An epistemic argument for tolerance

Abstract

In this paper I first take a critical look at Grube’s allegiance to the idea that bivalence should be rejected as it can serve the cause of religious toleration. I argue that bivalence is not what Grube says it is, and that rejection of bivalence comes at a very high price that we should not be willing to pay. Next I analyze Grube’s argument for religious toleration – an argument that does not involve the rejection of bivalence. I argue that the argument is unconvincing because there exists no relation between epistemic justification and toleration. (I also note problems with the notion of ‘justification’ as used by Grube.)

Keywords:

• toleration
• epistemic justification
• bivalence
• Dirk-Martin Grube

Many years ago the beat poet Steve Turner wrote the poem ‘Modern Thinker’s Creed’. One of its stanzas says:

We believe

all religions

are basically the same.

At least

The one that we read was.

They all believe in love and goodness.

They only differ in matters of

creation, sin, heaven, hell, God and salvation.¹

I think Dirk-Martin Grube, to whose paper this is a response,² should like this stanza, as it expresses, in an ironical fashion, the ‘pluralist’ view of the relationship between the world’s religions that he opposes. At the same time, it hints at some problems of the view that Grube also mentions. Let me explain.

The core of the pluralist view is that the world’s religions are, somehow, ‘one’, and that whatever differences exist between the religions, are inessential and hence unimportant. On this view, religious differences should not be focused on in the inter religious dialogue. The one-ness or unity of the world’s religions, on the pluralist view, is the basis for religious toleration. In Hick’s pluralist view, for example, the world’s religions are all geared toward ‘the Real an sich’, but as ‘the Real an sich’ is unfathomable, none of the world’s religions, or rather their adherents, can or should lay claim to absolute truth. They should moderate their claims. And this moderation is presented as the basis for religious toleration.

Grube has problems with the pluralist view, but not with the pluralist’s call for religious toleration.³ Rather, his problem is with the pluralist’s argument for that call. He therefore aims to provide an alternative argument for toleration. In this short reply to Grube’s paper, I will proceed as follows. I first recount what I take to be Grube’s key problem with the pluralist view: its commitment to the logical principle of bivalence. I shall argue that the rejection of bivalence is deeply problematic – especially in the context in which he wants to put that rejection to work. Next I will examine Grube’s own argument for religious toleration, and point to some problematic features.

Bivalence as presupposition of pluralism

Pluralism is problematic, Grube rightly says, because it ‘breeds the pressure to minimize difference’ (421). And he wants to reject it not just on the ground that it breeds this attitude, but because there is something inherently problematic in pluralism. And what he thinks is inherently problematic in pluralism is that pluralism presupposes bivalence. In the wake of his Doktorvater Joseph Margolis, he then goes on to reject bivalence. (Now I should think that giving up bivalence is a council of despair; it is the very last thing one should do. If one wants, like me,⁴ to reject pluralism one should first explore other ways. The reason for this is that our common practices of argument, deduction, and reasoning all rely on the validity of this principle. We do not really know what it would be like to think, to argue, or to do science without relying on bivalence.)

The logical principle of bivalence, Grube says, entails that declarative statements ‘have exactly one and not more than one truth value’ (421). Bivalence, he says, is the thought that declarative statements are either true or false.⁵ As a characterization of bivalence, however, this is not exactly right. For first there are declarative statements that have no truth value. Think of such statements as ‘When I stood up from my chair, my lap vanished’, or ‘ Her dream weighed 50 kilo’. These statements have no truth value because they make no sense. But this does not contradict bivalence. In order for a statement to have a truth value, it must at the very least make sense. There are other kinds of statements that have no truth value, for example the statement ‘John sat to the right of Nico’.⁶ This statement has no truth value because it is incomplete: it does not contain the crucial information from whose point of view John sat to the right of Nico. But this does not contradict bivalence. For a statement to have truth value, it must be complete. Furthermore, some statements are both true and false. Think of such statements as ‘Hilary Clinton is still running’. This statement is true – Hilary Clinton is still running for presidency. At the same time this statement is false – she is not (let us suppose) exercising, but sound asleep. This statement is both true and false, because it is ambiguous. But this does not contradict bivalence. For a statement to have one truth value, it must be unambiguous. A better phrasing of bivalence, then, is this: every declarative statement that is meaningful, complete and unambiguous is either true or false.⁷ I take it that Grube agrees with this.

