2.1 Pyrrhonism

Sextus was a skeptical doctor whose floruit seems to have been in the early third century CE. His substantial extant writings are our main source for ancient Pyrrhonism. This brand of skepticism is defined by the following attitudes, practices, and problems: wide-ranging suspension of judgment, engagement in open-minded inquiry into truth, emphasis on the existence of widespread and entrenched disagreements, the problem of the criterion of truth and that of the regress of justification, and taking what appears (or the way one is appeared to) as the guide to practical decisions and philosophical investigation.Footnote ³

In the first book of his Pyrrhonian Outlines (henceforth PH, the initials of the Greek title in transliteration), Sextus expounds three sets of “modes” of argumentation by means of which suspension of judgment is supposed to be induced: the Ten Modes (PH I 35–163), the Five Modes (PH I 164–177), and the Two Modes (PH I 178–179). The Five Modes are attributed to Agrippa by Diogenes Laertius (DL IX 88) and are the most lethal weapons of the skeptical arsenal due to their apparent intuitiveness and universal scope. They are disagreement, infinite regress, relativity, hypothesis, and reciprocity. Sextus expounds them as follows:

The modes of hypothesis, reciprocity, and infinite regress constitute what in contemporary philosophy is known as “Agrippa’s trilemma.” Much of recent epistemology is devoted to responding to the trilemma – mainly under the label “the epistemic regress problem” or “the problem of the regress of justification.” Each of those three modes targets a specific justificatory strategy: the mode of hypothesis targets the view that some beliefs are basic or self-justifying; the mode of reciprocity targets the view that a belief is justified provided it is part of a system of mutually supporting beliefs; and the mode of infinite regress targets the view that a belief is justified provided it is supported by an infinite chain of non-repeating reasons. Although Sextus says or implies that each of the three modes can induce suspension separately, the immediately following passage (PH I 170–177) – in which he explains the way the Agrippan modes bring about suspension regarding every object of investigation – makes it clear that suspension can be induced more effectively when they work in tandem (see Reference BarnesBarnes 1990). This is reasonable because, although one can imagine that someone adopting one of the above strategies may suspend judgment after being confronted with the corresponding mode, he will more probably appeal to the other strategies to find an alternative way to justify his beliefs.

With respect to the mode from disagreement, the mere existence of a disagreement does not of course justify suspending judgment, because one may come to the conclusion that one of the conflicting views is to be preferred to the others due to its superior epistemic credentials. This is why, when presenting that mode, Sextus speaks of “undecidable” or “unresolvable” dispute: it is the fact that one has so far been unable to resolve the dispute about whether p that leads one to suspend judgment about whether p.Footnote ⁴ Now, it may be argued that one’s inability to resolve the dispute about whether p is the result of the use of the Agrippan trilemma to show that any attempt to rationally justify one of conflicting views on p fails. If that were the case, the disagreement mode should be regarded as a two-step strategy: (i) presentation of a dispute and (ii) application of the trilemma to show that the dispute cannot be settled. When applied to a disagreement, the modes of hypothesis, reciprocity, and infinite regress work in tandem: when someone attempts to escape from one of those modes in his effort to justify his view on the disputed matter, he falls into one of the other two. The trilemma shows that the conflicting views fare equally badly as far as their justificatory standing is concerned. Since all the disagreeing parties get caught in the trap set by those modes, one must suspend judgment about which of the conflicting views, if any, is correct.

Note that, even if there is no disagreement about whether p, one may ask how the shared belief that p (or that not-p) is justified and then attack, by means of the trilemma, the different strategies purporting to provide the desired justification. If one realizes that one is unable to justify the belief that p, then one is (rationally or psychologically) constrained to suspend judgment about whether p. The interpretation of the mode from disagreement as a complex argumentative strategy discussed in the previous paragraph already presented the Agrippan trilemma as the core of that strategy, with disagreement only providing the material upon which the trilemma works. The point made in this paragraph is stronger inasmuch as it calls attention to the fact that suspension can be induced in the absence of a dispute. This may lead one to view disagreement “as a psychologically useful aid to the sceptic” rather than as “an epistemologically necessary condition for the generation of scepticism” (Reference BarnesBarnes 1990: 116). In the absence of disagreement, one might erroneously think that “there was no room or reason for doubt, that [one was] justified in assenting to the opinion insofar as there was no dissentient voice. Hence the observation of disagreement is pertinent to Pyrrhonism: it draws attention to the fact that assent should not be given without ado – doubt might be raised because doubts have been raised” (Reference BarnesBarnes 1990: 116, italics in the original).

