I am not yet so lost in lexicography, as to forget that words are the daughters of earth, and that things are the sons of heaven.

—Samuel Johnson, Dictionary, Preface.

Neither a Gifford Lecturer nor anyone else could give a course of lectures without making use of language, and if his subject is theology his language will be largely theological. That there are peculiar problems concerning theological language was discovered by English-speaking philosphers in the nineteen-thirties, though it had been a commonplace among theologians for centuries. And, although few philosophers today would adhere to the sweeping dismissal by the logical positivists of all metaphysical, ethical and theological statements as neither true nor false but strictly meaningless, most of them would still be unwilling, at any rate in English-speaking circles, to accept theological statements as assertions of matters of fact. It will, I think, therefore be useful, at this early stage in this course, to devote some time to the question of theological language.

The first point which I wish to stress is that, although there are special problems concerning theological language, there are serious and difficult problems concerning language as such. There has indeed been no lack of discussion about language and languages among analytic philosophers in recent years, from the time of Carnap and the Vienna Circle onwards. And behind this, round about the beginning of the century, there was the work of the mathematical logicians, with their disquieting discovery of the famous paradoxes; their heirs today are such persons as Dr Kurt Gödel, with his theorem stating that, in any logical system elaborate enough to be of practical use, it is possible to formulate statements whose truth or falsehood cannot be determined within the system. These philosophers, however, brilliant as they are, have been occupied almost entirely with the internal logical structure of languages (often highly artificial languages at that) or with the behaviour of words and sentences, and have paid remarkably little attention to the fact that human beings, unlike other living creatures, make use of language in order to clarify, order and communicate their thoughts. Their pursuit has in fact been rather like that type of sociological study which limits itself to the observable behaviour of human beings and ignores the whole realm of their mental life, their curiosity, their anxieties, their aspirations, loves and hates. It has interested itself chiefly with the behaviour of words and, even in the phase when its methodology was expressed by the maxim ‘Don't ask for the meaning, look for the use’, the word ‘use’ referred simply to the way in which words were used rather than to the intention of the user. Such a predominantly descriptive concern with language has, of course, its place and, properly used, it can deliver us from a great deal of confusion; but it condemns itself by its very method to leave out the most significant fact about languages, namely their function in the life of intelligent, self-conscious beings, the fact to which Sir Julian Huxley has drawn attention when he describes the evolution of language as ‘the invention of words as symbols for things, in place of sounds as signs for feelings.’1

Right at the start of any discussion about language we must take account of the awkward fact that we have to make use of language whenever we wish to discuss it. Thus any doctrine about language which denies its basic competence as an instrument of intelligent enquiry and communication simply saws off the branch on which it is sitting. The late George F. Woods warned the doctrinal critic ‘to have in mind the perpetual problem of the adequacy of language as an instrument for use in examining the adequacy of language’.2 This warning, salutary as it is for the doctrinal critic, is even more important for the linguistic critic, who, even when treating of the language which he studies in a purely behavioristic way, must claim more than a behavioristic status for his own assertions. It will not do for him to say that his own assertions are expressed in a metalanguage which is not itself subject to the judgments that it makes about the first-order language which is the object of its own study; for, even if it cannot be judged by itself, there is a meta-metalanguage in which it can be judged. We can easily get involved in an infinite regress of a familiar type, but I do not think the difficulty can be resolved by saying that, although there is no archetypal language in which all languages including itself can be judged, any particular language can be judged in the language which is immediately above it in the sequence. For unless, somewhere inside or outside the sequence, language ceases to be simply a dance performed by marks on paper or vibrations in the air and becomes the instrument of intelligent thought it fails to state anything at all and so loses the very end for which it exists. I do not want to involve myself in the intricacies of the higher logic, but simply to urge that, legitimate and illuminating as the logical study of language is, it must not be allowed to exclude the recognition that language is essentially an instrument and a medium of human thought and communication. I do not think that even the later Wittgenstein and his successors have clearly seen that the nature and function of language are not exhausted by a discussion of either its internal logical structure or the speech-behaviour of people who use it. Time and again, when reading discussions about the nature and function of language, one finds oneself wondering whether the writer has ever seriously applied his own theories about language to the language which he himself uses in formulating and defending them.

