“For If There Is No Resurrection of the Dead, Then Christ Has Not Been Raised Either”: Wittgenstein and the Cognitive Status of Christian Belief Statements

Abstract

The article engages in a critique of Wittgensteinian non-cognitivism about Christian belief statements but argues that Wittgenstein himself can only partially be classified under the non-cognitivist label. The article has three parts. First, it argues, contrary to a common view in the literature, that Wittgenstein moved not only from strong non-cognitivism to moderate non-cognitivism but also from general non-cognitivism towards partial cognitivism about Christian belief statements. Second, it examines three separate non-cognitivist Wittgensteinian claims about Christian belief statements and argues that these claims are untenable not only from a broad perspective, but also from the point of view of Wittgenstein’s Philosophical Investigations. The three claims are: (1) “The cognitive aspect of Christian belief statements is superstition only”, (2) “The cognitive aspect of Christian belief statements is parasitic on religious attitude only”, and (3) “The cognitive aspect of Christian belief statements is cognitive in surface grammar only”. Third, the article considers the case of believing Christian martyrs and follows the view that their religious belief might be cognitive only at the surface grammar level, through to its ultimate consequences. The article argues, using Wittgenstein’s Philosophical Investigations, that the non-cognitivist position ends up with a cognitive point about the correctness or incorrectness of the Christian martyr’s depth grammar.

1. Introduction

Non-cognitivism about Christian belief statements can take several forms. A radical version denies that they have any cognitive content at all. We can call this strong non-cognitivism. It must be distinguished from a more moderate version, which concedes that Christian belief statements can have cognitive content but denies that their grammar includes truth-aptness, as do other cognitive beliefs. Consequently, both strong and moderate non-cognitivism imply that Christian belief statements are not eligible for the “truth-game” (Pichler and Sunday Grève 2024), i.e., truth-aiming and truth-capable discourse, and do not contribute to it either. Wittgenstein wrote in 1937 that “historical proof (the historical proof-game) is irrelevant to belief” and that “the historical accounts of the Gospels might, in the historical sense, be demonstrably false, & yet belief would lose nothing through this” (Wittgenstein 1998, pp. 37e–38e = Ms-120,42r-42v).1 Bremer calls this “irrationalisation” and “immunisation” of religious belief against “the truth-oriented historical and scientific perspective” (Bremer 2013, p. 150; my translation).2 In this article, non-cognitivism that applies to all Christian belief statements is called general non-cognitivism, while non-cognitivism that is restricted to only some Christian belief statements is called partial non-cognitivism. The same distinction between general and partial is applied to cognitivism, resulting in general cognitivism and partial cognitivism.

Pichler and Sunday Grève (2024) have argued, as many have before, that general non-cognitivism about Christian belief statements, whether strong or moderate, is mistaken. Moreover, they have argued, that Wittgenstein—while advocating strong non-cognitivism in Tractatus (Wittgenstein 1922) and the Lecture on Ethics (Wittgenstein 1993a)—later abandoned it and instead gave expression to moderate non-cognitivism in the Lectures on Religious Belief (Wittgenstein 1966) from 1938–39.3 Again, this point has been argued before (see, for example, Jäger 2003, pp. 238–39; Schroeder 2007, p. 443). Most importantly, however, Pichler and Sunday Grève (2024) have argued, contrary to a dominant view in the literature (see Schönbaumsfeld 2023), that Wittgenstein eventually left even general moderate non-cognitivism behind.

This article goes on to argue, especially in Section 2.3 and Section 3, that general non-cognitivism about Christian belief statements is mistaken. In addition, the article continues to engage with the question of the extent to which Wittgenstein endorsed non-cognitivism about Christian belief statements and presents additional evidence for the view that later Wittgenstein moved not only from strong non-cognitivism (“Christian belief statements have no cognitive content”) to moderate non-cognitivism (“While Christian belief statements can have cognitive content, they do not participate in the truth-game”), but also from general non-cognitivism (“All Christian belief statements are non-cognitive”) towards partial cognitivism about Christian belief statements (“Some Christian belief statements are cognitive and do participate in the truth-game”).

In short, the article argues for two main points: First, it argues that Wittgenstein eventually became open to the possibility that some Christian belief statements are cognitive and do participate in the truth-game, or, alternatively, that it is possible for the same Christian belief statement to be non-cognitive in one context and cognitive in another. Second, it argues that partial cognitivism about Christian belief statements, and therefore also the kind of cognitivism attributed to Wittgenstein in this article, is correct.

2. Wittgensteinian Non-Cognitivism

2.1. Non-Cognitivism About Christian Belief Statements

It is difficult to define “non-cognitivism” about Christian belief statements in absolute terms. Therefore, the controversy will be introduced through examples. According to strong non-cognitivism, the statement “We will meet again after death” has no cognitive content. Christian belief in continued or, after a pause re-established, personal existence after death is interpreted by strong non-cognitivism in non-factual, antirealist and purely attitudinal, emotivist, expressivist, or other non-cognitive terms. Such an interpretation is raised by C. Lewy at one point during Wittgenstein’s Lectures on Religious Belief from 1939: If someone says “We might see one another after death”, “you might only mean that he expressed a certain attitude” (Wittgenstein 1966, pp. 70–71; Wittgenstein 2017, p. 131, dated by the editors 11.12.1939).

In distinction to strong non-cognitivism, moderate non-cognitivism holds that Christian belief statements can have cognitive content. This is a position endorsed by Wittgenstein in the same Lectures. There, Wittgenstein, in response to C. Lewy’s remark, denies that someone who says “We might see each other after death”, is merely expressing “a certain attitude”. Rather, Wittgenstein says, first, that he might be able to “understand him entirely”, and second, that saying “We might see one another after death” “isn’t the same as saying ‘I’m very fond of you’—and it may not be the same as saying anything else. It says what it says. Why should you be able to substitute anything else?” (Wittgenstein 1966, pp. 70–71; Wittgenstein 2017, p. 126) Thus, in December 1939, Wittgenstein contradicts the view that all Christian belief statements are to be interpreted purely non-cognitively as expressions of feelings, sentiments, values, or attitudes, and thereby clearly distances himself from a position of general strong non-cognitivism. The meaning of the statement “We might see one another after death” is not “I’m very fond of you” but precisely “We might see one another after death”.

A separate question, however, is whether the cognitive content now conceded to at least some Christian belief statements should also be treated as accessible to and eligible for participation in the cognitive discourse of justification and knowledge, and thus, whether the statement “We might see one another after death” can, should, and may be treated as subject to rational argument and the truth-game. On this point, Wittgenstein, in the Lectures on Religious Belief, continues a non-cognitive stance: Christian belief statements have no place in the cognitive and argumentative discourse and are not truth-apt. In this respect, ordinary belief and Christian belief are completely different for Wittgenstein. A key difference lies for him in how disagreement about the belief is expressed and addressed in each case. Disagreement with and within Christian belief manifests itself very differently from disagreement with and within other kinds of beliefs. In ordinary belief, if one believes that p, that belief is negated by simply denying that p; this is, according to Wittgenstein, not the case with Christian beliefs. Wittgenstein further suggests that it is entirely unclear what the negation of a Christian belief would even look like (see Wittgenstein 1966, pp. 53–59).

Thus, even non-cognitivism in its moderate version endorses a sharp epistemological or grammatical distinction between Christian belief and ordinary belief. Unlike other factual belief statements, apparently factual Christian belief statements are held to be unrelated to truth-making facts (other than psychological and sociological ones). Therefore they cannot be true, nor can they be false. Wittgenstein writes in 1947 that religious belief can ultimately only be a passionate commitment to a system of religious reference points and thus, rather than a belief, “a way of living, or a way of judging life” (Wittgenstein 1998, p. 73e = Ms-136,16b). Hyman comments on this: “[…] if a religious belief is something like a passionate commitment to a system of reference—as opposed to a passionate commitment to the truth of an empirical proposition—then a religious belief cannot be true or false” (Hyman 2001, p. 6). The statement “We will meet again after death”, the statement “God created the earth”, or the statement “God is one in three persons” as well as other Christian belief statements are therefore, according to the non-cognitivist interpretation, neither true nor false, nor do they help to determine the truth or falsity of other statements. Christian belief statements are treated as exempt from the standard epistemological expectations and discourses that apply to ordinary and scientific cognitive belief statements, such as “I believe it rains in Bergen today” or “I believe there is life on Mars”. Therefore, with Citron (2012), one could also speak of the “non-scientific” interpretation of Christian belief statements.

The non-cognitivist interpretation typically shifts the focus from the truth of the statement (for example, correspondence to facts, eligibility for non-subjective evidence, argument, and reasoning) to the truthfulness of the believer (for example, congruence with their attitude, conscience, sentiments, values, personality, or worldview): “it is not crucial whether the things they say are true or not, namely if there is a correspondence between their propositions and reality. What matters is that their lives are true” (Venturinha 2019, p. 103). Consequently, according to the non-cognitivist position, it does not matter for Christian belief whether Christ rose from the dead; what matters is that the Christian lives as if he did. To his close friend M. O’C. Drury who said that the New Testament would lose “its significance if it wasn’t an account of what really happened”, Wittgenstein replied: “It would make no difference if there had never been a historical person as Jesus is portrayed in the Gospels” (Wittgenstein 1984a, p. 101). To understand Christian belief correctly, one must move away from its seemingly factual, historical, and empirical dimensions altogether and keep it entirely outside the truth-game. According to non-cognitivism, the factual truth of the Gospels is irrelevant to Christian belief, and true Christian belief does not, or at least should not, treat the historical accounts of the Gospels in terms of historical truth. The truth-game, as in ordinary and scientific belief, has no role here; one could say that only faith does. However, this view is at odds with the way many religious believers believe. Many “would maintain that these events took place exactly as described in the New Testament” (Bailey 2001, p. 129).

