RELIGIOUS LANGUAGE

Utterances made in religious contexts are of many sorts. In the performance of public and private worship men engage in acts of praise, petition, thanks, confession, and exhortation. In sacred writings we find historical records, dramatic narratives, proclamations of law, predictions, admonitions, evaluations, cosmological speculations, and theological pronouncements. In devotional literature there are rules of conduct, biographical narratives, and introspective descriptions of religious experience. Philosophical discussions of religious language have concentrated on a restricted segment of this enormous diversity, namely, theological statements, that is, assertions of the existence, nature, and doings of supernatural personal beings.

There are two reasons for this emphasis. First, the crucial problems about religious language appear in their purest form in theological statements. If we consider a petitionary prayer or a confession, what is puzzling about it is not the act of petition or confession, but the idea of addressing it to God, and God answering it. It is the concept of communication with a supernatural incorporeal person that seems unclear. And this lack of clarity is most apparent in the statement that there exists a God who communicates with men in various ways. We may say that the difficulties in understanding other forms of religious language all stem from obscurities in statements about God.

The second reason for philosophical concentration on theological statements lies in the fact that the philosophy of religion is primarily concerned with questions of justifiability, significance, and value. And it has generally been supposed that whether religion is a justifiable form of human activity largely depends on whether there are sufficient grounds for accepting the theological statements on which it is based. Christianity is a justifiable institution if and only if we are warranted in accepting the proposition that the world is created and governed by an omnipotent, perfectly good personal deity who has revealed himself to men in the Bible. Thus the philosophy of religion is largely taken up with examining the grounds of religious statements. And it is when we do this that we become most acutely aware of the puzzling aspects of religious language. When we make a determined effort to decide whether it is true that God created the physical universe, it is difficult to avoid realizing how unclear what we are saying is, what implications it has, what it logically excludes, and what would count for or against it. Thus the philosophical investigation of religious language focuses on those indeterminacies in theological statements that hamper attempts to find rational grounds for acceptance or rejection.

Meaning of Theological Predicates

Most philosophers who have concerned themselves with the problem have located the difficulties of religious language in the predicates of theological statements. (What does "good" mean in "God is good"?) It may seem that we should start with the subject of the statement, with the concept of God. But there is really no alternative to starting with the predicates. For the only way to make clear what one means by "God" is to provide an identifying description, such as "the creator of the universe"; and to understand that phrase one must understand the predicate "created the universe" as applied to God. Theological predicates can be divided into negative (infinite, nontemporal, incorporeal) and positive. The positive predicates can be concerned either with attributes (good, wise, omniscient) or with actions (makes, forgives, speaks, watches over). Negative predicates present no special difficulty, but in themselves they are clearly insufficient to give any positive conception of the deity. Of the positive attributes we shall concentrate on attributions of action, partly because action terms pose more severe problems, partly because other attributes are dependent on them. (To say that God is wise is to say that he acts wisely; if we cannot understand what it is for him to perform one or another action, we cannot understand the attribution of wisdom to him.)

derivation and application

When one reflects on the use of predicates in theological statements one comes to realize two fundamental facts: (1) this use is necessarily derivative from the application of the predicates to human beings and other observable entities; (2) the theological use of predicates is markedly different from the application of predicates to human beings.

Theological predicates are derivative primarily because it is impossible to teach theological language from scratch. How would one teach a child what it means to say "God has spoken to me" without first making sure that the child knows what it is for a human being to speak to him? In order to do so one would have to have some reliable way of determining when God was speaking to him, so that when this happens one could say to him, "That is what it is for God to speak to you." And even if we admit that God does speak to people from time to time, there is no way for one person to tell when God is speaking to another person unless the other person tells him, which would require that the other person have already mastered the theological use of language. Hence there is no alternative to the usual procedure of teaching the theological use of terms by extension from their application to empirically observable objects.

As for the difference in the use of predicates as applied to God and to human beings, there are many ways of seeing that the terms cannot have quite the same meaning in both cases. If, as in classical Christian theology, God is conceived of as not in time, then it is clear that God's performance of actions like speaking, making, or comforting is something radically different from the temporally sequential performance of actions by human beings. St. Thomas Aquinas in his famous discussion of this problem based the distinction between the application of predicates to human beings and the application of predicates to God on the principle that God is an absolute unity and that, therefore, various attributes and activities are not distinguishable in God as they are in men. But even if we allow God to be temporal and straightforwardly multifaceted, we are left with the corporeal-incorporeal difference. If God does not have a body, it is clear that speaking, making, or comforting cannot be the same thing for God as for man.