Now, just because language is an instrument and a medium of thought and communication—that is to say, just because it is something that we use in order to clear our minds and to talk to other people—language can never be adequately understood except with reference to the social and linguistic context in which it is used. Even the most specialised discussion of mathematical topology, atomic nuclear structure or the coding of genetic characters in a living cell assumes the existence of a group of intelligent persons having a common conceptual framework and a common interest; the man who constructs a purely private language in which to talk only to himself must be a very rare person outside mental hospitals. I do not mean to suggest that, without, for example, achieving a perfect identification with the life and mentality of ancient Babylonia, I cannot in any degree understand the Code of Hammurabi, but I do mean that I am likely to misinterpret statements emanating from a culture other than my own unless I can achieve some degree of imaginative identification with it. It is, to give one example, his power to evoke the climate of thought and fife in fourth-century North Africa in the mind of the modern reader that gives Mr Peter Brown's study of St Augustine a value that is absent from many equally learned but less inspiring works on the same subject. To achieve the desired end will, of course, need a good deal of sheer factual information; to understand the Decree of the Council of Chalcedon I must find out as much as I can about the fifth-century context and usage of such terms as prosopon, hypostasis and physis. Nevertheless, more than a purely lexicographical mass of information is necessary; we need to understand the way that the minds of fourth—and fifth-century Christians were working, and the way that their problems were puzzling them, if we are really to understand why they used the kind of language that they did. No arrangement of computers, however intricate, bombinating in vacuo can perform this task of understanding, though they may perhaps assist it; it involves the living thought of living intelligent beings. Nor is this task necessary only in studying the documents of the past; it is just as necessary in trying to understand our own contemporaries to the extent that they live in a different cultural and mental setting from our own. To take one instance, the difficulty which most Europeans and Americans have in taking seriously many of the utterances of the Russian and Chinese governments may arise largely from this cause. More is needed to remove the sense of frustration which accompanies this than the study of Russian and Chinese grammar-books and dictionaries; the trouble is that, as we sometimes significantly say, we do not ‘know how their minds work’.

If what I have been saying is true, there is a real sense in which meaning is fundamentally mental and is logically prior to language. To adapt a famous phrase of Pope John XXIII, meaning is one thing and its verbal formulation is quite another. And, for reasons which I shall shortly indicate, no linguistic statement can do justice to the complexity of the situation to which it refers. This does not mean that all statements are equally and indifferently inadequate, so that it does not really matter what we say about anything. They are indeed all inadequate but some are less inadequate than others. And their degree of inadequacy will at least partly depend on the degree to which those who utter them share a genuine mental and linguistic community. It is said that there was once a group of American linguistic scholars who were convinced that the structure of Indo-European languages was so different from that of the real world as to make them useless for accurate philosophical purposes. Some of the Amerindian (Red-Indian) languages were alleged, in contrast, to be almost ideal. It might seem strange, in view of this, that the scholars in question, instead, as might have seemed reasonable, of writing books in Amerindian languages about Indo-European languages, wrote books about Amerindian languages in English. No doubt the reason for this was that they were writing for English-speaking people and not for Red Indians; it therefore seemed preferable to write somewhat misleading books for their fellow-scholars than logically perfect books for a non-existent public. I do not know whether this story is true, but I have seen an article by an English philosopher in which it was alleged that the English language was, in virtue of its sentential structure, unsuitable for theological purposes, and that theology could be conducted much better, if not absolutely ideally, in Classical Chinese. The writer fell indeed into a simple logical fallacy on the second page of her argument, but this might no doubt be taken as confirming her thesis of the unsuitability of English for theological purposes; though in fact the fallacy occurred when the English had been re-stated in the symbolism of Principia Mathematica and would have been unlikely to occur if the writer had been content with the ordinary language of the palace and the bus-queue. The significant fact, however, is that her article was written in the disreputable English language and not in the respectable Chinese. This was, no doubt, because she was writing in the context of an English-speaking linguistic community; she was trying to convert the English rather than to confirm the Chinese. And this does, I think, emphasise the fact that language cannot in practice be dissociated from its social context.