The Wittgensteinian non-cognitivist approach to Christian belief statements was early labeled “fideism”.4 In contrast to fideism, most non-Wittgensteinian philosophers and theologians have endorsed the cognitivist position that evidence, argument, truth, historicity and the empirical have their proper place in religious belief. So have also prominent and paradigmatic Christian believers themselves. For example, J.H. Newman, an influential Christian believer who was also esteemed by Wittgenstein (see Kienzler 2006; Kfouri 2023), says:

Nothing would be more theoretical and unreal than to suppose that true Faith cannot exist except when moulded upon a Creed, and based upon Evidence; yet nothing would indicate a more shallow philosophy than to say that it ought carefully to be disjoined from dogmatic and argumentative statements.

(Newman 1909, p. 13)

St. Paul says and argues, or the Church has him say and argue, that Christ has factually risen from the dead:

He was buried, and he was raised from the dead on the third day, just as the Scriptures said. He was seen by Peter and then by the Twelve. After that, he was seen by more than 500 of his followers at one time, most of whom are still alive, though some have died. Then he was seen by James and later by all the apostles. Last of all, as though I had been born at the wrong time, I also saw him.

(Holy Bible 2015, 1 Corinthians 15:4–8)

The following three statements may serve as examples for philosophers of religion and theologians critical of Wittgensteinian philosophy of religion:

My account of what D.Z. Phillips has been claiming over many years in fidelity to Wittgenstein’s few explicit remarks on the subject, is that religion is a self-contained practice (of prayer, worship, public and private conduct, and the way we think about things), commitment to which involves no metaphysical or historical beliefs different from those of people who do not practise the religion. As an account of the Christian religion, as it has been practised by so many over two millennia, this seems manifestly false […] Wittgensteinian philosophy of religion suffers from a very one-sided diet of examples.

(Swinburne 2001, pp. 16–17)

[Wittgenstein’s] remarks on religion do little more than assist us in identifying the interpretation of religious discourse favoured by Wittgenstein: they do not give us any substantial grounds for accepting that this interpretation adequately represents the content of religious statements as these are used by most people who think of themselves as religious believers.

(Bailey 2001, p. 122)

The main problem with non-cognitive reinterpretation proposals is that they are not adequate in terms of religious phenomenology.

(Löffler 2019, p. 124; my translation)5

One anonymous reviewer of this paper has challenged the cognitivist interpretation of the above passage from 1 Corinthians. They argue that this interpretation would—by ascribing to St. Paul that his statement is meant as factually and literally as, for example, “he was buried” and “his body was stolen”,—only beg the question. What one should do instead, especially from a Wittgensteinian point of view, as also another anonymous reviewer has argued, is to pay close attention to the meaning of the words in their contexts and thus also to contextual factors. Any potential cognitive claims of the passage will depend on this. This is certainly a valid point. However, just as the cognitivist would have to show that the statement makes a factual claim, the non-cognitivist would have to show that it does not. All that needs to be argued at this point is that the factual reading is possible, and, more importantly, that there are in fact Christian communities who, without or with attention to exegetical factors, read the statement in the factual way and who hold and practice the belief that Christ rose from the dead in the same sense of “factual” that his body could simply have been stolen. Any question or further dispute about whether St. Paul himself believed in a bodily or “merely” spiritual resurrection of Christ, whether he and Christians believe in their future bodily (even if only with a “transformed” body, see Wittgenstein 1998, p. 38e = Ms-120,54v; Wittgenstein 2015b, p. 50) or “merely” spiritual resurrection, whether there was a bodily or “merely” spiritual resurrection of Christ or none at all, and whether there will be a bodily or “merely” spiritual resurrection or none at all only proves the cognitivist point that there is a logical space for factual religious beliefs in Christian belief grammar. Throughout the history of Christianity, and even today, it matters to Christians whether the belief that Christ rose from the dead is factually true, or rather the belief that his body was stolen or decomposed in the tomb. There are Christians who, if they find that it is more likely that Christ’s body was stolen than that he was raised from the dead, will leave their Christian faith behind. Therefore, with one of Phillips’s interlocutors, one can argue that taking St. Paul’s statement in the “naïve” factual sense of the words is indeed a legitimate part of Christian belief grammar (which is not to say that it is the only way in which it is and has been taken by Christian communities):

Is it only a conceptual or logical confusion—as Wittgenstein seems to suggest—that a Christian can come to question his or her entire religious faith because he or she becomes convinced, on historical grounds, that the historical Jesus was utterly unlike the figure pictured in the four Gospels? Wittgenstein seems to argue in one place that it would make no difference to Christianity whether the Gospel narratives were true or false, that Christian faith is simply accepting a narrative (whether fictional or not) and orienting one’s life by it. Wittgenstein may think this but I do not believe it ‘does justice’ to the way in which historical and religious beliefs actually are related in Christian piety.

(“I” in Phillips 2005b, p. 203)

Child puts the point in these words:

[…] it seems extremely plausible that religious beliefs of this sort [for example, the belief in Christ’s resurrection or the belief in a Last Judgment day] do generally involve straightforwardly factual components. Whatever else the religious belief in a Judgement Day involves, it seems clear that it involves an empirical belief about the occurrence of a future event. If there will, in fact, never be any process of judgement in which a supernatural being holds people to account for what they have done, the religious belief is mistaken.

(Child 2011, p. 225)

2.2. A Development in Later Wittgenstein

One point of this article is to give further support to the view that the later Wittgenstein left behind even general moderate non-cognitivism, to which he apparently still adhered in the Lectures on Religious Belief and, instead of remaining merely agnostic about those Christian beliefs that he no longer regarded as belonging to strong or moderate non-cognitivism, moved further in the direction of cognitivism. The later Wittgenstein began to express doubts about the correctness of the validity of a sharp distinction between Christian belief statements on the one hand and ordinary and scientific belief statements on the other, and he, for example, became open to the possibility that ordinary disagreements might be found in the case of Christian belief as well. Consequently, while much of the criticism of Wittgenstein quoted above may apply to Wittgensteinian philosophy of religion, it may not be justified with regard to Wittgenstein himself.

In 1937 Wittgenstein had banned rational understanding from the realm of Christian belief. The “paradox” and the “irritation” (“Reizung”) and “tickling” (“Kitzeln”) that arise when one takes Christian belief statements cognitively and tries to understand them are to be avoided. Instead of believing that Christ did this and that, and believing in Christ (“Glaube an Christus”), he recommended “love of Christ” (“Liebe zu Christus”):

I believe: the word “believing” has wrought horrible havoc in religion. All the knotty thoughts about the ‘paradox’ of the eternal meaning of a historical fact and the like. But if instead of “belief in Christ” you would say: “love of Christ”, the paradox vanishes, that is, the irritation of the intellect. What does religion have to do with such a tickling of the intellect. (For someone or another this too may belong to their religion.)

It is not that now one could say: Yes, finally everything is simple—or intelligible. Nothing at all is intelligible, it is just not unintelligible.—(Wittgenstein 2023, pp. 106–107 = Ms-183,238–239)

To Norman Malcolm Wittgenstein had expressed the view “that a way in which the notion of immortality can acquire a meaning is through one’s feeling that one has duties from which one cannot be released, even by death” (Wittgenstein 1984b, p. 59, undated). However, in 1943, he is reported to have said to Rush Rhees, that “he could understand how anyone should believe that he would have to stand with a queer sort of body (einem verklärten Körper) before a judgment when his life was done” (Wittgenstein 2015b, p. 50; also see Wittgenstein 2009b, iv-§23). In a letter to Arvid Sjögren from 1947, Wittgenstein writes:

There is, strange as it may sound, something that can be called religious knowledge, or understanding, and of which one can possess much without possessing much religion, which is, after all, a way of life. Some who, up to a certain degree, begin to become religious start with such an understanding: religious concepts and expressions begin to mean something to them.—But some others come to religion from another side. For example, they become more and more helpful, selfless, insightful, etc., and, eventually, the religious words also begin to mean something to them.

(Wittgenstein 2011, Von Ludwig Wittgenstein an Arvid Sjögren, 9.10.1947; my translation)

Wittgenstein’s remark to Sjögren makes two notable points: First, Wittgenstein distinguishes here between religious knowledge, or religious understanding (“religiöses Wissen, oder Verständnis”) on the one hand and religion and a religious way of life on the other. Second, he says that one can come to religious knowledge and understanding, and thus also to an understanding of religious belief, in other ways than through religious experience or experience that is taught by a religious way of life.

Moreover, in a remark from 1944–1946, Wittgenstein considers the possibility that a religious person might be dissuaded from their belief by being pointed to a mistake (“Fehler”) in their reasoning “which they cannot religiously digest” (Ms-130,11; my translation).

Wittgenstein’s tightrope walker simile from 1948 (Wittgenstein 1998, p. 84e = Ms-137,67b) serves as crucial support for the view that the later Wittgenstein became open to the possibility that religious belief includes the cognitive aspect (see Pichler and Sunday Grève 2024). There, Wittgenstein compares the “honest religious thinker” to a tightrope walker and suggests, as Pichler and Sunday Grève (2024) have argued, that the honest religious believer must navigate well the cognitive tensions between their religious and other beliefs, be they factual or moral. While in 1937 he had described the “paradox” experienced by the honest religious thinker as something negative, he accepted it with the tightrope walker simile as the natural condition of the honest religious thinker. The honest religious thinker must balance between cognitive tensions, and the evidence for their religious belief is as thin as the tightrope of the acrobat. Wittgenstein had already noted in Lectures on Religious Belief that the religious believer bases “enormous things” on evidence that is “exceedingly flimsy” (Wittgenstein 1966, pp. 57–58).