Even though I agree that Agrippa’s trilemma can induce suspension independently of the existence of a dispute, I do not think that disagreement is merely a psychological aid. There are at least three reasons to view disagreement as one of the keys to the generation of skepticism. First, for some epistemologists, one is under no obligation to give reasons for one’s belief that p in the absence of a concrete challenge to its epistemic credentials. The existence of a disagreement about whether p can be taken to constitute such a challenge. For example, if two persons who take themselves to be roughly evidential and cognitive equals with respect to whether p discover that they disagree about whether p, they can take this disagreement as higher-order evidence that they may have made a mistake when assessing the first-order evidence bearing on whether p (more on this in Section 5). The disagreement can then be taken as a challenge for the disputants to provide reasons for their conflicting beliefs about whether p. In such a scenario, one can have recourse to the Agrippan trilemma to show that, in the end, neither disputant can justify his belief about whether p and, hence, that their beliefs are epistemically on a par.

Second, the mode from disagreement can lead to suspension without the application of the trilemma. Faced with the disagreement about whether p, a person may assess the arguments for and against p and find them equally strong, thereby suspending judgment about whether p, without realizing that both the attempt to justify belief in p and the attempt to justify belief in not-p lead to the epistemic regress problem. In several passages in which he does not appeal to the trilemma, Sextus says that the Pyrrhonist is compelled to suspend judgment in the face of an unresolvable disagreement. For example, in concluding his exposition of the Tenth Mode, which deals mainly with moral disagreements, Sextus remarks that, given that we are unable to say what each object is like in its nature but only how it appears in relation to various factors, we must suspend judgment about what things are really like (PH I 145–163). The mode from disagreement should not therefore be deemed to be a complex argumentative strategy that necessarily relies on the Agrippan trilemma – contrary to the interpretation sketched above.

Third, when dealing with certain general epistemological theories that offer solutions to Agrippa’s trilemma, the Pyrrhonist cannot make use of the trilemma to reply to those theories without begging the question, but he can still appeal to the mode from disagreement. Let me explain. Some epistemologists have objected that the trilemma rests on controversial presuppositions: in their view, it is based on a conception of justification that is not universally shared or it distorts our everyday epistemic practices (see, e.g., Reference Williams and Sinnott-ArmstrongWilliams 2004 and Reference Klein and MachucaKlein 2011). Faced with such objections, the Pyrrhonist would remark that there are long-standing and entrenched disagreements between the proponents of the various epistemological theories – including the dispute about what exactly is questionable about the Agrippan trilemma – and would press them to explain how those disagreements are to be resolved. How are we supposed to rationally settle the dispute between, for example, foundationalists, coherentists, and infinitists, given that there seem to be no discernible epistemic differences between the three groups taken as a whole? For the members of the different groups seem to be competent epistemologists who are familiar with the relevant arguments and theories concerning justification, and they are all aware of the pertinent conceptual analyses and distinctions. To make matters worse, within each group the members are far from agreeing, so that we also need to find a reliable touchstone for choosing between the different variants of the same general theories. Likewise, confronted with the claim that the epistemological ideas underlying the Agrippan trilemma distort some aspects of our ordinary epistemic procedures, the Pyrrhonist would argue that we are faced with two conflicting conceptions of justification (philosophical and ordinary), and that this conflict cannot be resolved by assuming arbitrarily that our everyday epistemic practices are to be preferred. Thus, awareness of disagreement is still epistemically significant and may lead us to suspend judgment, even without appeal to Agrippa’s trilemma.

In sum, the mode from disagreement poses a serious challenge to the epistemic credentials of our beliefs independently of Agrippa’s trilemma and should not be considered a mere psychological aid. This is the reason why ancient Pyrrhonists laid so much emphasis on the existence of widespread and persistent ordinary and philosophical disagreements. They also called attention to the fact that, when trying to find a way to resolve a given disagreement, we encounter further disagreements, sometimes high-order ones that concern even trickier questions. Disagreement is a phenomenon that keeps reemerging at different levels.Footnote ⁵

2.2 Academic Skepticism and Medical Empiricism

Plato’s Academy had a skeptical phase, whose main representatives were Arcesilaus (316/5–241/0 BCE) and Carneades (214–129/8 BCE). Academic skeptics are said (i) to have denied that things are apprehensible (i.e., knowable) or that the truth can be found,Footnote ⁶ and (ii) to have suspended judgment about everything or withheld all assent.Footnote ⁷ These two stances are sometimes related in our sources: across-the-board suspension or total withholding of assent stems from the recognition that everything is inapprehensible (Acad. I 45, II 78; Praep. evang. XIV.4.15; C. Acad. II.5.11, 6.14). This may be interpreted in the sense that suspension about all first-order matters follows from the second-order view that the truth about such matters cannot be known: if it is not possible to know what things are like either because we are cognitively limited (Acad. I 44, Praep. evang. XIV.4.15) or because truth itself is hidden (Acad. I 44–45, C. Acad. II.5.12), then one cannot but suspend judgment about what any given thing is like.Footnote ⁸ Another idea underlying the view that one should suspend judgment in the face of inapprehensibility is that assent should be given only to what is known inasmuch as holding mere opinions is rash, vicious, and dangerous (Acad. I 45, II 68; C. Acad. II 6.14). It is also sometimes reported that, for the Academics, suspension about whether p results from the fact that the arguments for and against p are of equal strength (Acad. I 45, Praep. evang. XIV.4.15; cf. DL IV 28, Acad. II 124, 133).