In the theological realm we have, I think, come to recognise that more is involved in understanding the classical statements of the Christian Church than the kind of knowledge that can be acquired from grammar-books and dictionaries. I will illustrate this by a passage from the Gifford Lectures of the late Leonard Hodgson, which is instructive both positively and negatively. He wrote as follows:

I have described the history of human thought as God seeking to make himself known to man through minds inevitably conditioned by the forms of thought and linguistic usage of their age and culture. This conditioning has to be taken into account in our study of the books of the Bible just as much as in that of the conciliar creeds, the patristic writings, and the works of scholastic and Reformation divines. We have to learn all we can about their authors’ ways of thinking and linguistic self-expression in order to discover in what way their insight into truth was coloured by this outlook and to what extent there was miscolouring which needs to be discounted. We ourselves have to think and speak as twentieth-century Western Europeans, in terms of the thought and language of our time and place. As we study the writings of the past the question we have always to be asking is; what must the truth have been and be if men who thought and spoke as they did saw it and spoke of it like that?3

Hodgson clearly attached great importance to the last sentence of this passage, since, with slight variations, he repeated it three times in another place;4 and I agree that he was making a valuable point. I am, however, doubtful whether he appreciated its full implications. For, if we try to answer the question ‘What must the truth have been and be if men who thought and spoke as they did saw it and spoke of it like that?’, in what terms are we to answer it? It would, of course, be possible to try to answer it in terms of thirteenth-century Western scholasticism or of eighth-century Hinayana Buddhism, and this might indeed be a useful academic exercise, though we who were performing it would be for the most part twentieth-century Europeans and not thirteenth-century scholastics or eighth-century Buddhists and we should be in danger of producing only a rather self-conscious pastiche. This is, in any case, not what Hodgson had in mind. As he himself recognised, we can only answer his question in terms of the thought and language of our own particular time and place, as twentieth-century Western Europeans.5 And what guarantee have we that those terms are more adequate than those of some earlier Christian period? If the suggestion is that the ways of thinking and speaking of all times and places are equally adequate, though those of our time and place are more suitable for us, this would surely need to be shown; prima facie it does not seem likely that the language of, say, an aboriginal tribe in Australia today will be as adequate a vehicle for Christian theology as the Greek of St Gregory Palamas. Or is the suggestion that, although they are not absolutely adequate and final, our twentieth-century thought and language are so much more developed than those of other times and places that, for all practical purposes, we can take them as if they were? Again this would need proving and it would be surprising if it were true, since our modern Western European words and concepts were not devised with the needs of Christian or any other theology in mind. Whatever may be the limitations of the traditional theological forms of thought and expression and however difficult it may be to make them intelligible and inspiring to our contemporaries, they were not simply adopted by the Church as given and fixed categories but were transformed and moulded by the Church for the purposes for which it needed them. This should be recognised even by writers like Dr Leslie Dewart, who holds that in the interplay between Christianity and Hellenism Christianity came off second best. Or did Hodgson perhaps mean that truth itself, and not merely its linguistic and conceptual expression, is relative and fluid, as Dewart has alleged in his books The Future of Belief and The Foundations of Belief? One of the variants of Hodgson's question might seem to suggest this: ‘If the truth about God's revelation in Christ be such that those men saw it and wrote of it like that, what must it be for us?6 I find it difficult to believe, in view of his writings as a whole, that Hodgson was quite such a relativist as that. It seems to me more likely that, in his anxiety to express the unchanging truth in terms that would appear intelligible and relevant to his contemporaries, he tended over-hastily to give our twentieth-century modes of thought and expression a specially privileged status. Another way in which to make my point is by remarking that, if Hodgson's question is legitimate, it must also be legitimate to ask ‘What must the truth have been and be if men who think and speak as we do think and speak of it like this?’ And this is patently unanswerable by us, except in a sheer tautology, since we have only our own thought and speech by which to answer it. The real lesson to be learnt from Hodgson's question is, I suggest, that, while we have only our own human languages in which to assert truths, the truths themselves are something other than our assertions of them and are not in themselves linguistic at all.7 That is to say, a statement is true to the extent that, in the social and linguistic context in which it is uttered, it accurately describes some feature of the real or the logical world, and it can perform this function even if it can never reproduce with complete adequacy the feature's total complexity. Language has inevitably a certain looseness of fit at both the objective and the subjective poles. In extreme cases this may render it practically useless; the statement ‘A plane leaves at 3.15’ will be minimally informative if it is divorced from all reference to the day, the place of departure and the destination; there are far too many situations to which it could apply. And the word ‘gut’ will fail to identify its object unless we know whether it is used in the context of German or of English discourse. This inevitably imprecise character of language should act as a warning, but it does not deprive language of its function as a means of thought and expression; I can state with perfect accuracy that a battle was fought at Waterloo in A.D. 1815, even if I am unable to specify the identity of the opposing forces, the details and the outcome of the conflict and the day of the week on which it was fought.

