing in Soskice's account of metaphor entails that language about God must be nonliteral. The claim All language about God is metaphorical is not metaphorical. The idea that no metaphor can be translated into or replaced by literal terms is false. Consider Soskice's example of an expression of hope that a soldier will be pardoned eliciting, "That's blowing on cold coals." "There's no chance of that" is a literal translation. "God is a rock" seems replaceable by "God is utterly reliable." If it is not, this is a matter of the associations of "rock" in biblical and theistic literature being multiple. It does not follow that any of the things that "rock" suggests are nonliteral. It just suggests that there are a variety of possibilities, more perhaps than we can list, each of which may be perfectly expressible without remainder in literal fashion.
A basic assumption is that no literal description can be true of God. As is typical, we are referred to certain ideas: We cannot comprehend (know all there is to know) about God; descriptions of God based on religious experience are defeasible; certainty about claims concerning God is unattainable; and it is always possible that we will have to modify our concept of God. But there are an infinite number of truths concerning a golden retriever, seeing the golden is defeasible, certainty about it is unavailable, and we may have to revise our concept thereof. But it is not beyond literal description. Further, God can be misdescribed (e.g., "God is a cantaloupe"), which even the most deluded of empiricist positivists presumably will recognize as false. But then what, in principle, precludes God from being correctly described?
William P. Alston's major essays concerning religious language are collected in Divine Nature and Human Language (1989). In "Irreducible Metaphors in Theology" he says that "in the typical metaphorical statement the speaker is 'building on' the relevant meaning of the predicate term in two ways … he is presenting the thing to which the term literally applies as a model of the subject [and] … he has in mind one or more resemblances between model and subject and he abstracts from these resemblances what he means to be attributing to the subject" (1989, p. 23). The resemblance may be either general or specific. Everything resembles everything else in some way. Any metaphor based on this fact corresponds to a literal way of expressing the similarity. Regarding metaphors intended to express truths, he writes: "Though irreducible metaphors seem to promise a way of combining the denial of predication in theology with the preservation of significant theological truth claims, this fair promise dissipates on scrutiny like mist before the morning sun. Either the panmetaphoricist abandons the aspiration to significant truth claims or he revokes the ban on literal predication" (p. 37).
"Can We Speak Literally of God?" considers predicates that apply to personal agents ("P-predicates") in their application to God. These include mental and action predicates. These have been understood on a private paradigm model (one knows what "depression" means by being depressed) and functionally ("being depressed" refers to a state that functions efficaciously in a causal system to yield a distinctive range of behavior). The idea of basic actions that involve no bodily movements, and of nonbasic actions that involve only mental actions that bring about effects, are both intelligible and applicable to incorporeal beings. "Literal" does not mean "empirical."
"Functionalism and Theological Language" and "Divine and Human Action" consider functional accounts of mental concepts to argue that these concepts can apply to God. We can "form the conception of a being (a 'system') in which some factors depend on their relations to others for being what they are, even though there are no temporally successive processes for formation of any subjection to laws. More specifically, we are to think of God as realizing a complex structure of attitudes, knowledge, tendencies, executive intentions, and volitions in the 'eternal now' …" (p. 99). The stability of this system is to be understood, not by way of there being laws that hold regarding it, but by way of essential properties of the system. But this gives us only a description of God as a system of items that bear various dependence or causal relationships, not of a personal agent. Insofar as the relevant concepts are strictly functionalist, they do not entail even consciousness. When we turn to religious discourse about God, the functionalist account is not nearly enough: "For the religious life, we need to go beyond that in ways that launch us into the still not sufficiently charted seas of the figurative and the symbolic" (p. 103).
"Referring to God" distinguishes between direct reference and reference by description. Reference by description offers a description that is true only of the referent; direct reference names an object of one's experience. Direct reference to God can occur only if someone experiences God (Alston takes it that some people do). Others who do not themselves experience God can then refer to the being that others have referred to; reference thus spreads throughout a religious community. Direct reference is more basic than descriptive, because if one refers to a being both descriptively and directly, and one learns that the description is false of the being directly referred to, it is the latter that determines what was the actual object of reference. Nonetheless, Alston admits that "reference could always take place via a description (p. 107). A consequence is that it is possible that someone who thinks of God as an omnipotent, omniscient spirit, and one who thinks of God as an impersonal force, may refer to the same being. Alston says that it may be that both are "worshiping the one true God" (p. 116). If so, worship does not require much by way of actual knowledge of God.
There was never any reason to think that a causal theory of reference wedded to a functionalist account of P-predicates would yield significantly positive results regarding description of God. It seems fair to say that in spite of the sophisticated and helpful discussions provided by Soskice and Alston, accounts of religious language that are philosophically articulate and allow for seriously realistic accounts in theology remain more matters on the agenda than they are accomplishments of current work in the field.