Grube says that bivalence ‘implies that, if proposition A is true and proposition B differs from A, B must be false’ (421). But this is not the case: bivalence does not imply this. Take, for example, for A: ‘Bayern München won’ and for B: ‘A German team won’. These propositions clearly differ from each other. Yet it is not the case that if A is true, B must be false. As a matter of fact, in this case, if A is true, then B must be true too! Or to take another example, suppose A is ‘Goethe wrote Die Leiden des junges Werthers’ and B is ‘Werner Bergengruen wrote Der spanische Rosenstock’. Then A and B are clearly different. Yet it is not the case that if A is true, B must be false. As a matter of fact, both propositions are true. And no one who adopts bivalence thinks otherwise: no one thinks that bivalence implies that if propositions A and B are different, then if A is true, B must be false. What bivalence does imply is that if A entails not-B, then if A is true, B must be false. I take it that Grube agrees with this as well.

But then he will also have to agree that his statement that ‘bivalence implies an equation between difference and falsity’ (421) is misleading, and as stated simply false. What is correct is the following modified statement: ‘bivalence implies that if proposition A entails proposition not-B, then if A is true, B must be false’.

Since Grube thinks his original statement is true, I assume that he will think my improved, modified statement is certainly true: bivalence has the implication stated. But since he rejects the implication (viz. that, with my added qualification, difference = falsity), he boldly reject bivalence, at least in the domain of religious discourse.

But what does it mean to reject bivalence? It must mean, given Grube’s own formulation of bivalence, that it is not the case that if proposition A is true and proposition B differs from A, B must be false. And if we take the improved formulation of bivalence that I offered, it must mean this: it is not the case that if proposition A entails proposition not-B, then if A is true, B must be false. Applied to the domain of religious discourse this works out as follows: ‘God is omniscient’ and ‘God doesn’t know everything’ can both be true, even though the former proposition entails the denial of the latter. Or in another example: ‘There is no salvation but through Christ’s death and glorious resurrection’ and ‘humans work out their own salvation through the laws of karma and reincarnation’ can both be true, even though the former proposition entails the falsity of the latter. But this way lies madness, also in the domain of religious discourse. If we reject bivalence in religious discourse and elsewhere, we legitimize nonsensical talk.⁸

And then there is this: it is an empirical fact that by far most religious people take bivalence for granted. That is why a Christian says that the Muslim is wrong in rejecting the doctrine of the Trinity – and a Muslim returns the compliment when he says that the Christian is wrong in accepting that doctrine.⁹ The Muslim and the Christian hold incompatible beliefs about God. Someone who rejects bivalence must be willing to say that these incompatible beliefs can both be true. And it is an empirical fact that most Christians and most Muslims are not prepared to do that. The rejection of bivalence jeopardizes honest communication between adherents of different religions every bit as much as does the acceptance of pluralism. If, as Grube avers, pluralism sweeps serious religious differences under the carpet and thereby precludes honest communication, then the rejection of bivalence cannot be presented as a cure, for it equally sweeps serious religious differences under the carpet, and thereby also precludes honest communication.

An argument for religious toleration based on the rejection of bivalence, will have a paralyzing effect on every religious believer. For he is then urged to hold not only that his own religious beliefs are true but that someone else’s beliefs that are incompatible with them, are true as well. And this, in turn, means that he is urged to hold that his beliefs are both true and false. These are the effects that the serious rejection of bivalence will work. But those effects are madness. Hence, if there is an argument for religious toleration, it better not be based on the rejection of bivalence.

An epistemic argument for religious toleration?