We may add at this point that it is because of the contextual character of language that there is danger in making use of the compendia of documents from past ages which the more erudite and diligent scholars compile for the benefit of their less learned or lazier colleagues; such works, for example, as Denzinger's Enchiridion, Mr J. Stevenson's New Eusebius and Creeds, Councils and Controversies, Dr C. K. Barrett's New Testament Background and Mr H. Bettenson's Documents of the Christian Church. In spite of the helpful notes with which they are furnished, these indispensable and time-saving works can easily mislead the unwary user. For, especially in the case of the briefer items, they are inevitably read out of their context; both the emphasis and sometimes the very meaning of a verbal formula may thus be completely misunderstood. Thus, to give one example, the doctrine of a limbus infantium appears in a very different light if it is recognised that it was directed not against people who held that babies dying unbaptised might share in the joys of heaven but against those who held that they were inevitably condemned to the flames of hell. Furthermore-and this is an even more dangerous limitation-the user is entirely at the mercy of the principle, or lack of principle, of the compiler in selecting his material; however conscientious the latter may be, his own views as to what is important will influence his choice. Thus it has been said, not without truth, that what in practice decides the influence of a papal or conciliar decree in the Roman Catholic communion is not, in the last resort, the authority with which it was promulgated but the value which it appeared to have in the eyes of Denzinger and his successors and which determined its inclusion in or exclusion from the indispensable Enchiridion. The selection of material and its arrangement in such a compendium inevitably provides it with a new and artificial context different from that in which it originated; both the overall picture and the nuance of an individual item may thus be seriously distorted.

We must look a little more closely at the question of the relation between a statement and the objective situation which it describes. It has sometimes been supposed that if a sentence is true it must picture precisely and in detail the situation to which it refers and that a language can be adequate only if sentences can be formed in it which picture precisely and in detail the whole universe or some part of it. Because no existing language measures up to this requirement philosophers have sometimes tried to construct, or at least to lay down principles for constructing, ideal languages which would satisfy it. Such an ambition was characteristic of the sense-datum epistemology of Russell and Broad in the period between the wars and, in a different way of the work of such members of the Vienna Circle as Carnap and Neurath, though it could never be more than an ambition. Wittgenstein held a similar view when he wrote his celebrated Tractatus Logico-Philosophicus, though he abandoned it later.8 He thought that, because a language consists of discrete elements, the world must consist of discrete ‘atomic facts’, to which the language's elementary statements would be in biunique, or one-to-one, correspondence; only if this were the case would it seem possible for language to ‘picture’ the world. It is, however, pretty clear that the world and language are not related in this way. So far from consisting of atomic facts which can be put in one-to-one correspondence with atomic linguistic statements, the world (at least as it appears to us) is a continuous, multidimensional, dynamic entity, while language is discrete, unidimensional and static. And even if we interpret current physical theory as showing that the universe consists of discrete ‘fundamental particles’, probably finite in number, so that its ostensible continuity is only an appearance, it will still be of such vast complexity that it cannot possibly be pictured biuniquely by any set of statements that we could ever manage to formulate.

What I am in fact arguing is that it is a mistake to suppose that the function of language is, even ideally, to picture reality by being in biunique correspondence with it, except in certain limited and highly specialised cases. Its function is rather to be an instrument by which intelligent beings are able both to clarify their own thinking and to communicate their thoughts to one another. Many of these thoughts, of course, have as their content the world in which the thinker lives and the people whom he meets in it; he is not concerned merely with his own subjectivity. Language is therefore essentially an activity of human beings living in a real world and in community, and it can be understood only as such. Any attempt to reduce it either to a set of labels attached to facts or to a set of noises emitted more or less systematically by human bodies in response to stimuli ignores its basic nature as an invention and instrument of intelligent persons. What is necessary for the fruitful use of language is the sharing of some degree of social and cultural life on the part of those who try to communicate by means of it. This calls for a genuine effort of understanding and not merely a grammatical, syntactical and logical expertise; it is, as I have said, largely through the lack of such understanding and of the effort to attain it that many political discussions in the contemporary world end in deadlock and frustration. It is significant that this situation is often described by saying that the several parties ‘simply do not speak the same language’; this means more than that they are uninstructed in one another's grammar, syntax and vocabulary. It is parallel to the phrase previously quoted, that ‘we do not know how one another's minds work’. And it is because it does make this much-to-be-desired effort at mutual understanding that so much value can be seen in such a work as the Marxist scholar Roger Garaudy's From Anathema to Dialogue, with the appended essays by the Catholic priests Karl Rahner and Johannes Metz.