This leaves us with a serious problem. We must show how the theological use of these terms is derived from their nontheological use. Until we do, it will be unclear just what we are saying about God in such utterances. The usual way of dealing with this problem is by cutting out the inapplicable portions of the original meaning of the terms, leaving the remainder for theology. Thus, since God is incorporeal, his speaking cannot involve producing sounds by expelling air over vocal cords. What is left is that God does something that results in the addressee having an experience of the sort he would have if some human being were speaking to him. The nature of the "something" is deliberately left vague. Since God is a pure spirit, it will presumably be some conscious mental act; perhaps an act of will to the effect that the addressee shall have the experience of being told such-and-such. More generally, to attribute any interpersonal action to God is to attribute to him a purely mental act that has as its intended result a certain experience, like the one that would result from such an action on the part of a human being.

This account may throw some light on the content of statements about God, but religious thinkers have become increasingly dissatisfied with it. For one thing, it represents theological statements as metaphysical speculations and does little to illuminate the ways they fit into religious activity. Having postulated a pure immaterial substance performing mental acts that, miraculously, have effects in human experience, how do we go about getting into communication with this immaterial substance? Why should it be worshiped at all, and if it should, why in one way rather than another? Moreover, this line of reasoning is not helpful in our efforts to verify theological statements. It offers no hints on how we might determine whether our statements are true, or even whether there is such a being that performs the actions in question.

Verifiability of Theological Statements

Recent discussions have concentrated on the problem of verifiability. In the last few decades a great many philosophers have come to accept some form of the "verifiability theory of meaning," according to which one is making a genuine factual assertion, a real claim as to the way the world is, only if it is possible to conceive of some way in which what he is saying can be shown to be true or false by empirical observation. Applying this theory to theology, it has been argued that since an empirical test is in principle impossible to carry out for statements about a supernatural incorporeal personal deity, these statements cannot be regarded as straightforward factual assertions, but must be interpreted in some other way.

John Wisdom in his influential essay, "Gods," analogizes the function of theology to the following situation. Two people return to a long-neglected garden and find some of the old flowers still surviving among the weeds. One suggests that some gardener has been caring for the plot, and the other expresses doubt about this. On investigation, it turns out that no one in the vicinity has ever noticed anyone working on the garden. Moreover they discover that gardens left to their own devices often take this form. But the first man does not abandon his hypothesis. Instead he expresses his belief that someone who is not discernible by the senses comes and cares for the garden, carrying out designs he and his companion do not fully grasp. At this point the first man has modified his "gardener" hypothesis to the point at which it is no longer susceptible to empirical confirmation or refutation. No matter what is or is not discovered empirically, he will continue to hold it. In this case it seems plausible to say that he is no longer expressing a belief about actual objective events. If he were, he would be able to imagine, however inadequately, some way in which the existence or nonexistence of these events would be revealed to our experience. He is, rather, expressing a "picture preference." It is rewarding to him to think of the situation as if a gardener were coming to take care of the flowers. If beliefs about God are equally refractory to empirical test, it would seem to follow that they too must be interpreted otherwise than as straightforward matters of fact. (Wisdom, however, does not commit himself to this conclusion.)

In considering the "verificationist" challenge to theology, we must scrutinize both premises of the argument: (1) theological statements are not susceptible to empirical test; (2) if they are not empirically testable they cannot be construed as factual assertions that can be assessed as true or false.

are they empirically testable?

The question of whether theological statements are subject to empirical test is quite complicated. If we rule out mystical experience as a means of observation, then it is clear that statements about God cannot be tested directly. But science is full of hypotheses about unobservable entities—electromagnetic fields, social structures, instincts—which verificationists accept as meaningful because they can be tested indirectly. That is, from these hypotheses we can draw implications that can themselves be tested by observation. The question is whether directly testable consequences can be drawn from theological statements. We can phrase this question as follows: Would we expect any possible observations to differ according to whether there is or is not a God? It would clearly be unreasonable to require of the theologian that he specify a set of observations that would conclusively prove or disprove his assertions. Few, if any, scientific hypotheses could meet that requirement. The most that could reasonably be demanded is that he specify some observable states of affairs that would count for or against his assertions.