In that connection, it should be noted that the Academics called attention to the skeptical implications of disagreement – specifically, of philosophical disputes – as is attested mainly in Cicero’s Academica II 114–146. After observing that the champion of each philosophical system claims that one must endorse it because only the doctrines in that system are true, and that such a claim is a sign of arrogance (Acad. II 114–115), Cicero reviews disagreements in physics (Acad. II 116–128), ethics (Acad. II 129–141), and logic (Acad. II 142–146). Regarding physics, he remarks that one needs to decide which of the arguments put forward by the physicists to believe, since it is not possible to choose more than one, and he focuses on the “extreme disagreement among the leading men” about “the principles of things out of which everything is constituted” (Acad. II 117; my translation). Cicero’s review of that disagreement makes it clear that he thinks that it is not possible to resolve it. Part of the reason may be that they are all “leading men,” and so that they are epistemic equals or that the epistemic differences between them are negligible (cf. Cicero, De natura deorum [ND] I 14). Now, at one point, Cicero remarks that he assents to neither of the rival views on whether god created everything for our sakes because sometimes one of them seems more persuasive and sometimes the other does (Acad. II 121, cf. II 134). Although one might interpret this remark in the sense that the Academic skeptic successively believes the view that strikes him as more persuasive at each moment, Cicero’s point seems to be that one cannot believe any of the conflicting views because, at the end of the day, none of them takes precedence over the others as far as their persuasiveness is concerned. In this connection, when referring to the disagreements about the nature of the mind and the body, Cicero points out: “Many things are said on either side. One of them seems certain to your sage, but to ours it does not occur which one is most persuasive, so equally weighty are, in most cases, the opposite reasons” (Acad. II 124). And when reviewing ethical disagreements, Cicero asks whether he should pick one of the conflicting views even when the arguments “on either side seem to [him] acute and equally weighty” (Acad. II 133). These passages indicate that, when confronted with disagreements between equally persuasive views, the Academic withholds assent – or is in doubt, as Augustine remarks at Contra Academicos III.7.16 (see also ND I 1, 14).

Other passages indicate that the Academic also draws a negative epistemological conclusion from the existence of disagreements between equally persuasive views, namely, that it is impossible to apprehend the truth about the disputed matters. For example, in the passage immediately following Academica II 121 (referred to above), Cicero observes: “All these things are hidden, Lucullus, concealed and surrounded by thick darkness, so that no human intellect is acute enough to penetrate the sky or enter into the earth” (Acad. II 122). And when reviewing the disagreement among the physicists about the first principles, Cicero says that perhaps the doctrines of one of them are true, for he allows that there is something true, but that he does not accept that they are apprehended (Acad. II 119). The reason why the possibly true doctrines are not apprehended is clearly that the truth cannot be known. The inapprehensibility of truth is then the reason why one cannot choose among the conflicting views: supposing that one of them is true, since one cannot know that it is true, one cannot accept it and reject the others. Note also that Augustine tells us that, in defense of the view that the truth cannot be found, the Academics appeal to, among other things, the disagreements among philosophers (C. Acad. II.5.11; see also III.10.23).

How are we to reconcile the claim that the recognition of disagreement between equally persuasive positions leads one to suspend judgment about whether p with the claim that it leads one to affirm that the truth about p is unknowable? The two claims are not incompatible, for the former is first-order while the latter is second-order. Note that the Academic skeptic may well suspend judgment in the face of a specific disagreement when the conflicting positions strike him as equally strong, and also suspend judgment on account of the second-order general view about the inapprehensibility of the truth about the disagreed upon issues. Both may well be sufficient reasons to suspend judgment. In this connection, at Acad. I 45 Cicero tells us that “by arguing against everyone’s opinions, [Arcesilaus] led most people away from their own opinions, given that, when on the same matter equally weighty reasons were found for the opposite sides, assent was more easily withheld from either side” (emphasis added). This passage may be interpreted as saying that, even though suspension is required by the view of the inapprehensibility of truth, the equal weight of the conflicting views on whether p – which is, after all, nothing but a sign that reveals the inapprehensibility of the truth about p – makes it psychologically easier to suspend judgment.