At latest in 1949 Wittgenstein is said to have changed his view of Newman’s practice of arguing for the validity of the content of his religious beliefs whereas “[t]wenty years ago he would have regarded Newman’s action as incomprehensible, as insincere perhaps” (see Wittgenstein 1986, pp. 33–37; also see Wittgenstein 1984b, p. 59).

In a remark made in 1950, we find another example of Wittgenstein’s doubts about the validity of the sharp distinction between religious belief on the one hand and ordinary and scientific belief on the other, and between disagreement in matters of religious belief on the one hand and ordinary and scientific disagreement on the other—and thus about the validity of even moderate non-cognitivism. Having said in Lectures on Religious Belief that the Christian believer and the non-believer cannot contradict each other (Wittgenstein 1966, pp. 53, 55), Wittgenstein now admits that Catholics—while they, according to Wittgenstein, believe that “all evidence proves the contrary”—would contradict if someone were to say that the Eucharistic wine is not the blood of Christ (Wittgenstein 1969, §239 = Ms-175,15r-16r; my emphasis). This opens up the possibility of treating the disagreement expressed by Catholics as a disagreement in the ordinary sense. Note that Catholics would not completely agree with Wittgenstein’s remark either; they would deny that all the evidence proves the contrary of the wine being the blood of Christ. Wittgenstein’s own Catholic friend G.E.M. Anscombe says in “On Transubstantiation”: “[…] we do not believe that contradictions and absurdities can be true” (Anscombe 2008, p. 87).

Likewise, Wittgenstein now, in 1950, acknowledges that religious believers “who offered such proofs [of God’s existence] wanted to analyse & make a case for their ‘belief’ with their intellect [my emphasis], although they themselves would never have arrived at belief by way of such proofs” (Wittgenstein 1998, p. 97e = Ms-174,1v).

Therefore, it seems that Wittgenstein must eventually have concluded that not even moderate non-cognitivism is tenable as a general position and that partial cognitivism should be adopted instead. However, even late in his life, Wittgenstein continued to make remarks that express non-cognitivism about Christian belief statements. For example, in a manuscript possibly written in 1948–49 he asks: “Isn’t belief in God an attitude?” (Wittgenstein 1992, p. 38e = Ms-169,61r) In September 1950, Bouwsma recorded Wittgenstein saying that he “could make nothing” of most of Christianity’s dogmas and had “no use” for cognitive religious language: “Whatever this use is, it is different from the use of ordinary sentences describing the world”. He adds: “I cannot pray”—prayer apparently being one of the few legitimate uses of religious language in his eyes (Wittgenstein 1986, p. 57; note, however, that Wittgenstein was not incapable of praying—see, for example, Wittgenstein 2023, pp. 85–86 = Ms-183,183–184). Additionally, as late as 1951, he suggests that statements about God might be “incapable of falsehood” (Wittgenstein 1969, §436 = Ms-176,25r; also see Wittgenstein 1986, p. 34).

While partial cognitivism allows us to recognize that Wittgenstein remained partly a non-cognitivist (Pichler and Sunday Grève 2024), there still seems to be a tension in later Wittgenstein between cognitivism and non-cognitivism. This question can be approached in at least three ways: First, one might argue that we must distinguish between “Wittgenstein the philosopher” and “Wittgenstein the man” and that while the former, recognizing legitimate cognitive elements in Christian belief statements, accepted partial cognitivism about them, “Wittgenstein the man” remained largely non-cognitivist, preferring and living “the kind of religious belief that he personally found appealing, comprehensible, intellectually respectable and morally attractive” (Schroeder 2007, p. 444).6 Second, while at the same time acknowledging that Wittgenstein continued to oscillate between cognitivism and non-cognitivism, one might argue that “Wittgenstein’s lasting ambivalence should, where necessary, be reconciled” with the cognitivist position of Wittgenstein’s mature philosophical work, the Philosophical Investigations (Pichler and Sunday Grève 2024). Third, one could argue that Wittgenstein may have believed that while Christian belief statements include genuinely cognitive elements, these elements are epistemologically and ontologically void and illegitimate and should therefore be replaced with non-cognitive grammar (see Section 4).

The article will now proceed as follows: After Section 2.3 highlights a key flaw in general non-cognitivist approaches to Christian belief statements, Section 3 will present three objections that non-cognitivism might want to make to continue defending its position. While these objections are partly rooted in ideas that can be traced back to Wittgenstein himself, Wittgenstein also provides counterarguments against them. Along with a systematic critique of these three objections, the section will continue to offer evidence for the view that Wittgenstein, in his later years, became open to a form of partial cognitivism. More specifically, it will show that it is especially the ideas, methods and approaches of the Philosophical Investigations that make this analysis possible. Indeed, it is a central point of this article that, for a Wittgensteinian philosophy of religious belief, it is advisable to refer to and use Wittgenstein’s Philosophical Investigations in addition to the scattered remarks on religious belief collected in Culture and Value and the student notes edited in Lectures on Religious Belief. This approach does not diminish the value of the remarks in Culture and Value and the Lectures on Religious Belief, but rather balances them against the much more mature, much more tempered, much more edited, and also much more authoritative Philosophical Investigations (see Pichler 2023, ch.1.1; also see Carroll 2014, ch.1).

It is worth noting that there are several other non-cognitivist arguments that could and should be considered, or again other approaches to interpreting Wittgenstein’s treatment of religious belief, but which cannot be discussed here in more detail due to space constraints.7 Similarly, this article does not comprehensively address all the arguments and aspects of potential counterarguments to the selected three non-cognitivist objections. For example, one central avenue that is hardly explored here is the notion and method of language games as it is presented and developed in Philosophical Investigations (see Pichler 2024, pp. 276–77; Pichler and Sunday Grève 2024).

2.3. Augustine’s Mistake

Clearly, as defenders of cognitivism will concede, many Christian belief statements can be seen as non-cognitive or used in non-cognitive ways. This may be especially true for statements used to express existential feelings and life attitudes or appeal to ethical commitments and other dimensions that can be subsumed under aspects of “form of life”, rather than purporting to express beliefs that can be true or false.8 To prove general non-cognitivism wrong, the cognitivist only needs to show the fallacy of the view that all Christian belief statements are genuinely non-cognitive, either in the sense of strong or in the sense of moderate non-cognitivism (also see Citron 2012). Cognitivism can argue that general non-cognitivism commits the same mistake that Wittgenstein in Philosophical Investigations attributes to Augustine’s description of language and language learning: “Augustine, we might say, does describe a system of communication; only not everything that we call language is this system” (Wittgenstein 2009a, §3).

It is evident that not only philosophers, theologians, and other scholars but also religious believers themselves treat a significant number of religious beliefs in cognitive terms similar to other beliefs and subject them to the truth-game as other beliefs. For Christian belief language and practice, it does indeed (still) matter whether it is true that there was a historical person named Jesus Christ and whether it is true that Christ rose from the dead and thus that Christian belief participates in the truth-game. Belief statements that refer to and enforce cognitive, factual, historical, and empirical factors can be found in in the New Testament itself (see Pichler 2024, pp. 256–58; Pichler and Sunday Grève 2024). Therefore, while some Christian beliefs, or some aspects of Christian belief, are undeniably non-cognitive, this does not seem to be the case for all Christian beliefs. The same applies to Muslim beliefs and various other religious beliefs. It would seem, then, that while non-cognitivism is correct for some religious belief statements, it is not correct for all of them and that general non-cognitivism therefore amounts to an illegitimate generalization. The cognitivist position, on the other hand, does not need to defend general cognitivism; it only needs to show that some religious beliefs or aspects of them are unquestionably cognitive or treated and practiced as such.

The non-cognitivist position may concede that some Christian belief statements seem, at least superficially, to have an undeniably cognitive character and role that cannot be straightforwardly explained or translated into non-cognitive terms. However, it might want to question whether these beliefs also are genuinely cognitive. Here are three such avenues that non-cognitivism might pursue:

• Christian belief statements, or aspects of them, that cannot be described or translated into non-cognitive terms are superstition rather than genuine religious belief.
• Christian belief statements, or aspects of them, that cannot be described or translated into non-cognitive terms are parasitic on attitudes, thus preserving their cognitive content but exempting them from the cognitive truth-game.
• Christian belief statements, or aspects of them, that cannot be described or translated into non-cognitive terms are cognitive only so at the level of surface grammar, not depth grammar.

In sum, according to these three possible non-cognitivist objections, the cognitive element identified by, and also conceded to, cognitivism could either be superstition rather than genuinely religious; or it could be cognitive only in the sense of being parasitic on attitude rather than independent of attitude; or it could be cognitive only at the surface grammar level rather than also at the depth grammar level. In short, what seems to be a cognitive Christian belief statement, or seems to be a cognitive aspect of the Christian belief statement, might result from a misinterpretation and misclassification of this statement. Let us consider each of these objections separately. To this purpose, a hypothetical non-cognitivist interlocutor is introduced below. While this character does not represent any particular author, it is believed to effectively represent actual or potential Wittgensteinian non-cognitivist objections to the arguments for cognitivism.