One thing that makes this problem difficult is the fact that on this point religious belief differs at different times and places. Supernatural deities have often been thought of as dealing in a fairly predictable way with contingencies in the natural world and human society. Thus in many primitive religions it is believed that the gods will bring abundant crops or victory in battle if they are approached in certain ways through prayer and ritual. Even in as advanced a religious tradition as the Judeo-Christian, it is believed that God has certain fixed intentions that will result in prayers being answered (when made in the right spirit and under proper conditions) and will result in the final victory of the church on earth.

It would seem that such expectations provide a basis for empirical test. Insofar as they are fulfilled, the theology is confirmed; insofar as they are frustrated it is disproved. However, things are not that simple. Even in primitive communities such tests are rarely allowed to be decisive; the empirical implications are hedged around with a variety of escape clauses. If the ritual dances are held and still the crops fail, there are several alternatives to abandoning traditional beliefs about the gods. Perhaps there was an unnoticed slip somewhere in the ritual; perhaps devils were conducting counterrituals. More sophisticated explanations are employed in the more advanced religions. For example, God will answer prayers, but only when doing so would be for the true good of the supplicant.

Moreover, as science develops, religion comes to be more concerned with the personal life of the worshiper and less concerned with prediction and control of the course of events. Among religious intellectuals today such predictions as are still made are clearly not testable in practice, because of their lack of specificity ("all things will work together for the good for those who love God"), their enormous scope ("everything in the world contributes to the development of moral personality"), or their inaccessibility ("after death we shall see God face to face"). Nevertheless, it seems that within religion there are strong barriers to completely divorcing belief in God from the expectation of one event rather than another; and so long as there is some connection of belief with testable predictions, however tenuous, it would be a mistake to think of religious statements as absolutely unverifiable in principle.

are they assertions of fact?

As to whether a statement that cannot be empirically tested must not be construed as an assertion of fact, a theologian might well challenge the application of the verifiability theory to theology. If God is supernatural, we should not expect his behavior to be governed by any laws or regularities we could hope to discover. But then we could never be certain that, for example, the statement that God loves his creatures would ever imply that a war should have one outcome rather than another. This would mean that, according to the verifiability theory, it would be impossible for us to make any statements, even false ones, about such a being. But a theory that would prevent us from recognizing the existence of a certain kind of entity, if it did exist, would be an unreasonable theory.

Nonassertive Interpretations

Be that as it may, a number of philosophers have been so impressed by these difficulties over verifiability that they have tried to construe theological utterances as something other than straightforward factual assertions. Attachment to the verifiability theory is not the only motivation behind the development of such theories. There are those, like George Santayana, who, without holding that theological sentences are factually meaningless, are convinced that as factual assertions they are false, but still are unwilling to abandon traditional religious discourse. They feel that somehow it has a valuable function in human life, and in order to preserve it they are forced to reinterpret it so that the unwarranted factual claims are expunged. Still another motivation is the hope that this will contribute to the resolution of the problem mentioned earlier, that of specifying the way predicates are used when they are applied to God. As we saw, attempts to give an illuminating definition of theological predicates have not been wholly successful, and this can be taken to indicate that a different sort of approach is needed.

One such line of investigation takes sentences as its units rather than words. It focuses on the kind of linguistic act performed when theological sentences are uttered, rather than on the meaning of words in theological contexts. Instead of asking what "forgives" means when applied to God, we ask what linguistic action is performed when one uses the sentence "God forgives the sins of those who truly turn unto him." It is this sort of question one is asking when one wonders whether theological sentences make factual assertions and, if not, what they are used to do. If we could answer this question we would have made sufficiently clear how words are being used in theological sentences without having to define special senses for constituent words.

Nonassertive interpretations can be divided into four groups. Statements about God have been interpreted as (1) expressions of feelings of various sorts; (2) symbolic presentations of a variety of vital aspects of experience, from natural facts to moral ideals; (3) integral elements in ritualistic worship; (4) a unique kind of "mythical" or "symbolic" expression, not reducible to any other use of language.

expressions of feeling

Theological utterances have been interpreted as expressions of feelings that arise in connection with religious belief and activity. Thus we might think of "God made the heavens and the earth" as an expression of the sense of awe and mystery evoked by grandeurs of nature; of "God has predestined every man to salvation or damnation" as an expression of a pervasive sense of helplessness; and of "God watches over the affairs of men" as an expression of a sense of peace, security, at-homeness in the world. This is "poetic" expression rather than expression by expletives. It is like expressing a sense of futility by saying "life's a walking shadow" rather than like expressing futility by saying "Ah, me." That is, the feeling is expressed by depicting a situation that might naturally evoke it; a sense of security, for instance, is evoked by some powerful person looking after one.

