A plurality of reasonable yet incompatible comprehensive doctrines is the normal result of the exercise of human reason. (John Rawls, Political Liberalism)
Most people think that it is possible for reasonable people to disagree – for two people to disagree about some question, even though both of their beliefs about the question are reasonable. Indeed, most people not only think that it is possible for reasonable people to disagree, but that it is possible for people to recognize that they are parties to a reasonable disagreement. Epistemologists have puzzled over how this can be possible. The aim of this paper is to explain how it is possible for reasonable people to recognize that they are parties to a reasonable disagreement.
I'm guided by the idea that the possibility of mutually recognized reasonable disagreement is a piece of contemporary common sense. The epistemological view that can explain the possibility of mutually recognized reasonable disagreement, therefore, has something in its favor. It has an advantage over theories that suggest that mutually recognized reasonable disagreement is impossible. Consider the view that parties to a “peer” disagreement ought to suspend judgment (cf. Feldman Reference Feldman and Hetherington2006), and thus cease to disagree, but also the view that the objectively reasonable party to a “peer” disagreement, but only that party, may reasonably maintain her position (cf. Kelly Reference Kelly2005). Both of these views seems to preclude the possibility of mutually recognized reasonable disagreement.
I'll proceed as follows. First, I'll articulate a view that I call “naïve liberalism” (§1): the view that there can be mutually recognized reasonable disagreements in religion and politics. Then, I'll articulate what I'll call the “entitlement claim” (§2), and argue that this can explain the truth of naïve liberalism (§3). Finally, I'll offer some arguments for the entitlement claim (§4), and consider some objections (§5).
1. NAÏVE LIBERALISM
Richard Feldman (Reference Feldman and Antony2007) describes a common sense epistemological view, held by most of his students in a course on religion:
Although a wide variety of different religious views were represented in the class … almost all the students had a great deal of respect for the views of the others. They “agreed to disagree” and concluded that “reasonable people can disagree” about the issues under discussion. (194)
What Feldman's students thought was that “while they had their own beliefs, the others had reasonable beliefs as well” (200). And Feldman rightly points out that this view is “widespread” in the contemporary world. This liberal epistemological view, in other words, is a piece of contemporary common sense, at least in societies where political liberalism holds sway. Here's a sympathetic formulation:
Naïve liberalism: Some ordinary religious and political disagreements are (or could become) mutually recognized reasonable disagreements.Footnote 1
Four comments on this. First, note that naïve liberalism is committed to the possibility of mutually recognized reasonable disagreement. It is a corollary of naïve liberalism that some ordinary religious and political disagreements are (or could become) reasonable disagreements. But naïve liberalism says more: it says that the parties to certain reasonable disagreements can sometimes come to recognize each other's reasonableness. Second, note that we concern ourselves with ordinary religious and political disagreements. I mean the kind of religious and political disagreements with which we are familiar in our non-academic lives; the kind of disagreements to which almost everyone living under political liberalism has been a party. Third, a clarification: we are concerned with disagreement in belief, and in particular we shall confine ourselves to cases in which one person believes that p and another believes that ~p. Fourth, for this reason, our formulation is confined to religious and political disagreements. Its scope could possibly be expanded; but we shall bracket moral disagreements (because they are less obviously disagreements in belief) and philosophical disagreements (because it is not obvious that philosophers believe the views that they advance, nor is it especially intuitive that they ought to believe the views that they advance).
Feldman wonders “what this respectful and tolerant attitude can sensibly amount to,” (Reference Feldman and Antony2007: 194; see also 198) and argues that naïve liberalism is false (Reference Feldman and Hetherington2006: 222–35; Reference Feldman and Antony2007: 204–10). My aim here is to defend naïve liberalism; my strategy (cf. §4) will be to appeal to a relativistic claim about reasonable belief (§2) and show how this vindicates the possibility of mutually recognized reasonable disagreement (§3). I will assume, then, that naïve liberalism is true, and show how the truth of this view can be explained.
We need to say something about the relevant notion of reasonable belief. Feldman sets aside a number of senses of “reasonable” that are distinct from the relevant sense. For example, we aren't using “reasonable” in any of the following senses:
• S's believing that p is reasonable iff S is a generally reasonable person (Reference Feldman and Hetherington2006: 220; Reference Feldman and Antony2007: 204).
• S's believing that p is reasonable iff S “is not being flagrantly unreasonable” (Reference Feldman and Antony2007: 202) or “crazy” (Reference Feldman and Hetherington2006: 221) in believing that p.
• S's believing that p is reasonable iff S's believing that p has “practical” or “prudential or moral” value (Reference Feldman and Hetherington2006: 221) or if S's believing that p is “prudentially or practically … reasonable” (Reference Feldman and Antony2007: 203).
So in contrast with the “practical” (etc.) sense of “reasonable,” the relevant notion of reasonable belief is that of epistemically reasonable belief (more on which below, §5.4). Like Feldman, in his work on disagreement, I won't offer a theory or account of reasonable belief. Three comments on the notion of reasonable belief, however, will make it sufficiently clear to proceed.
First, to say that S's believing that p is reasonable is to say that S is permitted to believe that p, and to say that S's believing that p is unreasonable is to say that S ought not believe that p. This focus on what we ought to believe is standard in the literature on the epistemology of disagreement: Feldman (Reference Feldman and Antony2007) writes that “one should give up one's beliefs in the light of the sort of disagreement under discussion” (213 and passim), David Christensen (Reference Christensen2007) asks “[h]ow should I react when I discover that my friend and I have different beliefs on the same topic?” (188; see also his Reference Christensen2009a: 231 and Reference Christensen2009b), Roger White (Reference White2005) discusses the issue of whether my “conviction should be undermined” in various cases (450 and passim, using throughout the language of “permission” and “obligation”) and elsewhere says that “it appears that I ought to be very skeptical of much of what I believe” (Reference White2009: 234 and passim), and Thomas Kelly writes that his issue is “how an awareness of disagreement should affect our beliefs” (Reference Kelly2005: 171 and passim), and asks “how (if at all) should we revise our original views?” (Reference Kelly, Feldman and Warfield2010: 112 and passim).
Second, the relevant sense of “reasonable belief” is distinct from that articulated by the following reliabilist theory of reasonable belief:
• S's believing that p is reasonable iff “S's believing p … results from a reliable cognitive belief-forming process (or set of processes)” (Goldman Reference Goldman and Pappas1979: 97; cf. Plantinga Reference Plantinga2000: 110–13, Bergmann Reference Bergmann2009: 337–8).
We need an alternative notion of reasonable belief to vindicate an intuition about the victim of a deceptive demon, whose beliefs seem reasonable – in some sense – even though they do not result from reliable belief-forming processes. This intuition forms the basis for the “new evil demon problem” (Cohen and Lehrer Reference Cohen and Lehrer1983; Cohen Reference Cohen1984). If we appeal to the reliabilist theory of reasonable belief, we get the wrong result when it comes to the victim of a deceptive demon: we get the result that her beliefs are not reasonable. But, intuitively, there is some sense in which they are reasonable – the demon victim is doing all the right things in her attempt to believe truths, but her environment is not cooperating. This is the alternative sense of “reasonable belief” with which we are concerned here. This is not to say that we should dispense with the notion articulated by the reliabilist, but I will assume that the “new evil demon problem” shows, at least, that the reliabilist's theory needs to be supplemented.Footnote 2 In this paper I'll articulate a claim about reasonable belief (§2.1) according to which some beliefs, resulting from apparently but not actually reliable sources, are reasonable.