3. Three Possible Objections to Cognitivism

3.1. “The Cognitive Aspect of Christian Belief Statements Is Superstition Only”

We have seen that non-cognitivism, in response to the charge of unjustified generalization, may concede that some Christian belief statements seem to follow cognitive grammar. But non-cognitivism may argue that their cognitive element, rather than being genuinely religious, aligns with the grammar of superstition. Non-cognitivism may even claim that historically and empirically or, more generally, cognitively understood Christian belief statements are superstition or that the cognitive interpretation turns them into superstition (see Koritensky 2002, p. 205; Schönbaumsfeld 2024, p. 81). Wittgenstein seems to have made this charge in the 1931 parts of his Remarks on Frazer’s Golden Bough (Wittgenstein 1993b) against J.G. Frazer. According to Wittgenstein, Frazer treated, in his The Golden Bough, primitive religion and magic as cognitive precursors of science, viewing, for example, beliefs in the power of rain magic as erroneous cognitive beliefs (Wittgenstein 1993b, p. 119 = Ms-110,178).9 Against Frazer, Wittgenstein argues from a non-cognitivist position about such beliefs that “[n]o opinion serves as the foundation for a religious symbol. And only an opinion can involve an error” (Wittgenstein 1993b, p. 123 = Ms-110,181). He makes the same charge against O’Hara who wrote that Christian truths “are reached by the very same intelligence that is operative in science and with the same certainty”. In his “Science and Religion”, O’Hara had not only listed alleged Christian truths what might still have been acceptable to Wittgenstein, but had also placed them in a cognitive and scientifically argumentative framework:

Is any religion true? Is it not based on imagination or sentiment or feeling? Christianity, most emphatically, at least, asserts the contrary. It holds that its truths are reached by the very same intelligence that is operative in science and with the same certainty. These truths are that God exists, that He created the world, that He created man with an immortal soul and a free will, and finally, that God came into this world as man. Is this last fact true? Christianity appeals to the historical records contained in the New Testament, and asserts that these records are trustworthy, that the events there narrated did happen, even when judged by the severest scientific criticism.

(O’Hara 1931, p. 112)

Thus, in Wittgenstein’s view, O’Hara had done precisely what Wittgenstein found most offensive—namely, he attempted to bring religious belief and practice into a theoretical and scientific system: “The symbolisms of Catholicism are wonderful beyond words. But any attempt to make it into a philosophical system is offensive” (Wittgenstein 1984a, p. 102). Sometime between 1943 and 1945, Wittgenstein stated: “If immortality is something that can be proved, then I don’t want it” (Wittgenstein 2015b, p. 49). While both Wittgenstein and O’Hara (see O’Hara 1931, p. 108) were concerned with preventing religion from degenerating into superstition, Wittgenstein must have thought that it was precisely O’Hara’s program of connecting science and religion—which is not unusual for Christians, at least not in the Catholic tradition to which O’Hara belonged—and in particular his claim that Christian truths “are reached by the very same intelligence that is operative in science and with the same certainty” that turned religious belief into superstition by making it “appear to be reasonable” and “a question of science” (Wittgenstein 1966, pp. 57–59).10

From a non-cognitivist position, one might want to describe anyone who believes that Jesus was dead and later seen alive by dozens of people as being superstitious. Similarly, from that perspective, thinking that Christ’s resurrection could be subject to empirical falsification, and thus potentially a false belief, turns the belief into superstition or bad science. Indeed, according to non-cognitivism, to make the resurrection of Christ the subject of reasoning about it and to test the historical and empirical claims of the Christian faith on precisely those terms must be bad science or superstition. However, this position is clearly not consistent with Christian belief itself, of which opinion and the possibility of historical and empirical error are a part. M. O’C. Drury tried to make this clear when he wrote to R. Rhees in 1971:

For a Catholic surely it is an essential part of the mass to believe that the rite was instituted by Christ on Maundy Thursday. And here it seems to me you have ‘opinion’ and ‘hypothesis’, and the possibility of error.

(Drury 2017, p. 247)

Against the charge that treating religious beliefs in cognitive and truth-apt terms turns them into superstition, it can firstly be argued that non-cognitivism’s definition of the difference between religious belief and superstition runs the risk of becoming stipulative. Secondly, it can be argued that the charge does not adequately account for the differences between the cognitive religious belief statement grammar on the one hand and the superstition grammar on the other hand (Pichler 2024, pp. 264–76). One difference is that “superstition” is typically used in an evaluative sense, whereas “religious belief” is not. Ascribing to someone or to oneself superstition can therefore have a revisionist character: “[…] the concept itself has a primarily critical force, and criticism implies the need for change: superstition is, after all, something to be avoided” (Mulhall 2001, p. 110). Another important point is that the religious believers themselves distinguish between cognitive religious beliefs and cognitive superstition beliefs, whether their own or those of others’, something that the view that the grammar of cognitive religious belief merges with the grammar of superstition does not seem to adequately address (also see Harvey 2005, p. 189).

Finally, cognitivism can argue that it is important to distinguish between two senses and therefore also potentially two different grammars of “superstition”. We clearly have the cognitive superstition grammar, such as in believing that a black cat crossing my path will bring bad luck (thus, superstition in the sense of false belief), but also the more practical superstition grammar, such as in performing the touch-wood ritual. This does not preclude the two grammars from occurring in combination. While there are significant and profound differences between cognitive Christian belief statement grammar on the one hand and “practical” superstition grammar on the other, there appear to be smaller differences between Christian belief statement grammar and “cognitive” superstition grammar. Indeed, it seems that cognitive Christian belief and superstition grammar partly coincide if “superstition” is the superstition of “cognitive false belief”. Cognitively understood Christian belief is then, to the extent that it is seen to merge with superstition, superstition in the sense of erroneous belief and thus cognitive belief. It was possibly in this sense that Wittgenstein attributed superstition to O’Hara. In the same vein, he might also have wanted to call his Catholic friends G.E.M. Anscombe and Y. Smythies, of whom he said, “I could never induce myself to believe all the things that they believe” (Wittgenstein 1984b, p. 60), superstitious. He also said of Newman, “How a man of such learning and culture could believe such things!” (Wittgenstein 1986, p. 34). The point that attributing superstition to someone is typically done from an evaluative perspective becomes particularly clear in the case of superstition as a false belief. This attribution is made from an external perspective and thus marks an important difference from Christian belief, which one can positively attribute to oneself, for example, at Sunday Mass when professing the creed.

Using “superstition” in the sense of a false belief has been extremely common throughout human history and has been used to label the false religious beliefs of others or also one’s own earlier false religious beliefs (Augustinus 1888, VI 2. Kap.; Newman 1914, p. 329). Notably, Wittgenstein himself applies this label not only to what he regards as false religious beliefs but to any major false philosophical belief—for example, the belief in the causal nexus (Wittgenstein 1922, 5.1361), the belief in the power of explanation (Wittgenstein 1993b, p. 129 = Ts-211,319) and the belief that “Language (or thinking) is something unique” (Wittgenstein 2009a, §110). It seems, then, that the objection from superstition may not ultimately succeed. On the one hand, there are significant differences between cognitive religious belief grammar and superstition grammar. On the other hand, to the extent that cognitive religious belief aligns with superstitious belief, it does so primarily in the sense of cognitive false belief. This clearly keeps Christian belief potentially within the grammar of cognitive belief. (Pichler 2024, p. 281)

Moreover, cognitivism can argue that non-cognitivism does not seem to adequately acknowledge that believing is a “family resemblance” concept that exhibits a wide variety of features and at the same time also possesses a unity (see Wittgenstein 2009a, §66–67). With a family resemblance concept (and phenomenon), none of the concept’s various features need to be present in every instance nor absent in any instance. Thus, while the unity of the concept of believing is not defined by an essence, it is not to be ruled out that the cognitive dimension, while not common to all forms of belief, can be common to cases of religious as well as ordinary and scientific belief (Pichler 2024, p. 278).

Cognitivism, on the other hand, does not need to argue that the cognitive element or any other kind of element must run through all forms of religious belief, nor does it need to argue that religious belief cannot share cognitive (and other) elements with superstition as it does with other forms of belief. Nor does it claim, contrary to what is often suggested in strawman arguments, that there are no significant differences between religious belief on the one hand and scientific and historical belief on the other; that evidence, historicity, argument, and scientific proof are the most important part or sufficient for religious belief; that a curious or extremely extraordinary event is all that there is to a miracle; or that religious belief cannot accept mysteries. Newman said: “Now I have no intention whatever of denying the beauty and the cogency of the argument which these books [of Natural Theology] contain; but I question much, whether in matter of fact they make or keep men Christians” (Newman 1908, p. 74). What cognitivism does claim is that argument, evidence, and truth can be genuine and crucial parts of religious belief, just as they are of other belief, and that they therefore must not be neglected at the outset of a philosophy of religion. These are precisely some of the lessons of Wittgenstein’s elaborations on family resemblance in Philosophical Investigations.

The objection from superstition does not seem to make room for the fact that religious belief can genuinely take on cognitive dimensions without losing them to superstition, false science, or scientism. According to non-cognitivism, faith and belief must be kept strictly separate. But these seem to be dogmatic and revisionist stipulations rather than truths. Indeed, the compartmentalizing approach of Wittgensteinian non-cognitivism has been identified as a central flaw of non-cognitivism:

[…] if language-games and forms of life are insular, inscrutable, self-regulating, and self-providing for norms of meaning and if, consequently, Wittgensteinian religion is thuswise disconnected from other ingredients of human life, then Wittgenstein must be wrong in his depiction of the religious arena.

(Biletzki 2010, p. 150)11

By splitting “believing” into two distinctly different conceptual essences—the essence of ordinary and scientific believing on the one hand and the essence of “faith” on the other—the compartmentalization approach can be seen to run counter to the family resemblance nature of believing. G.E.M. Anscombe remarked:

Thus belief in God would now generally be called ‘faith’—belief in God at all, not belief that God will help one, for example. This is a great pity. It has had a disgusting effect on thought about religion. The astounding idea that there should be such a thing as believing God has been lost sight of.

(Anscombe 2008, p. 2)

It should come as no surprise that the Catholic Church has designated Thomas Aquinas as a “doctor ecclesiae”, who, as one anonymous reviewer of this paper pointed out, emphasized that the meanings of terms appearing in both religious and non-religious contexts are not univocal (i.e., the same throughout) or equivocal (i.e., completely different throughout), but “irreducibly analogous”, and thus opposed to their strict compartmentalization (see, for example, Summa Theologica (Aquinas 1981): I, Q.13 and Q.16).

