If the philosophical friends of religion pride themselves on anything, it is on their conception of themselves as the guardians of the propositional element in religious belief. They are unhesitating in their criticisms of any analysis of that belief which seems to them to deny that propositional element. In contrast to any such an analysis, they claim to have a robust conception of truth.

I want to discuss one frequent way in which the propositional view of religious belief is expressed. The claim is that the point of religious practice is logically dependent on the truth of a proposition said to be independent of that practice. It matters little whether that proposition is said to be arrived at by the assessment of probabilities, by some form of trust, by being immediately formed as the result of an experience, or by being presupposed; but what does matter is that the proposition expresses and captures the fact of the matter.

Wittgenstein said that one of his aims was to teach us how to move from disguised nonsense to patent nonsense (Wittgenstein 1953, para. 464). In an influential article, ‘What Nonsense Might Be’, Cora Diamond criticises what she calls a natural view of nonsense in favour of one which she thinks can be found in Frege and Wittgenstein (Diamond 1991). Her article has been discussed illuminatingly by Lars Hertzberg in ways which bear on our discussion in this lecture. With Hertzberg’s help we can be brought to see why it is tempting to hold that religious practice is founded on the truth of the proposition ‘God exists’ which is said to be independent of it.

Hertzberg asks us to consider two sentences; ‘Scott kept a runcible at Abbotsford’ and ‘Caesar is a prime number’. On the natural view of nonsense, the first sentence is nonsense because the word ‘runcible’ has not been given any meaning. The second sentence is said to be nonsense because although all the words have a meaning, the meaning of the sentence is wrong. Hertzberg says that the first sentence has too little meaning, whereas the second sentence, as it were, has too much. He writes, ‘Cora Diamond rejects the latter part of this claim …It is due, she argues, to overlooking Frege’s principle that we cannot discuss the meaning of a word in isolation. Only as it occurs in a sentence does a word have logical properties’ (Hertzberg 2001, p. 91). In the second sentence ‘Caesar’ cannot mean what it does in ‘Caesar crossed the Rubicon’, and ‘prime number’ cannot mean what it does in ‘53 is a prime number’. But that means that the sentence has no meaning. Diamond argues that the only reason we think otherwise is ‘that we fail to take seriously another of Frege’s strictures, that of always distinguishing between the psychological and the logical. From the fact that in hearing (the sentence “Caesar is a prime number”) most of us will automatically think about the founder of Imperial Rome it does not follow that this is what the word “Caesar” must refer to in this sentence. What the word means there depends on how it is used in the sentence, not on what anyone happens to be thinking about’ (Hertzberg 2001, p. 91). Diamond then goes on to say:

In support of this view, Diamond cites Wittgenstein, ‘When a sentence is called senseless it is not as it were its sense that is senseless. But a combination of words is being excluded from the language, withdrawn from circulation’ (Wittgenstein 1953, para. 500).

The way in which Hertzberg takes up the discussion at this point is highly relevant to our concerns. He points out that whereas Diamond insists that a word taken in isolation does not make sense, she does not take the same view of sentences. Hertzberg argues, ‘It seems to me that considerations analogous to these that might persuade one to reject the natural view of nonsense should also make one doubtful about the possibility of asking whether a sentence, taken by itself, does or does not make sense …After all, the sentence “Caesar crossed the Rubicon” might as well describe, say, the dealings between a Mafia operator, Caesar, and a crime syndicate known as the Rubicon’ (Hertzberg 2001, p. 92).

We can make the same point about the sentence ‘God exists’. It need not have a religious application in a given context. Let us continue Hertzberg’s use of examples from the underworld. Imagine a feud between underworld factions who think it time for the Godfather (‘God’ for short) to step aside. One faction becomes impatient and decides to assassinate him. They think they have been successful when, in fact, they have botched the attempt. Later, news of the failure has to be conveyed to the head of the faction by one of his henchmen. He says to him, in ominous tones, ‘God exists’. So the sentence taken in isolation does not convey its sense without further ado. This is as true of ‘God exists’ as it is of ‘Caesar crossed the Rubicon’.

Having reached this conclusion, a question still remains. Why do philosophers think it a truism to say that religious practice depends on the truth of the proposition ‘God exists’, a proposition said to be logically independent of the practice—a view which clearly entails that the proposition has a meaning taken in isolation? Once again, Hertzberg’s discussion proves helpful.