The link between the observation of unresolvable disagreement and the affirmation of the inapprehensibility of truth is more explicit among adherents of medical Empiricism, which was one of the three main Hellenistic and Imperial medical sects and had a close connection with Pyrrhonism. In On the Sects for Beginners, Galen (129–216 CE) remarks that the Empiricists maintain that the disagreement among the dogmatists about non-evident things is unresolvable, and that inapprehensibility is the cause of unresolvable disagreement and that this kind of disagreement is, in turn, the sign of inapprehensibility (11–12, ed. Helmreich). Similarly, in the preface to his work On Medicine, Celsus (25 BCE–50 CE) points out that, according to the Empiricists, nature is inapprehensible and that it cannot be apprehended is plain from the disagreement among philosophers and doctors about non-evident causes and natural functions – a disagreement that cannot be resolved (Praefatio 27–28).

Why would the existence of unresolvable disagreements be a sign that the disagreed upon matters are inapprehensible? One possible answer consists in interpreting the thesis of inapprehensibility as the conclusion of an inference to the best explanation: given that the disagreement about whether p has existed for a long time and remains undecidable despite all the attempts at resolution by meticulous and intelligent thinkers, the most plausible explanation is that it is impossible for us to know what is really the case regarding the question whether p. This may be understood as referring either to a constitutive cognitive limitation of humans or to the fact that the matter under consideration is of such a nature that it cannot be known.

The view of Academics and medical Empiricists is no doubt skeptical inasmuch as it calls into question our beliefs about p. The Pyrrhonist’s skeptical stance is more cautious inasmuch as he regards the inapprehensibility thesis as one possible explanation of the existence of the long-running unresolvable disagreement about whether p that is not more plausible than, for example, the view that one of the disputants is right while the others are blind to the evidence because of the influence of certain biases, or the view that none of the disputants is right but that the dispute will be settled at some point because the truth about p will be discovered. When contrasting the Pyrrhonian outlook with that of the Academics Carneades and Clitomachus, Sextus remarks that the Pyrrhonist “expects it to be possible for some things to be apprehended” (PH I 226). Thus, the difference between Pyrrhonists, on the one hand, and Academics and medical Empiricists, on the other, concerns the higher-order question of whether knowledge of the truth about disputed matters is possible. For, given that they all regard the disagreement about the first-order question whether p as unresolvable in the sense that one is (at this point) unable to take a stand on that question, they all agree that one is to suspend judgment about whether p. The Pyrrhonist limits himself to saying that one has so far been unable to take a stand on the question whether p; he makes no claims about what might happen in the future. It is his meta-agnosticism about the possibility of knowledge that enables him to engage in continuing open-minded inquiry into truth (PH I 1–3, II 11).

I hope that the brief presentation of ancient skeptical approaches to disagreement will give a rough idea of stances that are not common in present-day discussions of disagreement and that are worth exploring as alternative reactions to that phenomenon. One key difference is that the disagreement-based skepticism found in ancient philosophy is much more wide-ranging than the disagreement-based skepticism found in contemporary philosophy.

5.1 Epistemic Peerhood

As we saw at the end of Section 4, Reference GuttingGutting (1982) understands the notion of epistemic peerhood solely in terms of cognitive parity. He thinks this is enough to raise an epistemic challenge to one’s belief about a disputed matter. Kvanvig understands it in the same way, but he remarks that, since epistemic peerhood only means equality with respect to intellectual virtues but not with respect to the possession of relevant evidence or of reliable methods for acquiring evidence, the mere fact that an epistemic peer disagrees with one does not give one a reason to question one’s belief (Reference KvanvigKvanvig 1983: 53–54 and n. 17).

The fact that such a conception of epistemic peerhood is less threatening to the epistemic credentials of one’s beliefs about controversial matters explains why, in current epistemological discussions of peer disagreement, epistemic peerhood is typically understood in terms of both cognitive and evidential parity. Thus, two individuals are epistemic peers with regard to the question whether p if and only if (i) they are equally intelligent, perspicacious, reasonable, thorough, and unbiased, and (ii) they are equally familiar with the available evidence relevant to the question whether p (evidence including data that directly bears on the question, background knowledge, and arguments for and against the conflicting views on whether p). Evidential parity is reached after full disclosure, that is, after the disputants have completely shared with each other their reasons for their respective beliefs about whether p.

There is, though, another conception of epistemic peerhood to be found in the peer disagreement literature: two individuals are epistemic peers with regard to a given question if and only if they are, prior to the disagreement, equally likely to be right about that question. It could be argued, however, that one’s assessment of such antecedent likelihood depends on one’s assessment of both the disputants’ cognitive abilities and their familiarity with the relevant evidence.

One problem with the conception of epistemic peerhood in terms of perfect cognitive and evidential equality is that it is idealized, and so that it cannot be applied to real-life disagreements. From a more realistic point of view, two individuals are epistemic peers with regard to a given question if and only if (i) they possess similar cognitive abilities, and (ii) they are familiar with the available evidence bearing on that question to roughly the same degree.

5.2 Conciliationism vs Steadfastness

Two main views have been defended in the literature on the epistemic significance of peer disagreement, which are commonly labeled “conciliationist” and “steadfast.” In this section, I offer a rough characterization of them and their most important varieties.