In summary, general non-cognitivism does not seem to fit in with the Philosophical Investigations’ recommendation of a non-revisionist philosophy that emphasizes the description of actual grammar. Moreover, it also seems to contradict the specific lessons the Philosophical Investigations teach about the behavior of family resemblance concepts. It seems then that the Wittgenstein of the Philosophical Investigations would himself oppose the non-cognitivist objection from superstition.

3.2. “The Cognitive Aspect of Christian Belief Statements Is Parasitic on Religious Attitude Only”

Non-cognitivism can raise a second objection: The cognitive aspect of religious belief is not independent but only an “epiphenomenon” of its truly non-cognitive nature. In other words, it is entirely based on and “parasitic” upon non-cognitive attitudes, experiences, and “religious or proto-religious emotions” such as “conscience”, “moral despair”, and “a sense of the futility of life” (see Schroeder 2007, p. 452), or “wonder” (see Cahill 2021, p. 168). In terms of cognitive belief per se, it would be cognitively completely unsupported. Thus, the cognitive would be inextricably bound to the precognitive attitude only and not also to facts, evidence and arguments. Without a pre-existing religious attitude, the cognitive would not only be indefensible, but it would also not have come into being at all. Wittgenstein writes in December 1937 that talk of Christ’s return and the Last Judgment can only mean something to him “if I were to live quite differently” (Wittgenstein 1998, p. 38e = Ms-120,54v). Approximately one month earlier, he had distinguished between different levels “of devoutness” and stated that the content of a religious belief is not available to the nonbeliever who is “at the lower level”; understanding its meaning depends on a religious attitude. Not even as a picture (“Bild”) can it play a positive role for the unbeliever; it will simply be “non-sense” to them (Wittgenstein 1998, p. 37e = Ms-120,4r-4v). This remark aligns with Wittgenstein’s earlier non-cognitivist labeling of religious talk as lacking sense. During the Tractatus period and still in the Lecture on Ethics, Wittgenstein made a strict distinction between meaningful, fact-oriented and truth-capable cognitive propositions on the one hand and meaningless and nonsensical sentences on the other. According to Tractatus, religious (as well as ethical, aesthetic, and other) discourse falls outside the realm of sense and the cognitive (see, for example, Wittgenstein 1922, 6.41–6.421; Wittgenstein 1993a, p. 44 = Ts-207,10). Furthermore, a religious belief sentence would never be fact-stating but rather express a way of seeing the world and its facts, an attitude (see Child 2011, p. 220). Even in 1950, as we have seen above, Wittgenstein would state that a proof of God’s existence can only mean something to someone who already believes and has the right religious attitude, and that religious belief cannot grow from cognition but only from “a certain upbringing”, from “experiences” (especially sufferings), and “life” itself (see Wittgenstein 1998, p. 97e = Ms-174,1v). First, there is the religious attitude, and only from that religious attitude can religious belief grow, and evidence come in. What can count as evidence for the rightness of the religious belief is then ultimately based on the prior existence of the religious attitude rather than on attitude-independent evidence. This view completely reverses, for religious belief, the standard justification chain used for justifying a belief from “evidence—[justifies]→belief—[justifies]→attitude” to “attitude—[justifies]→belief—[justifies]→evidence” (Schroeder 2007, p. 447).

There is reason to think that this view is incorrect. The cognitivist will ask why it should be ruled out that the believer’s cognitive religious beliefs can be as independent of attitude as their other cognitive beliefs. Church history, biblical exegesis, and theological and philosophical disputes have shown that the cognitive content of at least some religious beliefs, like other beliefs, is open to critical and rational examination of its evidence, independent of the attitudes of the believer and the examiner. Wittgenstein himself considers the possibility that the religious can become irreligious by being made aware of a mistake in their thinking which they consequently cannot “digest” religiously (Ms-130,11; 1944–1946—see citation above). Let us focus on the concept of “miracle” as an example of religious belief grammar that challenges the reversed justification chain view of religious belief. Indeed, miracles serve as a litmus test for cognitivism versus non-cognitivism.

Wittgenstein states in 1949 (Ms-172,20) that even those who do not believe in miracles can perceive them as “curious events” (“seltsame Begebenheiten”; translation by Pichler and Sunday Grève 2024). Twenty years earlier, in the Lecture on Ethics from 1929, he had said: “[…] we all know what in ordinary life would be called a miracle. It obviously is simply an event the like of which we have never yet seen” (Wittgenstein 1993a, p. 43 = Ts-207,9). However, to speak of miracles in this way is to speak of them in the relative sense, from which the use of the word “miracle” in the absolute sense must be distinguished. This distinction parallels the use of the word “good” in a relative and an absolute sense. Wittgenstein then goes on to say: “And I will now describe the experience of wondering at the existence of the world by saying: it is the experience of seeing the world as a miracle [in the absolute sense]”. However, just as when “good” is used in the absolute sense, so the use of “miracle” in the absolute sense—as we have it in the miracles of religious experience—will go beyond “significant language” and thus be nonsensical. Therefore, there is for Wittgenstein no way to connect the empirical, the factual, and the scientific discourse with the discourse of the miraculous: “The truth is that the scientific way of looking at a fact is not the way to look at it as a miracle” (Wittgenstein 1993a, p. 43). “[L]ooking at the existence of the world as a miracle means setting aside (suspending, as it were) any (scientific or metaphysical) claim to explanation; or, perhaps better, looking at the world and its existence without feeling the need or desire for an explanation” (Perissinotto 2024, p. 209). Wittgenstein ends the Lecture on Ethics by saying that while the sentences of ethics (and, we must assume, of religion and religious belief), because of their nonsensicality, do “not add to our knowledge in any sense”, they are “a document of a tendency in the human mind which I personally cannot help respecting deeply and I would not for my life ridicule it” (Wittgenstein 1993a, p. 44 = Ts-207,10).12

The difference between those who do not believe in miracles and those who do, is that, for the latter, miracles additionally represent “interruptions of the course of the world that are an intervention by a higher being” (“Durchbrechungen des Gangs der Welt, die ein Dreinsprechen eines höheren Wesens sind”; Ms-172,20). For the latter, miracles are divine interventions that involve something unexpected and transformative. Those who do not believe in miracles are “similar to someone who cannot perceive an ‘expression of emotion’ as such”, i.e., they are “simply someone who does not react naturally to this phenomenon” (Ms-172,20; my translations). However, to perceive an event as a curious or extremely extraordinary event should still be possible even for someone who does not believe in miracles. Even someone who does not believe in miracles, and therefore does not perceive the extraordinary phenomenon as a symbol or a “gesture” of God, should be able to perceive the sign of it and be “impressed” by it (see Wittgenstein 2023, p. 48 = Ms-183,82-84 and Wittgenstein 1998, p. 51 = Ms-128,47). Therefore, one should also be able to subject it as such to the truth-game and discuss it in purely factual and epistemic terms (cf. Schönbaumsfeld 2023, p. 27; Schönbaumsfeld 2024, pp. 79–80). As we will see below, this is precisely what Catholics do: for them, the fact of the miracle depends as much, if not more, on the truth-game as on the attitude.

It should therefore be possible for the non-believer to recognize events such as Christ’s feeding of the multitudes or the healing of the blind as extraordinary phenomena without seeing them as miracles. Recognizing the extraordinary event also as a miracle is a different matter; the New Testament reports that many simply stopped at the curious and extraordinary nature of the event and did not take the step of also experiencing the event as divine intervention. However, experiencing the extraordinary event as a miracle still does not require a specific pre-existing religious attitude. Rather, according to the Church’s treatment of alleged miracles, and according to what many religious believers believe and practice, miracles have led people to religious attitudes and have changed their lives. The Catholic Church holds that miracles provide external (and thus attitude-independent) evidence for the truth of its faith. This means that the concept of miracles, at least as it operates in the Catholic tradition, implies that miracles can precede conversion to religious belief and attitude.

Moreover, even the followers of Jesus who already had a religious attitude asked for evidence for what they later would recognize as miracles. The belief in the miracle of the resurrection is again an example. Luke’s Gospel tells us that the disciples (who already had a religious attitude) did not believe Mary Magdalene that Jesus had disappeared from the tomb. John’s Gospel tells us about doubting Thomas (Holy Bible 2015, John 20,25). In the Gospel of Matthew (Holy Bible 2015, Mt 28,17) and the (extension to the) Gospel of Mark (Holy Bible 2015, Mk 16,11 and 13), we hear of the disciples’ continued disbelief. They all asked for evidence: “But the story sounded like nonsense to the men, so they didn’t believe it. However, Peter jumped up and ran to the tomb to look. Stooping, he peered in and saw the empty linen wrappings; then he went home again, wondering what had happened” (Holy Bible 2015, Lk 24,11-12). Thus, miracles and the accompanying evidence need not be parasitic on a prior disposition and attitude; rather, they can affect a shift to a religious disposition and attitude. This shift does not occur with everyone who witnesses the event.

From this, we should conclude that the miracle does not hinge on a prepared and predisposed religious attitude, life experience, and expectation on the part of the person experiencing the miracle, as non-cognitivism would have it. Rather, it allows for the possibility that the miracle and its empirical evidence can directly impact the lives and minds of unprepared non-believers and lead them in a religious direction. Similarly, empirical evidence can cause people who are already religious to leave their religion behind (cf. Schönbaumsfeld 2024, p. 79). The Catholic procedure of canonization—including the collection of evidence and counterevidence to determine whether the person in question is eligible for sainthood—and the actual veneration of saints provide striking examples of the non-attitude dependent character of the miracle. This is significant because the narratives and concepts of the Bible and the authoritative practices of the Church are paradigmatic of the grammar and practice of Christian belief and should be treated as such by non-cognitivism as well. Moore contends that if non-cognitivism is correct, then Jesus himself was “making a philosophical mistake here, that he is confused over the grammar of ‘God’—that is, of the word ‘God’ as it is used in the Christian tradition”. However, that does not make sense because what Jesus said about God is itself the model of the Christian belief: “Within that tradition, what Jesus says is sayable, by definition” (Moore 2005, p. 221).