We must in fact avoid two opposite extremes. One is that of assuming that the ideal to which language should strive is simply that of picturing facts by statements logically isomorphous with them; this was the view of the earlier Wittgenstein. For it, an ideal language would be one in which every object and relation had been given a label and all the thinking could be done by a computer. The other extreme is that of minimising the function of language as expressing and communicating truth, or of redefining truth so that it no longer consists of the conformity of thought with reality but of an ever-changing vocalisation of man's subjective experience as a changing inhabitant of a changing world. This is the view defended by Dr Dewart with his repeated denunciations of ‘Hellenism’. There are, needless to say, elements of truth in both the extreme positions. There are occasions and areas of reality to which the proto-Wittgensteinian ideal is proper, such as that of mathematics and much of the positive sciences. There are also legitimate uses of language—some of which are indeed of great importance—in which it expresses emotions, announces decisions of action, recommends policies and so forth. But all these uses are subsidiary to, and emergent from, man's basic character as being who is capable of acquiring knowledge and so of apprehending truth. Mr J. R. Lucas has argued, quite convincingly in my opinion, that the essential functions of intelligent mentality cannot, even in principle, be simulated by any machine, however elaborate.9 And, next to his possession of intelligence itself, language is man's most important endowment both for extending his own knowledge and for communicating it to others. He is, of course, neither infallible nor impeccable; he may unwittingly make mistakes and he may deliberately tell lies. If it is alleged, as it has been by some, that language has taken the form that it has simply because of its survival-value for man in the evolutionary process and not because of its capacity to express truth about reality, the first part of the assertion may be accepted but the implied denial in the second part is to be rejected. For unless man's thinking and the language which he used to clarify and express it had in fact developed in conformity with reality, their survival-value for him would indeed have been small. We need only ask ourselves what chance man would have had of survival if his thought and speech had borne no relation to the facts of the world around him. On the other hand, his speech would presumably have been extremely rudimentary, if not non-existent, had he lived in complete isolation and not in company with others of his kind. Down to the present day it is by living as members of a speaking community that children learn to speak. And all this may be briefly summed up in the simple statement that man makes use of language because he is an intelligent and social being.

This has been a far from exhaustive discussion of the nature of language and the problems that it raises, but it must suffice for the present. All that I have tried to do has been to defend the common-sense position that, in spite of these problems, language can, if due care is taken, be a reliable, though limited, medium of thought and communication. And I repeat that anyone who denies this position will have to use language in order to argue against it. I must devote the remainder of this lecture to considering the type of language with which we shall be specially concerned, namely theological language.

Much of the language which theology employs does not differ in any relevant respect from that of ordinary discourse and it raises no special problems. Some of it, however, purports to describe, or at least to refer to, realities beyond those that are immediately perceived by the senses; in this it has a feature in common with metaphysics, in the traditional sense of that term. It goes, however, even beyond metaphysics in that the reality with which it claims to be specially concerned—what it denotes by the word ‘God’—transcends not merely the realm of the immediately sensible, but the realm of all finite objects. The grounds on which it has been alleged that statements containing the word ‘God’ are meaningless, and that therefore there can be no being, God, to which they apply, are well known to philosophers and theologians at the present day; they range from the crude complaint of the logical positivists, that theological statements fail to conform to what is in fact an arbitrary definition of meaningfulness, to Professor Flew's oft-repeated assertion that such statements must be meaningless since the people who affirm them refuse to admit that they could be shown to be false by any conceivable circumstances whatever. I shall not attempt to recapitulate the vast mass of writing on this matter. It is difficult to separate discussion of the meaning-fulness of theological assertions from discussion of the arguments put forward in support of their truth; I shall have more to say about this later on. I will, however, mention in passing two very useful works on the subject, Dr Frederick Ferré's Language, Logic and God and Dr John Macquarrie's God—Talk, though I must add that the most penetrating discussion which I know is contained in a brief, and unfortunately not easily accessible, article by Fr W. Norris Clarke, S.J., entitled ‘Analytic Philosophy and Language about God’, in the volume Christian Philosophy and Religious Renewal.10 The position which I myself wish to maintain is the very simple one that, since meaningfulness means the capacity to be understood, the only way in which to discover whether a statement or a concept is meaningful is to see whether people can understand it. The apparently tautologous character of this definition of meaningfulness is in fact its strong point; for any attempt to define it in terms of a more fundamental concept must perforce assume that that concept is itself meaningful. The famous attempt of Professor A. J. Ayer to define meaningfulness by his verification principle11 died the death of a simple suicide by a thousand qualifications, for, in spite of all its death-writhings, it never managed to conform to its own standard of meaningfulness.12

