Unpublished 252. Logic and Language (typed)

[1956. Logic and Language, The Gordon Review Vol. II No. 1 Feb.]



In contrast with the Greek philosophies of Plato and Aristotle, in contrast also with the earlier modern systems of Spinoza and Hegel, the late nineteenth and twentieth centuries have seen the development of avowedly irrationalistic viewpoints. Nietzsche, Dewey, Heidegger, and Sartre are examples. The themes of this contemporary irrationalism have been applied to the problem of religion and have infiltrated even moderately conservative Christian thought.

Admittedly, Christian history in other centuries too has produced certain anti- intellectualistic strains of pietism, mysticism, and stubborn obscurantism. Logic has been called “cold,” and “human” reason inadequate. The syllogisms of Aristotle are pagan, or else they are dry bones devoid of living flesh. At the present time these sentiments received added support from the peculiarly modern investigations of language and semantics and are presented to us in a new dress. For purposes of exposition, analysis and criticism, articles from recent issues of The Christian Scholar and the Gordon Review are here selected.

The Christian Scholar is published by the Commission on Christian Higher Education of the National Council of Churches. In the issue of September 1955, Geddes MacGregor has an article entitled “The Nature of Religious Utterance,” and John A Hutchinson writes on “The Religious Use of Language.” Their common point of view rather than any differences they may manifest is the present subject of study.

MacGregor opens with Croce’s assertion that “all language is metaphorical, or none is,” and soon follows with Urbans’ rejection of literalism and his conclusion as to “the inevitably metaphorical and symbolic character of all language.” MacGregor does not wish to be held to the position of Croce and Urban altogether, but he seems to accept the thesis that all religious language is metaphorical or symbolic. If this be so, religious utterances must be evaluated in a very different manner from the usual analysis of logical propositions.

In support of his position, MacGregor gives some examples of the use of language; and it will be our duty to determine whether or not these admitted cases require his conclusions. First, he refers to a college choir whose Jewish, Unitarian, and Quaker members, majoring in political science or anthropology, were singing a Medieval hymn. Few if any of them understood the concepts of the hymn, and yet their words communicated the concepts to those persons in the audience who had the proper understanding. Similarly, a child does not understand marriage when he reads the last sentence of a fairy tale.

These examples, particularly the second, are supposed to militate against a literal understanding of language because not even all adults have the same understanding of marriage. This word has “levels” of meaning, and some words have many levels, others fewer. There is no literal meaning. However, it does not seem to me that MacGregor’s examples prove what he intended. Granted, neither the child nor some adults know all the propositions that may be truly asserted of marriage as an institution. Some adults do not know all the propositions or theorems that may be truly asserted of a triangle. But the ignorance of these theorems does not entail an ignorance of the definition of triangle or marriage; much less does it prove that there is no literal meaning whatever to these words.

Another illustration MacGregor gives is that of a very ordinary preacher preaching a very ordinary sermon. But the sermon or a sentence of it becomes a vital message to someone in the congregation. The person is converted and his life is changed. Once again the words conveyed more meaning than the speaker intended, and hence, argues MacGregor, the meaning could not be literal. But why not? Could not the literal meaning of a sentence or two recall themes that had lain dormant in the hearer’s mind? Could not even the literal meaning itself point out a new way of life? How can such an instance be made to show that all religious language is metaphorical or symbolic?

Finally, MacGregor asserts that the theological proposition “God is omniscient” is never as satisfactory as the liturgical statement, “O my God, who knowest all things.” For MacGregor religious utterance “is always in the second person singular.” Of course, the third person sentence and the second person phrase, given here, are not precisely equivalent. But the differences do not derive from the person of the verb. If MacGregor had written the first sentence as “My God is omniscient,” he would have had a third person sentence which is the exact equivalent of a second person phrase. It may not be a “satisfactory” mode of address, for it is not a mode of address at all; but this is not to say that it is not entirely satisfactory for a creedal statement. Whatever may be the difference between second and third person verbs, it is not at all clear why second person verbs must be metaphorical rather than literal.

The Hutchinson article in the same periodical develops the theory somewhat more clearly and more profoundly. The thesis is that “religion in all its range and variety consists of symbols.” Where MacGregor hesitated, Hutchinson says expressly that “all language is metaphorical … Every common noun is a kind of dead metaphor. But religious terms of words are metaphorical in a further and distinctive sense.” That Jesus Christ is sitting at the right hand of God the Father Almighty and that the Lord is my shepherd are examples of examples of analogy or metaphor. MacGregor at least hints that the Virgin Birth is also a metaphor. To support his view, Hutchinson sketches a religious epistemology which is based on images – a sort of mental idolatry – and which is assimilated to art and mythology. God always (note the always) speaks to man through images, and “religious experience is a process of being hit by such images.”