The non-cognitivist view of the reversed justification chain for religious belief can only be maintained if one holds, as Schönbaumsfeld does that religious “attitude and content are one—neither can be understood, or made sense of, in isolation from the other […] in order to understand religious faith, one cannot separate ‘form’ (religious way of life; religious attitudes) from ‘content’ (doctrinal claims), as the form itself contributes to and transforms the meaning of the content. In this sense, form and content are one” (Schönbaumsfeld 2023, pp. 10, 19). However, one might ask: Why should the truth or falsity of a religious belief behave so very differently from the truth or falsity of the belief that there is life on Mars? If the truth of the latter does not depend on my attitude, why should the truth of my religious belief depend on my attitude, either in terms of my specific attitude toward the belief or in terms of my way or form of life? It seems that if God does not exist, if there was no resurrection and if there will be no Last Judgment, then the belief in the existence of God, the resurrection of Christ, and the Last Judgment are plainly wrong.

With Child, the question can again be linked to the question of the legitimacy or illegitimacy of the compartmentalization approach:

If religious beliefs are supposed to be true in the same sense as other beliefs, they cannot be judged by a distinctive set of standards that are internal to the religious world-picture to which they belong; they are answerable to the same standards of truth and rationality that we apply elsewhere. On the other hand, if religion is a self-contained practice, with no implications for our other beliefs—if, in particular, religious utterances and beliefs are not supposed to be literally true—then religion is not answerable to the standards of truth and rationality that are appropriate to non-religious beliefs.

(Child 2011, p. 229)

To defend cognitivism, it is not necessary to argue that true factual Christian beliefs are all literally true throughout but only that they are factually true in the same sense that other beliefs are factually true (or false) (Pichler 2024, p. 248). If the compartmentalization approach is correct, then there is no such possibility that (some) Christian beliefs can be factually true just in the same sense as other beliefs. However, as we have seen above, the compartmentalization approach has been met with widespread criticism, including from within the Wittgensteinian community.

Ultimately, the view that the truth value of the expressed religious belief statement is dependent on the attitude of the believer again leads to the position that the truth of religious belief coincides with truthfulness. This is a view that religious non-cognitivism has often endorsed. However, it is not a view that is consistent with what many Christian believers hold: they also care about truth, not just truthfulness and attitude. Moreover, this view seems to conflict with what Wittgenstein left us as his philosophical testament in his Philosophical Investigations. In Wittgenstein (2009a, §241), to which we will return below, Wittgenstein seems to follow, rather than this view, the opposite view strongly endorsed by G. Frege, who separated truth from “people’s taking-to-be-true” (Frege 1964, p. 15 (xvii)) and believed that “what is true is something objective and independent of the judging subject” (Frege 1964, p. 16 (xviii)). Additionally, Wittgenstein seems to depart from the “religious attitude and content are one” view when, in the 1947 letter to Sjögren, he states that the development of religious understanding does not depend on first having a religious attitude or way of life; rather, these can begin with understanding.

Constructing the grammar of miracles in terms of the view of a reversed justification chain then again invites the charge of revisionism. Non-cognitivism seems to offer a prescriptive reinterpretation rather than a descriptive analysis of the concept of the miracle. Trigg labels Wittgensteinian philosophy of religion as antirealism, though a “much more subtle form of antirealism […] that does not make blanket denials of what can exist, but which reinterprets what we mean when we use religious language” (Trigg 1997, p. 216). From the perspective of a Christian believer, the non-cognitivist analysis of miracles is indeed reversing things and putting the cart before the horse. For Newman, the claim that a miracle has occurred is as true or false as the claim that an earthquake has occurred (see Newman 1913, pp. 123–24). In evaluating miracles—or what is potentially a miracle—Newman, for whom “[r]eligious doctrine is knowledge” (Newman 1907b, p. 42) and probability “the guide of life” (Newman 1913, pp. 113, 121), looked at all the available empirical evidence and reflects on its relationship to a possible miracle character. He asserted that “Miracles to the Catholic are historical facts, and nothing short of this; and they are to be regarded and dealt with as other facts” (Newman 1913, p. 422). This astonished Wittgenstein, who was amazed at Newman’s beliefs:

On miracles, Newman cites the case of Christians, who taken by savages had their tongues cut out, and yet they could speak. He gives a natural explanation for this—if the tongue is only half cut off a man cannot speak, but if wholly cut off a man still can—but Newman then goes on to say that it may nevertheless have been a miracle.

(Wittgenstein 1986, p. 34; 22 August 1949)

Non-cognitivism then seems to stand in stark contrast to the grammar of “miracle” as it is understood and practiced by Christians who actively believe in the truth-, falsity-, and probability-aptness of reported extraordinary events that are candidates for miracle status. In contrast to non-cognitivism, however, it seems that, with the Philosophical Investigations as one’s guide, one should have no problem recognizing that the concept of a miracle not only involves possible “interruptions of the course of the world” (Ms-172,20) but also that such interruptions should be debatable and recognizable even to those who do not approach it with a pre-existing religious attitude, even if only as extraordinary events and not yet as miracles. This is possible thanks to the non-revisionist character of Wittgenstein’s Philosophical Investigations and its focus on the study of concepts, where and how they do their work.

3.3. “The Cognitive Aspect of Christian Belief Statements Is Cognitive in Surface Grammar Only”

In the previous sections, it was argued that a general non-cognitivist approach to Christian belief statements amounts to an illegitimate and revisionist reinterpretation of a sufficiently large number of them that it must be considered incorrect. In fact, rather than a descriptive analysis of those beliefs that do not fit the non-cognitivist scheme, non-cognitivism seems to amount to a revisionist critique of their content. Moore asks: “Is the person against whom Phillips is arguing making a philosophical mistake or is he rather expressing a religious viewpoint of which Phillips disapproves?” (Moore 2005, p. 220). Non-cognitivism can therefore be challenged not only from the perspective that it is incorrect but also from the specific perspective of Wittgenstein’s Philosophical Investigations, which endorses a non-revisionist approach to functional language.

However, as non-cognitivism might object, this point is only valid if the cognitive aspect of Christian belief statements is part of how language works in practice and thus is part of the real grammar of language, rather than merely of its surface structural grammar. Wittgenstein has shown that the real grammar, the “depth grammar” (Wittgenstein 2009a, §664) of our words and phrases can be very different from their surface grammar and thus that their real use and meaning can be very different from what we think they are. Wittgenstein speaks of “misleading parallels” (Wittgenstein 2009a, §571) on the surface of our language; it is not a superficial similarity between the language of religious belief and the language of other belief—where in both cases we can say “I believe that …”—that establishes the actual presence of cognitive elements in Christian belief statements. Rather, it is only when it can be shown that the cognitive elements of Christian belief statements are also active and functional at the level of their depth grammar that non-cognitivism is shown to be false. The charge of revisionism against the non-cognitivist position is then valid only if not only surface grammar but also depth grammar is cognitive. What counts “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; see Wittgenstein 2009a, §254). Moreover, “[y]ou can only say what someone believes if you know their reasons; ‘we must look to his grounds,’ we must find ‘what things are connected with what he says,’ and ‘we may find all sorts of different grounds’ (Wittgenstein 2003, 404)” (Brusotti 2024, p. 95; except for the quotations from “Wittgenstein’s Saturday Discussions: 1946–1947”, my translation). Therefore, our question becomes: Are the cognitive elements of religious belief language merely surface grammar elements? For the later Wittgenstein, the level where the depth grammar, the real meaning of an expression is located and at work, where the concept itself appears, is the realm of our practices. Thus, the real meaning is the role and function of the expression in the language game, i.e., the unity of language and the activities and actions with which language is interwoven (Wittgenstein 2009a, §7).

While religious language is certainly not exempt from discrepancies between surface and depth grammar, Wittgenstein himself did not provide the same depth grammar analysis as he did—or at least aimed to provide—, for example, for “to mean”, “to understand”, “the inner”, “simple”, “game”, “agreement”, “have in common”, “rule”, “rule-following”, and others. Even if he had devoted himself more substantially to the project of analyzing the depth grammar of religious belief language, his results might still be a matter of disagreement, as is his analysis of the depth grammar of mental concepts. Wittgensteinian philosophers of religion also do not seem to have been successful in their attempts to get the depth grammar of Christian belief statements right. At any rate, their analyses are far from winning the general approval of those who are the practitioners of Christian belief, and who, also from a Wittgensteinian point of view, should therefore have a say in the matter. Indeed, they have often met with strong criticism. Not infrequently, the way in which the relationship between surface grammar and depth grammar is presented by non-cognitivism is questioned as such, or the actual distribution between what belongs to depth grammar and thus also to functional practice on the one hand—and what belongs only to the surface on the other—is disputed. For example, Harvey strongly objects to the way Phillips (2005a) distinguishes “between experience [depth grammar] and gloss [surface grammar]” (Harvey 2005, p. 193; see entire pp. 191–93). Eventually, the discussion often goes round in circles, remaining repetitive and unresolved, with the cognitivist or the believer accusing the Wittgensteinian of either ignorance or revisionism, and the Wittgensteinian going on to tell them that what they regard as functional depth grammar is only surface grammar. The following two dialogues are instructive:

D: Why not? Wittgensteinians always claim to tell us what we really mean. Why not ask Christians what they do mean? If the majority says they mean such-and-such, that settles the matter. You can do this kind of philosophy by Gallup poll. C: No you can’t, because when reference is made to what people mean, the reference is to the role the words play in their lives, not to the account they would give if asked. Notoriously, in giving that account our own words can lead us astray.