The need for an alternative notion of reasonable belief is good news for naïve liberalism. The naïve liberal thinks that it's possible for two people to reasonably disagree. Suppose we have such a case: Jill believes that p and Ted believes that ~p. What's the naïve liberal's view? That both Jill's belief and Ted's belief are true? That's incoherent; if we can avoid attributing such incoherence, we should. That Jill knows that p and Ted knows that ~p? Again, incoherent when “knows” is factive. That Jill formed her belief in a reliable way and Ted formed his belief in an equally reliable way? Perhaps. But the naïve liberal also thinks that some reasonable disagreements could become mutually recognized reasonable disagreements. Is it plausible to think that Jill and Ted could think that both their beliefs were formed in an equally reliable way, and yet reasonably continue to believe that p and that ~p, respectively? In the absence of any consideration that favors Ted's position over Jill's, it seems unreasonable for Ted to believe that p while conceding that methods, as reliable as his, have also yielded the negation of his belief. Such a case seems like one in which Ted ought to suspend judgment about whether p (Feldman Reference Feldman and Hetherington2006: 233–5). This doesn't preclude the possibility of a vindication of naïve liberalism by appeal to a reliabilist theory of reasonable belief, but it makes the task look difficult (cf. §4). Here I shall pursue an alternative route.
Third, in line with my first two comments, if S's believing that p is reasonable, then S is blameless in believing that p. This jibes with the idea that reasonable belief is permissible belief, and with the idea that the demon victim's beliefs are reasonable, since, intuitively, she cannot be blamed for her (false, unreliably formed) beliefs. So I assume that reasonable belief entails blameless belief. I leave open whether blameless belief entails reasonable belief.Footnote 3
A final clarification on the articulation of naïve liberalism. Note well that the view does not maintain that disagreement never requires suspension of judgment. Suppose that Paddy and Roddy are discussing football when it becomes clear that they aren't on the same page when it comes to the outcome of last night's match: Paddy remembers reading that Hibs won; Roddy remembers reading that Hearts won. Intuitively, they should suspend judgment about the outcome of the match. So naïve liberalism is not equivalent to the view that you generally ought to be “steadfast” or to “stick to your guns” in response to disagreement. An appealing feature of the account defended here is that it will explain the conditions for mutually recognized reasonable disagreement (§3.2), and thus explain the difference between cases in which mutually recognized reasonable disagreement is possible and those in which such disagreement does not seem possible (e.g. the case of Paddy and Roddy).
2. THE ENTITLEMENT CLAIM
In this section, I'll articulate a relativistic claim about reasonable belief, inspired by the view that we enjoy a non-evidential entitlement for some of our beliefs. That view is defended by Crispin Wright (Reference Wright2004a, Reference Wright and McManus2004b) and Michael Williams (Reference Williams1996), and suggested by Ludwig Wittgenstein (Reference Wittgenstein1969) and Thomas Reid (in his 1785 Essays on the Intellectual Powers of Man). On my proposal, we enjoy a non-evidential entitlement to trust certain sources of information. Later on, we'll return to naïve liberalism (§3) and to arguments for the proposed claim (§4).
2.1. Entitlement to trust and reasonable belief
First, I'll define the notion of “entitlement to trust a source of information.” Second, I'll connect entitlement to trust and reasonable belief.
In general, the sources of information I trust will make a difference when it comes to how I ought to respond to a piece of evidence. Someone trusts a source of information when she is disposed to form beliefs on the basis of that source, i.e. if she is disposed, other things being equal, to believe that p when the source says that p.Footnote 4 Whether I take e to be evidence that p, and, if I do, how strongly I think e supports the proposition that p, will depend on the sources of information I trust. The horoscope for Pisces in today's Sun says: “You will amaze people as the moon reveals the hidden side of your personality.” Someone who trusts astrological predictions will be disposed to treat this as evidence that I will amaze people as the moon reveals the hidden side of my personality. Someone who puts no stock in astrological predictions will not be so disposed.
It seems that trusting a source of information can be warranted or unwarranted. It would be irrational for me to trust astrological predictions, as I have much evidence that they are not reliable (e.g. from their poor track-record), as well as much evidence against the theories from which they are derived (e.g. from the inconsistency of astrology with our best current scientific theories). At least for this reason, were I to trust astrological predictions, my trust would be unwarranted.Footnote 5
Trusting a source of information seems warranted if you have evidence of its reliability (cf. §5.2). So one species of warranted trust is trust based on evidence of a source's reliability. But is such evidential support necessary for warranted trust? Is there another species of warranted trust? The view I'm proposing here says that evidential support is not necessary for warranted trust. It is based on Wittgenstein's (Reference Wittgenstein1969) idea that “some propositions are exempt from doubt” and “are as it were like hinges on which [our doubts] turn.” (§341) Developing this idea, Wright (Reference Wright2004a: 188–97; Reference Wright and McManus2004b) argues that we are “entitled” to some beliefs, namely, those:
whose rejection would rationally necessitate extensive reorganisation of – or more, might even just throw into confusion – our highly complex conception of what kind of thing should be taken as evidence for what kind of proposition. (Reference Wright and McManus2004b: 41–2)
Beliefs of this kind will have as their content “presuppositions” of our “cognitive projects,” where:
p is a presupposition of a particular cognitive project if to doubt p (in advance) would rationally commit one to doubting the significance or competence of the project. (Reference Wright2004a: 191)
Such presuppositions are what Williams (Reference Williams1996) calls the “methodological necessities” for a given species of inquiry (123). Entitlement, for Wright, involves non-evidential permission to believe: I'm permitted to believe that p despite lacking evidence that p.Footnote 6 And the reason that I'm permitted to do so is that doing otherwise would throw into confusion my conception of what counts as evidence for what. Why think this? The reason has to be something like this: that I'm sometimes permitted to believe something, without evidence, because doing otherwise would be cognitively disastrous, i.e. it would be very bad, from the point of view of inquiry and the aim of acquiring knowledge. That's the idea I want to take away from Wright.
Assume, for the sake of argument, that that idea is correct.Footnote 7 If so, then it's plausible that I'm permitted to trust a source of information if not trusting it would cut me off from a significant set of truths. If it's plausible that we're entitled to believe the presuppositions of our cognitive projects, because doubting these propositions would undermine those projects, then we are likewise entitled to trust sources of information such that not trusting them would cut us off from a significant set of truths.Footnote 8 If the threat of evidential confusion can generate entitlement to believe some proposition, absent evidence in favor of that proposition, as Wright argues, then the threat of being cut off from the truth can generate entitlement to trust some source of information, absent evidence of the reliability of that source.
Let's articulate this idea. We'll say that S is entitled to trust a source of information F iffFootnote 9
(i) not trusting F would result in S losing access to a significant set of (apparent) truths,
(ii) S doesn't have evidence sufficient for her to believe that F is not reliable (call this the open question condition on entitlement to trust), and
(iii) S has not viciously avoided acquiring evidence relevant to the question of whether F is reliable (call this the anti-vice condition on entitlement to trust).