After Grube has expressed allegiance to the rejection of bivalence, he offers an argument for toleration that does not rest on the rejection of bivalence! I now wish to examine his argument, the crucial notion of which is ‘justification’. Now ‘justification’ is an epistemological term of art. When you ask a person in the street whether a particular belief of his is ‘justified’ he will be puzzled, as he has no clear idea what is being asked. Perhaps the best way to introduce the term is by giving examples, and then hope for the best. Someone living in the seventh century could not have been justified in believing that the earth revolves around the sun, even though it was already true in the seventh century that the earth revolves around the sun. So here we have truth without justification. But it is also possible to have justification without truth. For example: John Milton, let us suppose, believed that the Morning star and the Evening star are two different celestial bodies; his belief was justified but false. What these examples show is that truth and justification can come apart: a belief’s truth does not entail that belief in it is justified, and the fact that a belief is justified does not entail that it is true.

Still, many epistemologists argue (or at least: hope), there is a relation between truth and justification.¹⁰ Some have stated this relation as follows: the more justification you have for one of your beliefs, the more likely it is that that belief is true.¹¹

This way of putting things indicates something else as well: justification is degree-sensitive. One person can have more justification for believing proposition p, than someone else. It brings out something else as well: justification, or the strength thereof, is person-relative. Suppose I heard via a usually unreliable source, Bild perhaps, that Bob broke his back; then, if I believe that Bob broke his back, my belief has much less justification than, say, you who shares my belief but were an eye witness to Bob’s dismal fall from his horse. I assume that when Grube says that ‘justification is plural’ he means the same as what I have just said, viz. that justification is person-relative and degree-sensitive.

I said that ‘justification’ is the crucial notion in Grube’s case for religious toleration. The case itself, or so it seems, is as follows: adherents of different world religions, Christians, Muslims, Jews, Hindus etc. surely believe different things – but their believing attitudes, given their respective epistemic circumstances, can all considered to be epistemically justified. And because they can all be so considered, adherents of the world’s religions should be tolerant towards each other.

Earlier I added the qualifier ‘or so it seems’, because Grube nowhere makes his case for religious toleration fully explicit. In the opening section of his article he does explicitly say that he is going to offer a ‘framework’ that allows for religious toleration. The ‘framework’, as it turns out, just is the idea that adherents of the world’s religions, given their different epistemic circumstances, are all epistemically justified in their religious beliefs. So I will take Grube upon his word: he does intend to offer an argument for religious toleration.¹² Hence, I take Grube’s argument for religious toleration to run as follows:

1. The adherents of the world’s religions are justified in their religious beliefs.

2. If the adherents of the world’s religions are justified in their beliefs, they should be tolerated.

3. Therefore, the adherents of the world’s religions should be tolerated.

(And this entails that the adherents of the world’s religions should tolerate each other.)

Is this argument convincing? It is at least puzzling for what it does not say. It does not say that religious toleration should be extended to people whose religious beliefs are not justified. For premise (2) formulates only a sufficient condition for toleration, not a necessary one. On the basis of this argument toleration can only be extended to persons whose religious beliefs are justified. If it is to be extended to persons whose beliefs are not justified, an additional argument is needed – an argument Grube has not offered. So, if his argument is successful, it at best offers a partial case for toleration. But is even the partial argument successful – is it plausible?

I am reluctant to say ‘yes’. One reason is that, on the basis of what Grube has said about justification, I have no clue as to why we should think that premise (1) is true. Are adherents of the world’s religions all epistemically justified in holding their religious beliefs? Well, what is justification? For all the work this notion is supposed to do for the argument, Grube leaves the reader in the dark as to when a person’s belief is justified. He offers no conditions, neither necessary nor sufficient ones, that must be satisfied in order for someone’s belief to be justified. Sure, he does say that the justificatory status of a person’s belief depends ‘on her perspective’, ‘on what information is available to her at a given moment’, ‘on her epistemological context’, ‘on her grounds’, ‘on whether she has carefully mustered the evidence’, ‘on the reasons she has for assuming her belief to be true’ (422–423). And sure, he contrasts the notion of ‘justification’ with Wolterstorff’s notion of ‘entitlement’ according to which a person is entitled to believe p if that person has fulfilled certain epistemic obligations. But these indications are radically insufficient to say with any assurance whether or not Christians (or any particular Christian), Muslims (or any particular Muslim) etc. in the contexts in which they happen to find themselves, are justified in their religious beliefs. A lot of recent epistemological discussion has gone into the question just what, if anything, justification is – and some have expressed grave doubts that there is one thing called ‘justification’.¹³ It bears pointing out that some accounts of justification are such that on their basis it could be argued that no one can be justified in her religious belief; other accounts are such that on their basis it could be argued that some but certainly not all religious believers can be justified in their beliefs; and yet other accounts are such that on their basis it could be argued that virtually all religious believers can. Let me give examples of each of these accounts.