(Phillips 2001, p. 150)

C: As Peter Winch has said, we can give a date for the resurrection, but what happened on that date, the resurrection itself, cannot be determined historically, any more than that the person crucified was the Son of God. I: But this is the trouble. Speaking in the way C just has, simply does not accord with what Christians believe. C: But what do Christians believe? You can’t find that out by asking them, since their answers will themselves be a form of philosophical or theological reflection. I: I’m not suggesting that we do ask people, but I am saying that we need to examine the doctrine.

(Phillips 2005b, p. 205)

However, just as it is the task of cognitivism to show that the cognitive elements of Christian belief statements are part of the believers’ depth grammar, i.e., “the role the words play in their lives”, it also must remain the task of the non-cognitivist position to show that they are not. Non-cognitivism, however, does not yet seem to have shown that the cognitive element is not present at the level of depth grammar, and mainly only claims that it is not or could not be. This is often seen as paternalistic: “The philosophers who follow Wittgenstein, such as Phillips, say that religious believers often do not understand the ‘grammar’ of their language. […] philosophers can, it seems, see what is actually going on in a language-game in a manner participants cannot” (Trigg 2014, p. 41). Wittgenstein himself might eventually have agreed with Trigg’s analysis: “A dispute about religious belief cannot exist for you since you don’t know what the dispute is about (aren’t acquainted with it)” (Wittgenstein 2023, p. 73 = Ms-183,159). He knew that he did not participate in Christian belief practice and should therefore—although the difference between religious and ordinary belief statements “must be recognized by both those who have a use for them and by those who do not” (Wittgenstein 1986, p. 57) according to his own philosophy—potentially refrain from making judgements about its depth grammar. (Pichler 2024, p. 284)

Newman was a thinker whom Wittgenstein appreciated but also struggled with precisely because of Newman’s cognitive affirmation of certain “great dogmas of the faith” and confidence “in the truth of a certain definite religious teaching” (Newman 1913, pp. 150, 151) that Wittgenstein could not share and therefore had to label at least as a false belief or as “superstition”. Newman even argued “that certain doctrines are necessary to be believed for salvation”. In his “Prophetical Office” Newman wrote the following about the “Christian solution of the problem of life” (cf. Wittgenstein 2023, p. 77 = Ms-183,160):

Besides other points in common we both [the Anglican and the Catholic] hold, that certain doctrines are necessary to be believed for salvation; we both believe in the doctrines of the Trinity, Incarnation, and Atonement; in original sin; in the necessity of regeneration; in the supernatural grace of the Sacraments; in the apostolical succession; in the obligation of faith and obedience, and in the eternity of future punishment.

(Newman 1913, p. 169)

Religious truths, he argued, are communicated and maintained not only in cognitive beliefs and statements of faith but also through practices—or, as Newman put it, “by liturgies, rites, ceremonies, and customs, by events, disputes, movements, and all those other phenomena which are comprised under the name of history” (Newman 1859, §2). Religious depth grammar, then, for Newman—as his own journey from the Anglican to the Roman Catholic Church testifies—includes “disputes, movements”, and thus practices of cognitive disagreement and argument. However, non-cognitivism might still want to challenge that these practices are part of the depth grammar of religious belief.

In response, cognitivism, using Newman and other paradigmatic Christian believers and communities as touchstones, could provide cases for the view that Christian practices are grounded in and interwoven with cognitive religious beliefs. It seems fair to say that the separation of the Protestant and Roman Catholic churches was motivated in part by differences in cognitive dogma, as were differences and schisms in the modern, medieval, and early Christian churches. The following examples seem to illustrate the ways in which differences in Christian practice are rooted in differences in cognitive dogma: Different conceptualizations of the Holy Trinity are interwoven with different practices of baptism and prayer. Similarly, the varying interpretations of the nature of Jesus Christ have resulted in distinct practices, such as the veneration of icons or the Eucharist. Additionally, the differing conceptions of salvation and associated concepts, including heaven, hell, and purgatory, are interwoven with divergent practices, such as confession and sacraments. Furthermore, there are differences in practices related to daily life, such as personal Bible study, missionary work, clergy organization, and penance and prayer. Additionally, differences in views on the role of Mary and the saints are interwoven with practices such as praying and motivate believers to relate their entire life to paradigms of Christian faith. Moreover, differences in the dogmatic conception of the sacraments have led to the formation of distinct practices related to baptism, communion, marriage, liturgy, and numerous other rituals. The dogmatic contents and differences can become the subject of specific creeds, which are confessed and subscribed to by Christians at the Holy Mass. A notable contemporary illustration is the divergence in societal structures arising from disagreements concerning concepts such as “marriage” and “family”.

The above examples illustrate the potential for discrepancies in Christian dogma to manifest themselves in variations in practices, rituals, and the organization of society. Consequently, it can be argued that the cognitive factors of Christian belief influence the development of religious practices and the structure of Christian depth grammar and that Christian praxis cannot be accounted for without seeing how it is significantly shaped by cognitive belief: “It is belief in the existence of God that underlies and justifies religious practices, in the eyes of believers and non-believers alike” (Moore 2005, p. 211). Cognitivist or even rationalist philosophers of religion, of course, have never had a problem acknowledging that there are “credal claims which underlie the practice” (Swinburne 2001, p. 5).

From a non-cognitivist perspective, one might want to object further that cognitive practice in depth grammar, even if it appears to be so, need not be related to the cognitive elements in surface grammar. This could be demonstrated by showing that the religious person is, for example, dishonest and only pretending that there is a connection between practice and the cognitive surface. It is also possible that the religious believer does not truly adhere to the cognitive beliefs they profess. This may be true for a significant portion of Christians. It is also possible to argue that the person professing their beliefs genuinely holds and believes them but is nevertheless deceiving themselves given the lack of evidence to support their belief. Their belief would not be sufficiently grounded in reality even though they believe it is and present the matter as such. Their belief may simply indeed be bad false belief and thus superstition. After characterizing O’Hara as superstitious, Wittgenstein elaborates as follows:

But I would ridicule it, not by saying it is based on insufficient evidence. I would say: here is a man who is cheating himself. You can say: this man is ridiculous because he believes, and bases it on weak reasons.

(Wittgenstein 1966, p. 59)

In what follows, let us take seriously the suggestion that there may indeed be no cognitive depth grammar corresponding to the cognitive surface grammar because the connection between the surface and the depth grammar is broken—either because the believer is dishonest or because they deceive themselves or for some other reason—and explore some of its consequences. Regarding dishonesty, as we have seen, Wittgenstein emphasized in his tightrope-walker simile the “honest religious thinker” (my emphasis). It seems indeed true that numerous Christians who regularly profess their religious cognitive beliefs at Mass do not actually hold those beliefs. There are also those who honestly profess their cognitive beliefs but lack the practices with which their professed beliefs should be interwoven with and thus lack precisely that part of the depth grammar. Most Christians fall far behind in living what they state as their cognitive Christian beliefs. However, there seem to be also those who strive to and do live their beliefs, and whose lives and thoughts are not perceived as alien and external even by many of those Christians who do not, or do not sufficiently, and who play an exemplary role also in the lives of the latter. The lives and beliefs of Christian martyrs are regarded as such examples and are therefore considered central to Christianity, and their practices are considered exemplary also by Christians who do not follow their example.

4. Can Depth Grammar Be Mistaken? A Cognitive Wittgensteinian Answer

What would it mean for martyrs who believed in continued personal existence after death to have been dishonest about this belief? Does it mean that they dishonestly held on to their belief but were still willing to accept martyrdom, knowing that they did not actually believe in continued existence after death? Does it mean that they were basing their martyrdom on beliefs that they did not believe in? Does it make sense to say that St. Paul, who suggests in 1 Corinthians that Christians are liars and cheaters if Christ is not risen from the dead (but his body has been stolen, for example), was himself dishonest and only faked his belief in Christ’s resurrection or, alternatively, cheated himself into his own martyr death in Rome because of insufficient evidence? What would it mean that the martyr cheats themselves into giving their life for their belief?

Wittgenstein states in the Lectures on Religious Belief that religious believers, while he “wouldn’t call them unreasonable”, base “enormous things” on proportionally “exceedingly flimsy” evidence (Wittgenstein 1966, p. 58). Christian cognitive believers might concede that there is some truth in this, but would they also agree to Wittgenstein’s view that “they are certainly not reasonable” and “don’t treat this [their belief and the evidence for it] as a matter of reasonability” (Wittgenstein 1966, p. 58)? What about martyrs who not only believed that their deaths were for a good cause but also believed in salvation and believed that their martyrdom would make them better persons not only in this life but also in eternal life after death? What could be more reasonable in their eyes than to die for their belief? Were they not genuinely “employing their religious concepts”; was their martyrdom not grounded in their belief (cf. Mulhall 2001, pp. 105 and 108; Brusotti 2024, pp. 95–96)? Moreover, honesty need not—also in Wittgenstein’s eyes—be proven by martyrdom. Wittgenstein thought that Newman—who happens to be recognized as a saint by the Catholic Church—was honest: “Kingsley accused him of insincerity. But Newman was sincere” (Wittgenstein 1986, p. 34). So were and are many others in their religious beliefs: “[…] ‘the Christian solution of the problem of life’ seems to require ‘salvation, resurrection, judgment, heaven, hell’ […] apparently many people honestly believed those things (Dostoevsky, for instance)” (Engelmann 2024, p. 75).