symbolic presentations

Symbolic interpretations of religious doctrines have been common for a long time. The story of Noah and the Flood has been regarded by many Christian thinkers not as an account of actual historical occurrences, but rather as a symbolic way of presenting certain religiously important points—that God will punish the wicked, but will also, under certain conditions, show mercy. Many of the traditional ways of speaking about God have to be taken as symbolic. God cannot literally be a shepherd or a rock. The shepherd functions as a symbol of providence and the rock as a symbol for God's role as a refuge and protection in time of trouble. A symbol in this sense is some (relatively) concrete object, situation, or activity that can be taken to stand for the ultimate object of discourse through some kind of association, usually on the basis of similarity. We speak symbolically when what we literally refer to is something that functions as a symbol.

In the traditional use of symbolic interpretation it is, necessarily, only a part of theological discourse that is taken as symbolic. For if we are to hold that the symbolic utterances are symbolizing facts about God, we will have to have some way of saying what those facts are; and we cannot make that specification in symbolic terms, on pain of an infinite regress. But we are now considering views according to which all theological discourse is symbolic, which means that if we are to say what is being symbolized it will have to be something in the natural world that can be specified in nontheological terms. The most common version of such a view is that theological utterances are symbolic presentations of moral ideals, attitudes, or values. This position has been set forth most fully and persuasively by George Santayana, and in a more up-to-date form by R. B. Braithwaite. According to Santayana every religious doctrine involves two components: a kernel of moral or valuational insight, and a poetic or pictorial rendering of it. Thus the doctrine that the physical universe is the creation of a supremely good personal deity is a pictorial rendering of the insight that everything in the world is potentially usable for the enrichment of human life. The Christian story of the incarnation, sacrificial death, and resurrection of Jesus Christ is a way of making the point that self-sacrifice for others is of supreme moral value. It is worthwhile embodying these moral insights in theological doctrine because this vivid presentation, together with the systematic cultivation of feelings and attitudes that accompanies it, provides a more effective way of getting across the insights than would a bald statement.

The way in which interpretations of the first two kinds throw light on the theological use of predicates is analogous to the way in which one explicates the use of words in poetic metaphors. If we consider the metaphor in "sleep that knits up the raveled sleeve of care," it is clear that "knit" is not used simply to refer to a certain kind of physical operation. This utterance has quite different kinds of implications from "she knit me a sweater," in which knit does have its usual sense. In the metaphoric statement, knit is used in its usual sense to depict a certain kind of situation that, as a whole, is presented as an analogue of the effect of sleep on care. The only way of effectively getting at the function of the word knit is by seeing how the whole phrase "knits up the raveled sleeve" is used to say something indirectly about sleep.

In the first two of the four kinds of nonassertive interpretation we are examining, theological statements are essentially metaphors. And if they are correctly so regarded, we get nowhere if we extract the word made from the sentence "God made the heavens and the earth" and try to say what it means by itself. What we have to do is take the picture presented by the whole sentence and see how it functions as a way of expressing a feeling of security, or as a way of presenting the insight that everything in the world can be used to enrich human life.

ritualistic interpretation

The ritualistic interpretation of theological discourse can best be introduced by citing the reply of an intellectually sophisticated high-church Anglican to a question from an agnostic friend. The question was, "How can you go to church and say all those things in the creed?" The reply: "I don't say them; I sing them." In the view under consideration, the corporate practice of worship is the native soil from which talk about God springs. Talk about the attributes, doings, and intentions of a supernatural personal being has meaning as a part of the practice of worship and is puzzling only when it is separated from that context. If we think of an utterance like "God made the heavens and the earth" as the expression of a belief about the way things in fact originated and then wonder whether it is true or false, we will be at a loss. To understand it we have to put it back into the setting where it (or rather a second-person correlate, such as "Thou, who hast made the heavens and the earth") does its work. In that setting, these words are not being used to explain anything, but to do something quite different.

