Trapped for nine hours in a lift with Karen Armstrong?
25 Dec 2011

Herman Philipse feels your pain.

1. Introduction

In this paper I attempt to substantiate the thesis that the core-beliefs of religions are irrational. These core-beliefs are the monotheist contention that there is one God or the polytheist opinion that there are a number of different gods. Outside mathematics, the word ‘irrational’ may signify two different things. Either it means that a sentient being is not endowed with reason, for instance if one speaks of ‘irrational animals’ such as slugs. Or it means that a belief or an action is contrary to reason, that is, unreasonable, utterly illogical, or absurd. I claim that all religious core-beliefs are irrational in this second sense. And of course, irrationality should be avoided.

It will be objected to my thesis that beliefs cannot be accused of being unreasonable unless they are situated within the province of reason. Could one not argue that religious beliefs are not located within this province because, as Pascal said, ‘the heart has reasons which reason does not grasp’? According to some religious authors, the domain of reason is somehow limited, and faith must be situated entirely, or in part, beyond the limits of human reason. I shall argue that even if faith transcends reason in this manner, the core-beliefs of religions are unreasonable.

2. Strategy

While believers are always partial or biased in religious matters because they prefer their own religion to the others – even a syncretistic religion is a specific faith that is logically incompatible with other religions –, the atheist must be epistemically impartial in that he rejects religious favouritism. His arguments should refute each and every religious existence claim, and they have to hold against all interpretations of each religion, orthodox and more liberal ones. Furthermore, atheism as such is restricted to the epistemic aspect of religions: the atheist rejects the idea that a god or that gods exist, hence he might appreciate from an aesthetic or moral point of view many different religious forms of life, taking an anthropological stance. This neutrality with regard to specific religions puts the atheist in a difficult argumentative position. The number of religions is large, and the set of actual and possible interpretations of each of them is perhaps an infinite one. How will the atheist be able to argue against all religious existence claims at once?

The most promising strategy is to proceed by way of dilemmas. If the atheist is able to construe a dilemma concerning all religions, which exhausts the entire field of religious possibilities because its two horns are each other’s contradictories, and if the atheist is {268}able to show that each horn gives rise to atheism, he will have won the battle. I shall briefly sketch such an argumentative strategy, which I have further developed elsewhere. It consists of a series of dilemmas, starting with the overarching dilemma that religious faith either (A) transcends reason or (B) finds itself within the province of reason. Let me call this the dilemma of faith and reason. I shall argue that the horns of the dilemma of faith and reason imply atheism.

Before I develop my argument, two clarifications will be useful. First, the dilemma might be applied both to large clusters of religious tenets and to individual claims to religious truth. For instance, in the Scholastic tradition some elements of Christian faith, such as the claim that there is an infinite god, were considered to be sustainable by reason, whereas others, such as the dogma of the monotheist unity of Father, Son, and Holy Spirit, were often thought to transcend reason and to be known by revelation only. It is up to believers to determine which religious tenet they want to put on which horn of the dilemma. Incidentally, since the dilemma applies to each and every religious tenet, its two horns are contradictories instead of contraries.

Second, the atheist is free to define the notion of reason as he pleases, provided that he uses his notion consistently. For the sake of argument, I define ‘reason’ as the methods of empirical research and critical discursive thought. It is no objection to my overarching dilemma that the term ‘reason’ might be defined differently, for instance as Hegelian Vernunft as opposed to Hegelian Verstand. The only legitimate objections are either that at least one of the horns does not lead to atheism (‘grasping the dilemma by the horns’), or that the two horns do not exhaust the field of possible religious positions (‘escaping between the horns’), or that the believer is able to launch a counter dilemma which destroys atheism (‘rebuttal’).

3. Faith Beyond Reason

Let us start, then, by discussing the first horn (A) of my overarching dilemma, the horn according to which faith transcends reason. In other words, the first horn says that religious belief is a-rational. This idea is old and venerable. Allegedly, the object of faith is too sublime to be grasped by a down-to-earth capacity such as human reason. But accepting this horn leads one inexorably to an atheist conclusion of a peculiar kind, as the following reflections will show. Hence, if religious belief is a-rational, it is irrational.

Everyone should agree that religious belief is impossible without a propositional content, for one cannot believe without believing that some proposition is true. Indeed, believing is always believing that p, and believing that p is accepting as true that p (these are observations on the logical grammar of ‘to believe’). The minimal content of a religious belief is the proposition that gods, or God, exist or exists, and according to my precising definition of ‘religion’ there simply is no religion without acceptance of such a proposition. Propositions are expressed by meaningful declarative sentences. Furthermore, declarative sentences cannot be meaningful if one of the words used in the sentence is meaningless. Hence there cannot be religious belief unless the word ‘god’ or ‘God’ has been assigned a meaning.

It is not up to the atheist to define the proper name ‘God’ or the common noun ‘god’, {269}except in the minimal sense of a supernatural entity. The atheist leaves this task to the religious believer. From the impartial point of view of the atheist, believers have an immense room for choice here, since innumerable descriptive definitions of gods have been provided in the history of mankind, and believers might conjure up indefinitely many new definitions. As we will see, however, the set of all possible definitions is exhaustively divided into two subsets by the dilemma that either faith transcends reason or faith is located within the province of reason.