It seems that non-cognitivism must—but fails to—recognize the paradigmatic role that the accounts of the lives and beliefs of the apostles, the early Christian theologians, and especially the Christian model saints and martyrs play for Christian belief regardless of the historical accuracy and authenticity of these accounts. What Christians believe and practice is shared and demonstrated by saints and martyrs as their model exemplars; the beliefs and lives of the saints are examples for other Christians to trust and follow. If non-cognitivism concedes that the cognitive belief of saints and martyrs can extend into their depth grammar and form their lives and practices in such a way that it is “regulating for in all his life” (Wittgenstein 1966, p. 54) up to including death, and if it also concedes that a sufficiently large number of martyrs not only held normative beliefs (ethical ideas that they wanted to uphold through their deaths) but actually believed in a continuation and judgment of their existence after death that enabled them to consciously and in their right minds accept and embrace death, then there would seem to be only one objection left. Non-cognitivism might agree that the saint’s reported life and the martyr’s willing death are expressions of and interwoven with their deeply held cognitive beliefs in practice—so the cognitive element is part of their depth grammar. However, the saint and the martyr, while not faking, may in fact be unconsciously deceiving themselves at the same time in the sense that their Christian depth grammar may itself be mistaken.

The objection is that, even if it is now conceded that there is both cognitive surface and depth grammar which are mutually supportive and interwoven, there might simply be no reality that corresponds to them, and hence to the depth grammar of the saint, the martyr, and of Christianity as a whole. All that exists would be Christian beliefs and practices on this planet, sometimes validated by martyrdom, but without the Christian facts believed in actually obtaining—whether past (like Christ’s resurrection), present (like God’s continued creation), or future facts (like the Last Judgment). These facts and what Christian depth grammar refers to would simply not exist: God does not exist; God did not create the earth and man; God does not continue to create; Christ has not risen from the dead; there were and are no miracles; there is no personal existence after death; there will not be a last judgment … Hence, Christianity’s cognitive depth grammar, with its saints and martyrs as functional and trusted paradigms, would be fundamentally false and have no correlate in reality.13

Is this scenario conceivable? In other words, is it conceivable that not only the beliefs and practices of a single believer and martyr, but those of an entire community, are cognitively mistaken? Non-believers would not hesitate to assert that this is not only conceivable and possible but actually the case. While the question deserves serious consideration outside the debate between cognitivism and non-cognitivism, this article will only attempt to situate it within that debate and to relate it to the later Wittgenstein’s thought on the relation between agreement and shared practices on the one hand and correctness on the other. Our questions then are: Is this possibility conceivable for later Wittgenstein, and if so, does it, according to later Wittgenstein, undermine cognitivism about Christian belief statements? Later Wittgenstein’s answer seems to be “yes” to the first question and “no” to the second, precisely because of what he says in his Philosophical Investigations about the relation between the grammar of correctness and the grammar of agreement.

Can, according to Wittgenstein, the depth grammar of an entire community be mistaken? We readily condemn practices with which we disagree as morally wrong. We also condemn forms of life and depth grammars of an entire society if we find them objectionable (see, for example, Biletzki 2024). Wittgenstein seems to have no problem accepting this practice of moral condemnation; so at least shared moral depth grammars can be clearly wrong also according to Wittgenstein. But can, according to Wittgenstein, also shared cognitive depth grammars be wrong? Can the Christian martyr’s depth grammar and thus also the martyr’s example be cognitively wrong? Again, we have no problem admitting that people can and still do today sacrifice their lives for wrong cognitive beliefs and wrong depth grammar, including wrong cognitive religious beliefs. With specific regard to Christianity, many have diagnosed belief in the Christian God as false. But does it make sense, from the point of view of Wittgenstein’s later philosophy, to say that the entire Christian community’s cognitive depth grammar and practice, with the cognitive beliefs and corresponding lives and deaths of the saints and martyrs as its paradigms, can be epistemically and ontologically mistaken or, to use a Wittgensteinian phrase, “confused”?14

It may appear challenging to reconcile the idea that an entire community is practicing a cognitively mistaken depth grammar with Wittgenstein’s On Certainty, where the concepts of right, wrong, true, and false seem to be bound up with and contingent upon the framing of shared practices, and where these concepts thus seem to apply only within the agreement that a shared practice and depth grammar, and ultimately a shared world picture, provide (also see Child 2011, p. 228). “But then, if we say that a practice itself is confused, how is this to be shown? If ‘practice’ is the last court of appeal, as it were, how can there be a confused practice?” (Phillips 2005a, p. 174) The Philosophical Investigations, however, seem to offer a different view. For the Philosophical Investigations, correctness does not depend on shared practices and world pictures. Rather, in Wittgenstein (2009a), §241, Wittgenstein distinguishes between the grammar of agreement on one side and the grammar of correct and false on the other (also see Wittgenstein 2009a, §136). Thus, it seems that with the Philosophical Investigations as one’s point of view, it does indeed make sense to say that people can agree in their cognitive depth grammars and practices and still be mistaken. Not only can depth grammar and practice be confused, but they can also be epistemically and ontologically mistaken. Gottlob Frege puts it this way: “There is no contradiction in something’s being true which everybody takes to be false” (Frege 1964, p. 13 (xvi)).15 Is this view consistent with the position of the Christian believer? Is the possibility of such a serious cognitive error within the scope of Christian belief and faith? It seems that it is. St. Paul himself argues: If there will be no resurrection of the dead, then Christ did not rise from the dead either, and consequently, the entire Christianity’s depth grammar can be considered epistemologically and ontologically false:

For if there is no resurrection of the dead, then Christ has not been raised either. And if Christ has not been raised, then all our preaching is useless, and your faith is useless. And we apostles would all be lying about God—for we have said that God raised Christ from the grave. But that can’t be true if there is no resurrection of the dead. And if there is no resurrection of the dead, then Christ has not been raised. And if Christ has not been raised, then your faith is useless and you are still guilty of your sins. In that case, all who have died believing in Christ are lost! And if our hope in Christ is only for this life, we are more to be pitied than anyone in the world.

(Holy Bible 2015, 1 Corinthians 15:13–19)

St. Paul’s belief statement certainly does not appear to be non-cognitivist. At the same time, it shows that, in Wittgenstein’s words, the honest Christian believer, both cognitively and existentially, indeed “risks things on account of it which he would not do on things which are by far better established for him” (Wittgenstein 1966, p. 54). But this is at the same time nothing out of the ordinary because “we [all] act on trust every hour of our lives” (Newman 1907a, p. 191). Despite the supposedly disproportionately small evidence for them, the Christian believer can live and act on their Christian beliefs as reasonable and firm convictions or, to use Newman’s terms, as “certitudes” and “real assents” (Newman 1903):

For when we consider the subject attentively, how few things there are which we can ascertain for ourselves by our own senses and reason! After all, what do we know without trusting others? We know that we are in a certain state of health, in a certain place, have been alive for a certain number of years, have certain principles and likings, have certain persons around us, and perhaps have in our lives travelled to certain places at a distance. But what do we know more? Are there not towns (we will say) within fifty or sixty miles of us which we have never seen, and which, nevertheless, we fully believe to be as we have heard them described? To extend our view;—we know that land stretches in every direction of us, a certain number of miles, and then there is sea on all sides; that we are in an island. But who has seen the land all around, and has proved for himself that the fact is so? What, then, convinces us of it? the report of others,—this trust, this faith in testimony which, when religion is concerned, then, and only then, the proud and sinful would fain call irrational.

(Newman 1907a, pp. 194–95)

5. Postscript

While this article criticizes some of Wittgenstein’s remarks about the nature of Christian belief and Christian belief grammar, it aims to align with the mature Wittgenstein of the Philosophical Investigations. One anonymous reviewer of this paper has expressed concern that Wittgenstein might not have approved of the cognitive/noncognitive dichotomy at play in this article, and that Wittgenstein would never have said, “I defend cognitivism with respect to …”. Regarding the dichotomy, it must be said that it is created by general non-cognitivism (including Wittgenstein where he himself endorsed this position) in the first place, either by explicitly denying the cognitive aspects of Christian belief grammar, or by first superficially acknowledging them and then reinterpreting them in ultimately non-cognitive terms. This article aims to do what the Philosophical Investigations recommend and gives examples of: bringing unjust philosophical dichotomies and positions, including Wittgenstein’s own, back to where their terms are at home and operative in communal language use (see Wittgenstein 2009a, §116) and thereby doing “justice to the facts” (Wittgenstein 1993b, p. 129 = Ts-211,319; see Pichler 2023, p. 14). A study of Christian belief statements shows that non-cognitivism’s exclusion of cognitive aspects from Christian belief grammar, and thus the strict separation of its cognitive and non-cognitive aspects, does not do justice to the facts of Christian belief grammar and is therefore false, and that the dichotomy constructed by non-cognitivism is therefore artificial.

The Wittgenstein of the Philosophical Investigations would surely agree with this article that the dichotomy does not exist in Christian belief language and practice itself. This does not mean, however, that the terms as such are useless in philosophical discussions or for the description of the grammar of Christian belief language. Just as the Philosophical Investigations continue to use conceptual frameworks like internal vs. external, mental vs. non-mental, correct vs. incorrect, grammatical vs. empirical, fallible vs. infallible, meaningful vs. nonsensical, and others, the terms “cognitive” and “non-cognitive” can do their work. While, regarding the fact that Wittgenstein would never have said, “I defend partial cognitivism with respect to …”, it is true that he would not, it is also true that he did defend, or became open to defending, a trivial view that in this article is called “partial cognitivism about Christian belief statements”. The problem lies in the assumption of a dichotomy between the cognitive and the non-cognitive aspects of Christian belief grammar. The cognitivist position itself that Christian belief language does indeed contain genuinely cognitive elements, is a trivial view and should be uncontroversial (see Wittgenstein 2009a, §128), and the dichotomy and the problem arose only because non-cognitivism fundamentally denied this triviality.