Some accounts of justification have it that a person’s belief is justified provided the belief is supported by scientifically validated methods of research. On this basis it has been argued that no one can be (and in fact no one is) justified in any religious belief she might have.¹⁴

Other accounts have it that a belief is justified provided the belief is produced by a de facto reliable belief-forming mechanism (and a belief-forming mechanism is reliable when the preponderance of beliefs formed by that mechanism are true).¹⁵ On this account, someone may be justified in holding a belief without knowing that she is: that person’s belief may be generated by a de facto reliable mechanism, without the person knowing that it is thus generated. On this basis it can be argued that if A’s religious belief that p is generated by mechanism-1 and B’s religious belief that not-p by mechanism-2, at least either A of B is not justified in her belief – and then either mechanism-1 or mechanism-2 must be unreliable.

Yet other accounts have it that a person is justified in holding a certain belief only if the degree of that person’s belief is proportional to the degree to which it is probable given the evidence that is available to him.¹⁶ In order to be able to address the questions whether a religious believer is justified in her belief, we need to know a lot about her evidence. In the absence of that, we cannot estimate whether she is justified in holding her belief.

On yet another account a person is justified in holding a belief provided her belief coheres with other beliefs that she holds.¹⁷ If we do not know what other beliefs a person has, we cannot estimate whether a particular belief of her is justified.

All of this is goes to show that premise (1) of Grube’s argument cannot be properly discussed unless we know a lot more about what Grube means by ‘justification’. It also goes to show that whether one’s religious belief will be counted as ‘justified’ is fully dependent on what one takes ‘justification’ to be.

What about premise (2)? Should we accept it? Again I am reluctant to say ‘yes’. The worries I expressed about premise (1) carry over to (2) as well, of course. But (2) faces a new problem, even if my worry about what justification is would be alleviated. The problem is this: why should we even think that having a justified religious belief is sufficient for a believer to merit toleration? It would certainly seem possible that a person has a justified belief but still should not be tolerated. And this is something Grube himself acknowledges. Think of the quasi-religious Nazi-beliefs that some people have. These people should not be tolerated in this account.

So my question is: why should we even think there is a meaningful relation between a person’s belief being justified, and that person’s meriting toleration? As we have seen, near the end of his paper Grube himself admits that

1. If a person’s religious belief is not justified, that person may still merit toleration.

And he also admits that

1. Even if a person’s religious belief is justified, that person may still not merit toleration.

Having admitted these points just is admitting that having justified belief is neither necessary nor sufficient for meriting toleration. And this means that there simply is no meaningful relation between justification and toleration. And this, in turn, means that we have not been given a solid argument for toleration.

To assess the plausibility of (2) in conjunction with the qualifications (i) and (ii), compare the following two statements that are mimicked after (2), (i) and (ii):

• (2*) If the adherents of the world’s religions wear green shoes, they should be tolerated. [But (i) some adherents of the world’s religions that don’t wear green should also be tolerated; and (ii) some adherents of the world’s religions that do wear green shoes should not be tolerated!]

• (2**) If the adherents of the world’s religions are vegetarians, they should be tolerated. [But (i) some adherents of the world’s religions that are not vegetarians should also be tolerated; and (ii) some vegetarian adherents of the world’s religions should not be tolerated.]

