Wittgenstein's Language Games

Wittgenstein's language games successfully defeats Ayer's verification principle and Flew's falsification challenge, but in doing so it risks generating a fideist insulation of religious belief that many theologians and ordinary believers find unsatisfactory.

Wittgenstein argues that the meaning of a statement is its use within a language game, not its capacity for empirical verification. Religion is a distinct language game — a form of life encompassing prayer, ritual, moral commitment, and community — and the rules governing meaning within that game are internal to it. The statement "God is love" is meaningful within the religious form of life because it shapes how believers live, pray and relate to others; it does not need to be verified by laboratory experiment any more than the statement "checkmate" needs to be verified by physics. This directly answers Ayer: the verification principle is itself just a rule of the scientific language game, and there is no neutral meta-game from which to declare all other games invalid. It also answers Flew: if religious believers do not allow empirical evidence to count against "God loves us," this is not, as Flew claims, a sign that they are asserting nothing — it is a sign that they are playing a different game in which empirical evidence is not the relevant move.

However, the very feature that makes language games effective against positivism — the insulation of each game from external criticism — also generates the charge of fideism: the view that religious belief is immune from rational assessment and criticism. If no one outside the religious game can legitimately criticise religious claims, then religious believers are equally unable to criticise other games — including morally abhorrent ones. As the PEPED analysis notes, "you cannot externally criticise that use" — but this means that a religious community that uses "God commands genocide" cannot be criticised by secular ethics, since ethics is a different game. Furthermore, Kai Nielsen charges that language games leads to relativism: if meaning is entirely internal to each language game, then the religious claim "God exists" is true within religion and false outside it — which seems to make truth entirely relative to community, rather than tracking anything real.

D.Z. Phillips, Wittgenstein's most prominent religious advocate, responds that fideism misunderstands the theory: language games does not insulate religious belief from all criticism, but only from inappropriate external criteria — it does not follow that religion is beyond internal rational assessment. A community's own game can be criticised from within its own rules: "this prayer is incoherent," "this reading of the text is inconsistent," "this practice contradicts the community's own stated values" are all legitimate criticisms that language games permits. On relativism, Phillips argues that "God exists" is not merely a community convention but an attempt to express genuine commitment to ultimate reality — the game is not arbitrary but reflects deep truths about human existence and value.

Phillips' response partially defuses the fideism charge — internal criticism is preserved — but does not fully resolve it. If external criticism is genuinely prohibited, then the language games framework cannot support the kind of inter-religious dialogue and ethical assessment of religion that contemporary theology and society require. The relativism problem is more serious: as the PEPED analysis concludes, "language games theory fails to take seriously what ordinary religious believers take themselves to be saying" — most believers intend their religious claims to be universally true, not merely true within their own game. Language games thus defeats positivism but creates new problems about rational accountability and the realist aspirations of religious belief.

Wittgenstein's language games therefore resolves the logical positivist issue effectively, but the fideism and relativism charges reveal that this resolution comes at a significant cost to the rational credibility and universalist aspirations of religious language.

The most serious limitation of language games as a resolution to the issues of religious language is that it misrepresents how ordinary religious believers understand their own claims — most intend them to be cognitively true, not merely functionally meaningful within a community.

Wittgenstein, and especially D.Z. Phillips, treat religious statements as non-realist expressions of a form of life rather than truth-claims about an independently existing God. Phillips famously argues that "God exists" is not a claim about a being whose existence could be confirmed or denied by investigation, but an expression of ultimate commitment and orientation towards life. This makes language games a non-cognitive theory: religious statements are meaningful because of what they do — inspire, orient, unite — not because of what they describe. The theory therefore "resolves" the positivist challenge by accepting its terms on one important level: it concedes that religious language is not empirically descriptive, and argues that this is not a deficiency.

However, as Peterson et al. argue and PEPED confirms, this non-realist interpretation does not accurately represent what most religious believers take themselves to be saying. When a Christian says "God raised Jesus from the dead," they typically intend this as a historical and metaphysical claim — that something actually happened in the world, not merely that this statement plays a certain role in community life. When a Muslim affirms "there is no god but Allah," they intend this as a claim that is universally true, not true-within-Islam in the same way that "checkmate" is only valid within chess. Language games, by reducing religious language to its functional and communal role, effectively evacuates the cognitive content that religious believers themselves consider central — and in doing so, it resolves the problem of religious language by redefining it into something believers would not recognise.

Wittgenstein can respond that his theory is not a revisionist programme but a descriptive philosophy: he is not telling believers what they should mean, but clarifying what they actually do when they use religious language, which is more complex and form-of-life-embedded than they consciously realise. Furthermore, the alternative — insisting that religious language is cognitively truth-apt and assessed against empirical standards — leads straight back into the logical positivist trap, where religious language fails the verification test. Language games at least preserves genuine meaningfulness, even if at the cost of cognitive realism.

This response highlights the genuine dilemma that language games exposes: religious language cannot be both fully cognitively realist (and thus subject to Ayer's and Flew's challenges) and empirically safe (and thus non-cognitive). Wittgenstein chooses the non-cognitive route and preserves meaningfulness, but the PEPED verdict stands: language games is "only moderately successful" because it emphasises the functional aspect of religious language at the expense of its informative element. Hick's eschatological verification offers an alternative that preserves both: religious statements are cognitively meaningful and in principle verifiable — not empirically now, but eschatologically after death. This means Wittgenstein does not fully resolve the issues so much as trade one set of problems for another.

Language games therefore fails to offer a fully satisfactory resolution because it accurately describes the communal and practical dimension of religious language while misrepresenting — or simply setting aside — the cognitive, realist dimension that most believers consider essential.