Conciliationism maintains that all the parties to an acknowledged peer dispute are rationally required to significantly revise their beliefs because of the epistemic symmetry between them. Thus, upon learning that a peer disagrees with one about whether p, one cannot rationally continue to believe that p or to hold the belief that p with a high degree of confidence.Footnote ²¹ According to the conciliationist, discovering a peer disagreement makes one wonder whether one has not made a mistake in evaluating the available evidence bearing on whether p. For example, if I am a doctor who diagnose a patient with a given disease based on his symptoms and the results of various tests but then discover that a colleague whom I consider my epistemic equal has reached a different diagnosis after evaluating the same data, I gain a reason to think that I have incorrectly evaluated the evidence – and so does my colleague if he considers me his epistemic equal. The defeater provided by peer disagreement may be rebutting if one thinks that one must conciliate to the point of adopting one’s opponent belief, or undercutting if one thinks that one must conciliate only to the point of suspending judgment about which of the competing beliefs, if any, is true.

The most prominent conciliationist position is what Adam Elga calls the “Equal Weight View”:

This view can be interpreted in two different ways depending on whether one adopts a coarse- or a fine-grained approach to doxastic attitudes. On the former approach, the disputants must suspend judgment about whether p when they learn of the peer disagreement, since there are only three possible doxastic attitudes that may be adopted, namely, belief, disbelief, and suspension. On the latter approach, the disputants must split the difference in the degrees of confidence in their respective beliefs.Footnote ²³

The above version of the EWV refers to a requirement, commonly called “Independence,” which might be formulated thus:

What conciliationists seek to avoid with this principle is any dogmatic or bootstrapping move by means of which one could dismiss out of hand one’s peer’s dissenting opinion simply because it is different from one’s own. Some have rejected Independence on the grounds that, in many cases, one’s peer’s disagreement over a given question shows that he has not responded appropriately to the first-order evidence bearing on that question.Footnote ²⁵ This is so when, for example, my peer disagrees with me about whether 12 × 5 = 60 or about whether there is a person sitting two feet in front of us. In these cases, it is argued, I begin with an extremely high level of justified confidence in the truth of my belief and the reliability of my faculties, and it is therefore absolutely clear that my peer is suffering from some kind cognitive malfunctioning or else being insincere. One may reply, however, that such cases can be accounted for without appeal to the belief about the disputed issue or the reasoning behind it, but to more general considerations. For it could be argued that, since it is highly unlikely that two persons thinking lucidly about the kinds of issues in question hold contrary opinions, the most probable explanation of their disagreement is that one of them is confused, disingenuous, or cognitively deficient (see Reference ChristensenChristensen 2007: 198–201, Reference Christensen2011: 8–12).

Another key thesis endorsed by at least the great majority of conciliationists is the so-called “Uniqueness Thesis”:

This thesis claims that, given E, there is a unique doxastic attitude towards p that it is fully rational to adopt or a unique level of credence in p that it is fully rational to possess.Footnote ²⁷ It is plain why this thesis is endorsed by at least most conciliationists: the reason why, in the face of peer disagreement, one is rationally required to abandon one’s belief is that at most one of the beliefs held by the disputants may be right. If conflicting beliefs about the same matter were supported by the same evidence and the disputants were therefore fully rational in their beliefs, then there would be no need for them to revise their beliefs.

Those who reject UT endorse epistemic permissivism, which might be formulated thus:

According to the permissivist, there are some cases in which the evidence supports conflicting beliefs, so that each of the disputants is rational to hold the belief he or she holds. In those cases, the evidence may be depicted as complex and ambiguous. If so, is it really rationally permissible for different persons to hold conflicting beliefs once they acknowledge the disagreement between them and the complexity and ambiguity of the evidence? Permissivism can be interpreted in terms of there being equally strong evidence in favor of conflicting beliefs about p, or in terms of conflicting beliefs about p being underdetermined by the total available evidence. If so, it seems that the permissivist should recognize malgré lui that the evidential situation is such that suspension is the rationally required response to the disagreement over p. From the external vantage point of someone who has not formed any belief about the matter and is looking for answers, it seems arbitrary to defend any single view: why would he prefer one view to any other when he is aware that they are all (supposedly) equally well supported by the same body of evidence? The only rationally required attitude for him to adopt appears to be suspension. And from the vantage point of the disputants themselves, it seems that, if they become aware that the evidence is complex and ambiguous, they should conclude that they cannot actually rationally stick to their guns, but should instead suspend judgment. Otherwise, they would suffer from a doxastic tension: on the one hand, they would believe that the views of their dissenters are equally well supported by the available evidence, but, on the other, they would believe that their own assessment of the evidence is correct. Their own view might actually be true: if a new piece of evidence were found, we would perhaps realize that the new body of evidence supports only one of the conflicting views, or that it supports one of them more than it supports the others. But in the current evidential situation any choice between the conflicting views seems capricious.Footnote ²⁹ The same line of thought applies if it were argued that, on the basis of the same evidence, conflicting beliefs about p are rational because the disputants have different epistemic standards or prior probability distributions or cognitive goals (Reference Kelly, Steup, Turri and SosaKelly 2014; Reference SchoenfieldSchoenfield 2014). For if the disputants become aware of this fact and cannot affirm that one epistemic standard or probability distribution or cognitive goal is epistemically to be preferred to the others, should they not suspend judgment about whether p? Alternatively, if one is epistemically to be preferred to the others, then, even though some beliefs about p are reasonably held in the sense that they are required by the mistaken epistemic standards or probability distributions or cognitive goals, only one belief about whether p is strictly rational.

