Interpreting Religious Language: A Wittgensteinian View

Abstract

This paper outlines a view of religious language that revolves around the notion of informed interpretation. The view can be summed up by saying that there is no fact of the matter independently of context and informed interpretation as to whether some religious statement or expression has cognitive content, or what that content may be. Where informed interpretation of religious language is impossible, we can give no answer to the question of what the content of a given statement or expression is. Equally, there can be no answer to the question of what that statement or expression presupposes or implies. If this is correct, then the idea that there can be a general and abstract philosophical analysis or theory of religious language should be called into question.

1. Introduction

God grant the philosopher insight into what lies in front of everyone’s eyes.

(Wittgenstein 1998, p. 72e, MS 135 103: c. 27.7.1947)

My discussion takes its cue from recent discussions of remarks by Wittgenstein that do not seem to rule out, and may even support, pluralist interpretations of religious language, and that suggest a view of religious language that I take to be independently attractive. Roughly put, this is the view that religious statements or expressions have a range of different functions and characteristics, some of which are best seen as expressive. Importantly, however, some of these expressive functions and characteristics are best understood as ways of expressing propositions. If so, some religious statements or expressions are used to express something that is, straightforwardly, either true or false. Others presuppose or imply something that is, straightforwardly, either true or false, even if these statements or expressions themselves are not used to express propositions.

In my view, such uses are common, and it would be surprising if Wittgenstein were simply to deny that. In fact, his own interpretations of religious language turn out to be more nuanced than they may at first appear. In any case, it is important to acknowledge that religious statements or expressions have a range of different functions and characteristics. There is no reason to doubt that they are normally bound up with a multitude of different attitudes, emotions and propositions that are either true or false. This is no more or less than what we should expect given their significance and their pervasiveness in human life. Any adequate interpretation of religious language must do justice to these basic facts.

I will not attempt to show that the view that I outline was one that Wittgenstein endorsed at any given point in time. I will discuss some passages that seem to me to point in that direction, but I will remain largely agnostic with respect to Wittgenstein’s own views. What I hope to show is that the view that I discuss is independently attractive and that there is a point in calling it a Wittgensteinian view.

The structure of the paper is as follows. Section 2 and Section 3 introduce the main idea and clarify some of the relevant concepts. Section 4 engages with two recent discussions of remarks by Wittgenstein that do not seem to rule out, and may even encourage, pluralist interpretations of religious language. Using Wittgenstein himself as an example, Section 5 and Section 6 examine notes that Wittgenstein put down in private notebooks when he was on duty on the battlefield in World War I, with a view to possible interpretations of religious language. Section 7 and Section 8 examine the results of this discussion in the light of some remarks dating from the 1930s, where Wittgenstein explores the uses of religious language from a different point of view. Section 9 builds on a typescript passage that refers to Gottfried Keller, who devised a playful paradox in his novel Green Henry. Section 10 considers three objections to my view. In light of my response to those objections, Section 11 examines lecture notes concerning Frazer’s (1922) Golden Bough, focusing on the idea of religious disagreement. Section 12 sums up the results.

2. Differences

As Drury reports, Wittgenstein considered using a quotation from King Lear as a motto for his book: “I’ll teach you differences” (Drury 1981, p. 171).1 While this quotation is familiar, there are two interesting aspects to it that are worth spelling out: First, if this quotation is considered as a motto, the “differences” in question must be philosophically important. Second, if the “differences” in question must be taught, they cannot be obvious. Perhaps they are elusive or easy to overlook. So seeing them for what they are must be philosophically important.

Before he mentions Shakespeare, Wittgenstein contrasts his own approach to philosophy with one that aims to find unity in what appears to be diversity: While some seem to want to say “that things which look different are really the same”, he says that his interest “is in shewing that things which look the same are really different” (Drury 1981, p. 171). His view of philosophical enquiry, then, is one that aims to find diversity in what merely appears to be unity or uniformity. As readers of his later work are well aware, one important focus of that work is a diversity in ways in which language is used where “surface grammar” is misleading (Wittgenstein 2009, § 664). This article pursues the question of what such an approach to philosophy can offer when it comes to interpreting religious attitudes, beliefs and related ways of using religious language.

In what follows, I will focus on religious language in a Christian context. It is clear that Christian practice and belief possessed a special personal significance for Wittgenstein. It was also, partly for that reason, of special philosophical interest to him. But it is also clear, and one of the more obvious applications of a slogan like ‘I’ll teach you differences’, that religious language as it is used in a particular Christian tradition at a given time is only one example of a way of using language in an area of life that shows great divergence—a multitude of forms of life, as Wittgenstein might put it. It is important to remember at the outset that an expression like ‘religious language’ already abstracts from many of these differences and should not be taken to suggest a unity or uniformity that the reality of human life, especially religious life, does not in fact bear out. We will come back to this point.

Later on, I will consider some of Wittgenstein’s remarks on Frazer’s Golden Bough, but it is best to start with elements of Christian practice and belief. What are we to make of seemingly factual statements or expressions that are seen as central to Christianity? Consider some examples: ‘Christ is risen!’ answered by ’Indeed he is risen!’; ‘God wills it’; ‘I believe in God, the Father almighty, creator of heaven and earth’; ‘There will be a resurrection of the dead’; ‘Christ died for our sins’. Someone who uses these and similar expressions would normally be taken to believe a number of things, some of which are presupposed or implied by what he says: for example, that Jesus lived and died and was resurrected, which must mean that, in some sense, he is now alive again; that there is a God, that he has a will, and with it, other attributes of persons; that God acts or has acted; in particular, that he made heaven and earth, which would not exist without him. Again, some of these expressions seem to be expressions of belief that at some future point in time, those who are dead will be resurrected, and that Jesus died for a specific reason, which is to relieve us of our sins. This belief in turn seems to entail that there is such a thing as sin, a state of being fallen or deficient by some measure of perfection, and that the death of Jesus makes some crucial difference to this state. Many other beliefs of this kind will be in the background, forming narratives surrounding those beliefs.

It is natural to think that these are examples of belief in a familiar, ordinary sense. If so, their content is a proposition—that which is believed, so that what is believed will be either true or false. To hold any such belief is, in some sense, to believe that it is true.2 Others may accept or reject these beliefs, thereby, once again, taking that which is believed by some other person to be either true or false. Some may interpret statements like the ones I mentioned differently. But even in that case, we would normally expect there to be something else that could be identified as the content of belief, a proposition that a person might believe, and so as being either true or false.

In my view, in a majority of cases, this is the correct account of religious beliefs. But it is important to say that this is compatible with saying something else: a person using these expressions normally does more than merely express a proposition that is either true or false. Many different things may be involved, and it should be clear that religious statements are not simply used to, as it were, state religious facts. A person takes a stance; expresses a commitment to a set of values; sees the world from a particular perspective, and takes up an attitude towards it or a part of it. Again, a person may engage in rituals or perform a part in ceremonies that have several layers of meaning; she may use language that resonates emotionally with her as well as with other people, building and affirming bonds with them. She may be using symbols and similes and find herself unable to translate these into plain, prosaic forms of speech. She may be engaged in ethical reflection; she may be summoning courage, making an appeal, reminding herself of her limitations, giving comfort or expressing hope in the face of deep despair.

This, then, is another way in which an adequate account of religious language must be sensitive to differences: where the uses of religious language are entangled with a range of propositions, for example, because something is believed to be or said to be the case, or because certain assumptions are presupposed or implied by what is being said, there are likely to be many other functions or characteristics of religious language that can be distinguished and described. These go beyond the function or characteristic of religious language to express beliefs about religious facts or to dispassionately assert, deny, or doubt that something is the case.

Given that there is a multitude of functions, it seems plausible to think that there may be uses of religious language that are not entangled with a wider range of propositions and related attitudes in this way at all. That may be so because this language has a very different function in the context of a given person’s life or plays a different role in a given practice. In such a case, presumably some other dimension of religious language would come to the fore. Given that there are those additional dimensions, it is also plausible to think that there could be cases in which cognitive and factual dimensions of religious language are not merely less developed but not developed at all. In certain cases, they may simply not be present. In other cases, they may have been erased, edited out by critical reflection.

Accordingly, interpretations of religious language can be offered as descriptions, as interpretations of the grammar of ‘God‘ or of ‘belief’, as accounts of how these words are used. But they may also be revisionary, offering new ways of understanding and of using language that perhaps once had a different role and may still have a different role in the lives of other people. It seems likely that at different stages of his life, Wittgenstein engaged in the second project. It seems likely, too, that he did not always clearly separate the second project from the first.