Unfortunately, proponents of this view have never been very clear about what this "something different" is. The clearest suggestion they give is that the talk about God serves to provide an imaginative framework for the conduct of worship. It articulates one's sense that something important is going on, and it helps to indicate the appropriateness of one response rather than another. In speaking of the sacrament of communion as the reenactment of the self-sacrifice of an omnipotent personal deity who took on human form, and in conceiving of it as a cleansing and renewing incorporation of the substance of such a deity, one provides for the activity a pictorial framework that records and nurtures the felt solemnity of the occasion and the attitudes and aspirations kindled by the ceremony. This position presupposes, contrary to the usual view, that ritual worship has an autonomous value, apart from any theological foundation. It is generally supposed that a given ritual has a point only if certain theological doctrines are objectively true. But in the ritualistic interpretation, theological doctrines are not regarded as statements about which questions of truth or falsity are properly raised. Since these doctrines depend for their significance on the ritual, it is supposed that the ritual has some intrinsic value in forming and giving expression to valuable sentiments, feelings, and attitudes.

myths

Ernst Cassirer has developed the notion that the basis of religious discourse lies in a unique "symbolic form" that he terms "mythical." He maintains that it is found in purest form in the myths of primitive peoples and is based on a way of perceiving and thinking about the world that is radically different from our accustomed mode. In the "mythical consciousness" there is no sharp distinction between the subjective and the objective. No clear line is drawn between symbol and object, between wish and fulfillment, between perception and fantasy. Again, no sharp distinction is made between the object itself and the emotional reaction it evokes; emotional response is taken to be an integral part of the environment. As a result none of our familiar standards of truth or objectivity are applicable. What is most real is what arouses the greatest intensity of emotional response and, particularly, what is felt as most sacred. (The sacred-profane distinction is the fundamental contrast.) The mythical consciousness carries its own special organizations of space and time. For example, there is no distinction made between a position and what occupies it; every spatial position is endowed with a qualitative character and exerts influence as such.

It is the view of Cassirer, and of followers such as Susanne Langer, that sophisticated theology represents an uneasy compromise between mythical and scientific modes of thought, and as such cannot be understood without seeing how it has developed from its origins. It is basically a mythical view of the world, given a "secondary elaboration" in a vain attempt to make it acceptable to the rationalistic consciousness; judged by rationalistic standards it is not only groundless, but meaningless.

Mysticism

Philosophers and theologians in the mystical tradition have put forward versions of this fourth kind of interpretation that do not regard theology as a manifestation of cultural lag. To the mystic the only way to communicate with God is through mystical experience, and this experience reveals God to be an ineffable unity. He can be directly intuited in mystical experience, but since there are no distinctions within the absolute unity of his being, and since any statement we can make predicates of him one thing rather than another, for example, wisdom as distinguishable from power, no statement can be true of him. The most we can do in language is to direct our hearers to the mode of experience that constitutes the sole means of access. Proponents of this view sometimes speak of theological language as "symbolic," but this differs from our second type of theory in that here there is no way to make explicit what it is that the theological utterances symbolize, and it is therefore questionable whether we should use the term symbol. A symbol is always a symbol of something. In fact it is difficult to make clear just what, on this view, religious utterances are supposed to be doing. They are said to "point to," "adumbrate," or "indicate" the ineffable divine reality, but all too often these expressions remain uninterpreted.

In recent years two interesting attempts have been made to develop this position further. W. T. Stace, in his book Time and Eternity (1952), considers the chief function of religious language to be the evocation of mystical experience, or faint echoes thereof. This seems at first to be a subjectivist account, with the deity omitted, but, as Stace correctly points out, it is an axiom in the mystical tradition that no difference can be found in mystical experience between subject and object, and on these grounds Stace refuses to make the distinction. Although Stace goes along with the mystical tradition in regarding mystical experience as ineffable, he departs from this official position to the extent of giving some indications of the aspects of this experience that different theological utterances evoke. "God is truth" evokes the sense of revelatoriness, "God is infinite" the sense of all-inclusiveness, "God is love" the blissful, rapturous character of the experience, and "God is one" the absolute unity of the experience and the sense of the dissolution of all distinctions.

Paul Tillich, although not squarely in the mystical tradition, is faced with similar problems in the interpretation of religious language. He holds that theological doctrines "symbolize" an ultimate reality, "being-itself," about which nothing can be said literally except that it is metaphysically ultimate. In attempting to clarify the function of religious language, Tillich develops the notion that it is an expression of "ultimate concern," a complex of devotion, commitment, and orientation, focused on something nonultimate—a human being, a nation, or a supernatural deity. Religious statements, which literally refer to such relatively concrete focuses of ultimate concern, express the sense of the sacredness such objects have as "manifestations" of being-itself. But just what it is for such an object to be taken as a "manifestation" or "symbol" of being-itself, Tillich never makes clear.