Conciliationism is rejected by steadfasters, who maintain that peer disagreement does not always provide a reason for the disputants to revise their beliefs, since in at least some cases it is perfectly reasonable to retain one’s belief with the same degree of confidence.Footnote ³⁰ While some steadfasters maintain that only one of the parties to a peer disagreement can reasonably retain his belief, others maintain that both parties can do so.Footnote ³¹ This difference depends on whether the reasonableness of the belief is understood in terms of which of the competing beliefs is best supported by the shared first-order evidence, or rather in terms of whether from a first-person perspective each of the disputants has legitimate reasons for preferring his own belief to his rival’s.Footnote ³² All steadfasters reject Independence, since they maintain that a person can prefer his own belief to that of his rival by appealing to the fact of the disagreement itself – as the mathematical and perception cases referred to above show.

Some views on the epistemic significance of peer disagreement occupy a middle ground between conciliationism and steadfastness. Such is the case of the Total Evidence View (Reference KellyKelly 2010). In line with steadfastness, that view maintains that, in evaluating how one should react to the discovery of a disagreeing peer, one should take into account not only the higher-order evidence (i.e., each disputant’s opinion) but also the first-order evidence (i.e., the evidence directly pertaining to the disputed issue). In some cases, one can reasonably privilege one’s own belief over that of one’s peer, namely, when one has correctly evaluated the original, first-order evidence. However, in line with conciliationism, the Total Evidence View claims that one should always give at least some small weight to one’s peer belief, and hence that one’s confidence in one’s own belief should always be at least slightly diminished when confronted with a disagreeing peer, even if one has responded appropriately to the first-order evidence.

5.3 First-Person Perspective, Self-Trust, and Personal Information

This section discusses two related views that can be regarded as versions of steadfastness inasmuch as they (a) maintain that one can sometimes rationally retain one’s belief in the face of peer disagreement and (b) deny that one should always give at least some small weight to one’s peer’s belief. The first view appeals to the ineliminability of the first-person perspective and the self-trust that comes with it.Footnote ³³ The second appeals to the asymmetry between the information one possesses about one’s own epistemic situation and the information one possesses about one’s rival’s epistemic situation, and to the high level of justified confidence in the correctness of one’s belief about the disputed issue.Footnote ³⁴

It is plain that the first-person perspective cannot be completely eliminated. When engaged in a disagreement with someone, it is I who judge whether that person is my epistemic peer, superior, or inferior, or it is I who judge that I am (currently) unable to determine what that person’s epistemic standing is in relation to mine. Hence, it is I who determine whether I am rationally required to revise my belief about the disputed matter and, if so, to what extent. Even if I adopt a third-person perspective to analyze the epistemic implications of the disagreement, the first-person perspective cannot be entirely eliminated, since the analysis is ultimately conducted from a first-person vantage point: it is I who determine how the disagreement would look from an allegedly neutral vantage point. In each case, I trust the results of my deliberations, and so there seems to be an inescapable degree of trust in my own opinions and assessments.

Given the ineliminability of the first-person perspective and the self-trust that comes with it, it seems that, when I believe that p, I can legitimately take my opponent’s belief that not-p as evidence that he is mistaken and, hence, that he is less reliable than I am with respect to the topic at hand. From my own first-person vantage point, my reason for doing so is not that he believes not-p whereas I believe p, but rather that he believes not-p whereas p (Reference EnochEnoch 2010: 982). That is, the reason is not that his belief is different from mine, but rather that it is false (as I believe). Thus, at least sometimes I may legitimately take the disagreement itself as a reason to demote my opponent from peerhood (Reference EnochEnoch 2010: 979–981).

Note, first, that the above line of argument begs the question against my peer, and in a problematic way (contra Reference EnochEnoch 2010: 980–981). For the subject of our dispute is precisely whether that p or that not-p. That is, we engage in a dialectical exchange in order to determine which is true: my belief that p or rather his belief that not-p. In such a context, it does not seem legitimate to simply take for granted that p.