Three comments on this definition. First, in clause (i), I have defined entitlement to trust in terms of a person's losing access to a significant set of apparent truths. In the relevant sense, these are truths that appear true to the relevant S (i.e. to the subject of the would-be entitlement), where the proposition that p appears true to S iff S believes that p or would believe that p immediately upon reflection. Furthermore, implicit in the definition is the following idea: that whether someone is entitled to trust a source of information at a particular time depends (in part) on what propositions appear true to her at that time. What is relevant in determining entitlement to trust, at present, is whether the relevant propositions appear true, at present, and not whether the relevant propositions would appear true in any counterfactual situation.
Second, clause (ii) says that entitlement requires that a person not have evidence sufficient for her to believe that the relevant source is not reliable. It does not, however, require that a person have evidence sufficient for her to believe that the relevant source is reliable. For entitlement to exist, it must at least be an open question whether the relevant source is reliable; if that question has been closed in the negative, then entitlement does not exist. The idea is that it would obviously be unwarranted for you to trust a source of information whose non-reliability has been established.
Third, in clause (iii), I have defined entitlement to trust a source of information in terms of the vicious avoidance of acquiring evidence. The idea is that epistemic virtue sometimes requires us not to avoid acquiring certain evidence, by requiring inquiry into the reliability of a source. Suppose that you get most of your information about foreign affairs from the New York Times. However, a credible report emerges that the Times isn't a reliable source of information about foreign affairs, having made a number of significant false and uncorrected assertions about foreign affairs in the last year, as a result of systematically sloppy investigative reporting. It would be epistemically vicious for you to ignore this report, and to not inquire further into the reliability of the Times' reporting on foreign affairs. Alternatively, it would manifest epistemic vice were you to engage in biased inquiry in connection with the question of the reliability of the Times' reporting – for example, if instead of consulting the opinion of the independent and non-profit Institute for Journalism Ethics, you were to consult the opinion of the Centre for Journalistic Excellence, a think-tank owned by the same corporation that owns the Times. Entitlement is incompatible both with a vicious refusal to inquire into the reliability of a source and with viciously biased inquiry into the reliability of a source.
We'll turn now to the relationship between entitlement to trust and reasonable belief. I'll defend the view that it is sometimes reasonable for someone to believe that p, on the basis of the fact that some source of information says that p, in virtue of the fact that she is entitled to trust that source. When is this reasonable? I assume an open question condition on reasonable belief, analogous to the open question condition on entitlement to trust: reasonably believing that p requires that you not have evidence sufficient for you to believe that ~p. And I assume an anti-vice condition on reasonable belief, analogous to the anti-vice condition on entitlement to trust: reasonably believing that p requires that you not have viciously avoided acquiring evidence relevant to the question of whether p. Imagine that John has a prejudicial belief that women are bad at sports, and seeks out evidence that women are bad at sports, while actively avoiding the acquisition of evidence against this claim. John may end up in possession of strong evidence that women are bad at sports, but his prejudicial belief is still unreasonable, given its history of vicious formation. Alternatively, consider a case of vicious lack of inquiry: people keep pointing out that my understanding of Aristotle is completely mistaken, and that I really need to sit down and read his work, but I refuse, and continue to hold on to my misinterpretation. Reasonable belief is incompatible both with viciously biased inquiry and with vicious refusal to inquire. So I propose:
Entitlement claim: It is reasonable for S to believe that p, on the basis of the fact that some source of information F says that p, if
(1) S is entitled to trust F,
(2) S does not have evidence sufficient for her to believe that ~p, and
(3) S has not viciously avoided acquiring evidence relevant to the question of whether p.
The entitlement claim offers a sufficient condition for reasonable belief based on the testimony of a source of information. The entitlement claim isn't a theory or account of reasonable belief. It doesn't offer necessary and sufficient conditions for reasonable belief. Nor does it offer a reductive account of the notion of reasonable belief. This is essential, as clause (2) makes implicit appeal to the notion of reasonable belief: plausibly, someone has evidence sufficient for her to believe that p precisely when it is reasonable for her to believe that p on the basis of that evidence. In any event, the possibility of other species of reasonable belief is left open. (As well, I have not offered a theory or account of warranted trust, and have left open the possibility of species of warranted trust other than entitlement.)
Two comments on the formulation of the entitlement claim. First, according to clause (i) of the definition of entitlement to trust a source of information, S is entitled to trust F only if not trusting F would result in S losing access to a significant set of (apparent) truths. You might wonder why I didn't opt for the simpler:
(i*) not trusting F would result in S losing access to a significant set of truths.
We might then connect entitlement to trust with reasonable belief, with the resulting notion of reasonable belief resembling the reliabilist notion of reasonable belief we discussed above (§1). We'll return to this issue, below (§4, §5.4).
Second, Wright (Reference Wright2004a) requires, for entitlement to believe some presupposition, that there be no independent way to justify such belief (191–2). In my formulation, I have not included an analogue of Wright's requirement: that there be no F-independent way of assessing the reliability of F. I have required only that S not have not viciously avoided inquiring into F's reliability. If there is no F-independent way of assessing the reliability of F, surely no one is vicious for failing to seek independent assessment of F's reliability. But my formulation leaves open the possibility that someone might enjoy entitlement to trust a source of information, without inquiring after its reliability, even if such inquiry is possible, so long as she has not been vicious in failing to so inquire. The motivating idea for the entitlement claim is still present: you're entitled to trust a source of information when not trusting it would be (apparently) cognitively disastrous.Footnote 10 Note also that my arguments for the entitlement claim (§4) stand, even absent Wright's stronger requirement.
2.2. An objection forestalled: astrology
You might have worried that entitlement to trust is too easy to come by. For example, not trusting astrological prediction would result in the committed astrologer losing access to a significant set of (apparent) truths. So it might seem that the committed astrologer is entitled to trust astrological prediction, and might in virtue of that come to reasonably believe the deliverances of astrological prediction. And you might think that “anyone who believes in astrology believes in something that is simply indefensible” (van Inwagen Reference van Inwagen, Jordan and Howard-Snyder1996: 141; see also Feldman Reference Feldman and Antony2007: 211). The problem here is the apparent licensing of dogmatism or conservativism: the entitlement claim seems to imply that, just in virtue of being committed to astrology, the astrologer's astrological beliefs count as reasonable. Four replies to this.
First, given the open question condition on entitlement to trust (§2.1), the astrologer is not entitled to trust astrological prediction unless she lacks evidence sufficient for her to believe that astrological prediction is not reliable. As I argued above, I presently have evidence – its poor track-record and its incompatibility with contemporary science – sufficient for me to believe that astrological prediction is not reliable. Any astrologer with evidence sufficient for her to believe that astrological prediction is not reliable is not entitled to trust astrological prediction. But many, if not most, contemporary proponents of astrology do have such evidence, and so the entitlement claim does not imply that their astrological beliefs are reasonable.Footnote 11
Second, given the anti-vice condition on entitlement to trust (§2.1), the astrologer is not entitled to trust astrological prediction unless she has not viciously avoided acquiring evidence relevant to the reliability of astrological prediction. Today's Daily Mail advises Sagittarians that “[t]he eclipse in your opposite sign heralds a decisive drama.” Suppose, however, that for Sagittarians no “decisive drama” ensues. The committed astrologer might explain this away by creatively interpreting the course of events: the lack of drama of course was the drama, or some such nonsense. But this manifestation of vicious bias is incompatible with entitlement to trust. So many, if not most, contemporary proponents of astrology will not satisfy the anti-vice condition on entitlement to trust, and so the entitlement claim does not imply that their astrological beliefs are reasonable.