What makes us think that the proposition ‘Caesar crossed the Rubicon’ does not need to be placed contextually, but has a meaning which is immediately apparent? Hertzberg replies, ‘Part of what makes us overlook this may be a peculiarity of the example. Julius Caesar is one of those individuals, like Napoleon or Shakespeare, or phenomena like the weather, that one can bring up at the start of a conversation almost anywhere and at any time without having to prepare the ground for it. If the example had been instead “Smith is a prime number”, the dependence on what was being said on the context would be more immediately striking. Wittgenstein speaks, in Philosophical Investigations, §117, about the mistake of regarding the sense of the word as an atmosphere that it carries with it into every kind of application. In these terms, it might be said that a name like ‘Caesar’ comes as close as any word can to carrying its context with it like an atmosphere’ (Hertzberg 2001, p. 92, f.n. 3). Thus, even in a sentence like ‘Caesar is a prime number’ we think that ‘Caesar’ carries its normal meaning, despite the fact that the sentence has no application, because that meaning is thought of as the atmosphere which the name ‘Caesar’ carries with it.

If this confusion is present with the name ‘Caesar’, how more tempting is the confusion where the word ‘God’ is concerned. Thus, in thinking that ‘God exists’ makes sense, independently of any religious practice and application, it is confusedly assumed that ‘God exists’ carries its meaning with it, independent of practice and application, as though it were an atmosphere accompanying the word ‘God’ or the sentence ‘God exists’. It is in this way that a confused claim in contemporary philosophy of religion can be thought to be so obviously the case as to be a truism by those who propound it. It is in this way, too, that those who oppose the confused thesis are regarded with incredulity for doing so.

How does all this bear on the question whether we would be better off if religion were taught by means of pictures without any developed set of propositions to which one had to assent? I can imagine certain philosophers saying that this would be logically impossible, because the pictures would presuppose that certain propositions were being assented to prior to painting them. I do not think this is true. In Part Two of the Investigations, p. 178, Wittgenstein writes,

So if the pictures can perform the same service as the words, it cannot be argued that the pictures presuppose the spoken doctrines. What is true is that the pictures could not be what they are without a language or culture in which they are recognised as pictures, but it does not follow that there would have to be in the language a parallel in verbal or written form of what the pictures say. At this point, I’m going to purloin comments made by Peter Winch in his paper ‘Meaning and Religious Language’, which appeared originally in Reason and Religion edited by Stuart Brown, but was reprinted later in his own collection, Trying to Make Sense. His example has to do with a tribe who have nothing which could be identified as verbal expressions of religious belief, but who practise certain rituals with respect to the mountains in which they dwell. I want to take over his comments but substitute rituals in relation to pictures for the practices he discusses.

Let us suppose that when a member of a tribe dies, certain practices are carried out at his funeral. The rituals may also include contemplation and prostration before certain pictures. Similarly things happen at other important moments in the life of the tribe – at marriage, on the occasion of a birth – perhaps every child is given a picture, when an adolescent is initiated into adult life, and so on.

I want to say that in relation to the pictures the tribe expresses something like reverence or religious awe. I would be able to recognise different degrees of devoutness in relation to the pictures. Maybe certain pictures are rejected on their production and perhaps there are punitive measures against those who keep them and continue to look at them.

Suppose now that we add talk of gods to these practices. Talk of the pictures as pictures of God or holy pictures. The talk I want to say, is not presupposed by the practice, but grows out of it. We should be wary of saying that the talk explains the practices with respect to the pictures, as the motives of important people in the tribe would explain how portraiture emerged among them. Here, ‘because people wanted their likeness painted’ explains the practice, because the existence of the motives can be explained independently of the practice. In fact, the primitive picture practice I am imagining might condemn the attempted emergence of portraiture as a desecration, and this would be shown in the destruction of any such efforts and the penalties involved in hoarding an image of oneself.