If that is our view of religious language, we now face a question: How can we tell what is being said or done in a given case? The contention of this paper is, in brief, that careful and informed interpretation is required. What a person says, believes and does in any given instance has to be interpreted in context. In part, this will be a context of traditions, rituals and practices, and formulae that have a history. That history will typically involve a set of shared activities and attitudes, and these must be understood against a background of familiar, often shared, human experience. But more immediately, what a person says, believes and does when she engages in religious practice or employs religious language must be understood against the background of what else that person says, believes and does, or would be prepared to say, believe and do in light of what she says. This seems to be generally true: If we want to understand what a person believes, or says, or does, what we see or hear must be interpreted against the background of her own view of the language that she uses. Thus, we may ask what else she is prepared to say, believe or do; what she thinks she should say or believe or do; or what she would say or believe if certain circumstances changed.

In what follows, I will try to develop and discuss this approach to religious language in a little more detail. I will also try to illustrate it, using Wittgenstein himself as my example. I do not claim that as it stands, the view that I describe is one that Wittgenstein himself articulated or endorsed. But there are remarks that point in this direction, and I will try to develop these thoughts further.

Conversely, this discussion might help to illuminate some of Wittgenstein’s remarks. There is no doubt that much of what he says about religious language and religious practices is interesting, illuminating and important. Moreover, much of it is less dogmatic, perhaps more personal and more exploratory, than it may at first appear.

3. Belief

A final point needs to be made with respect to differences relating to the matter of cognition and belief. In Remarks on Colour, Wittgenstein says this:

311. “I know that he arrived yesterday”—“I know that 2 × 2 = 4”—“I know that he was in pain”—“I know that there is a table standing there”.

312. In each case I know, only something different? Of course—but the language games are far more different than we are conscious of, considering these sentences.

(Wittgenstein 1977, p. 58e, translation revised)

One point of these remarks appears to be to bring out differences: As Wittgenstein interprets it, ‘2 × 2 = 4’ should be understood as expressing a rule. Accordingly, ‘I know that 2 × 2 = 4’ expresses knowledge, or acceptance, of that rule. Given this interpretation, knowing that 2 + 2 = 4 does not involve relations between me and certain items in some shadowy Platonic realm or an intuition of mathematical truth coming in a mysterious way from a mysterious source that can explain knowledge. Again, ‘I know that there is a table standing there’ is a report based on immediate experience, whereas ‘I know that he was in pain’ is not, both because the other person’s pain is not experienced and because it is a case of knowledge that concerns the past. There are many other differences involved.

Now, a second point of these remarks seems to be this: Despite the differences, it is perfectly acceptable and need not be misleading to speak of knowledge as we do, so that it is quite simply true that “in each case I know, only something different”. Moreover, it is natural to add: If there is ‘something’ I know, then that ‘something’ must be a proposition, so there is cognition, and these propositions can of course become the contents of belief. Again, it seems to me that this is generally true and need not be misleading as long as the differences between those language games remain in view.

If this is true, it has important implications for an adequate interpretation of religious language. It can sometimes seem that there is a dispute between two parties: one party insists that religious statements should be taken at face value, where this is taken to mean, among other things, that religious statements have cognitive content, that they are embedded in a network of beliefs, and that these beliefs are either true or false, while the other party might seem to deny that. But this oversimplifies the situation. The important question is what a person does, or thinks she does, when she uses an expression like ‘I believe in God, Father almighty, creator of heaven and earth’. What she does and thinks she does, however, is not settled just by pointing out that she uses the expression ‘I believe’ or that ‘God exists’ is a proposition. The question is how such expressions work in a given context. What is it to ‘believe’ in God in the relevant sense?3 What is it to think of ‘God exists’ as being either ‘true’ or ‘false’, or to say that there must be some ‘fact of the matter’, even if we cannot know it?

In my view, these questions must be answered by competent, informed interpretation—by considering what else the person says, believes and does; what else the person is prepared to say, believe and do; or what the person thinks she should say or believe or do. There is certainly belief in religious matters—but saying this, and nothing more, does not yet give us any interesting account of that belief. Everything depends on what we say next. Some language games look very similar to others on the surface but are in fact quite different, and that is also true of language games involving words like ‘true’, ‘fact’, and ‘belief’.

4. Three Stages of Development

In a recent article, Pichler and Sunday Grève have identified three stages in the development of Wittgenstein’s views of religious language (Pichler and Sunday Grève 2025; see also Pichler 2025). According to their line of interpretation, the first stage culminates in the Tractatus, includes the ‘Lecture on Ethics’, and lasts until about 1931. The view of religious language that is prominent in this phase could be called ‘strong non-cognitivism’. Its defining characteristic is the claim that religious statements or expressions of belief have no cognitive content whatsoever—that is to say, there are no substantial truth conditions for these statements or expressions. Wittgenstein claims that in this respect, religion is on par with ethics or aesthetics. There are no states of affairs that could be pictured by such statements or expressions; these statements or expressions do not represent, so they are neither true nor false. Rather, they belong to the realm of that which can only be “shown”, as Wittgenstein puts it in the Tractatus (cf. Wittgenstein 1971, 6.41–6.522).4

According to Pichler and Sunday Grève, the second stage in Wittgenstein’s development is characterised by a movement away from this view and towards a view that is at least more open to cognitive elements in religious language. This could be labelled a ’moderate’ form of non-cognitivism. Importantly, it would still be a form of non-cognitivism, because it would interpret notions like ‘belief’ in some special way, as expressing something very different from what ‘belief‘ would normally express in other contexts.

What this might mean comes out clearly in a quote from Culture and Value:

Christianity is not based on a historical truth, but presents us with a (historical) narrative & says: now believe! But not believe this report with the belief that is appropriate to a historical report,—but rather: believe, through thick & thin & you can do this only as the outcome of a life. Here you have a message!—don’t treat it as you would another historical message! Make a quite different place for it in your life.—There is no paradox about that!

(Wittgenstein 1998, p. 37e, MS 120 83 c: 8.–9.12.1937)

Moreover, Wittgenstein construes belief of this kind as being compatible with believing (in a different sense) that the historical accounts of the Gospels are false:

Queer as it sounds: the historical accounts of the Gospels might, in the historical sense, be demonstrably false, & yet belief would lose nothing through this: but not because it has to do with ‘universal truths of reason’! rather, because historical proof (the historical proof-game) is irrelevant to belief. This message (the Gospels) is seized on by a human being believingly (i.e., lovingly): That is the certainty of this “taking-for true”, nothing else.

(Wittgenstein 1998, pp. 37e–38e, MS 120 83 c: 8.–9.12.1937)

At a push, this could be taken to mean that there are religious truths, such as the truth that there will be a resurrection, even if the historical account of the resurrection given in the Gospels turns out to be false. Again, there could be symbolic truths expressed by historical falsehoods. In such a case, ‘belief’ could still be belief in its familiar sense, involving truth conditions, and there would be no contradiction in holding the belief that while a particular historical account is false, there is something true about the resurrection that can be the object of belief. But what Wittgenstein is getting at seems to be more radical than that. It seems that a perspective is adopted, a message seized upon, and that is all there is to “taking it for true”—no further questions about truth, historical or otherwise, arise. So as we take cognitive content away, ‘belief’ begins to turn into something like an attitude, a way of life, or a way of looking at life. Belief could be a framework that no longer needs a basis in specific facts about the resurrection, or the existence of God, or life after physical death, or anything else of that kind.

That way of understanding it would naturally lead to a form of non-cognitivism about religious belief. Seen from that perspective, the ‘historical proof-game’ as well as the ‘truth-game’ would become irrelevant to religious belief, and that would certainly set religious ‘belief’ apart from the way ‘belief’ is understood in ordinary discourse. Apart from passages in Culture and Value, this kind of view seems to be expressed in ‘Lectures on Religious Belief’, dating from the later 1930s:

If you ask me whether or not I believe in a Judgment Day, in the sense in which religious people have belief in it, I wouldn’t say: “No. I don’t believe there will be such a thing.” It would seem to me utterly crazy to say this. […]

I think differently, in a different way. I say different things to myself. I have different pictures. […]

Anything that I normally call evidence wouldn’t in the slightest influence me.

(Wittgenstein 1966, pp. 55–56)

There are many other passages like that.

What Pichler and Sunday Grève see as the third phase is perhaps the least developed. As they point out, a cognitivist view of religious language is suggested by remarks from the late 1930s and the 1940s. It also fits the overall approach to philosophy that emerges in Philosophical Investigations. Pichler and Sunday Grève argue that non-cognitivist interpretations are implausible as general accounts of religious belief. Given that philosophy as Wittgenstein conceives of it in Philosophical Investigations “leaves everything as it is” (Wittgenstein 2009, § 124), they conclude that non-cognitivist interpretations of religious language as it is actually used are at best an “unjustified reinterpretation” (Pichler and Sunday Grève 2025, p. 64).