If one opts for the horn that faith transcends reason, the descriptive definition of ‘god’ or ‘God’ which gives meaning to the religious proposition ‘God exists’ has to meet specific requirements, supposing at least that one wants to believe that God (or gods) exist(s). For if faith transcends reason in the sense defined above, that is, empirical inquiry and discursive thought, it must be a priori impossible that a religious core-belief be refuted by empirical data. In other words, the definition of the word ‘god’ has to meet the postulate of empirical irrefutability (apart from the logical requirement of consistency) in the following sense: the belief that God, or a god, as defined by the definiens, exists in fact, must be immune to all possible empirical refutations.

The postulate of empirical irrefutability implies another postulate, which I call the postulate of factual emptiness. Although there are events that according to current science cannot be investigated by direct empirical research, such as the events of which special relativity theory says that they are outside our ‘light cone’, one cannot exclude a priori that these events will ever be open to empirical investigation of a more indirect kind. Moreover, the idea that there are facts or events which are a priori outside the empirical domain presupposes that there may be true synthetic a priori propositions, a presupposition that is now generally rejected by philosophers. Consequently, as there can be no facts of which one might guarantee a priori that they are not open to empirical investigation, the postulate of empirical irrefutability implies the postulate of factual emptiness.

Accordingly, the believer who claims that faith transcends reason has to give a descriptive definition of ‘God’ or ‘god’ such that the proposition ‘God exists’ is devoid of factual content. This postulate of factual emptiness has been endorsed by religious philosophers of the twentieth century such as the early Wittgenstein and the later Heidegger. Defining the world as the totality of facts, Wittgenstein claimed that ‘God does not reveal himself in the world’ (Tractatus, 6.432), and Heidegger held that Seyn radically transcends the world of Seiendes.

But clearly, the postulate of factual emptiness destroys the possibility of defining ‘God’ or ‘god’ altogether, because a descriptive definition of ‘x’ such that ‘x exists’ is devoid of factual content is impossible: a fact is precisely what obtains if a descriptive statement is true. We now see the devastating implications of the first horn (A) of our overarching dilemma. If faith transcends reason, we cannot give meaning to the word ‘god’ or ‘God’. Consequently, the phrase ‘God exists’ is meaningless, and the claims that God exists (faith), that we do not know whether God exists (agnosticism), and that God does not exist (traditional atheism) are also meaningless. I call this implication semantical atheism, for if one cannot give meaning to the thesis that God exists, religious belief is impossible.

We must conclude that the first horn of the dilemma is self-refuting. If the very idea that faith transcends reason implies, by a chain of arguments, that faith is impossible if it {270}transcends reason, faith cannot transcend reason. I said that my overarching dilemma divides the set of all possible definitions of ‘god’ into two subsets. Clearly one of these subsets equals the null class, for no definition of ‘god’ can satisfy the requirement of factual emptiness.

As a consequence, the would-be believer who claims that faith transcends reason is landed in a second dilemma. Either he continues using the word ‘god’ without having given meaning to it, which, I suspect, is the case of many popular religious authors, or, if he has provided a descriptive definition of ‘god’ or ‘God’, this definition has descriptive content. In the latter case, his claim that God exists will have factual implications. Hence it is in principle refutable by empirical research or discursive argument and the believer has not succeeded in transcending reason. This brings me to the second horn of the overarching dilemma, the horn that faith is located within the province of reason.

4. Faith Within the Province of Reason

The believer who locates his faith within the province of reason is confronted by a great number of dilemmas that destroy his position, for now his faith is answerable to reason. It will suffice here to point out three interconnected dilemmas from which the believer cannot escape. If faith is located within the domain of reason, one should raise the question as to what explains the fact that the believer has faith. This question triggers a first dilemma: either (C) one gives a religious explanation or (D) one gives a secular explanation. Assuming that each of these explanations points to a cluster of causes of faith that is sufficient to explain the presence of faith, the religious and the secular explanation exclude each other because of Occam’s razor. Even believers who reject natural theology, holding that reason cannot bring us before God, might locate their faith within the province of reason if (a) they provide us with a religious explanation of their faith and (b) they are prepared to bring their religious explanation into competition with secular explanations and to adjudicate between these competing explanations by using accepted criteria of theory choice.

For instance, believers might explain the fact that they have faith by saying that God’s grace bestowed faith upon them, and that this grace is a sufficient condition for faith. Such is the traditional Christian explanation given by Paul, Luther, and many others. Clearly this explanation not only explains the presence of faith in believers but also justifies it: if God caused faith in us, God must exist, and the belief that He exists is true. For this reason we might call religious explanations of faith self-justifying. The question is, however, whether such a religious explanation is acceptable according to the usual criteria of theory choice. Should one not prefer a secular explanation? Secular explanations all belong to the class to which Freud’s theory of projection belongs. They typically start from the assumption that the belief that God or gods exist(s) is not true and try to explain the fact that the believer has the illusion that God or gods exist(s) by pointing to psychological, sociological, or other secular mechanisms. Clearly, then, the religious and the secular explanations are mutually incompatible, because they contradict each other.

A second dilemma shows that according to accepted criteria of theory choice, one should always prefer an explanation of the secular type to religious explanations of the {271}presence of faith in believers. This dilemma starts from the fact that there is a plurality of religions and that the religious contents of these religions contradict each other at many points. For example, monotheistic faith contradicts polytheistic beliefs. The dilemma arises for the believer who wants to provide a religious and self-justifying explanation of his own faith. Either (E) this believer provides a religious and self-justifying explanation for his own faith only, explaining the faith of other religions by a theory of projection. But this is an illegitimate move. It is a case of special pleading, unless the believer is able to argue convincingly that his own faith is true and the beliefs of competing religions are false, which is unlikely. Or (F) the believer chooses to explain the faith of all religions by means of a religious explanation. However, this second horn triggers a third dilemma: which religious explanation should he prefer?

Either (G) the believer tries to explain the faith of all religions by supposing that the self-justifying explanation of his own faith also explains the faith of other religions even though these other religions may contradict his own. This is the theory of the Catholic Church, which claims that Catholic faith is absolutely true and is caused by the Catholic God in Catholic believers. The Catholic God would also have caused (or at least ‘permitted’) Hindu beliefs in Hindu believers and polytheist Germanic beliefs in the Germans of the Edda epoch. Why would the veracious Catholic God do such a weird thing? Why would He cause religious beliefs in non-Catholics, beliefs that must be false to the extent that they are incompatible with the absolute truth of Catholicism? The Catholic solution to this embarrassing problem is that the Christian truth is hidden in all other religions, and that believers of these other religions are ‘on the way’ to the Catholic Truth even though they are not quite ready to receive Christian grace. This clearly is an ad hoc solution which shipwrecks the attempt to explain all religions by the self-justifying explanation of one’s own faith.

Should one then (H) try to explain each faith by supposing that the self-justifying explanation of each and every religion is true? That is, should one suppose that the Catholic faith is to be explained by claiming that the Catholic God caused this faith in Catholics, and that the Hindu faith is to be explained by supposing that each of the innumerably many Hindu gods caused faith in him- or herself in Hindu believers? This possibility is ruled out by the fact that these explanations contradict each other; according to Christian faith, there is only one God. We must conclude that the attempt to explain faith by religious explanations runs into insuperable difficulties. Only a secular explanation is able to account for all occurrences of religious faith. Irrespective of their individual scientific credentials, then, secular explanations must be preferred, because they are a priori more empirically adequate than their religious rivals. And because secular explanations of religion start from the assumption that each religious faith is false, those who want to advance in the endeavour of explaining the phenomena of faith must be professional atheists.

I conclude that the overarching dilemma destroys all possibilities for faith. If faith transcends reason, semantical atheism is the result. If faith is located within the province of reason, we must all become traditional atheists. This conclusion substantiates the thesis of the irrationality of religion.

{272}5. Discussion

How is the atheist to discuss religion with the faithful? It is wise to start by saying nothing at all. But if a believer advances the claim that God or a god exists, or engages in a language game such as prayer which presupposes that there is a god, the atheist might ask the believer by which descriptive definition the latter gives meaning to the word ‘god’. Typically, such a definition, conjoined to the existence claim, will have factual implications, and the atheist might point out the dilemmas which the believer has to confront (B-H). However, if the believer reacts to these dilemmas by eliminating the factual implications from his definition, the atheist will show that by this move the believer has destroyed the content of his existence claim and that, eo ipso, he has ceased to be a believer (A).

A meticulous analysis of religious language games in contemporary Western culture will reveal that believers keep oscillating between these two options and refuse to choose consistently, because they are dimly aware that every consistent choice will shipwreck their religion. This is what we should expect. For religion is the product of man’s longing for there being more to life than in fact there is. As soon as one tries to formulate this ‘more’ in meaningful propositions, one perceives that these propositions are very probably false.

I have called this argument for the irrationality of religious belief a ‘plea for atheism’, and so it is. One might object that it would be more prudent to become an agnostic. Is atheism not as irrational as religious belief? Indeed, is it not a kind of dogmatic belief itself? But it is easy to see that the overarching dilemma of faith and reason destroys agnosticism as it destroys faith. Agnosticism is the position that we should refrain both from believing that there is a god and from believing that there is no god, because the arguments in favour of one of these positions are not stronger than those in favour of the other. Accordingly, agnosticism does not make sense unless the phrase ‘God exists’ makes sense. But as we saw, this phrase is meaningless if (A) faith transcends reason. And if (B) faith is located within the domain of reason, the arguments for atheism are stronger than those in favour of a particular religious belief.

— Herman Philipse, ‘The Irrationality of Religion – A Plea for Atheism‘, in Rationalität und Irrationalität: Akten des 23. Internationalen Wittgenstein-Symposiums, 13. bis 19. August 2000, ed. by Berit Brogaard and Barry Smith (Vienna: öbv & hpt Verlagsgesellschaft, 2001), pp. 267-272.