Third, given the open question condition on reasonable belief, the entitlement claim will never imply that it is reasonable for someone to believe that p when she has evidence sufficient for her to believe that ~p, pace some critics of appeals to entitlement (Brueckner Reference Brueckner2007; cf. Wright on predicting horse races with dice, Reference Wright2004a: 195). You might think that it can never be reasonable for someone to believe that p when she has evidence sufficient for her to believe that ~p, e.g. strong and undefeated evidence that ~p. The entitlement claim is consistent with this. What the entitlement claim does imply is that it is sometimes reasonable for someone to trust a source of information without having evidence of its reliability.Footnote 12
Fourth, recall the anti-vice condition on reasonable belief. The astrologer for the Daily Star maintains that I have been “getting [my] finances in order over the last few months.” But you might think that anyone who believed this had viciously avoided inquiring about my finances. To believe that people born in late February and early March have in general recently gotten their finances in order, without seeking out the available evidence on such matters, would manifest a willful and vicious refusal to inquire when inquiry is called for. So many, if not most, contemporary proponents of astrology will not satisfy the anti-vice condition on reasonable belief, and so the entitlement claim does not imply that their astrological beliefs are reasonable.
However, note well that the view I've proposed places no a priori restriction on the contents of propositions that might be reasonably believed in virtue of entitlement to trust. One encounters the idea that some beliefs – e.g. an astrological worldview – are unreasonable simply in virtue of their “absurd” or “crazy” content. My view rejects this idea (cf. §5.3). But it doesn't implausibly vindicate dogmatism or conservativism.
My view is a form of relativism about reasonable belief.Footnote 13 The sources of information I am entitled to trust will make a difference when it comes to what it is reasonable for me to believe. This is what will allow us to say something (§3.1) about the possibility of mutually recognized reasonable disagreement.Footnote 14
3. MUTUALLY RECOGNIZED REASONABLE DISAGREEMENT
In this section I'll argue that the entitlement claim (§2.1) can explain the truth of naïve liberalism (§1). In other words, I'll argue that the entitlement claim can explain the possibility of mutually recognized reasonable disagreements in religion and politics. So, in this section, I'll assume the truth of the entitlement claim. Below (§4), we'll turn to arguments in its favor.
3.1. A case of mutually recognized reasonable disagreement
Let us first describe a case of mutually recognized reasonable disagreement. Julia and Maria were raised in, and continue to be members of, different communities: Julia belongs to a community of Roman Catholics; Maria belongs to a community of secular humanists. Julia and Maria's membership in their respective communities comes down to various logistical, emotional, and cognitive relationships with other people. Julia spends much of her time with other Catholics, her extended family and her intimate friends are Catholics, she puts a high degree of trust in their testimony, and she is disposed to believe what they believe. Maria spends much of her time with other humanists, her extended family and her intimate friends are humanists, she puts a high degree of trust in their testimony, and she is disposed to believe what they believe. One outcome of their having been raised in different communities is that Julia and Maria have incompatible worldviews: an interconnected set of beliefs about metaphysics and ethics. Julia is a theist, believes that God sometimes miraculously intervenes in human affairs, takes the Christian bible to have significant authority on moral matters, trusts the testimony of priests as her church, and so on. Maria is an atheist, believes that miracles never happen, rejects the moral authority of the bible, distrusts the testimony of said priests, and so on.
Imagine further that Julia and Maria's worldviews are inherited from their communities. What I mean by this is that the best explanation of why Julia and Maria have their different worldviews is their historical and ongoing membership in their respective communities. We should not assume a fortiori that their worldviews are unreasonable; we have said something about the causes of Julia and Maria's beliefs, but this alone does not threaten the reasonableness of their beliefs. But the source of Julia and Maria's worldviews is largely the testimony of their respective communities (in a broad sense of “testimony”). Each has subjected her worldview to critical scrutiny (at college, say), but each found the arguments against her view, and in favor of incompatible views, unconvincing. The case is therefore neither a case of religious disagreement involving religious experience or the sensus divinitatis (e.g. one person has had some religious truth revealed to her, and the other hasn't) nor a case involving scholars or academics (e.g. an academic debate between a theist and an atheist where both take themselves to have sound non-testimonial arguments for their respective positions).
Now consider some particular proposition such that that proposition is part of Julia's worldview, and its negation is part of Maria's worldview. At La Villa de Guadalupe north of Mexico City there is a piece of cloth called the tilma of San Juan Diego, which some people believe to have a miraculous past. It is believed, among other things, that in 1521 an image of the Virgin of Guadalupe appeared on Juan Diego's tilma (a kind of cloak) and that the tilma was miraculously impervious to an anarchist bombing in 1921 (which destroyed the shrine in which the tilma was kept). Let's consider the proposition that God protected the tilma in 1921, and suppose that this proposition is widely believed in Julia's community (and that Julia believes it as well), and that it is widely disbelieved in Maria's community (and that Maria disbelieves it as well).
Imagine, finally, that Julia and Maria meet, and come to discuss the question of whether God protected the tilma in 1921. They offer arguments for and against that claim. They discuss – in some detail – their respective communities and their respective personal histories. They discuss the fact that the proposition that God protected the tilma in 1921 coheres with Julia's worldview, and that its negation coheres with Maria's worldview. They discuss the fact that their respective worldviews are inherited, in the sense described above. They share everything they know that is relevant to the issue of whether the tilma was miraculously protected, and they share everything they know about how they came to believe what they respectively believe about that proposition.Footnote 15 However, since neither Julia nor Maria scores a decisive argumentative victory, neither suspends judgment about the proposition that God protected the tilma in 1921. Julia continues to believe that proposition, and Maria continues to disbelieve that proposition.
The application of the entitlement claim (§2.1) to this case is relatively straightforward. Julia is entitled to trust the testimony of her community, because not trusting that testimony would result in her losing access to a significant set of (apparent) truths, namely, various truths about metaphysics and ethics. If she were to give up trusting the testimony of her community, it would result in a significant change in her cognitive practices, and the loss (so it seems to her) of her religious knowledge. And Julia satisfies the open question and anti-vice conditions on entitlement to trust. Julia might, of course, have inquired further into the reliability of her community's testimony – she might have engaged in some philosophical theology, attempting to verify the metaphysical and ethical claims of her community – but her failure to do so does not seem epistemically vicious. Therefore, she is entitled to trust the testimony of her community, which says that God protected the tilma in 1921. Moreover, so we can easily imagine, she does not violate the open question and anti-vice conditions on reasonable belief. To put it in a completely ordinary way, we can say that her belief is reasonable, given her background. The same, mutatis mutandis, for Maria. I conclude that both Julia and Maria's beliefs about the tilma are reasonable.
We have made things simpler by treating trust as an all-or-nothing affair (§2.1), and by imagining that Julia trusts the testimony of her community, and not that of Maria's community, and that Maria trusts the testimony of her community, and not that of Maria's community. More realistically, we should have said that Julia trusts her own community's testimony to a greater degree than that of Maria's community, and the same for Maria, mutatis mutandis. Given a suitable formulation of the entitlement claim, in terms of degrees of trust, we would still get the desired result: a difference in reasonable belief, or in reasonable degree of belief, in virtue of a difference in entitlement to trust.