Some passages suggest, and some even emphasise, that religious belief can conflict with belief in other areas. It is natural to think of this not merely as a case of conflicting attitudes but as a conflict of cognitive content. One example from On Certainty is this:

I believe that every human being has two human parents; but Catholics believe that Jesus only had a human mother. And other people might believe that that there are human beings with no parents, and give no credence to all the contrary evidence. Catholics believe as well that in certain circumstances a wafer completely changes its nature, and at the same time that all evidence proves the contrary. And so if Moore said “I know that this is wine and not blood”, Catholics would contradict him.

(Wittgenstein 1969, p. 239)

Pichler and Sunday Grève also point to an interesting remark from 1948, where Wittgenstein compares a figure that he calls “the honest religious thinker” to a tightrope walker (Wittgenstein 1998, p. 84e, MS 137 67b: 5.7.1948). This remark can be interpreted in a number of different ways. It is possible to read it as an image that relates belief (in its familiar sense, involving truth conditions) to the available evidence for that belief, which is extremely slim—“And yet it really is possible to walk on it” (ibid.). But is that all there is to it? It seems to me also possible to see the image as an image of commitment, courage, doubt, and a number of related attitudes. More importantly, it seems to me possible to see it as an image that combines belief in its familiar sense in the absence of convincing evidence and courage in the face of doubt. An honest religious thinker is fully aware of his own doubts and the dangers of despair, but he holds on to his faith and hopes that it will keep him going.

As I have said, I will remain largely agnostic with respect to Wittgenstein’s own views of religious language.5 I agree that, when they are advanced as general, descriptive theories, neither ‘strong’ nor ‘moderate’ versions of non-cognitivism offer satisfactory accounts of religious statements or expressions of belief. I also agree that a plausible account of religious statements or expressions of belief may involve “mixed strategies” (Scott 2022). A pluralist account of religious language will allow that religious statements or expressions have a range of different functions and characteristics, some of which are best seen as expressive. At the same time, it insists that religious statements or expressions are frequently used to express propositions, and that belief in religious matters often is belief in its familiar, ordinary sense. Such a view is certainly consistent with, and perhaps encouraged by, the methods of Philosophical Investigations.

In what follows, I assume that for most religious beliefs, a pluralist account is correct. One of my questions is how such a view can account for expressive elements or aspects of religious language that do not fit the model of a cognitivist analysis, and how to identify these elements or aspects in a given case. This is where the notion of informed interpretation turns out to be helpful. Conversely, this notion can help to identify cognitivist elements and support a pluralist analysis of religious language.

To develop this idea, I now turn to an interesting case where informed interpretation proves to be quite difficult and thus highlights the importance of interpretation: Wittgenstein’s own references to God in the private diaries he wrote in World War I.6

5. The Gospel in Brief

When Wittgenstein was young, he was introduced to Catholicism, but according to various reports, he lost his faith when he was young (McGuinness 2005, p. 43). Russell reports that when Wittgenstein studied in Cambridge, he seemed hostile to religion. In fact, Wittgenstein struck him as being “more terrible with Christians” than Russell himself (Monk 1990, p. 116). By all accounts, there was a significant change when Wittgenstein served as a soldier on the Eastern Front.

Consumed by fear and faced with death, he kept reflecting on his life, his character and failings. Wittgenstein read Tolstoy’s Gospel in Brief, which he had found in a bookshop, and was deeply moved by it (Tolstoj 1892). He carried the book around with him and recommended it to others. In fact, Wittgenstein soon became known to his comrades as “the man with the gospels” (Monk 1990, p. 116). He told Ludwig von Ficker later that the book virtually kept him alive: “If you are not acquainted with it, then you cannot imagine what an effect it can have upon a person” (Monk 1990, pp. 116, 132).

However, what effect the Gospel had on Wittgenstein is not easy to pin down. The first reference to the book is found in a notebook entry in the early autumn of 1914: “Yesterday I began to read Tolstoy’s Gospel in Brief. A wonderful book. But I haven’t yet found what I expected from it” (Wittgenstein 2022, 2.9.14). What did he expect from it? Clarity? Salvation? Soon after, references to God appear and become intense: “God give me strength. Amen! Amen! Amen!” (12.9.14); “God be with me! May the spirit enlighten me” (15.9.14); “Man is helpless in the flesh and free through the spirit. And only through the spirit” (16.9.14).7

At this time, Wittgenstein was posted on the SMS Goplana on the river Dunajec, a tributary of the Vistula. He manned a searchlight at night, and his unit had been close to Russian fire: “The night passed quietly. In the morning heard heavy cannonade & rifle fire. We are, in all probability, inescapably lost. The spirit is still with me but will it abandon me in my utmost need?” (16.9.14) In passages like these, it is unmistakable that the spirit is portrayed as active, as something that acts upon the helpless individual and can give it hope and strength. Its presence may be felt, and it acts on other human beings. So, the spirit is portrayed as an independent force, coming from somewhere outside, that affects the human spirit and connects all humans through the spirit. Given that it acts, and so exists, independently of the individual, its will and hopes and fears, the helpless individual must be faithful to that spirit, and can only hope not to be abandoned by it (cf. McGuinness 2005, pp. 220–21).

References to God and the Holy Spirit in these notebook entries are frequently tied to ethical reflections and Wittgenstein’s demanding, critical examinations of his character: “Now I would be given the chance to be a decent human being since I am standing eye to eye with death. May the spirit enlighten me” (15.9.14).8 There are exhortations to keep the faith: “Only one thing is necessary”—and that is faith and courage holding on to his commitment, his belief and trust in God (17.9.14).9 Other entries look towards the future, frequently invoking God or the Holy Spirit as a higher power, a source of hope and faith, transcending human understanding and desire: “Thy will be done!” (30.9.14); “Do not worry!! Just now the order came to take off for Russia. So it’s becoming serious again! God with me” (6.10.14).10

At times, the entries take on a distinctly stoic tone, sometimes with a view to present experience, sometimes with a view to an ethical ideal: “One should enjoy the good hours of life gratefully as a blessing, and otherwise feel indifferent toward life. God is with me!” (12.10.14); “Only let the spirit live! He is the safe haven, protected from the desolate, boundless gray seas of happenings” (13.12.14). Again: “I am all spirit and therefore I am free” (13.10.14); “Only be true to one’s own spirit! And leave everything to God” (30.11.14); “Everything is in God’s hands” (7.12.14). While the “spirit” sometimes seems to be the human spirit, there is no doubt that it is often thought of as external and possessing agency, as something that affects, or is present in, the human spirit. This is how Wittgenstein conceives of it when he considers obstacles within himself that hinder his work in philosophy: “Worked without success. I am quite in the dark as to how my work will continue. Only through a miracle can it succeed. Only if the veil over my eyes is removed from somewhere outside of me. I have to submit entirely to my fate. Whatever is imposed on me, will be. I live in the hands of fate” (25.1.15).11 In this passage, the idea of fate begins to merge with the idea of God, and both are linked to the idea of a world that is changing, constantly in motion, and quite fundamentally beyond one’s own control.

This motive is both Christian and stoic. It becomes quite prominent in entries dating from April 1916 onwards. Wittgenstein was now stationed as an ordinary soldier at the Eastern Front, taking up the observation post. His unit was repeatedly engaged in battle, and Wittgenstein came under fire. In a way, this seems to be what he had wanted, hoping for a transformation of himself: “Perhaps the proximity to death will bring me the light of life! May God enlighten me! I am a worm but through God I become a man. May God stand by me. Amen” (4.5.16) In this existential crisis, entries in the notebooks sometimes sound like pointed written prayers: “God deliver me and grant me peace! Amen” (7.4.16); “God will help” (18.4.16); “God be thanked. Thy will be done! Go thy ways! Thy will be done!” (26.4.16); “To the observation post in the afternoon. We were shot at. Thought of God” (29.4.16). “In danger, He will not desert me!!” (10.5.16); “I will be sleeping tonight under infantry fire, will probably perish. God be with me! For ever and ever Amen. God be praised in eternity, Amen. I give my soul to the Lord” (16.5.16)12; “Being shelled. As God wills it” (25.5.16).

Wittgenstein survived.

6. Perspectives of Interpretation

How are these references to God, the Holy Spirit, fate, enlightenment and the soul to be understood? Should we take them as expressing belief in its familiar sense? Or should we read these utterances differently? If so, how?

It is certainly tempting to read notebook entries from this time in the light of what came after. Wittgenstein divided pages of his notebooks into left and right and kept their contents largely separate. Entries on the left are written in code. They concern personal matters, and the passages that I have quoted in the previous section all come from that private, coded side. Wittgenstein did not code entries on the right. These concern his work, developing his thoughts about language, truth, and logic. There are many references to “work” in the coded notes, but for a long time, coded and uncoded passages merely run in parallel. In July 1916, the lines of thought begin to merge. Wittgenstein’s reflections about matters of religion, ethics and aesthetics now become a part of his work in philosophy, and that is how they eventually find their way into the Tractatus.

Starting with an entry dated 4 July, we read thoughts like the following:

What do I know about God and the purpose of life?

I know that this world exists. […]

That something about it is problematic, which we call its meaning.

That this meaning does not lie in it but outside it.

That life is the world. […]

The meaning of life, i.e., the meaning of the world, we can call God.

And to that meaning we can connect the image of God as a Father.

To pray is to think about the meaning of life.

I cannot bend the happenings in the world to my will; I am completely powerless.

I can only make myself independent of the world—and so in a certain sense master it—by renouncing any influence on events.

It is clear that these ideas are a direct continuation of the thoughts recorded on the private, coded side and influenced by them, but also that they now form part of his new vision for philosophy. A few days later, we read, “What cannot be said, cannot be said!” (7.7.16). And on the next day, Wittgenstein comes back to his line of thought about God, fate and the meaning of life:

To believe in a God means to understand the question about the meaning of life.

To believe in a God means to see that the facts of the world are not the end of the matter.

To believe in a God means to see that life has a meaning. […]

Wittgenstein refers to a feeling “that we are dependent upon an alien will” and adds that “what we are dependent upon we can call God” (8.7.16).

Once again, this does not sound like an appeal to a belief in God in its familiar sense: we read that to believe is “to understand the question about the meaning of life” or to “see that life has a meaning”. Moreover, ‘God’ does not seem to refer to a higher being, a personal God as the Christian tradition understands it: “In this sense God would simply be fate, or, what is the same thing: the world—which is independent of our will” (8.7.16). It is therefore doubtful that religion, as it is portrayed in passages like these, has much in common with belief as the Christian tradition understands it. In fact, it is doubtful that religion as it is portrayed in these passages has much to do with belief as that is understood in ordinary discourse: it is not so much a matter of holding something to be true, but rather of “understanding” or of “seeing” something, of looking at life and one’s own place in the world in a certain way, and living life accordingly.

This seems to me to be correct as far as it goes, but it is important that it does not follow from it that the notebook entries already express that view. After all, as the notebooks show, this connection was an insight. It seemed to Wittgenstein that he had turned a page, achieved a breakthrough in his own thoughts and his way of doing philosophy—and if it was a breakthrough, it was something new. That observation brings us back to the question of what came before. The later entries by themselves do not answer that question—they bring it back to the fore.

What are the possibilities? There are more than I can mention here, but I will make a few tentative suggestions.13 It is possible that appeals to God and the Holy Spirit are symbolic or metaphorical. When Wittgenstein uses formulae like “God be with me” or “Thy will be done”, their use could be primarily ethical in character—a recognition of the fact that events are beyond his control, that he faces grave danger and that in this situation, he needs to be strong, honest, and courageous and should live and feel accordingly. All this is compatible with an interpretation that retains belief in its traditional sense, so that there is belief in God and the Holy Spirit, acting as a higher and external force, somewhere in the background. In fact, it could be in the foreground: that could be the case if Wittgenstein became religious in a traditional sense but then reinterpreted his own belief in terms that are expressed in notes to the effect that God and world and fate are one, and that to “pray is to think about the meaning of life”, or that to “believe in a God means to see that the facts of the world are not the end of the matter”.

More intriguingly still, it could be unclear and underdetermined which interpretation is correct, even if we had more information than we do. Moreover, it could be unclear and underdetermined even from Wittgenstein’s own point of view if someone had asked him and if he had attempted to give an honest account of his religious beliefs. His beliefs themselves may have been unclear, transient, shifting, or conflicting. It is therefore possible that a clear, consistent view of them was not available to anyone, including Wittgenstein himself—not because of epistemic limitations, but because there was nothing clear and consistent to be seen.14

Once again, I will not speculate about Wittgenstein’s personal beliefs. Just as I remain agnostic with respect to his considered views of religious language, I remain agnostic with respect to his own expressions of belief. The point that matters for my purposes is that what is needed to make sense of these expressions is what I have called informed interpretation. This involves careful analysis of practices of contexts in which religious language is used. It also involves careful analysis of what a person using it would say about her actions and her attitudes, her uses of religious language, and her own view of the place her beliefs have in her life.

It is clear that many different answers to these questions could be given, and that we could consistently combine many of these answers, so that a statement like “Thy will be done” could serve to express an attitude or feeling, an insight or a recognition. What looks like a reference to a God and to His will could be metaphorical or symbolic, and its import could be mainly ethical. Perhaps it serves as a reminder not to falter in the face of mortal danger when faced with an enemy in combat; perhaps it serves as a reminder of the value of Stoic ataraxia, being willing to accept one’s fate, finding peace within a world that is not only constantly in motion but also fundamentally beyond one’s own control.

So understood, belief in God can be noncommittal with respect to the truth of propositions such as that there is a higher being like the Christian God, that He has a will, and that His will is the one that matters, even where it is fundamentally at odds with our own. But the reference to God and His will can also be taken much more literally and involve belief in propositions of this kind. In this case, there would be a question of the truth of those beliefs, and that is what would make them similar to other, more familiar and mundane, instances of what we call belief.

I now turn to three passages that illustrate that Wittgenstein seems to have fully recognised that religious utterances or statements can be understood in different ways. This discussion shows that different ways of interpreting the same expression can be located on opposite sides of the divide between cognitivist and non-cognitivist interpretations of religious language. Moreover, the passages suggest that there are different ways of combining the two.

7. A Book Written to the Glory of God

Wittgenstein wrote in his foreword to Philosophical Remarks, dated November 1930:

I would like to say ‘This book is written to the glory of God’, but nowadays that would be chicanery, that is, it would not be rightly understood. It means that the book is written in good will, and in so far as it is not so written, but out of vanity, etc., the author would wish to see it condemned. He cannot free it of these impurities further than he himself is free of them.

(Wittgenstein 1975, p. 7)

Why would such a remark “not be rightly understood” at the time of writing? Drury comments, “It implies that words which in one age could be correctly used can at a later stage be ‘the words of a cheat’; because if these words are constantly used in a superficial way they have become so muddied that the road can no longer be trod” (Drury 1981, pp. 93–94). What Drury seems to have in mind is that at some point in time, some expressions may have had a certain, quite determined meaning, but then lost it through constant superficial use, or overuse. This can be a process of secularisation: The French expression ‘adieu’, ‘a dieu [vous commant]’, ‘[I commend you] to God’ now simply means ‘goodbye’, and few will think of God when they use it. In fact, ‘goodbye‘ itself appears to be a contraction of ‘God be with ye’.15 Few will think of God when they use the word ‘goodbye’ today.

In this instance, however, Wittgenstein’s remark may point in the opposite direction. Note that Wittgenstein makes very clear how he wishes the expression ‘This book is written to the glory of God’ to be understood: as an expression of an attitude that is broadly speaking ‘ethical’, relating to ideas of purity, motives of vanity, goodwill, integrity and truthfulness in the pursuit of insight in philosophy. As we might expect, Wittgenstein does not invoke a literal, cognitive commitment to something like the existence of God. In this way, and again as we might expect, he seems to be offering a non-cognitivist interpretation of ‘This book is written to the glory of God’. Why, then, would the phrase “not be rightly understood”? Why would it be considered “chicanery”, indicating subterfuge, trickery, deception?

It is difficult to pin this down. One interpretation, perhaps closer to the views of Drury, would be that in his time, an expression of this kind of ethical attitude in these terms would sound hollow. Perhaps it would even sound presumptuous or vain. But a more plausible interpretation might be that in his time, the expression ‘This book is written to the glory of God’ would likely be heard against a background of an earlier, more literal interpretation. So understood, the expression could easily be heard as involving a cognitive commitment, a belief in the existence of God, and so precisely to the kind of cognitive commitment that Wittgenstein’s own interpretation conspicuously leaves out. After all, that cognitive commitment was part of the context of the use of that expression for hundreds of years, and that use continued and continues to this day. It was certainly a cognitive commitment for Johann Sebastian Bach, who wrote “S. D. G.” on many manuscripts of compositions—“Soli Deo Gloria”, “To the Glory of God Alone”. In the case of Bach, this clearly had an ethical dimension but was also an expression of religious faith, a dedication of his work to God in the literal, familiar sense.16

That context could be both the inspiration and the source of the misunderstanding that Wittgenstein is eager to avoid. For Wittgenstein to simply say, ‘This book is written to the glory of God‘ without any further comment could amount to ‘chicanery‘ precisely in the sense of pretending to have a kind of belief in the existence of God that Wittgenstein does not in fact have. He could easily be taken to be pretending to be dedicating the book to God in a literal sense—something that Wittgenstein, a highly conscientious person, does not want to do.

However that may be, the point for purposes of our discussion is that the expression ‘This book is written to the glory of God‘ can be interpreted in a number of different ways, and that Wittgenstein clearly acknowledges that. He mentions the possibility of misunderstanding, and he offers his interpretation of the phrase as applied to his work, thereby determining how he would wish it to be understood. This interpretation seems to keep it clear of a cognitive commitment to the existence of God, focusing on earnest ethical ideals that are important for his work and his attitude towards it. But it is also clear that cognitive commitments can be present in the background, or indeed the foreground, when expressions like ‘This book is written to the glory of God‘ are used by others, and in other contexts. That is why it is possible, and even likely, that such an expression will be misunderstood. In his exchange with Drury, Wittgenstein preempts this kind of misunderstanding of the phrase by saying what he means by it. And saying what he means by it determines what he means by it—whereas the expression by itself does not.

8. The Grammar of ‘God’

A similar conclusion can be drawn from a discussion of ‘God’ in lectures given in Cambridge in May 1933. Here are excerpts from notes taken by G. E. Moore:

I have always wanted to say something about grammar of ethical expressions, or e.g., of “God”.

One great trouble our language gets us into is that we take a substantive to stand for a thing or substance.

Ordinary grammar doesn’t forbid us to use substantives in this way: the origin of all use of substantives & verbs is in fact a simile for physical bodies moving. […]

Now: use of such a word as “God”.

It has been used in many different ways: e.g. sometimes for something very like a human being—a physical body.

Cf. “soul” which has sometimes been described as something “gaseous”.

But others haven’t meant by “soul” anything like this.

(Wittgenstein 2016, p. 318)

Just as it is very difficult to delineate the use of ‘soul‘ in a given context, it can be very difficult to delineate the use of ’God’. As Wittgenstein says, it can be used, and has been used, in many different ways. Some of these uses suggest anthropomorphistic conceptions of God; others do not. Some think of God along the lines of a thing or substance; others do not, and this will typically be shown in their use of ‘God‘. “Ordinary grammar” is compatible with either interpretation.

As we have seen, Wittgenstein himself was probably inclined to emphasise the possibility that belief in God does not involve commitments to the existence of a “thing or substance”—not to mention the existence of “something very like a human being” or even “a physical body”. But he acknowledges that these interpretations are correct as interpretations of certain instances of religious belief.17

That this is Wittgenstein’s view is also suggested by the passages that follow:

Fact that “I hope to God” is used by a man, may mean that he does believe in God in a sense.

When a man worships idols

(1) One possibility is that he believes idol is alive & will help him. Then man must have forgotten that he made it: but this can happen.

(2) In millions of cases, this will not happen, but e.g.,

(a) God dwells in the statue:

But what does “dwells” mean?

By asking what he would say, & what he wouldn’t, you can get at how he uses the word. Haeckel said “God is a gaseous vertebrate”, meaning that that’s what people meant.

This is like saying “Soul is a gaseous human being”; & answer is sometimes people so use this word, but sometimes not at all.

(Wittgenstein 2016, p. 319)

The upshot of this discussion seems to be that a practice like worshipping idols can be interpreted in different ways. Sometimes, there will be differences at the level of belief, especially when it comes to interpreting beliefs that are part of, or associated with, the practice. There could be a belief that the idol is alive, that the artefact itself is in some sense a living being, a God, and not in fact a sculpted piece of stone. Alternatively, we could find beliefs to the effect that, in some sense, God “dwells” in the statue, and this in turn could be spelled out in a number of different ways.

Again, the most important point for the purposes of my discussion is the one that Wittgenstein appears to make in passing. What might someone be thinking when he says that a ‘God dwells in the statue‘? “By asking what he would say, & what he wouldn’t, you can get at how he uses the word”. Putting it differently, if we want to understand what that person believes, we should try to assemble information for the purposes of interpretation. The more information we get, the better our chances of giving an adequate, informed interpretation of a given practice or expression will be.

This point seems to me to be correct and also to apply more generally: If we want to know what a religious person means when she says that she believes that there is a God, or that Jesus died to relieve us of our sins, or that there will be a resurrection, the best way to find out would seem to be to ask further questions and to see ‘what she would say & what she wouldn’t’. So if the question is whether a cognitivist or a non-cognitivist interpretation better fits a given statement or belief, the answer is that it depends on the details of the case, and to get an overview of those details is the task of informed interpretation. If the question is whether belief in God involves some kind of commitment to the truth of a proposition, for example, that there really is some higher being that fits certain relevant descriptions given in the Christian tradition, the answer is that “sometimes people so use this word, but sometimes not at all”.

9. An Air of Paradox

Let me now come back to a different part of the quotation from the lecture of 1 May 1933:

Fact that “I hope to God” is used by a man, may mean that he does believe in God in a sense.

Evidently, if ‘I hope to God’ can be interpreted in such a way that the person saying it “does believe in God in a sense”, we should expect there to be interpretations of ‘I hope to God’ that carry no such implication. This seems to be highly plausible and related to the point about ‘adieu’ and ‘goodbye’ that I made above: these expressions have become unmoored from their religious origins, and in the process, they have taken on a life of their own. The same seems to be true of ‘I hope to God’, which still can and sometimes should be taken to involve belief in God in the familiar sense. But ‘I hope to God’ can and often should be taken as equivalent to ‘I hope’, which normally carries no commitment of this kind.

The editors of Moore’s notes of the lectures point to a possible literary inspiration for the passage: Wittgenstein held a famous novel by the Swiss author Gottfried Keller in very high regard, and a fascinating figure in that novel is portrayed in such a way as to raise the question of how her references to God should be understood. The novel is Green Henry, first published in the 1850s and published in revised and final form in 1879–1880. A lovely irreligious figure that appears in Part IV of the novel is Dorothea Schönfeld, and Keller sometimes paints her musings in the colours of lighthearted and amusing paradox.18 She does not believe in immortality. Does she believe in God? Dorothea says: “Dear God, what can a poor thing like me know? With God, all things are possible, even that He exists” (Keller 1879–1880, IV, p. 251, my translation).

A first paradox arises because ‘Dear God’ is evidently open to interpretation. It need not, but it can be taken to involve and even express belief in God. If it is taken in this way, there is a subtle conflict with ‘what can I know?’, which is naturally understood as an expression of agnosticism. In this context, it suggests a suspension of judgment about the existence of God. The passage plays on the ambiguity and therefore seems to be ironically suspended between ‘I believe in God’ and ‘I do not believe in God’.

“With God, all things are possible” is a quote from the New Testament and naturally taken to express a belief in the existence of God.19 But of course, God has to be real in order to be able to do anything. So how can there now still be a question as to whether He exists? There also seems to be a wilful category mistake: What is ‘possible with God’ relates to his abilities. But existence cannot be an entry on a list of anyone’s abilities. The exercise of God’s abilities requires his existence. But “With God, all things are possible, even that He exists” seems to imply that His existence is unlikely. If anything, it is an expression of belief to the effect that God does not exist. That is the second paradox.

The upshot seems to be that Dorothea keeps herself behind a veil of irony. She probably does not believe that God exists but is not too worried by the question anyway.

The relevance of this to Wittgenstein is that there is a puzzling entry in a manuscript that refers to Keller and Green Henry:

Keller (in Green Henry) once writes about a man who, while he says that he does not believe in God, still uses all those familiar phrases that contain the word “God”. (“Thank God”, “would to God”, etc.) And Keller thinks that in doing so, the man would contradict himself. But there would have to be no contradiction at all, and one can say: I will learn what you mean by the word “God” considering which sentences with that word you use und which ones are meaningless to you. For I also use the word “sense of a proposition”, “meaning of a word” in certain contexts, yet I do not know any thing, ‘the meaning of a word’, and a shadow of an event‚ ‘the sense of a proposition’.20

(Wittgenstein 2015, TS 219-6 [6])

I have been unable to locate a passage that fits this description in Green Henry, and I take it that Wittgenstein either misremembered or thought of Dorothea Schönfeld and the paradoxes that I have described.21

The main point of this remark seems to be related to the one that came up in the lectures of May 1933: “One great trouble our language gets us into is that we take a substantive to stand for a thing or substance” (Wittgenstein 2016, p. 318). There is surely no such thing as “the meaning of a word” or “the sense of a proposition”, but expressions like these have a use, as Wittgenstein might put it. Again, ‘God’ can be used to stand for “something very like a human being”, or some higher, spiritual being, or something transcendent that is altogether incomprehensible. But phrases like ‘I hope to God’, ‘would to God’, ‘God willing’, and so on, do not have to be construed in this way at all.

Why is this relevant? The point that is most interesting for the purpose of our discussion is the one about the ways of finding out how a given phrase is used by a given person in a given context. What kind of evidence is needed? How is informed interpretation possible in a given case? The answer is the one that Wittgenstein gives in this passage: “I will learn what you mean by the word ‘God’ considering which sentences with that word you use und which ones are meaningless to you”. In other words, what a person says, and what she says she means by it, is typically the key to understanding what she means, and this is how we can identify not only what someone believes, but what other functions a religious attitude or practice or statement or expression may have in the given context of a person’s life.22

To see how this might work in practice, consider a remark from a discussion of Frazer’s Golden Bough:

If I, a person who does not believe that there are human-superhuman beings somewhere which one can call gods—if I say: “I fear the wrath of the gods,” that shows that I can mean something by this, or can give expression to a feeling which is not necessarily connected with that belief.

(Wittgenstein 1993, p. 131)

Here, Wittgenstein explains his use of the expression ‘I fear the wrath of the Gods’, emphasising its relation to a feeling, not to a belief. But he does this by speaking of belief in gods in a familiar sense and distinguishing his use of the expression from it. There is no doubt that we understand both elements and can distinguish them with ease.

In a passage written in 1950, Wittgenstein comes back to the subject of religious language and belief. The contrast here is one between a certain form of words and the use that people make of it, which points to a difference that certain expressions make in practice, in the context of a life:

Really what I should like to say is that here too what is important is not the words you use or what you think while saying them, so much as the difference they make at different points in your life. How do I know that two people mean the same thing when each says he believes in God? And just the same thing goes for the Trinity. Theology that insists on certain words & phrases & prohibits others makes nothing clearer.

(Karl Barth)

It gesticulates with words, as it were, because it wants to say something & does not know how to express it. Practice gives the words their sense.

(Wittgenstein 1998, 97e, MS 173 92r: 1950)

If we take this perspective seriously, there is no reason to exclude belief in a familiar sense from the religious sphere, much less a priori, by employing abstract models of interpretation that are motivated by some system of philosophy. We will have to “look and see” (Wittgenstein 2009, § 66)—and when we do that, we will see a wide variety of uses of words like ‘God’ and ‘belief’. Some of these will rightly be interpreted as expressions of belief in the existence of a higher being, where that which is believed is taken by believers to be true, while others may think that it is false. So understood, ‘God‘ will be used to refer to a higher being that is taken to exist. Belief in the existence of that higher being will be belief in its familiar sense, so that what is believed is a proposition that is either true or false.

Conversely, there is no reason to exclude other avenues of interpretation, much less a priori. In many cases, more than one dimension of religious language will be touched upon where words like ‘God’, ‘belief’ or ‘soul’ or ‘spirit’ are used. A person who says, for example, ‘God is with me’, will normally be taken to avow belief.23 At the same time, she may use these words to express an attitude, to try to summon courage, or to suppress a nagging doubt. She may pray or reflect on her life and her own failings; she may seek the comfort of a ritual, take part in a ceremony, or affirm her ties to a community, and many other things besides. All this is not only compatible with interpretations that ascribe belief in its familiar sense, involving truth conditions, but naturally conjoined with them.

Of course, it is also possible that no such element of belief in its familiar sense is present, and one way of bringing this out would be to ask the right kind of question. Furthermore, a non-cognitivist view of religious language may be offered not so much as a description of belief as we normally encounter it among religious people, but as a preferable way of interpreting the religious language that they use. In an openly revisionary spirit, this interpretation could be introduced as an alternative way of continuing the practices in which they are engaged.

In this case, there will certainly be questions as to whether such a resolute non-cognitivist reinterpretation really would provide a better way of understanding talk of ‘God’ or of ‘belief’, and what the relevant criteria for that kind of evaluation would be.24 Again, there will be questions as to whether the alternative interpretation would be robust enough to stand up to scrutiny, and whether the resulting practice would continue to be “stable under reflection”, to borrow a suggestive phrase from Bernard Williams.25 But there should be no doubt that such interpretations are possible. They are intelligible, at least in outline, and they are intelligible as they are intended.

Again, the way to find out how interpretations of religious language are intended would be to ask questions. Wittgenstein is right: “We will learn what you mean by the word ‘God’ considering which sentences with that word you use und which ones are meaningless to you”.

10. Three Objections

I will now consider three objections to the view that I have outlined. The first objection is that this approach puts too much trust in interpretation, both in general and with respect to self-interpretation in particular, and so runs a risk of missing what a proper philosophical analysis would reveal about the uses of religious language. The second criticism starts from a conception of meaning that connects it to the use of words and then asks to what extent an individual could be said to determine the meaning of the language that they use. The third objection grants that the view I have outlined is plausible, but that it is plausible because it is quite trivial and so of very little interest to philosophy.

The first objection can be motivated by a passage from the work of Stephen Mulhall. In his view, what counts philosophically “is not whatever religious believers are at first inclined to say about themselves—that is philosophical raw material, not its end-product—but rather how they in fact employ religious concepts in the practices which go to make up their lives” (Mulhall 2001, p. 108). The allusion here is to PI § 254, where Wittgenstein discusses “shades of meaning”:

What we are ‘tempted to say’ in such a case is, of course, not philosophy; but it is its raw material. So, for example, what a mathematician is inclined to say about the objectivity and reality of mathematical facts is not a philosophy of mathematics, but something for philosophical treatment.

(Wittgenstein 2009, § 254)

This example is instructive. It is highly tempting to conceive of mathematical facts as ‘objective’ and to spell out their objectivity in ways that are both similar and radically different when compared to the kind of objectivity that seems to be a characteristic of facts in the physical world. The similarity is, briefly put, that there seems to be some kind of reality that determines the relevant facts and determines them to be as they are independently of what we say or think about the matter. The difference is that we seem to have some idea of what the relevant reality is in the case of the physical world, but we have no clear idea at all of what the relevant reality might be in the case of mathematical facts. We are tempted to think of it as a kind of shadowy reality, a realm of abstract objects, a mathematical world, over and above the physical. But how could such a picture even give us an account of mathematics? All the same, there is a temptation to think that the structure of some shadowy reality or mathematical world determines the relevant facts and determines them to be as they are independently of what we say or think about the matter.

This would be a view not unlike Hardy’s, and if Wittgenstein is right, it is misleading and rests on a deep misunderstanding of the way mathematical language is used. Wittgenstein quips as early as 1929: “There is no religious denomination in which so much sin has been committed through the misuse of metaphorical expressions as in mathematics” (Wittgenstein 1998, p. 3e, MS 106 58: 1929).26 As is well known, Wittgenstein came to conceive of ‘truth’ in mathematics in terms of rules, and rules have a very different function than descriptions or representations of some supposed shadowy state of affairs.

Now, a first question is this: Is this clear from a cursory glance at the relevant expressions? Are people engaged in the practice of mathematics normally aware that mathematics should be thought of “as a collection of rules”, as opposed to, say, “the natural history of the domain of numbers” (Wittgenstein 1978, p. 230)? Clearly the answer is no, and Wittgenstein himself had to work hard to spell out and support this alternative view. Experts may misunderstand what they are doing, and they may be prone to being deceived precisely because they are deeply involved in the practice. If so, this will come out in what they say about the practice.

A second question is: How can we best make sense of the practice? How should we understand this area of discourse, given that we are prone to misunderstanding it? Wittgenstein’s aim in his philosophy of mathematics is to understand this practice properly, to give a convincing and conspicuous account of it. Sometimes “treatment” is required before new interpretations are available and can come to be seen as a more plausible conception of a practice that can easily appear to us at first in a very different light.

Considering those two sets of questions, informed interpretation still appears to be the key to understanding religious statements or expressions. The idea is not so much to simply take what people say or are inclined to say and to treat this as some kind of theory, although that may be helpful too. Rather, the idea is to make sense of what they say in the context of the practice, and this will involve informed interpretation of the things that people say and do and of how they understand that practice. My contention is that this will often reveal belief in a familiar sense, involving truth conditions, and that this will not be a merely superficial view of religious discourse that requires “treatment”, unless “treatment” is employed to fundamentally transform it.

Belief in a familiar sense can rarely be erased, or analysed away, without substantial changes in that practice. That this is so comes out precisely when those who are prepared to say that they believe in God or in an afterlife or in redemption through Christ are invited to spell out what they mean—what they think they are doing when they avow those beliefs, and what they would say if it turned out that in fact, there is no God, no afterlife, and no redemption through Christ. Informed interpretation gives us, and gives them, an answer to the question “how they in fact employ religious concepts in the practices which go to make up their lives”. In my view, it is highly unlikely that an interpretation of religious concepts and religious practices in purely non-cognitivist terms could be offered as anything other than a substantial revision of their own interpretation, and in my view, it is highly likely that such a revision would transform religious concepts and religious practices themselves.

I will now turn to the second criticism. Given that Wittgenstein approaches questions of meaning by considering the ways in which words are used, it can seem surprising to suggest that what a person says, and what she says she means by what she says, is typically the key to understanding what she means. But this should not come as a surprise. For one thing, we are not considering a ‘private language’, so there is no doubt that others are perfectly capable of understanding what a person says, just as the person can explain and interpret what she says in terms that are available to all. More importantly, it is not only possible but very common to “mean one thing rather another” when there is room for interpretation, and that is precisely where context and interpretation matter. Those are the kinds of cases where we often take what a person says she means to be the thing she means.

Of course, there are familiar misconceptions.27 Wittgenstein points out that ‘meaning someone’, for example, should not be construed as some kind of ‘inner pointing’ or as a matter of ’having a certain image in mind’:

If I say “I meant him”, a picture might come to my mind, perhaps of how I looked at him, and so forth; but the picture is only like the illustration to a story. From it alone, it would mostly be impossible to infer anything at all; only when one knows the story, does one know what the picture is for.

(Wittgenstein 2009, § 663, cf. §§ 661)

Again, what matters is the “story”—context, not a picture in the mind. But this is perfectly compatible with the idea that someone ‘meant‘ one person rather than another, using words or images that could be taken to represent either. Who was ‘meant’ can be articulated and explained, and it can be understood by others on the basis of the “story” being told.28

Of course, it is also possible to make a new connection, for example, by means of a definition:

But how come?—can’t I say “By ‘abracadabra’, I mean toothache”? Of course I can; but this is a definition, not a description of what goes on in me when I utter the word.

(Wittgenstein 2009, § 665)

Again:

“I meant this by that word” is a statement which is used differently from one about an affection of the mind.

(Wittgenstein 2009, § 676)

In a related discussion of ‘meaning’, Wittgenstein says, “All this points to a wider context” (Wittgenstein 2009, § 686). It seems to be a basic and important fact about religious language that it is deeply embedded in a context of attitudes, emotions, rituals, and performances, and that these permeate all areas of human life. There is certainly a multitude of “stories” to be told about the meaning of a given statement or expression, just as there is a multitude of “stories” to be told about a given ritual or practice, its significance and point.

If context is what matters, then a person may at times be wrong about the meaning or significance or status of something she says or does. Wittgenstein himself points this out in ‘Lectures on Religious Belief’:

Suppose somebody made this guidance for this life: believing in the Last Judgment. Whenever he does anything, this is before his mind. In a way, how are we to know whether to say he believes this will happen or not?

Asking him is not enough. He will probably say he has proof. But he has what you might call an unshakeable belief. It will show, not by reasoning or by appeal to ordinary grounds for belief, but rather by regulating for all his life.

(Wittgenstein 1966, pp. 53–54)

All the same, we will normally take what a person says she means to be enough, as long as she has said enough. What a person says she means will often serve as our criterion to establish what it is that she means.

If so, informed interpretation has to focus on the individual. ‘What exactly do you mean by that?’ is a familiar question, not just in philosophy. What a person says in response to this kind of question will quite often be the answer to the question of what she means: what she says she means is what she means, provided that her answer is sincere and is itself based on informed interpretation—in this case, her interpretation of her own beliefs, attitudes and actions. Even so, interpretations in this area will often be uncertain, subject to revision and reinterpretation. But that is what we should expect, and religious language is like other areas of language in this respect.

This brings me to the third worry, which is that the proposed view of religious language is vacuous, or trivial, or empty, and has little significance for a philosophical account of religious language. Does it amount to anything more than a familiar Wittgensteinian reminder that meaning is connected with use, and that use must be understood against the background of a practice, a context, something like a ‘form of life’? While this would clearly not be trivial, it would not be a distinctive view of religious language either.

There is a truth in this criticism that is worth bringing out, and doing so can help to bring out ways in which this view is, in fact, distinctive and philosophically significant. Moreover, it is philosophically significant in ways that Wittgenstein himself would recognise. To see why, it is helpful to distinguish different sets of questions that are located on different levels.

A first set of questions is this: How should we understand religious language? What is the meaning of religious statements such as ‘God exists’, ‘Christ died to relieve us of our sins’ or ‘There will be a resurrection of the dead’? How can we determine what their meaning is? The answer is that we have to engage in informed interpretation, consider context, ask the right questions, and listen. If we do this properly, we consider carefully what a person says she means when she avows belief, what she does when she is engaged in ritual or prayer, what she would say about what she is doing, and what other attitudes, beliefs and practices are bound up with what she says. A second set of questions is: Do religious statements or expressions have cognitive content? Is the belief that God exists like belief in its familiar sense, in that it involves substantial truth conditions? Or should we understand it in some other way, perhaps as an expression of an attitude, a stance towards the world, a certain way of living? Is belief of this kind sensitive to evidence, or does that question not arise? Can there be combinations of these different kinds of characteristics of different kinds of belief?

In light of the view that I suggest, answers to the second set of questions are in part determined by answers to the first. As a consequence of that, there may be very little to be said about the meaning of religious statements or expressions at the level of a general and abstract theory. If that is so, a general question about the cognitive content of religious belief will probably not have an interesting answer at the level of theory either. If we ask, ‘Is that belief in its ordinary sense, involving truth conditions?’, the answer will be: ’It depends’. It depends specifically on what a given person says and does, how she thinks of her own beliefs, how she acts on their basis, and what connections to her other attitudes or views or activities there are. To understand this properly will be the task of informed interpretation.

By contrast, such an understanding will not be achieved by appealing to a general theory of meaning or to a philosophy that offers sweeping generalisations where attention to detail and a sense for differences that lie below the linguistic surface are required. So it is correct to say that the appeal to informed interpretation is not exclusive to this area of language and also that it is an application of a Wittgensteinian approach to meaning. By itself, it does not commit us to a particular interpretation of religious statements or expressions.

It is therefore also correct to observe that by itself, this approach does not settle the question of whether religious statements or expressions have cognitive content or not, or what that content might be. But that is precisely the point. These are reminders, a “marshalling of recollections for a particular purpose”, as Wittgenstein says in Philosophical Investigations (Wittgenstein 2009, § 127). What we should question is an abstract, theoretical approach to religious language that attempts to give general answers to specific questions. These questions can only be answered by informed interpretation in a particular case. The truth may be that at the highest level of abstraction, there is not much of interest to be said.

There are some further consequences of this view that I will briefly mention. First, the appeal to informed interpretation can easily account for the great variety of religious practices and views and their different cognitive commitments. Sometimes, cognitive commitments are in the foreground, sometimes in the background, sometimes in the distance or in the surrounding area. Sometimes, cognitive commitments are quite vague, sometimes barely visible. Sometimes, they may simply not be there. All this is consistent once we give up the idea that there is a general interpretation of religious statements or expressions that philosophy would have to find.

Second, this approach can easily account for the fact that ‘mixed‘ strategies of interpretation often seem to be correct. A religious point of view frequently involves cognitive commitments, but evidently it involves much more than that. Belief that comes with truth conditions will be closely related to many other attitudes, emotions and beliefs. They are part of the background, or the context, of belief. Someone who says, ‘God is with me!’ will normally be taken to express much more than a proposition that is either true or false. At the same time, someone who puts forward an expression of belief explicitly as an expression of an attitude, as opposed to a belief in the familiar sense, is not making a mistake. Such a person tells us how we should interpret her beliefs. We understand what she is saying well enough, even though we may have further questions.

Finally, distinguishing the two levels of questions can help us to see more clearly where Wittgenstein himself takes a surprising turn. It also helps us focus on some aspects of his view of religious language that many find much less convincing than his overall approach. In closing, I will now consider some examples and discuss interpretations that Wittgenstein offers in his remarks on religious practice and belief.

11. Religious Disagreement

In his lecture of 1 May 1933, Wittgenstein continues his discussion of the “grammar” of “God” (Wittgenstein 2016, pp. 318, 320). He points out that religious language can be metaphorical and then goes on to say this:

There are many controversies about meaning of “God”, which could be settled by saying “I’m not using the word in such a sense as that you can say … ”

If parents believe that they have an answer to prayer, then in a sense they have an answer from God.

If we get a psychological explanation of how the idea came into their head, this doesn’t prove it didn’t “come from God”—that they were mistaken.

(Wittgenstein 2016, p. 321)

I take the first of the three remarks to be a general point about method. As such, it is reminiscent of informed interpretation as I have been using the term. Note that the idea is spelled out in the first person singular, not plural: “I’m not using the word in such a sense that you can say …”—as opposed to “We’re not using the word in such a sense that you can say …”, even though that will often also be appropriate to say.

The second and the third remarks, by contrast, seem to me more doubtful. It certainly seems possible, and even likely, that people think of prayer as genuine communication, involving some external agency. If so, something over and above mere psychological processes that ‘take place in the head’ must be involved if an occurrent thought is to be counted as an answer to a prayer. In any case, it rarely seems to follow from the fact that people believe that they have an answer to prayer that they have an answer to prayer. Of course, as before, it is possible to simply count a certain thought that occurs to a person as an answer to her prayer. If we do this, we require nothing over and above for such a thought to count as an answer to a prayer than what is undoubtedly present in this case—there is an occurring thought, and that thought is taken, seen, described, or treated in a particular way.29

But while this kind of interpretation is available, it seems to be unusual, to say the least—it leaves out something important, and that is probably the thought that in some sense, prayer is genuine communication, presupposing independent agents. At the very least, an interpretation of a thought as an answer to a prayer would seem to require some belief in an independent, higher agent and his agency to be present somewhere in the background, making some kind of difference to the life of the agent who prays. What Wittgenstein suggests simply does not fit what most people seem to think would be required for a thought to count as an answer to prayer, however vague their thoughts regarding those requirements may be.

Wittgenstein goes on to discuss the idea of ‘religious disagreement’, which may occur if two people get conflicting answers to their prayers. Again, Wittgenstein insists that it would be “wrong to doubt on this ground”: “Disagreement of 2 answers can quite properly produce a religious conflict; but this need not be religion of that kind which takes religious statements scientifically” (Wittgenstein 2016, p. 322). It is perhaps not fully clear what it would be to take religious statements “scientifically” in the relevant sense. In this context, the label seems to have to do with the idea of explanation, of causal inference, of testing, and of ruling out hypotheses, deciding between different accounts that are incompatible by means of an experiment. If we understand religious statements in this way, not more than one of two conflicting thoughts that both seem to be an ‘answer to a prayer’ can, in fact, be an answer to a prayer, while the other one must be considered an illusion, a mere psychological event. But once again, many people seem to think of religious statements and beliefs in something like this way, and it seems to be obvious that predictions, causal explanations, inferences and a wide array of factual assumptions are involved in almost all familiar forms of religious belief.30

Of course, it is possible to say that God might answer prayers differently and give inconsistent answers. Still, reference to God as a higher agency, to his intentions and decisions with respect to communication, cannot be erased completely without losing grip on the idea of an answer to prayer. Wittgenstein may well be right that a religious conflict need not be a conflict “of that kind which takes religious statements scientifically”. Some conflicts can be seen as conflicts of attitudes or as conflicts between different points of view that remain after thorough and honest reflection. But very often, there will be much more involved than that, and religious conflict will involve a conflict of belief in its familiar sense. Beliefs in this sense have substantial truth conditions. It follows that not more than one belief in a religious conflict of beliefs in this sense can be true.

Relations between science, explanation and religious or ritualistic practice are discussed at length in Wittgenstein’s remarks on Frazer’s Golden Bough. The topic is then touched upon in the lectures of May 1933:

In the Golden Bough, Frazer constantly makes one particular kind of mistake in explanation.

There have been 3 accounts of punishment (1) to deter (2) to improve (3) to take vengeance.

But if you ask “Why does a father punish his son?” there may be none of them, or all 3, or something between 2 of them.

(Wittgenstein 2016, pp. 325–26)

This is a criticism of reductive explanation, and I take it to be plausible as far as it goes.31 Wittgenstein examines practices, as described by Frazer, with a view to misunderstandings of the language that is being used. A proper survey of the language game can show the complexities of such a practice, its expressive functions, its relations to our own attitudes and emotions, and its motivations.32

As Wittgenstein points out, there is a parallel question about the explanation of a practice. Once again, there is a temptation to look for, and accept, some reductive explanation. In this case, the explanation would typically apply to a community, and it may postulate some origin of rituals or practices where their historical development remains unclear. As Wittgenstein points out, these explanations tend to be too simple: To the extent that we can offer explanations, they may have to be much more complex, and they may involve not only past events, beliefs and practices but also present attitudes:

Frazer talks of Magic performed with an effigy, & says primitive people believe that by stabbing effigy they have hurt the model.

I say: Only in some cases do they thus entertain a false scientific belief.

It may be that it expresses your wish to hurt.

Or it may be not even this: It may be that you have an impulse to do it, as when in anger you hit a table; which doesn’t mean that you believe you hurt it, nor need it be a survival from prehuman ancestors.

Hitting has many sides.

(Wittgenstein 2016, p. 326; cf. Wittgenstein 1993, pp. 123–25)

A ritual like this may have symbolic or expressive functions, and those functions would seem to be largely independent of historical accounts or causal beliefs or ‘scientific’ views about the efficacy of stabbing effigies. So understood, these rituals would not presuppose erroneous belief, and they would not be vulnerable to criticism in the name of some more modern and enlightened science. An understanding of the practice in terms of symbolic or expressive functions would consequently be compatible with a variety of different histories. There could be many narratives about the relevant beliefs and attitudes on the part of those who started it, or were engaged in it, in the past.

The relevance of this point for the purposes of our discussion is that none of this rules out that expressive or symbolic aspects of a practice are often entangled with belief, including historical belief and belief that is undoubtedly causal in character. How else are we to understand the thought that on the third day, Christ was resurrected, that God created Heaven and Earth, or that apostles performed miracles filled with the Holy Spirit? If we set aside all elements of belief in its familiar sense, or if we reject all that seems to be a factual account as superstition, we seem to be offering a radical revision of religion. It would certainly no longer be an instance of informed interpretation of what many, if not most, Christians believe.33

It should be pointed out, however, that this is not quite what Wittgenstein suggests in this passage about Frazer. What he says about those who ritually stab an effigy is that “only in some cases do they thus entertain a false scientific belief”; that doing so “may” express a wish, or “may” be motivated by an impulse to hurt the person, which is out of reach. So Wittgenstein would seem to agree that in some cases, false beliefs are still a part of the correct explanation, either with respect to the origins of a given practice or with respect to motives or to expectations of those who are currently engaged in it.

In other cases, we may not be sure, and informed interpretation turns out to be difficult. It is instructive that in ‘Lectures on Religious Belief’, he uses a scenario that is reminiscent of ’radical interpretation’:

We come to an island and we find beliefs there, and certain beliefs we are inclined to call religious. […] These statements would not just differ in respect to what they are about. Entirely different connections would make them into religious beliefs, and there can easily be imagined transitions where we wouldn’t know for our life whether to call them religious beliefs or scientific beliefs.

(Wittgenstein 1966, p. 58; cf. Davidson 2001)

In this respect at least, Wittgenstein seems to be less dogmatic than he may at first appear, and the passage illustrates how a given practice or belief should be understood will have to be decided by informed interpretation, given the facts of the case. A criticism of reductive explanation naturally serves as a reminder that religion is a highly complex, multi-faceted phenomenon. It certainly has “many sides”—but among the complex attitudes, emotions, symbols, metaphors and images, we will typically, and rightly, expect to find some solid bedrock of belief in its familiar sense, involving truth conditions. Everything is what it is, and not another thing—including the belief that God exists, that Jesus died for our sins, or that Apostles performed miracles while they were filled with the Holy Spirit.

12. Conclusions

I have outlined a Wittgensteinian view of religious language that revolves around the notion of informed interpretation. According to this view, there is no fact of the matter independently of context and informed interpretation as to whether some religious statement or expression has cognitive content, or what that content may be. This view naturally supports a pluralist interpretation of religious language. Typically, religious statements or expressions have a range of different functions and characteristics, and these go beyond a mere representation of religious facts. Attitudes, emotions, metaphors and symbols are involved, and there can be religious language games that are independent of all factual belief. But this is rarely true of religious belief as we know it.

Wittgenstein has rightly emphasised that interpretations of religious statements or expressions have to focus on the individual. We have seen that his own notes in his wartime diary can be read in a number of different ways, and it is at least possible to think that at the time, Wittgenstein himself would have been unable to give a clear account of what he did and did not believe, or of what it is to believe in the relevant sense. In other cases, it is fairly clear how a belief in God, in a Last Judgment or in miracles is to be understood. A Wittgensteinian conclusion we can draw from this is that there may be no such thing as an informative, general and abstract philosophical analysis of religious language.

With respect to Wittgenstein himself, we can conclude that many of the views we find expressed in lectures and in notes are more nuanced than they may at first appear to be. While he often takes a stance, Wittgenstein also highlights possibilities, corrects imbalances of interpretation, or criticises a reductive style of explanation. All this is compatible with an interpretation of most religious beliefs in broadly factual terms. In any case, a new conception of belief or a criticism of reductive explanation will not be effective as a criticism of the widespread view that religious statements or beliefs frequently express propositions and are frequently embedded in a network of propositions, all of which can be straightforwardly either true or false.