There is not space to analyze this unacceptable epistemology, but I would like to say that while Hutchinson may be describing his own religious experience, he is not describing mine. His sweeping generalization is simply not true to fact.

One objection, however, Hutchinson feels obliged to answer. If myth is unavoidable in religion, some explanation is required as to the choice of myths. One person chooses Greek mythology, another Christian mythology. Doubtless such choices are often made unreflectively; but Hutchinson thinks that it is possible to make a rational choice of myths. The basis for such a rational choice is the adequacy of the myth to explain the facts of existence as we confront them in daily life and action.

Before proceeding to the defense of an opposing view of religion and language, I wish to offer a preliminary criticism of Hutchinson and then to outline the views of a third exponent of non-literal language and non-Aristotelian logic.

It seems to me that Hutchinson fails rationally to justify a choice among myths. If myths were literal truths, one might be more adequate than another. The Greek myth of Zeus’ method of producing rain might be considered more adequate, or less adequate, than the myth about the windows of heaven, attributed to the Hebrews. But if these stores are both mythological and symbolic, simply symbolic of the literal fact that it rains, it is hard to judge what adequacy might require. A literal statement from Aristophanes’ Clouds might explain, but a myth explains nothing. Furthermore, if the language is symbolic, it seems clear that one symbol, before historic events have fixed its meaning, is as good as another. Today the swastika symbolizes National Socialism, and the hammer and sickle, Communism; but at the start there was no reason why the communists could not have chosen the swastika and Hitler the hammer and sickle. To push this preliminary criticism one step further, we might ask the question, What is a religious symbol the symbol of? The cross no doubt is the symbol of Christ’s crucifixion; but when the crucifixion is mentioned, most people understand it literally. Now, if all language is symbolic, it would be necessary to express the meaning of a symbol by another symbol, and this by another. How can this regress be of any value unless sometime, sooner the better, we come upon a symbol that symbolizes a non-symbolic meaning?

The third article with the same general viewpoint, “Language and Theology,” by Richard K. Curtis was published in the Gordon Review of September, 1955. The phenomena to which this article appeals are somewhat different from the preceding, and the scope of the whole is considerably wider. Curtis follows John Dewey in substituting the know-how of craftsmen and slaves for the know-why of Aristotle. He gives what appears to be a mechanistic theory of sensation. He ventures the judgment that Indo-European grammar emphasizes free-will, while the grammar of the Eskimos is fatalistic. With these matters I do not propose to deal. Nor am I concerned with the Aristotelian metaphysics that he believes with some plausibility to be inherent in Aristotelian logic. It is specifically the logic and the language on which attention is to be centered.

Language, for Curtis, at least literal, Aristotelian language, is a verbalization that is substituted for experiential knowledge. It is an unnatural way of thinking by which propositions must be severely strained to include everyday answers to questions. Curtis then gives certain instances, such as, all tomatoes are vegetables, all the works of Poe are magnificent, and all Socialists are Communists. The point of these examples is difficult to make out. Perhaps the three assertions are false; but the falsity of several particular sentences does not disqualify judgments of the all a if b type; nor would they prove that propositions must be severely strained in order to talk of botany, literature, or politics. The law of identity, a is a, and the principle of disjunction also fall under Curtis’ condemnation. For a reason he does not state, he objects to the assertion, If I am an historian, I either follow F. J. Turner or I do not. But apparently his conscience pricks him a little, for he admits that it is legitimate to say, Either I am a Christian or I am not. But by what principle can he reject disjunction in history and retain it in religion? If disjunction or excluded middle is a fallacious form of thought, it can never be used legitimately; but if it is legitimate in one case, all the misapplications of slovenly thinkers will not serve to invalidate its proper universality. Toward the end of his article Curtis again repudiates the laws of identity, contradiction, and excluded middle by making the somewhat vague assertion that they “have less and less meaning in an increasingly scientifically oriented society.”

Declamation without argument and the enumeration of alleged mis-applications do not refute Aristotle. To defend Aristotelian logic and literal language I wish to reproduce one part of Aristotle’s argument. Only after the opponent has clearly indicated where and how this argument breaks down, has he any right to discard the law faces in question. Neither Curtis. Hutchinson. nor MacGregor aces the challenge squarely.

Aristotle begins by asking his opponent to make any assertion he chooses. Perhaps he will say that three is an odd number or that Socrates is a man, or even that a religious term is symbolic. Whatever the assertion may be, it is always in the form of a is b. Now, may I ask the authors of these three articles, is this a mistake? Can they make any assertion whatever without using, perhaps disguising, but veritably using the form a is b? Since all of the assertions in their articles have this form, they have not yet demonstrated the possibility of avoiding it. Aristotle then continues: when a person says that a is b, he does not mean that a is not b. But in case this is too easy a dismissal of the opponent, let some attention be paid to the predicate term.

Suppose then that the opponent has said, three is odd, Socrates is a man, or a is b. Aristotle wishes to show that the predicate term, odd, man, or b. has a single, definite meaning. Of course, words in any language bear several connotations or meanings (called levels by MacGregor), and this ambiguity frequently produces blunders in the application of logical principles. This ambiguity is also the probable cause of most of the opposition to traditional logic. But the fact that a word has several meanings does not damage Aristotle’s contention, unless the meanings be infinite in number. If the word man has ten different meanings, a separate term can be invented for each: and then each term would have a single meaning. If, however, the term man had an infinite number of meanings, it would mean nothing in particular, and the assertion, Socrates is a man, would mean nothing. In other words, language, conversation, and argument would be impossible. For if a word is to convey a meaning, it must not only mean something; it must also not mean something. If it had an infinite number of meanings, if the term man had the meanings of all the words in the dictionary, it would be useless in speech. In fine, if man means not-man, the sentence, Socrates is a man, means nothing. But those who deny the law of contradiction identify man and not-man. Those who deny the law of excluded middle assert that Socrates is neither man nor not-man. What they say is nonsense. Nothing sensible can be said without using the laws they deny.

I have given here but a part of Aristotle’s argument. Those who wish to reject logic should point out, if they can, what flaws the argument contains. They should also state clearly what laws govern their own use of language. If they reject excluded middle and deny that three is either odd or even, let them tell what three is. If they go further and deny the laws of validity, so that the premise, No triangles are squares, would not imply the conclusion, No squares are triangles, let them tell whether any inferences are possible or not, and by what criteria they distinguish valid inferences from fallacies. And all this they must do without making surreptitious use of the principles they deride.

Unless the arguments of Aristotle can be squarely met, the traditional logic will serve as a basis for intellectualism and for a literal use of language in religion as well as in every other subject. Croce’s dictum that all language is metaphorical or none is, must be denied. It is a fallacy to suppose that the sentence, Jesus was born of a virgin, cannot be literally intended, on the ground that the other phrase, Jesus is seated at the right hand of God the Father Almighty, cannot be other than metaphorical. Contrary to Croce, a logician, one who thinks accurately, must maintain that metaphors exist only because words have literal meanings. To understand Milton’s leaden stepping hours of time, one must know that boots weighted with lead reduce the speed of anyone who wears them. The point of similarity may be wisely or stupidly selected; but unless a word has a literal meaning, the comparison which the metaphor attempts cannot be made. If there were no literal lions, the phrase, the lion of the tribe of Judah, would be totally meaningless.

One final consideration may also be adduced in defense of language and logic. The opposite viewpoint sometimes looks at language as a “social” phenomenon which, though it serves roughly for most daily necessities, is inadequate for accurate metaphysical, theological, or even scientific purposes. But whatever reason is put forward for the inadequacy either of language or its conceptual content, it would seem that Christian theism can and indeed must take a different position.

Christianity maintains both that language is a gift from God and that society is completely under his control. This is not to say that everybody uses language adequately or that nobody makes a blunder in logic. But whatever misapplications of principles are made, there is no inherent defect in language or logic that justifies their being regarded as inadequate for their purposes. And these purposes include the formulation of a detailed theology, literally intended. Strange as it may at first seem, the one sense in which all language may be called symbolic guarantees its adequacy. Words are conventional symbols. As sounds or letters they have no meaning of their own. There is nothing in a given word or sign to prevent it from standing for any reality you choose. The words God and Dieu are equally satisfactory signs or symbols, and in some unknown tongue Naig would do as well. But because all words are symbolic in this sense, their meaning in speech can be literal when desired. Admittedly, there is a problem respecting the communication of ideas, but it is a problem that St. Augustine discussed rather well without denying the possibility of literal meaning.

We have of course been born into a society where the conventional symbols have long ago been selected, and you and I have little opportunity to coin new terms. Neither did Moses in Egypt. Yet the doctrine of inspiration requires us to believe that God gave Moses an Egyptian education as well as a fluency in Hebrew, so that when the time came Moses was fully prepared, spiritually, mentally, and linguistically to express accurately and adequately the message God had for his people.

There are indeed metaphors and symbols in that revelation; but “Thou shalt not steal’ is literal.