(2*) and (2**), I take it, look silly. And they look silly because there just does not seem to be a meaningful relation between wearing green shoes and meriting toleration, nor a meaningful relation between being a vegetarian and meriting toleration. But (2) seems to be in the same league!

My conclusion of this section, then, is that we have not been given a successful argument for religious toleration on the basis of justified belief. Premise (1) needs a lot of work before it can be properly evaluated, whereas premise (2) as it stands is no more plausible than (2*) or (2**).¹⁸

Additional information

Notes on contributors

René van Woudenberg

René van Woudenberg is a professor of philosophy at the Vrije Universiteit, Amsterdam, where he teaches Epistemology and Metaphysics. He is the director of the Abraham Kuyper Center for Science and Religion. Email: [email protected] For publications consult his website renevanwoudenberg.nl.

Notes

1. Turner, Tonight We Will Fake Love.

2. Grube, “Justified Religious Difference.”

3. Grube lumps together toleration and respect. These are, however, quite distinct attitudes. I may tolerate your views in the sense that Wolterstorff has explained, so in the sense that I take it upon me to put up with your beliefs that I thoroughly dislike. (Wolterstorff, “Tolerance, Justice, Dignity”) But toleration does not require respect. In fact, it would seem that the more respect I have for your beliefs, the less occasion there is for me to have to tolerate your beliefs (or you). I therefore concentrate on toleration only.

4. See van Woudenberg, Toeval en ontwerp in de wereld, 153–68.

5. Grube subsumes under bivalence also the principle of excluded middle. But that is a different principle. Let ‘Tp’ represent ‘p is true’ and ‘T-p’ ‘p is false’ (and let ‘v’ represent the disjunction), then bivalence is Tp v T-p; excluded middle is T(p v-p). Since nothing in his argument hangs on this, I let it pass.

6. This is not to deny that in specific contexts sentences like the one given in de body of the text will usually have a determinate truth value: contextual clues indicate how the statement must be understood. Outside of any specific context, however, such sentences are incomplete.

7. Usually, bivalence is formulated in terms of propositions – and propositions are usually supposed to be meaningful, complete and unambiguous. See for example Audi, Cambridge Dictionary of Philosophy, s.v. ‘Principle of Bivalence’.

8. In the Dutch context Jan Riemersma has recently also defended that bivalence should be given up, by arguing that true contradictions exist, i.e. that there are propositions such that both they and their denials are true. For an excellent response to this effort, see De Ridder, “Riemersma over Plantinga.”

9. If two persons, say a Muslim and a Christian, have conflicting beliefs about God, this does not entail that they therefore refer to different Gods, or worship different Gods. Two persons can refer to the Dutch King, even if one of them believes he has three daughters, and the other that he has four. Likewise, two persons can refer to the only God that exists, but hold incompatible beliefs about God. See de Ridder & van Woudenberg, “Believing In, Referring To, and Worshipping the Same God.”

10. See e.g. Lemos, Introduction to the Theory of Knowledge, 13–17.

11. E.g. BonJour, Structure of Empirical Knowledge, 7–8.

12. But I note that Grube’s article seems to display shift of emphasis as it progresses: initially religious tolerance stands center stage, later on it is inter religious dialogue – which is quite another topic.

13. Alston, Beyond ‘Justification’. As Peels 2010 argues, Alston’s views are problematic. But what is not problematic is Alston’s point that ‘epistemic justification’ is a moving target.

14. See Philipse, “The Real Conflict Between Science and Religion”; for a criticism of Philipse’s proposal see Van Woudenberg & Rothuizen van der Steen, “Science and the Ethics of Belief.”

15. Goldman, “What is Justified Belief”; Plantinga, Warranted Christian Belief.

16. Conee & Feldman, “Evidentialism.”

17. Lehrer, Theory of Knowledge, 87.

18. For comments and advice on an earlier version of this paper, I am indebted to Kees van der Kooi, Rik Peels, and Jeroen de Ridder. Work on this paper was made possible by a grant from the Templeton World Charity Foundation; the views expressed in the paper are the author’s and do not necessarily reflect the views of the grant-giver.