Finally, we can see how Julia and Maria are poised to recognize the reasonableness of their respective beliefs, and thus how this ordinary religious disagreement could become a mutually recognized reasonable disagreement. Imagine that Julia recognizes that Maria is entitled to trust the testimony of her (Maria's) community, given Maria's perspective (i.e. her secular worldview), and that Maria recognizes that Julia is entitled to trust the testimony of her (Julia's) community, given Julia's perspective (i.e. her Catholic worldview). Each takes the other's belief about the tilma to be false, each takes the other's worldview to be largely wrong, and each takes the other's favored source of information to be not reliable. But this is compatible with each taking the other's belief about the tilma to be reasonable, since it is compatible with taking the other to be entitled to trust her favored source of information, and with taking her belief to be reasonable, in virtue of her being so entitled.
3.2. Conditions for mutually recognized reasonable disagreement
This suggests that, in general, mutually recognized reasonable disagreement will be possible when disagreement results from a difference in inherited worldview.Footnote 16 In such cases, difference in inherited worldview, i.e. difference that results from historical and ongoing membership in different communities, leads to entitlement to trust different sources of information.Footnote 17, Footnote 18
Disagreement resulting from difference in worldview is deep disagreement. Julia and Maria disagree not only about the tilma, but about theism, about the possibility of miracles, about the reliability of local clerical testimony. Other disagreements appear shallow by comparison. Recall Paddy and Roddy (§1), who agree about everything, apart from the outcome of the match. Their worldviews are the same; they just disagree on this one point. In such cases of shallow disagreement, suspension of judgment seems warranted, and thus, in such cases of shallow disagreement, mutually recognized reasonable disagreement will not be possible.Footnote 19, Footnote 20
3.3. Incommensurability
Wittgenstein (Reference Wittgenstein1969) compares Moore's certainty that he has hands – a paradigm “hinge proposition” – to certain religious beliefs, for example that Jesus did not have two human parents, and that in the Eucharist wine transforms into blood (§239). For Wittgenstein, hinge propositions are certainties immune from doubt; and thus the possibility arises of a conflict between people with conflicting hinge propositions. Thus:
239. […] And so if Moore said “I know that this is wine and not blood,” Catholics would contradict him.
609. Supposing now we meet people [who] instead of the physicist, they consult an oracle. (And for that we consider them primitive.) Is it wrong for them to consult an oracle and be guided by it? — If we call this “wrong” aren't we using our language-game as a base from which to combat theirs?
611. Where two principles really do meet which cannot be reconciled with one another, then each man declares the other a fool and a heretic.
612. I said I would ‘combat’ the other man, — but wouldn't I give him reasons? Certainly; but how far do they go? At the end of reasons comes persuasion. (Think what happens when missionaries convert natives.)
The picture of deep disagreement suggested by these passages is unappealing. It implies the radical incommensurability of worldviews, on which rational dialogue and mutual criticism, between people with different worldviews, is impossible. This is the view that Feldman has in mind when he criticizes the idea that “starting points are beyond rational scrutiny” (Reference Feldman and Hetherington2006: 225).Footnote 21 But the entitlement claim (§2.1) doesn't imply radical incommensurability. Two reasons for this. First, as we just saw (§3.1), mutually recognized reasonable disagreements are possible. Second, recall the open question conditions on entitlement to trust and on reasonable belief (§1). It is a necessary condition on entitlement to trust a source of information that you not have evidence sufficient for you to believe that said source is not reliable, and a necessary condition on reasonable belief that p that you not have evidence sufficient for you to believe that ~p. Our trust in a particular source of information, and our beliefs based on the testimony of that source, are open to rational scrutiny. Two people who trust different sources of information might have a rational discussion about the reliability of a particular source of information, a source that one, but not the other, trusts. Far from having to see disagreeing others as “fools” and “heretics,” we can see them, at least sometimes, as reasonable in their beliefs. And far from the inevitability of non-rational “combat,” we can hope for rational dialogue.
A kind of partial incommensurability is implied by the entitlement claim: people with different entitlements to trust may be unable to agree on how to respond to a particular piece of evidence, e.g. the fact that local priests believe that God protected the tilma in 1921. So there is no guarantee of fruitful dialogue between people with different entitlements to trust. Likewise, there may be no way to resolve many disputes about the reliability of sources – at the end of their discussion, Julia and Maria (§3.1) have come to no such agreement, and no such agreement may be possible for them.
Rational dialogue will be possible only when there is some overlap of entitlements to trust. (Mutual recognition of reasonableness, on the other hand, is possible even without overlap of entitlements.) Consider a simple case: Simon believes in astrology, Amanda doesn't, and she aims to convince him of this by testing the predictions of Simon's favorite spiritualist, Lupe. A careful record of Lupe's predictions is kept, and they are compared with what actually happens (which is noted using means independent of Lupe's predictions). Once Simon sees that Lupe is almost invariably wrong in her predictions, he eventually (if reluctantly) comes to doubt the reliability of astrology, and his entitlement to trust astrology evaporates. A relativistic diagnosis of what goes on in this case: Simon and Amanda share a commitment to trusting scientific induction, as well as whatever independent means are used to test Lupe's predictions, e.g. perception. And this is the basis for Amanda's critique: Simon's commitment to scientific induction and perception rationally requires him to believe, after enough failed predictions, that astrology is not reliable.
But, again, this kind of resolution is not guaranteed. If Simon has rejected induction and perception, and adopted astrology as his only guide, it may be that he reasonably remains unconvinced. Imagine, for example, that he has adopted an extreme version of astrology, on which all evidence against the reliability of astrological prediction is misleading. He is convinced, in effect, of an unfalsifiable conspiracy theory about the evidence against the reliability of astrological prediction. If this is the case, he comes to resemble a certain kind of philosophical skeptic, who maintains that “all our reasoning and judging powers are fallacious in their nature”; as Reid points out, “it would be impossible by argument to beat him out of this stronghold; and he must even be left to enjoy his scepticism” (Reference Reid1983: 276). The extreme astrologer, who rejects induction and perception, must likewise be left to enjoy her position. It is no mark against the entitlement claim that it does not imply that the extreme astrologer's astrological beliefs are unreasonable.
Partial incommensurability also arises from the anti-vice conditions on entitlement to trust and on reasonable belief. Whether someone has been epistemically vicious may itself be a matter of disagreement: when secularists talk about superstition, credulity, and enthusiasm, and religious people talk about faith and spirituality, they are talking about the same thing. If two people cannot agree about the relevant questions of epistemic vice, then they will be unable to come to mutually recognized reasonable disagreement.
4. ARGUMENTS FOR THE ENTITLEMENT CLAIM
I've argued that the entitlement claim can explain the truth of naïve liberalism (§3). Why accept the entitlement claim? Three arguments speak in its favor.
First, the entitlement claim (§2.1), like other appeals to non-evidential entitlement, vindicates a certain kind of anti-skeptical strategy. The skeptic's argument challenges the permissibility of trusting perception, for example, by pointing out that we are not in possession of evidence that favors the reliability of perception: we have no evidence that rules out the possibility that perception is not reliable (e.g. because we are deceived by a demon), and even the evidence we do have, in favor of the reliability of perception, itself comes from perception – which seems a dubious source to which to appeal in defense of the reliability of perception. The appeal to entitlement allows us to see why we are permitted to trust perception, in spite of all of this. We need not maintain that other anti-skeptical strategies are unable to resolve this problem, by showing that we are, for example, justified in believing that perception is reliable. But two things speak in favor of the appeal to entitlement. The first is that this allows us to see immediately why we are permitted to trust perception, even when we have no response to the skeptical challenge. The second is that the appeal to entitlement does justice to some of the intuitions that motivate the skeptical argument, by conceding that our warrant for trusting perception is “unearned,” i.e. not based on evidence. Thus we can see why the problem of skepticism moves us, while at the same time vindicating our refusal to be swayed by the skeptical argument.