Two points much be made before I complete the story of this primitive scene. First, it may be asked (as Sydney Schoemaker did ask of Winch) why things couldn’t be the other way round. Why couldn’t the stories of the gods in the pictures come first, without any of the practice in relation to the pictures? My answer is the same as Winch’s answer—What would make the stories or the propositions religious? Divorced from practice we wouldn’t know how to take them. This is the logical point of the first two sections of the paper. Second, the story about the relation of the prepositional talk to the practice surrounding the pictures, does not mean that the reality of the gods consists in the fact that people speak in this way, but that speaking in this way brings out the grammar or logic of that reality. Winch made this point to allay fears expressed by John Hick—fears which, I believe, he still harbours.

This brings us to the next stage of the discussion—even accepting the primitive scenario I have mentioned, does it follow that a picture-practice is spared from the confusions which may inflict those who emphasise the propositional element in religious belief? I do not see how it does, since, as I have said, the pictures have their significance in a surrounding culture, just as propositions have their sense in a surrounding practice. That being so, similar confusions may arise. If ‘culture’ stands to ‘picture’, as ‘proposition’ stands to practice, why should this not be so?

Let me develop Winch’s general scenario in my own further direction. Let us suppose that despite opposition, portrait painting does develop among the tribe. Some time later, scepticism, hitherto unknown, begins to develop in relation to the sacred pictures. Discussion of them beings to develop. The scepticism takes the form of questioning the status and validity of the religious pictures, by comparing them with the status of portraits. Whereas, in the case of the portrait, one could not only see the portrait, but also the person portrayed, this was not possible with the religious pictures. Scepticism mounted: were the pictures portraits at all? If not, what are they about? Are they just the creations of an artistic imagination?

Has this tribe died out? No, we live in he midst of them. They are called philosophers.

Wittgenstein refers to complaints that may be made against the religious pictures. In his Lectures and Conversations on Aesthetics, Psychology and Religious Belief he says, ‘The word “God” is amongst the earliest learnt—pictures and catechisms, etc. But not the same consequences as with pictures of aunts. I wasn’t shown [that which the picture pictures]’ (p. 59).

Someone may say, ‘Well, of course not. That’s because God is transcendent.’ But that (a) makes it look as if God can’t be pictured simply because he is too far away, which misses the logical point Wittgenstein is making about the kind of picture he is talking about; (b) if the protest is that since God is transcendent, he cannot be captured in the picture, then ironically, philosophers who pride themselves on being realists deny the eminently realistic claim that God is in the picture, by seeing what kind of practice it is. And here, as in the case of propositions, this can only be brought out by emphasising the surroundings in which the picture has its sense. Wittgenstein brings this out with respect to portraits.

The logical point being made is that this method of projection does not belong to religious pictures (obviously, I’m not referring to religious portraits). Hence, it doesn’t make sense to speak of them, apologetically, as though the method of projection does apply, but that there is an excuse to explain if it has not captured God—he is transcendent, too high to capture etc. As Wittgenstein says, the comparison is absurd. He imagines someone complaining about the religious pictures, ‘“Of course we can only express ourselves by means of pictures”’, and responds, ‘This is rather queer …I could show Moore the pictures of a tropical plant. There is a technique of comparison between picture and plant. If I showed him the picture of Michelangelo and said, “Of course, I can’t show you the real thing, only the picture” …The absurdity is, I’ve never taught him the technique of using this picture.’

If we concentrate on the negative part of these comments, we will come away with the point that religious pictures are not like portraits. But don’t forget the positive point. The complainant is saying, ‘I only saw the picture, not the real thing’. In rejecting this, Wittgenstein is insisting that in the picture you did see the real thing. You saw God in the picture. That would explain why the destruction of a religious picture might be regarded as a desecration.

Wittgenstein is referring to Michelangelo’s painting of the Creation of Adam. He says, ‘Take “God created man”. Pictures of Michelangelo showing the creation of the world.’ He adds, humorously, ‘In general, there is nothing which explains the meanings of words as well as a picture, and I take it that Michelangelo was as good as anyone can be, and did his best, and here is the picture of the Deity creating Adam’ (p. 63).

He then makes his logical point, ‘If ever we saw this, we certainly wouldn’t think this the Deity. The picture has to be used in an entirely different way if we are to call the man in that queer blanket “God”, and so on. You could imagine that religion was taught by means of these pictures’ (p. 63).