In defining meaningfulness, however, in terms of capacity to be understood, I am, of course locating it in the setting of a linguistic community. Apart from this context language is just marks on paper or vibrations in the air. God-talk, can of course, like any other talk, be mere psittacism; people can talk in their sleep and gramophones can talk in an empty room. Nevertheless, talk can be the medium of intelligent communication within a community, and only within a community can it be intelligible. For a linguistic empiricist to declare, with all the fervour of a Welsh evangelist, that he cannot give any intelligible meaning to the sentence ‘God exists’ may indicate nothing more than that he has never made a serious effort to enter into the linguistic community of those who affirm it. Religious people can, of course, talk nonsense without recognising it, when they are weary or off their guard, but then so too can linguistic analysts; nobody has yet made sense of the famous sentence ‘Pirits karulise elatically’, because there is (so far as is known) no linguistic community within which the words that compose it have been used in intelligent conversation. Nevertheless, I would maintain that, when religious people, including theologians, speaking carefully and responsibly, make statements containing the word ‘God’, they do understand what they are saying. This is, as they would themselves be the first to admit, very surprising and mysterious; and to outsiders who have some arbitrary and extrinsic criterion of meaningfulnessness it may be downright stupefying. Like other statements these may need a good deal of clarification. What, however, one should never do is to deny facts on the grounds of preconceived theories; and I maintain that, as a fact of experience, theological statements are meaningful, in the context in which theological talk occurs. Whether they are in addition true is, of course, another question. It will occupy our attention a great deal later on.

It is indeed remarkable that we should be able to make meaningful statements about a being who is alleged by the very people who make the statements to be altogether transcendent to the finite world and radically different from it; for all the words and concepts that we use in this discourse are, or are constructed from, words and concepts that are normally used in talking about finite beings of everyday experience. Christian philosophers have in fact elaborated a great mass of theory to deal with this surprising situation; it is known as the Doctrine of Analogy. I have devoted a book to the subject and I shall not attempt to recapitulate it here.13 It may just be worth while mentioning, for those who are interested in the technicalities of the subject, that a suggestion which I make in that book, to the effect that a satisfactory doctrine of analogy cannot be built up on analogy of proportionality alone but requires in addition analogy of attribution, seems to have been confirmed by more recent Thomist scholars, who find in Aquinas himself the notion of analogy of causal participation.14 I have in any case come to see more clearly that one is already loading the question if one puts it in the form ‘How can terms which in their normal and natural application refer to finite beings refer analogically to God?’ The primary datum, though many modern empirical philosophers may not be ready to admit this, is that the terms apply both to finite beings and to God, and the question is how this dual application can be explained. From an ultimate standpoint a theologian will indeed assert that their primary application is to God and only their secondary application is to finite beings, but this will come at the end of the story, not at the beginning. To anticipate future argument, I will merely say here that the explanation of this dual application of terms to God and to finite being is that God and finite beings are in a definite causal relation. If God and his creation were as totally unrelated as some opponents of natural theology affirm, no terms could apply to both, however analogically. I am not, of course, asserting that the dual application of terms can be made the basis of a proof of God's existence; with all due respect to supporters of the ontological argument, I do not think that one can make that kind of transition from language to reality. What I do think, on the other hand, is that only the notion of a God who is related to the world—and who is related to it in a very particular way—can make the fact of this dual applicability of terms intelligible.

A final point of importance is this. Believers not only believe that God exists; they also believe a number of things about him. In consequence, as Fr Norris Clarke has pointed out, the word ‘God’ is used with a good deal of latitude, so that it may be legitimately asked whether ‘God’ is a proper name or a description and, if the latter, what description it is.15 The word ‘God’ is not alone in this respect. Suppose I speak about ‘the Queen’ and somebody says to me ‘What do you mean by “the Queen”?’ In order to make it plain that I am referring to the Queen of England and not to the Queen of Holland or Greece, I may reply ‘the lady whose head appears on our fourpenny stamps’ or ‘the legitimate successor of King George VI’. If I adopt the former definition, then the statement ‘The Queen's head appears on the fourpenny stamps’ is an analytic proposition and ‘The Queen is the legitimate successor of King George VI’ is a synthetic one; and conversely if I adopt the latter. If I adopt some other definition, both statements will be synthetic.