Second, recall the “new evil demon” intuition (§1). Consider someone who is, according to the definition given above (§2.1) entitled to trust some source of information, where said source is, unbeknownst to her, not reliable. Is it reasonable for her to trust said source of information? The intuition that motivates both the “new evil demon problem” and the entitlement claim is that there is something right about a person's trusting a source of information that appears reliable, even if there is also something wrong about such trust, namely, that it will lead her to form many false beliefs. The person who trusts sources of information that she is entitled to trust is, like the victim of the demon, doing her best to get at the truth, given how things seem to her. We'll return to this point below (§5.4).
Third, the entitlement claim can explain the truth of naïve liberalism (§3.1). Since naïve liberalism is a common-sense epistemological view, this speaks in favor of the entitlement claim.Footnote 22 In particular, this is a relative merit of the entitlement claim by contrast with views on which mutually recognized reasonable disagreement is impossible. Note well the structure of this argument. The argument is not that naïve liberalism is a surprising corollary of the entitlement claim, which is independently plausible, and that we should therefore accept naïve liberalism. The argument is that the truth of the entitlement claim explains the truth of naïve liberalism, which is independently plausible, and that we should therefore accept the entitlement claim.
This third argument, like the previous two, is not decisive. This is for two reasons. First, you might find reason to reject naïve liberalism, and thus to revise common sense. In the same way, you might find reason to reject anti-skepticism, or to reject the “new evil demon” intuition. My arguments for the entitlement claim are based on certain assumptions – anti-skepticism, the “new evil demon” intuition, naïve liberalism – and these assumptions could be rejected.
Second, you might offer an alternative explanation of the truth of naïve liberalism. In the same way, you might give an alternative vindication of anti-skepticism and the “new evil demon” intuition. The success of these three arguments ultimately depends on an examination of the relative merits of alternative approaches, which is beyond the scope of the present inquiry.
However, it is important that contemporary epistemologists have, for the most part, struggled to come up with such an alternative explanation of the truth of the naïve liberalism.Footnote 23 Some (Kelly Reference Kelly2005, Feldman Reference Feldman and Hetherington2006, Reference Feldman and Antony2007) suggest the impossibility of reasonable disagreement in religion and politics, which implies the impossibility of mutually recognized reasonable disagreement in religion and politics. Others (Goldman Reference Goldman, Feldman and Warfield2010) defend the possibility of reasonable disagreement in religion and politics, but give no insight into the possibility of mutually recognized reasonable disagreement in religion and politics (see footnote 16). Given the inability of many other approaches to explain the truth of naïve liberalism, the fact that the entitlement claim can do this, it seems to me, is the most important and compelling reason to endorse it.
5. OBJECTIONS AND REPLIES
5.1. Suspension of judgment about reliability
The open question condition on entitlement to trust (§2.1) says that your would-be entitlement to trust a source of information is compromised if you have evidence sufficient for you to believe that said source is not reliable. You might argue that your would-be entitlement to trust a source of information is also compromised when you have suspended judgment about whether said source is reliable. Should the definition of entitlement to trust be amended to include this condition on entitlement to trust?
What speaks in favor of the proposed amendment? One possibility is that the motivation for the open question condition on entitlement to trust also supports the proposed amendment. The idea behind the open question condition (§2.1) was that it would obviously be unwarranted for you to trust a source of information whose non-reliability has been established. Is it likewise obviously unwarranted to trust a source whose reliability is an open question, i.e. a source such that you suspend judgment about whether it is reliable? It seems to me that there is a relevant difference between the case in which you trust a source whose non-reliability has been established and the case in which you trust a source whose reliability is an open question. In the former case, you have evidence against the reliability of the source in question, whereas in the latter case, you may very well have no such evidence. Suspension of judgment about whether p does not constitute evidence that ~p; nor does it constitute evidence that p. Suspension of judgment about whether p is not always or typically a response to evidence that ~p; nor is it always or typically a response to evidence that p. Suspension of judgment about whether p is sometimes a response to ambiguous or counterbalanced evidence relevant to the question of whether p, but it may also be a response to higher-order evidence, an undercutting defeater, to a lack of relevant or useable evidence, or to any number of situations which seem to preclude coming to form a reasonable belief about whether p. Trusting a source whose non-reliability has been established is unwarranted because in that case the would-be warrant for your trust is opposed by your evidence. But when you trust a source whose reliability is an open question, it may be that no such opposition exists.Footnote 24 So the motivation for the open question condition on entitlement to trust does not support the proposed amendment.
Another argument in favor of the proposed amendment appeals to the illegitimacy of bootstrapping. The entitlement claim (§2.1) allows that a person can form reasonable beliefs, based on a source of information, absent evidence that said source is reliable. This, so the argument goes, opens the door to an illegitimate argument from those beliefs to the conclusion that said source is reliable. However, this problem does not threaten the entitlement view in particular; it threatens reliabilism, phenomenal dogmatism, and any view on which a person may form a reasonable belief, based on a source of information, absent evidence that said source is reliable (cf. Cohen Reference Cohen2002). It's beyond the scope of this paper to defend that idea (but see Weisberg Reference Weisberg2010 for a promising strategy).
5.2. Disagreement as evidence against reliability
The fact that you disagree with someone is often evidence that you have made a mistake. This seems like the right thing to say about the case of Paddy and Roddy (§1): both Paddy and Roddy should be surprised to discover that they disagree about the outcome of the match, and they should respond accordingly, by suspending judgment in advance of further inquiry. As Christensen (Reference Christensen2007) puts it:
Actual disagreement … is informative because it [sometimes] provides evidence that a certain possibility – the possibility of our having made an … error – has been actualized. It makes what we already know possible more probable. (Reference Christensen2007: 208)
In this respect, disagreement is often evidence that you are in error. Given this, you might also think that disagreement is often evidence against the reliability of the sources of information that you trust. This threatens to undermine my account of mutually recognized reasonable disagreement (§3). Consider the case of Julia and Maria. If the existence of Julia and Maria's disagreement is evidence sufficient for each to believe that her respective sources of information are not reliable, then neither Julia nor Maria is entitled to trust said sources, and the entitlement claim does not apply to their beliefs about the tilma.
However, there are two reasons to think that the existence of Julia and Maria's disagreement is not evidence sufficient for them to believe that their respective sources of information are not reliable.
First, the existence of disagreement per se is neutral as to which of the disagreeing parties is in error; it speaks equally against both parties. Therefore, when the disagreeing parties trust different sources of information, the existence of disagreement is, at most, neutral as to whose sources are not reliable, and speaks equally against both parties. The existence of Paddy and Roddy's disagreement is neutral as to the outcome of the match, which is why suspension of judgment about the outcome of the match is called for. Likewise, the existence of Julia and Maria's disagreement is neutral as to whose sources are not reliable. The evidence provided by their disagreement is neither sufficient for Julia to believe that her sources are not reliable nor sufficient for Maria to believe that her sources are not reliable. At most, suspension of judgment about the reliability of their respective sources is called for. But suspension of judgment about the reliability of a particular source is compatible with entitlement to trust that source (§5.1).
Second, the case of Julia and Maria is not plausibly understood as a case in which disagreement is evidence against reliability. The reason is that each of Julia and Maria reasonably think that the testimony of the other's community is not reliable (§3.1), since this follows straightforwardly from their respective worldviews. This is compatible with Julia thinking that Maria is entitled to trust the testimony of her (Maria's) community, and with Maria thinking that Julia is entitled to trust the testimony of her (Julia's) community. Entitlement to trust is orthogonal to reliability (cf. §2.1). But given each of their beliefs about the reliability of the testimony of the other's community, neither Julia nor Maria has reason to think that the testimony of her own community is not reliable. Julia reasons as follows: “Maria is entitled to trust the testimony of her community, and thus her belief about the tilma is reasonable, given her background. But her belief is false, and the ultimate explanation of this is the fact that she trusts the unreliable testimony of her community. Our disagreement, therefore, is no evidence that the testimony of my community isn't reliable.” The same, mutatis mutandis, for Maria. So neither Julia nor Maria has evidence sufficient for her to believe that the testimony of her own community is not reliable. And what I have said here about the case of Julia and Maria applies, in general, to cases of deep disagreement (§3.2). In these cases, the parties to the disagreement will reasonably take the other party to be in error about a wide range of questions, and will in many cases, on this basis, reasonably take the other party's sources to be not reliable. In these cases, unlike the case of Paddy and Roddy, the existence of disagreement is unsurprising. In any event, in cases of deep disagreement, the open question condition on entitlement to trust will often not be violated.
All this is consistent with the fact that suspension of judgment about the reliability of a source of information often undermines one's warrant to trust that source. Imagine that your beliefs about Persian history are based primarily on the testimony of a book that you checked out of the local library, when you find out that the library carries two books of Persian history, one of which is reliable and the other of which is unreliable. Intuitively, you should abandon your beliefs about Persian history. This is because your trust in the first book was based on the premise that the book is reliable, which premise you believed based on evidence that, in general and for the most part, history books in libraries are reliable. The discovery of the existence of the second book defeats this evidence, and your belief that the first book is reliable is abandoned, in favor of suspension of judgment. The present case is a case of trust the warrant for which depends on its evidential basis (§2.1); entitlement to trust does not depend on the existence of such an evidential basis.
5.3. Subjectivity
You might object that the entitlement claim (§2.1) makes the notion of reasonable belief “subjective.” One worry might be that the notion is “subjective,” while Feldman's notion is “objective.” This may be, but this charge alone would not threaten my defense of naïve liberalism – for I have sought to make sense of naïve liberalism, by articulating a charitable defense of this common-sense epistemological view. If naïve liberalism is false when interpreted using an “objective” notion of reasonable belief, so much the worse for that interpretation, so long as there is a defensible “subjective” notion of reasonable belief available.
Another worry is that it is implausible to place no a priori restriction on the contents of worldviews that might generate entitlement to trust (§2.2), e.g. to not rule out entitlement to trust astrological prediction on the basis of the “absurdity” or “craziness” of astrology. However, I see no plausible way to place such a restriction, consistent with naïve liberalism (§1). For example, if astrology, on which the positions of the planets and stars makes a difference when it comes to the course of human affairs, is ruled out, then I see no way to avoid ruling out a worldview on which wine can transform into blood and on which someone was once resurrected from the dead. But ruling this out, it seems to me, would be to abandon the spirit of naïve liberalism, which I have assumed in this paper (§1).
A final worry is that it is a consequence of my account that someone could be reasonable in believing that p in the face of disagreement with someone who knows that p is false and who can support her view with very good arguments. But we should accept this consequence: as James Pryor (Reference Pryor2004) argues, a person's background beliefs and other commitments can make it the case that sound arguments, that would otherwise be very good arguments, are “dialectically ineffective,” i.e. unsuccessful, where “[a]n argument succeeds for a given audience to the extent that it presents the audience with a piece of reasoning they can rationally accept” (p. 369). Julia may know that God protected the tilma in 1921, and her arguments may be sound, but they are dialectically ineffective in the context of her disagreement with Maria, who cannot rationally accept their premises.
5.4. Non-epistemic reasonableness
The prudential value of a belief can make it seem ‘reasonable’ for a person to hold the belief even though the belief is not reasonable for another person having the same evidence. For example, a hostage and a neutral reporter on the scene may have the same evidence about the prospects for the hostage's release. The hostage, unlike the reporter, may have a motive for believing that he will be released. We might say that the hostage, but not the reporter, is reasonable in so believing, given this motive. But this is not an epistemic evaluation, and not my concern here. (Feldman Reference Feldman and Hetherington2006: 221)
Critics of appeals to entitlement sometimes argue that the entitlement in question is “prudential,” “pragmatic,” or “practical,” and not “epistemic” (Brueckner Reference Brueckner2007; Jenkins Reference Jenkins2007). But the entitlement enjoyed by Julia doesn't resemble that enjoyed by the hostage in Feldman's example. The hostage, by having a higher-than-objectively-warranted credence that she will be released, aims to secure for herself various goods: psychological comfort while she remains a hostage, which cool attitude may contribute to her eventually being released. These are uncontroversially non-epistemic goods. Entitlement to trust doesn't concern non-epistemic goods, but an uncontroversially epistemic good: true belief. For example, in trusting perception in the absence of evidence that perception is reliable, I aim not to secure psychological comfort or physical security, but true belief. It is in this sense that the reasonableness in question is epistemic.
The notion of the epistemic is obscure and difficult to articulate. It is often characterized (as above, by Feldman) by contrasting it with the prudential or pragmatic. It is also often characterized by saying that epistemic considerations essentially have to do with believing the truth. Richard Foley (Reference Foley1987) writes that:
To describe what it is rational for S to believe from a purely epistemic point of view is to describe what S himself from his own perspective on reflection would believe to be an effective means to his epistemic goal of now believing truths and now not believing falsehoods. (Reference Foley1987: 125)
My conception of reasonable belief is epistemic in the same sense that Foley's conception of rational belief is epistemic. There may be other ways of articulating the notion of the epistemic. The conception of reasonable belief employed in this paper deserves to be called “epistemic” for two reasons: first, it is distinct from prudential or pragmatic reasonableness, and, second, it retains the essential connection to true belief, captured in Foley's articulation of the epistemic.
On my view, it can be epistemically reasonable for a person to trust an unreliable source of information when that source appears reliable to her, and thus it can be epistemically reasonable for a person to form many false beliefs, on the basis of such an appearance. You might argue that the formation of true beliefs is an epistemic good, and that the formation of beliefs that seem true to you, by contrast, is not an epistemic good. Is this claim consistent with the claim that it can be epistemically reasonable for a person to form many false beliefs, on the basis of an appearance of reliability? To see that these two claims are consistent, we must draw a distinction between the good and the right. True belief, let's assume, is the sole epistemic good. My view does not imply that apparently true belief is an epistemic good. Rather, it implies that it is sometimes epistemically permissible to form apparently true beliefs. But this is only ever epistemically permissible when you are trying (in a sense that would require articulation) to acquire an epistemic good, namely, true belief. True belief is an epistemic good; trying to acquire this is what you epistemically ought to do. Believing what seems true to you, therefore, is what you epistemically ought to do; this is compatible with the assumption that apparently true belief is not an epistemic good.
5.5. Long-run epistemic benefits
Somewhat similarly, having an unreasonable faith in your favorite theory now may enable you to believe many important truths later on[.] This shows there can be a kind of long-run epistemic benefit to an unjustified belief. This, too, is not my concern here. (Feldman Reference Feldman and Hetherington2006: 221)
The intuitive appeal of naïve liberalism gives us reason to make sense of the idea that there can be mutually recognized reasonable disagreements. Above (§3) I argued that this idea can be made sense of, if we appeal to a conception of reasonable belief on which apparent “long-run epistemic benefits” of believing are relevant to reasonableness. This motivates the use of this conception of reasonable belief (cf. §4), and speaks against dismissing it in connection with naïve liberalism.
The entitlement claim (§2.1) implies that it is sometimes reasonable for you to believe that p, on the basis of the fact that a particular source of information says that p, absent evidence of the reliability of that source, in virtue of the fact that you are entitled to trust that source, and given that certain other conditions (the open question and anti-vice conditions on reasonable belief) are met. Entitlement is a matter of the “long-run epistemic benefits” of trusting a particular source of information. You might think that the sort of belief licensed by the entitlement claim always unreasonable or unjustified. Why think this?
You might argue that it is always unreasonable to believe something on the basis of the fact that so believing will have “long-run epistemic benefits.” But the entitlement claim is consistent with that, since it only licenses believing (that p) on the basis of the fact that a particular source of information says that p. The entitlement claim says that trust can be warranted in virtue of its “long-run epistemic benefits”; it doesn't say that it is ever reasonable to form beliefs on the basis of their “long-run epistemic benefits.” That in virtue of which someone's belief is reasonable (e.g. the fact that it is based on a source that she is entitled to trust) need not be, and typically isn't, her basis for believing.
Some cases of “long-run epistemic benefits” will violate the anti-vice condition on entitlement to trust, as when the existence of “benefits” is reason to suspect that the relevant source is unreliable, and thus when epistemic virtue requires inquiry into the reliability of said source. Suppose you are approached by a demon who threatens to scramble your brains – and so deprive you of access to a significant set of truths – unless you believe the contents of a book that you have no reason to believe (or disbelieve) is reliable. Intuitively, it would be unreasonable to believe the contents of the book. This is because you are not entitled to trust the testimony of the book. The demon's offer ought to make you suspicious of the reliability of the book; it would be vicious, at this point, not to inquire into its reliability.
This is why it is important that, in cases of deep disagreement akin to that of Julia and Maria (§3.1), there is no reason for suspicion – because disagreement is not evidence against reliability (§5.1). Intellectual virtue requires inquiry into the reliability of a book whose testimony a demon is trying to force you to trust; it does not require (further) inquiry into the reliability of the beliefs and testimony of your community in cases of deep disagreement akin to the case of Julia and Maria.
5.6. Uniqueness
Some epistemologists have argued against the possibility of mutually recognized reasonable disagreement by appeal to the following principle:
Epistemic uniqueness: For any proposition that p, given a body of evidence (relevant to the question of whether p), there is only one reasonable doxastic attitude to take towards the proposition that p (White Reference White2005: 445; Feldman Reference Feldman and Hetherington2006: 231; Reference Feldman and Antony2007: 207; Christensen Reference Christensen2007: 192; cf. Elga Reference Elga2007).
If entitlements to trust are a species of evidence, then the approach to disagreement defended here is compatible with epistemic uniqueness. We might say that Julia and Maria don't have the same evidence, because of their different entitlements to trust. On the other hand, if entitlements to trust aren't a species of evidence, then the approach to disagreement defended here must reject epistemic uniqueness. If this is right, then Julia and Maria have the same evidence, even after they have discussed the source of their disagreement, and their case is a counterexample to epistemic uniqueness.Footnote 25, Footnote 26
The resolution of this question awaits a theory of evidence. There are theories of evidence that entail that actual people rarely or never have the same evidence (and so uniqueness rarely or never actually applies). Feldman (Reference Feldman, Feldman and Conee2004) suggests (though see below) that a person's evidence consists of her token beliefs, perhaps along with some of her “experiences or perceptual states” (225–6), and defends the view that a piece of evidence is “available” to someone iff she is thinking about it (232). Strictly speaking, then, no two people could ever have the same evidence, on the assumption that no two people could ever have the same token beliefs, experiences, or perceptual states. Alternatively, consider Timothy Williamson's (Reference Williamson2000: 184–208) equation of a person's evidence with the propositions that she knows. On this view, two people could have the same evidence, so long as they know the same propositions. The same is true if we equate a person's evidence with the propositions that she believes. Something similar is defended by Feldman (Reference Feldman, Feldman and Conee2004) when he writes that the “total possible evidence a person has at a time includes all and only the information the person has ‘stored in his mind’ at the time” (226). Two people could have the same evidence, on this view, since their two minds could store the same information. On any of these views, Julia and Maria would not have the same evidence, and epistemic uniqueness would never apply in cases of deep disagreement.
This issue is related to the question of whether Julia and Maria are “epistemic peers” – and thus to the question of whether I have defended, in this paper, the possibility of mutually recognized “peer” disagreement. The resolution of this question awaits a definition of “epistemic peers.” We could say that two people are epistemic peers iff they are equals when it comes to epistemic virtue, or when it comes to the epistemic virtues relevant to some question (Gutting Reference Gutting1982: 83; cf. Bergmann Reference Bergmann2009: 336). We could appeal to “general epistemic virtues such as intelligence, thoughtfulness, and freedom from bias” (Kelly Reference Kelly2005: 175), sincerity in seeking the truth (Bergmann Reference Bergmann2009: 336), along with openmindedness, intellectual integrity, intellectual honesty, and so on (cf. Gutting Reference Gutting1982). Julia and Maria are (for all we have said) epistemic peers, in this sense, and might recognize each other as such, and so mutually recognized reasonable peer disagreement, in this sense of “peer,” is possible. Or we might understand mutual recognition of peerhood to require that “neither of us has any particular reason to think that he or she enjoys some advantage over the other … or that he or she is more or less reliable about the relevant domain.” (Kelly Reference Kelly, Feldman and Warfield2010: 112) Along the same lines, we could say that “you count your friend as an epistemic peer with respect to [some claim] if and only if you think that, conditional [on] the two of your disagreeing about the claim, the two of you are equally likely to be mistaken” (Elga Reference Elga2007: 499). Julia and Maria cannot recognize each other as epistemic peers, in this sense. Alternatively, we could say that two people are epistemic peers iff they are equals when it comes to epistemic virtue, etc., and equals when it comes to evidence, i.e. when they have the same evidence, or equally good evidence, relevant to some question (Kelly Reference Kelly2005: 174–5). Whether Julia and Maria are epistemic peers, in this sense, will (as above) depend on our theory of evidence.
On some definitions of “epistemic peer,” therefore, I have not defended the possibility of mutually recognized “peer” disagreement. Does this threaten my defense of naïve liberalism? It does only if naïve liberalism – i.e. whatever view of disagreement contemporary common sense holds – implies the possibility of mutually recognized peer disagreement. But there is no reason to think that it does, given that “epistemic peer” is a term of art invented by philosophers.Footnote 27
6. CONCLUSION
I have explained how mutually recognized reasonable disagreements in religion and politics are possible, by appeal to a relativistic claim about reasonable belief. People who inherit different worldviews will be entitled to trust different sources of information, which will lead to reasonable disagreement, and when the source of such disagreement is recognized, mutually recognized reasonable disagreement will be possible.Footnote 28