Wittgenstein is too confident by far when he says that if we ever saw Michelangelo’s paintings we wouldn’t take them to be portraits or diagrams. He had forgotten the philosophers, since that is precisely how J.L. Mackie does take them. He thought that the believer needn’t commit himself to every detail of Michelangelo’s depiction of the Last Judgement, but, he said, unless Michelangelo was close (remember Wittgenstein’s—he’s as good as anyone can be) Mackie wondered what the Last Judgement could be. Mackie says, in The Miracle of Theism, p. 3, ‘I am saying only that talk about a last judgement can be understood literally’. But if by ‘literal’ use we mean ‘standard’ use, why should we assume that the standard use is always the empirical, or, in this case, that of a portrait or a diagram? Imagine the absurdity of the claim that all paintings must be portraits or diagrams, and that this must apply to pictures of God.

The analogy between pictures and language continued to influence Wittgenstein in his later work, despite his criticisms of the picture theory of propositions in the Tractatus. He says in the Investigations I 522:

He was thinking of depictions such as a painting of the crowning of Napoleon as representational pictures, and say Cezanne’s painting of the Card-Players as a genre painting. He continues:

Wittgenstein replies in the next paragraph 523: ‘ I should like to say “What the picture tells me is itself”.’ And in the next paragraph, 524 he writes, ‘Don’t take it as a matter of course, but as a remarkable fact, that pictures …occupy our minds’.

Now let us go back to the following comment in the Lectures and Conversations: ‘It is quite clear that the role of picture of Biblical subjects and role of the picture of God creating Adam are totally different ones. You might ask this question, “Did Michelangelo think that Noah in the ark looked like this, and that God creating Adam looked like this?” He wouldn’t have said that God or Adam looked as they look in this picture’ (p. 63). But not because they look differently, of course.

So the pictures, like the propositions, may be misused – this is because although the pictures and the propositions may be in the foreground, their application is in the background – that is, easily forgotten. What is more, if the pictures are given the wrong application, Wittgenstein says that they can cause great harm – by which he didn’t mean philosophical trouble – but I’ll say no more of that now.

The genre picture says itself. Is a religious picture a representational picture or a genre picture? If it is the latter, it says itself. Well, what does Michelangelo’s Creation of Adam say? Winch replies, ‘Well, one might speak here of how the power of God the Father and Adam’s dependence are, on both sides, inseparably linked with love. God’s power is not simply combined with his love: it is his love. And likewise with Adam’s dependence on and love for his Creator: they are one’ (p. 209). One might add that this is captured in the nakedness of both God and Adam. Creation is God’s nakedness for our sake—the allowing of something other than perfection to exist. The reaching out for that perfection by human beings depends too on their recognition of nakedness – that they are nothing in their own right – that they owe it to a God who became naked at creation and whose Son dies naked on a cross.

One might dwell in this picture. It can provide the parameters within which one thinks of one’s life. Wittgenstein says in the Lectures, with respect to religious belief:

Mackie has no use for the picture. He never could if he thought that Michelangelo was painting a representational work. Others may find God in the picture. That is why the loss of the picture would be calamitous for the latter. Wittgenstein writes in the Lectures: ‘Of certain pictures we say that they might just as well be replaced by another—e.g. we could, under certain circumstances, have one projection of an ellipse drawn by another. [He may say]: “I would have been prepared to use another picture. It would have had the same effect …” But in other cases, religious cases included, he says, ‘The whole weight may be in the picture’ (pp. 71–2).

We have seen how propositions can mislead us if divorced from practice, but the same may be said with respect to paintings in relation to a culture. Nevertheless, reflection on genre paintings which say themselves may help us to realise that neither all propositions nor all paintings are of a representational kind. The whole weight may be in the other picture. It says itself. And what it says, if the picture is a religious one, may become that in which we live, and move and have our being. But isn’t that what we say of God? Precisely, but then if God is in the picture, to be absorbed by the picture would be to be absorbed by God at the same time. After all, why shouldn’t an omnipresent God be present in our pictures of him? This may wean us away from the view which we criticised at the outset, namely, that our practices depend on the truth of propositions logically independent of them. To counter that view, Wittgenstein says on p. 28 of Culture and Value:

Pictures can capture ‘consciousness of sin’, ‘despair’ and ‘salvation’. Religion could be taught by means of such pictures. To appreciate this is to recognise that the pictures capture and convey something real. Are we prepared to admit that they capture God—that God can be in the picture? If someone is looking for realism in the philosophy of religion, that could be a good point at which to begin.