This particular instance may appear to be trivial, but the matter is far from trivial when statements are made about God. Since, as I have said, believers believe a number of things about God, it is not always clear how many of these things are explicitly or implicitly included in the definition which they assume of the word ‘God’. In consequence, when they use the word ‘God’ they may not all be meaning precisely the same things by it. For one man the word ‘God’ may simply signify the ultimate ground of the universe, and the question whether that ground is personal or impersonal may seem to him of little importance; for another the question may be so important that he would refuse to apply the word ‘God’ to the ground of the universe unless he was convinced of its personal character. Thus to these two men the question ‘Is there a God?’ may have very different meanings, and if this is not recognised confusion and frustration may arise. As a matter of common honesty it is essential not to take a minimal definition of the word ‘God’ in arguing for God's existence and subsequently to slip over quietly into a notion of God that is stored with all the fullness of Christian belief and devotion. Consider the following dialogue:

Titus: ‘God’ means the ultimate ground of the universe.

Balbus: Agreed.

Titus: The universe must have an ultimate ground.

Balbus: Agreed.

Titus: Therefore God must exist.

Balbus: Agreed. But what is God like?

Titus: Well, he must be loving, wise and all-powerful.

Balbus: How do you know that?

Titus: Well, I wouldn't call him ‘God’ unless he was.

Balbus: Shouldn't you have thought of that before?

This kind of intellectual cheating is by no means fictitious, and it is all the more insidious for being usually unintentional and unconscious. Some critics have detected it in Dr J. A. T. Robinson's persuasive book Honest to God.16 There is, of course, no harm in using the word ‘God’ with different definitions in different contexts. When the Christian theologian uses it he will often mean by it the Holy Trinity of Father, Son and Holy Spirit, who are one divine unity of love. It is clear to me that St Anselm used a different definition of ‘God’ in the Proslogion from that which St Thomas Aquinas assumed in the Summa Theologiae; this does not mean that St Anselm and St Thomas believed different things about God. Two things are, however, essential. The first is to be quite clear about the meaning we are giving the word in any particular discussion; the second is to refrain from slipping from one meaning to another in the course of the same argument. If we fail in this we shall constantly confuse the two questions that need to be kept distinct: ‘Is there a God? and ‘What is God like?’

• 1.

  Evolution in Action, Penguin ed., p. 36, Cf. Naom Chomsky: ‘Thesestudies [sc. of animal communication] simply bring out even more clearly the extent to which human language appears to be a unique phenomenon, without significant analogue in the animal world’ (Language and Mind (1968), p. 59).

• 2.

  ‘Doctrinal Criticism’, in F. G. Healey, ed., Prospect for Theology, p. 83.

• 3.

  For Faith and Freedom, II, pp. 4f.

• 4.

  ibid., I, pp. 87, 88.

• 5.

  I suspect that Hodgson would have included citizens of North America under the description ‘Europeans’, as the South African government did until it discovered them making use with the Bantu of the facilities labelled ‘Non-Europeans’.

• 6.

  ibid., p. 88 (italicisation mine).

• 7.

  Mr J. A. Baker (The Foolishness of God (1970), pp. 364f) has criticised Hodgson's methodological question on the ground that ‘truth does not come to men clothed in words, it comes to them as words; and when as far as possible we know what the words meant to them, then as far as possible we know what the truth was to them.’ I am not quite clear whether this criticism is the same as mine or not.

• 8.

  Cf. J. O. Urmson, Philosophical Analysis: its Development between the Two World Wars (1956).

• 9.

  ‘Minds, Machines and Gödel’ in Minds and Machines, ed. by A. R. Anderson; and replies to objectors, ‘Satan Stultified’, The Monist, LII (1968), pp. ‘45ff; The Freedom of the Will (1970), pp. 124ff.

• 10.

  Edited by George F. McLean, O.M.I. (Washington, D.C., 1967), pp. 39ff.

• 11.

  Language, Truth and Logic, 2nd ed., 1946.

• 12.

  Dr Alvin Plantinga remarks, at the end of a careful investigation: