History of ideas

Words of God. Father Lequeux, priest of Buc, Île-de-France, and his puzzled congregation photographed in 1958. Photo by Edouard Boubat/Gamma-Rapho/Getty

Wittgenstein and religion

In the case atheists vs religious belief, Ludwig Wittgenstein is called to the stand. Whose side does his testimony serve?

Stephen Law

Words of God. Father Lequeux, priest of Buc, Île-de-France, and his puzzled congregation photographed in 1958. Photo by Edouard Boubat/Gamma-Rapho/Getty

Stephen Law

is the editor of the Royal Institute of Philosophy journal THINK. He researches primarily in the philosophy of religion. His books include The Philosophy Gym: 25 Short Adventures in Thinking (2003) and A Very Short Introduction to Humanism (2011). He lives in Oxford.

3,200 words

Edited by Nigel Warburton

Syndicate this Essay

Aeon for Friends

Find out more

When contemporary atheists criticise religious beliefs, they usually criticise beliefs that only crude religious thinkers embrace. Or so some people claim. The beliefs of the sophisticated religious believer, it’s suggested, are immune to such assaults. 

Those making this kind of response often appeal to the later work of Ludwig Wittgenstein (1889-1951) – in particular, to remarks he made in Lectures and Conversations on Aesthetics, Psychology and Religious Belief (1967) and Culture and Value (1970), both published posthumously. Wittgenstein made a number of interesting, if rather cryptic, comments about religious belief in these books, and did seem to suggest that such atheist criticisms miss their mark.

What follows is a brief guide to the leading ‘Wittgensteinian’ defences of religious belief, rooted in Wittgenstein’s later work. Note that it’s contentious what Wittgenstein’s later views about religious belief are. The views I discuss are not necessarily Wittgenstein’s own, but attributed to him. Examine these different positions more closely, and we find little to reassure most religious believers that their beliefs are ‘off limits’ so far as atheist criticism is concerned. This is not to say that contemporary atheist criticisms of faith are good – they might not be. It’s just that going Wittgensteinian provides little immunity to such attacks.

In Lectures, Wittgenstein said that, as a non-believer, he couldn’t contradict what the religious person believes:

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.
And then I give the explanation: ‘I don’t believe in …’, but then the religious person never believes what I describe.
I can’t say. I can’t contradict that person.

Indeed, Wittgenstein is widely interpreted as supposing that, not only can non-believers not contradict what the religious believe, they can’t refute those beliefs either. But why not?

It’s at this point that interpreters of Wittgenstein diverge. There’s a fairly broad consensus that Wittgenstein supposes the religious person and the atheist use sentences such as God exists, Jesus rose from the dead and There will be a Judgment Day very differently, and consequently with different meanings. However, this raises further questions:

  •  How exactly does the use differ?
  •  Why should we suppose that atheists must then fail to contradict – and indeed fail to refute – what the religious person expresses using such sentences?

It’s in answer to these questions that a number of different Wittgensteinian views are offered. I map out the three main contenders below.

According to what I’ll call the non-cognitivist Wittgensteinian, sentences such as God exists and Jesus rose from the dead are not used by the religious to make claims at all. But if no religious claim is made, then there is nothing for atheist critics to contradict or indeed refute.

If God exists is used not to make a claim but in some other way, how is it used? One suggestion is expressivism, which says that sentences such as God exists are used by religious people to express emotions or attitudes. For example, God exists might be used to express an intense form of optimism or a feeling of awe and reverence. Perhaps when religious people say God exists, they are in effect saying Oh wow! in amazement that the Universe exists at all. But if that’s how God exists is used, then to ask But how do you know that God exists? and What’s your evidence for God existing? is to ask questions that get no grip. Yet these are the questions to which atheist critics demand answers.

Non-cognitivist accounts of how religious language is used really do have the consequence that what religious people express using such language is not something that atheist critics might successfully contradict or refute. However, the downside to non-cognitivism is that it’s highly implausible as an account of how religious language usually functions. Wittgenstein himself encourages us not to assume how language is used, but to ‘look and see’. He also encourages us to focus on ordinary linguistic practice. But ordinary religious linguistic practice does not appear to be what the non-cognitivist claims. The vast majority of religious people – including most self-styled Wittgensteinians – appear to use religious language to make a variety of metaphysical, historical and other claims. It usually matters very much to them that these claims are true.

In particular, it matters very much to most Christians that, as a matter of historical fact, Jesus rose from the dead. Indeed, they often present evidence to support that historical claim, such as eyewitness reports of a postmortem Jesus. Most also appear to use God exists to make a claim about some sort of transcendent, perfect being – a claim they try to defend when presented with criticisms such as the evidential problem of evil (the argument that the depth of evil in the Universe provides good evidence against the claim that such an all-powerful, all-knowing and good deity exists). Why mount a defence of a claim if there is no claim to which they’re committed? It appears that, if there are religious folk who use religious language in a wholly non-cognitivist way, they form a tiny minority.

Given the obvious implausibility of non-cognitivism as an account of how religious language is generally used, many interpreters of Wittgenstein reject it. Their Wittgenstein allows for religious claims. Religious language isn’t purely expressive – or else, as the philosopher Genia Schönbaumsfeld points out in A Confusion of Spheres (2007), religious faith would boil down to nothing more than ‘some bizarre sort of rapture’. In his essay ‘The Tightrope Walker’ (2007), the philosopher Severin Schroeder concurs:

Contrary to [a] widespread view, Wittgenstein did not propound a purely expressivist construal of credal statements […] Wittgenstein stresses the importance of commitment, the practical dimension of religious faith, without denying that it is, or involves, also believing certain things to be true.

So, on this view, while (contrary to non-cognitivism) claims are made by religious folk who say God exists and Jesus rose from the dead, there are often further dimensions of meaning or significance that are lost on the atheist.

There are dimensions to the use and significance of religious language lost on the atheist naysayers

The religious philosopher John Cottingham in 2009 makes the Wittgensteinian point that: ‘Our language-games are interwoven with a web of non-linguistic activities, and cannot be understood apart from the context that gives them life.’ Cottingham then uses an analogy to explain what he supposes that atheist critics often miss out on. What the philosophical critic of religious belief attempts, he suggests, is the ‘fruit-juicer method’: an attempt to extract from religious belief the clear liquid of certain claims that can be examined in isolation, discarding the pulpy mush of context. However, a juice extractor does not, as might first be supposed, give us the true essence of a fruit. Rather,

what it often delivers is a not very palatable drink plus a pulpy mess. Someone who has tasted strawberries only via the output of the juicer, and has firmly decided ‘this is not for me’, may turn out to have a radically impoverished grasp of what it is about the fruit that makes the strawberry lover so enthusiastic.

Schönbaumsfeld similarly interprets Wittgenstein as suggesting that there are dimensions to the use and significance of religious language lost on the atheist naysayers. Just as someone who lacks a ‘musical ear’ will not be able to contradict the judgment of a music connoisseur – such a person lacks the musical sensibility even to understand what the connoisseur is saying – so, she argues, Wittgenstein supposes that non-believers cannot contradict what the religious person says, for they also lack the relevant concepts.

Of course, Schönbaumsfeld allows that both Wittgenstein and atheist critics can parrot the words said by religious people, in much the same way that someone lacking a musical ear can parrot the words used by a music connoisseur (they too might talk about who the great composers are, the laws of counterpoint, and so on). However, the most that those lacking a musical ear or religious sensitivity can acquire is a purely ‘intellectual’ understanding of the subject, ‘comparable to having learnt a code’. On Schönbaumsfeld’s view, in order to be able actually to contradict the music connoisseur or religious person, we require ‘the kind of understanding that makes the musical work or the prayer (the religious words) live for me, not the kind that allows me to parrot a form of words’.

There is some plausibility to this ‘Wittgensteinian’ view – let’s call it the juicer view – that there are rich layers of meaning and significance involved in the religious use of God exists etc that are often lost on atheists. It’s not far-fetched to suggest that many atheists have only an impoverished ‘juicer’ understanding of what the religious believe.

However, Schönbaumsfeld’s Wittgenstein would be wrong to conclude that, because there are these rich layers of meaning and significance lost on atheists, so atheists must then be unable to contradict what the religious believe.

Consider this analogy. Suppose Mary overhears Tom say Otto is a Kraut. In so saying, Tom clearly communicates his belief that Otto is German. However, Tom could also be doing much more than that. He might, in using these offensive words, intend to express his contempt for Germans, and thus Otto. The real point of Tom’s utterance might be to insult Otto.

Now suppose that Mary suffers from a condition that makes her, as it were, insult-blind. Consequently, Mary fails properly to understand the full use of the word ‘Kraut’. In particular, Mary fails to grasp the way in which Tom is using that expression on this particular occasion. The rich and varied use of the kind of vocabulary to which ‘Kraut’ belongs is entirely lost on Mary, who consequently thinks ‘Kraut’ just means ‘German’. And so, knowing that Otto isn’t German, she says to Tom: No, you’re mistaken: Otto is not a Kraut – he’s not German. Has Mary succeeded in contradicting Tom?

Surely she has. True, Mary might have only a thin, insult-blind, ‘juicer’ understanding of Tom’s Otto is a Kraut, but that doesn’t prevent her from successfully contradicting, and indeed straightforwardly refuting, what Tom said.

But then similarly, even if atheists have only a thin, religious-meaning-blind, ‘juicer’ understanding of what the religious person expresses using God exists etc, it doesn’t follow that they can’t contradict the religious beliefs expressed using such sentences. Nor, of course, does it follow that atheists can’t straightforwardly refute what the religious believe.

But perhaps what Wittgenstein and Schönbaumsfeld intend to suggest is that there really is no overlap at all between what the religious are committed to and that to which the atheist supposes they are committed?

What if the religious person’s meaning is metaphorical, for example? Suppose I say Tom is the moth to Jane’s flame. If metaphor-blind Mary overhears me and replies But Jane clearly isn’t a glowing body of ignited gas, and Tom lacks wings, it becomes clear that Mary has entirely misunderstood. There is little if any overlap between what Mary understood and what I actually meant. Certainly, Mary has entirely failed to contradict me.

But then perhaps what the religious mean when they say God exists is similarly wholly metaphorical, with the result that the atheist critic entirely misunderstands what the religious person is committed to? Call this the strong juicer view.

If I shout Checkmate! during a game of gin rummy, I will cause complete bafflement

The problem with the strong juicer view is that, like non-cognitivism, it’s implausible as an account of how religious language is generally used. When the religious person says: God exists, and atheists respond by pointing to the depth of evil in the world, the religious person typically tries to account for the evil by offering theodicies – explanations for the evil (This evil – war – is a result of acting freely. That evil is there to build our characters) – or by appealing to mystery (For all we know, God exists and has good reason to allow these evils). If what the religious person meant were entirely metaphorical, such religious responses would make as much sense as my responding to Mary by pointing out that the evidence that Jane is not a glowing body of ignited gas and that Tom lacks wings is less than decisive, or by insisting that for all we know Jane literally is a glowing body of ignited gas.

A variant of the strong juicer view starts with the Wittgensteinian thought that our use of language is embedded in certain ‘language-games’ and ‘forms of life’. Suppose that during a game of gin rummy I shout Checkmate! Given that ‘checkmate’ has its home in an entirely different game bound by very different rules to gin rummy, my remark will likely result in complete bafflement. Indeed, it’s not clear that I would have succeeded in saying anything at all. Some suggest that, similarly, utterances such as Jesus rose from the dead have their home in religious language-games, and when they are placed in different language-games, bound by different rules (eg, in a science lab or academic history class), they too are stripped of content. But then such utterances offer nothing for the scientist in her lab or the historian in her academic class to dispute. Scientific and historical criticism of religious belief is impossible.

Again, this variant strong juicer view is implausible as an account of how religious language is generally used. The religious do themselves regularly use God exists, Jesus rose from the dead and so on in scientific and historical contexts – using fine-tuning arguments for God’s existence, and historical arguments for the resurrection. And even if it is true that talk of God’s existence and Jesus’ resurrection have their home within a religious form of life to which non-believers fail to belong, it does not follow that those non-believers cannot reasonably contradict and reject such beliefs. After all, accusations made during the 17th-century Salem witch trials had their home in a religious form of life to which contemporary atheists fail to belong, but that doesn’t preclude contemporary atheists contradicting and indeed refuting 17th-century religious accusations of witchcraft.

A third view attributed to Wittgenstein regarding religious language is what I dub the atheist minus view. On this, the religious person is committed, not to more than the atheist supposes, but to less.

Here’s a simple illustration. Suppose Ted has taken the analogical representation of God as a father looking down on us from the clouds too literally. Ted thinks that when religious people commit to the existence of God, they commit to the existence of a physical person who is literally sitting on a cloud somewhere overhead, gazing downwards. And so, in order to refute that belief, Ted inspects every cloud and finds no such person on any of them.

Clearly, when Ted says There is no God, using ‘God’ as he understands it, Ted really does fail to contradict the religious person. Further, while Ted might have refuted what he understands God exists to express, he’s clearly failed to refute what our sophisticated theist uses God exists to assert.

Most atheist critics of religion are not, of course, guilty of such a ridiculous misunderstanding. But perhaps they’re guilty of a subtler one. Theologians often stress that God is not a ‘thing’: God is not some sort of extra item in addition to the Universe and its contents. Indeed, there is no such ‘thing’. Thus in Faith Seeking (2002), the theologian Denys Turner says to the atheist:

It is no use supposing that you disagree with me if you say: ‘There is no such thing as God.’ For I got there well before you.

Given that atheists understand God to be a ‘thing’ that exists in addition to all other existent things, then however else they might characterise God, in denying that ‘God’ exists, atheists will indeed fail to contradict what theological sophisticates such as Turner believe.

None of the Wittgensteinian accounts of religious language offers the religious a response to atheists

Interestingly, Wittgenstein himself says that, when it comes to God, what is at issue is not ‘the existence of something’. So perhaps Wittgenstein would make the same point that Turner makes: that the atheist’s denial of such a ‘thing’ as God is a denial of something no sensible religious person believes. However, this third ‘Wittgensteinian’ view of how religious language functions also faces significant objections.

First, the atheist minus view is implausible for many religious sentences. Consider Jesus rose from the dead. Are we to suppose that what the Christian means by it is rather less than what the atheist understands by it? Surely not: most Christians really are committed to a physical resurrection as a matter of historical fact, just as their atheist opponents suppose.

Second, even if atheist critics have misunderstood what’s meant by God exists – perhaps thinking of God as an extra ‘thing’ in addition to the Universe and its contents – it doesn’t follow that they haven’t refuted what the religious commit themselves to when they say God exists. Suppose for example, that Mary shows there’s no all-powerful, all-good deity. It’s no good the believer saying in reply But Mary, you understand ‘God’ to refer to an all-powerful, all-good being that lives on a cloud, and I believe no such thing! The fact is, Mary might have shown there’s no God, even as the believer understands that term, whether or not she’s guilty of any such misunderstanding.

Third, even if the atheist critic has, according to the atheist minus view, misunderstood what the sophisticated Christian means by ‘God’, that doesn’t mean she’s incapable of grasping what the Christian means. Indeed, if the atheist minus view is correct, grasping what the sophisticated religious person means is straightforward enough: all the atheist need do is cross off ‘is a thing’, ‘lives on a cloud’, etc from the list of divine attributes.

None of the above Wittgensteinian accounts of how religious language is used offers mainstream religious people much in the way of a response to atheist critics. Either the account fails to provide the kind of immunity for which many self-styled Wittgensteinians had hoped, or else the account is implausible as an account of how most religious people – including most self-styled Wittgensteinians – use religious language. As the philosopher John Searle once said: ‘You have to be a very recherché sort of religious intellectual to keep praying if you don’t think there is any real God outside the language who is listening to your prayers.’

Stephen Law

is the editor of the Royal Institute of Philosophy journal THINK. He researches primarily in the philosophy of religion. His books include The Philosophy Gym: 25 Short Adventures in Thinking (2003) and A Very Short Introduction to Humanism (2011). He lives in Oxford.
Syndicate this Essay
Get Aeon straight
to your inbox
Join our newsletter
Aeon is not-for-profit
and free for everyone
Make a donation
Thinkers and theories
Susan Sontag was a monster

She took things too seriously. She was difficult and unyielding. That’s why Susan Sontag’s work matters so much even now

Lauren Elkin

History of science
Natural philosophy redux

The great split between science and philosophy must be repaired. Only then can we answer the urgent, fundamental problems

Nicholas Maxwell