The basic weakness in these mythical and mystical interpretations is the failure to present any clear hypothesis concerning the function of religious language. Even Cassirer's ideas on "mythical thought" have never been developed to the point of clarifying what contemporary religious believers mean when they talk about God. The other positions are more intelligible, and they all base themselves on important aspects of the use of language in religion. But it seems that each, by inflating its chosen aspect to sole authority, has killed the goose that lays the golden eggs. There is no doubt that in talking about God, religious people express feelings of various sorts, present moral ideals, and articulate what is going on in ritual. But it is not at all clear that they would be using this kind of language if they were not convinced of the truth of the statements they make. Why should I express a feeling of security by saying "God made the heavens and the earth" unless I believe, or at least have some tendency to believe, that as a matter of objective fact the physical universe owes its existence to the creative activity of a supernatural personal deity? Still more, why should I take on the complex of attitudes and activities that goes along with this assertion unless I believe it to be true?

The statement-making function is the cornerstone on which all the other functions depend. And if one is convinced that theological statements are either false or meaningless and still wants to hold to traditional religious formulations, one may propose a reinterpretation of theological utterances as expressions of feeling or symbolizations of natural facts. But a proposal for adopting a certain interpretation must be distinguished from a claim that the proposed interpretation correctly reflects the way doctrines are commonly understood.

It would seem that talk about God is much more complex than is recognized by any of the existing theories. The brief discussion given above of empirically testable implications illustrates this point. Theological sentences perform a great many closely interrelated linguistic functions. In saying "God, who created the world, watches over the affairs of men," the believer is committing himself to a certain general view of the ultimate basis of the world, giving voice to certain, perhaps very indefinitely specified, expectations as to how things will ultimately turn out, expressing a basic sense of security in life, committing himself to approach God in prayer and ritual in one way rather than another. And these functions are intimately dependent on each other. What is needed is a description of the relationships among these functions, one sufficiently complex to match the complexity of the subject matter.

Bibliography

St. Thomas Aquinas's historically important discussion of the "analogical" character of theological terms is found in Question XIII of Part I of the Summa Theologiae. For further discussions in the Thomist tradition see Cajetan, The Analogy of Names, translated by E. A. Bushinski and H. J. Koren (Pittsburgh: Duquesne University, 1953), and E. L. Mascall, Existence and Analogy (London, 1949). Recent discussions by practitioners of analytical philosophy are to be found in Antony Flew and Alasdair MacIntyre, eds., New Essays in Philosophical Theology (New York: Macmillan, 1964), and in Basil Mitchell, ed., Faith and Logic (London: Allen and Unwin, 1957). See also Alasdair MacIntyre, "The Logical Status of Religious Belief," in Metaphysical Beliefs (London: SCM Press, 1957); John Wisdom, "Gods," in Essays in Logic and Language, edited by Antony Flew, first series (Oxford, 1951); C. B. Martin, Religious Belief (Ithaca, NY: Cornell University Press, 1959); and I. T. Ramsey, Religious Language (London: SCM Press, 1957). Various kinds of symbolic interpretations of religious statements are presented in George Santayana, Reason in Religion (New York: Scribners, 1905); R. B. Braithwaite, An Empiricist's View of the Nature of Religious Belief (Cambridge, U.K., 1955); W. M. Urban, Language and Reality (New York, 1939); Edwin Bevan, Symbolism and Belief (Boston: Beacon Press, 1957); and Philip Wheelwright, The Burning Fountain (Bloomington: Indiana University Press, 1954). Ernst Cassirer's views on mythical language are set forth most completely in Vol. II of Philosophie der symbolischen Formen, 3 vols. (Berlin: Cassirer, 1923, 1925, 1929), translated as The Philosophy of Symbolic Forms by Ralph Manheim (New Haven, CT: Yale University Press, 1952, 1955, 1957); and in more concentrated form in Sprache und Mythos (Leipzig: Teubner, 1925), translated as Language and Myth by S. K. Langer (New York: Harper, 1946). For other versions of the view that religious language constitutes an autonomous mode of discourse, see W. T. Stace's Time and Eternity (Princeton, NJ: Princeton University Press, 1952) and Paul Tillich's Systematic Theology (Chicago: University of Chicago Press, 1951–1963) and Dynamics of Faith (New York: Harper, 1957).