Second, an uninvolved observer would remark that, just as I can take that p (as I believe) as evidence against my opponent’s reliability, so too can my opponent take that not-p (as he believes) as evidence against my reliability. Thus, from the vantage point of an uninvolved observer, the disagreement cannot be resolved because there is a dialectical symmetry between the disputants. But even from my own vantage point, once I become aware that my opponent demotes me because that not-p (as he believes), I may wonder whether I am really entitled to demote him because that p (as I believe). It may be argued that the symmetry I am committed to is that between my view and my rival’s, not between p and not-p (insofar as I believe p), and so my reason for demoting my rival (i.e., that p) is not a reason I have for demoting myself (Reference EnochEnoch 2010: 982, 985, 987). But, once again, my opponent can reason in exactly the same way, something of which I am fully aware. The key point is that, from my own first-person perspective, I am aware both of the symmetry resulting from the fact that both my rival and I appeal to self-trust in an attempt to show that one is entitled to stick to one’s own belief in the face of disagreement, and of the fact that I regard my rival’s belief as false despite his relying on self-trust. I thus become aware that self-trust is no guarantee that one’s beliefs about controversial issues are true: if at most one of us may be correct, then self-trust does not prevent at least one of us from getting things wrong. The crucial question then is: what is so special about me that, unlike my rival, I cannot be mistaken in my belief despite relying on self-trust?

There seems to be a dialectical symmetry between my rival and me that has epistemic implications, the kind of symmetry that calls for suspension and that can be appreciated from a first-person point of view. Even though I am unable to eliminate entirely the first-person perspective, I am still able, from my own vantage point, to call into question self-trust as a reliable source of knowledge or justification. Ineliminability entails neither infallibility nor reliability: the fact that we cannot but use our cognitive faculties when inquiring does not entail that these faculties are either infallible or reliable, or that they are less fallible or more reliable than the cognitive faculties of our opponents. Hence, even if the first-person perspective is ineliminable, this ineliminability does not prevent one from subjecting one’s own beliefs about controversial matters to rigorous criticism, and so one can refrain from endorsing the verdicts of one’s cognitive capacities on those matters.

The second view to be considered maintains that the abundant and accurate information one has about one’s own mental states and the normal functioning of one’s own cognitive capacities enables one to avoid engaging in doxastic revision in many cases of real-life disagreement. Such information functions as a symmetry breaker because one lacks the same kind of information about the epistemic situation of one’s opponent. One can therefore legitimately hold one’s ground and dismiss one’s opponent’s opinion inasmuch as one has a high degree of justified confidence in the veridicality of one’s own phenomenology and the reliability of one’s own cognitive capacities.

Let us consider three examples given in support of that view. First, suppose that I ask my boss if I can leave early from work because I have a headache. In this case, I know full well that I have a headache – in this case, I cannot be mistaken about my ongoing conscious experience – even if my boss claims that I am faking it. In a case like this, the principle of Independence cannot properly be applied because it is legitimate to downgrade one’s opponent’s ability to correctly assess the evidence by appealing to the substance of the disagreement (Reference Sosa, Haddock, Millar and PritchardSosa 2010: 286). Second, suppose that I affirm but my roommate denies that a mutual friend is eating lunch with us at the dining room table in our apartment. In such a situation, not only does it clearly seem to me that my friend is sitting at the table, but I also know that I have never in my life hallucinated an object, that I have not been drinking or taking drugs, that I have my contact lenses in, and that my eyesight functions reliably when my short-nearsightedness is corrected (Reference Lackey, Haddock, Millar and PritchardLackey 2010: 306–307). Third, suppose that I disagree with a long-time neighbor about the location of a restaurant the two of us frequent. Although prior to the disagreement neither of us had any reason to suspect that the other’s memory was in any way deficient and we viewed each other as epistemic peers regarding knowledge of the city, I can retain my belief about the restaurant’s location with the same high degree of justified confidence I had before the disagreement occurred. The reason is that I have lived in the city for many years, I know the city extremely well, I have eaten at the restaurant many times, I have not been drinking or taking drugs, and I have ample evidence that my memory is functioning reliably (Reference Lackey, Haddock, Millar and PritchardLackey 2010: 308–309).

In disagreements like the three just described, I have introspective access to my current phenomenologically vivid experience or I have knowledge about the past and present normal functioning of my cognitive faculties. As a result, even if prior to the disagreement I had good reason to regard my disputants as epistemic peers, it is rational for me to stick to my beliefs. Also, given the extraordinarily high degree of justified confidence with which I hold those beliefs, the fact that my disputants disagree with me is best taken as evidence that they do not have access to relevant information or that something has gone wrong with them – they may be confused because they have been drinking, or they may be hallucinating because they have taken drugs, or they may be suffering from some kind of memory loss.

In reply, notice, to begin with, that it is far from clear that we are entitled to regard ourselves as reliable judges of our own stream of conscious experience, even in circumstances we consider normal or regarding mental states we deem transparent. Eric Reference SchwitzgebelSchwitzgebel (2011) offers persuasive and empirically informed arguments to the effect that we often err or are confused about our ongoing conscious experience. For example, we do not know whether we dream in color or black-and-white; we are not accurate judges of our visual experience of depth, our eyes-closed visual experience, our visual imagery, or our auditory experience of echolocation; we are in the dark regarding whether we see things double or single and whether consciousness is abundant or sparse; we are prone to go wrong in judging our ongoing emotional phenomenology; and we are at a loss as to whether there is a distinctive phenomenology of thinking. Now, if often enough we do not have reliable introspective access to our ongoing conscious experience, then it seems that we cannot appeal to the phenomenal given or to our current phenomenology as solid rock upon which we can construct a case for remaining steadfast in the face of certain kinds of disagreement.

Second, in the last two disagreements described above, both disputants can offer the same reasons for privileging their own beliefs over those of their rivals. From the vantage point of an external observer who cannot form an opinion on the disputed matters unless the parties reach consensus, neither disagreement can be resolved because both disputants claim to be highly and justifiably confident that their cognitive capacities are reliable and that their performance is accurate. But even from a first-person perspective, it seems that I should suspend judgment. For if I dismiss my rival’s opinion because I am not equally confident in the reliability of his cognitive capacities and the accuracy of his performance as I am in the reliability of my own capacities and the accuracy of my own performance, I then learn that he dismisses my opinion for the same reason, and I believe that at most one of us can be right, what is so special about me that I can simply rule out the possibility that it is my confidence that is unjustified? If my rival can be wrong despite his high level of confidence and his reliance on his own personal information, why can I not be wrong despite my high level of confidence and my reliance on my own personal information? If I can entertain the possibility that my disputant is hallucinating or suffering from some kind of memory loss, why can I not entertain the same possibility about myself? Both our own experience and the psychological literature teach us that people suffering from delusion or mnemonic confabulation are unaware of it. This is precisely what happens in the perception and the restaurant cases, since my rival is highly confident in his belief about the absence of our friend at the dining room table or in his belief about the location of the restaurant, even though I think that there is definitely something wrong with him. If I can entertain the possibility that something has gone awry with my opponent without his being aware of it, why can I not entertain the same possibility about myself? For example, how can I know that I have never hallucinated? Have I checked every auditory and visual perception I have ever had against someone else’s perception or against a sound or visual recording? If I am one of the many religious believers who have had visions or auditions, should I regard them as veridical or hallucinatory?

In the cases under consideration, there seems to be a dialectical-cum-epistemic symmetry between the disputants: learning that the strategy one’s rival follows for resolving the disagreement from his own first-person perspective is the very same strategy one follows for resolving the disagreement from one’s own first-person perspective should make one wonder whether the strategy in question is as reliable as one thinks it is. One finds further reasons to call into question such a strategy if, in one’s analysis of the disagreement from a first-person vantage point, one incorporates as well experimental evidence to the effect that we do not actually know much about the causes of our beliefs, choices, and behavior – an issue that will be tackled in Section 5.4. It thus seems that, in examining a controversial issue, to one’s first-order evidence about it, one should add the higher-order evidence consisting in the fact of the disagreement, the dialectical-cum-epistemic symmetry between the disputants, and the empirical findings provided by experimental research.

Before moving on, let me say something about Lackey’s notion of high degree of justified confidence in one’s own belief. The first thing to point out is that both parties may have a high degree of confidence in their respective beliefs, and so one must find a reliable way to determine which party is in fact justified in having such a degree of confidence. Here Lackey goes externalist. Regarding the perception case, she invites us to suppose that my roommate denies the presence of our mutual friend at the table because he is hallucinating, and that his hallucination is caused by the fact that, unbeknownst to him, he was drugged by someone. My roommate cannot realize that he is hallucinating because the drug produces no discernible signs. Lackey then argues that although she is happy to grant that, from a purely subjective point of view, my roommate is as reasonable in his belief as I am in mine, our beliefs are not equally justified inasmuch as they are produced by processes that are not equally reliable: in my case the belief is the result of a veridical perceptual experience, while in my roommate’s case the belief is the result of a hallucination (Reference Lackey, Haddock, Millar and PritchardLackey 2010: 320).

I find Lackey’s line of argument unsatisfying because, even if from the vantage point of an external observer who is fully informed about the epistemic standing of both disputants it is possible to determine which of the conflicting beliefs was produced by a generally reliable process that is functioning properly, this is much more difficult from the vantage point of the disputants themselves. I may claim that my belief was caused by a properly functioning reliable mechanism, but so too may my opponent. In order to legitimately affirm that he is hallucinating while my belief is based on a veridical perceptual experience, I need to offer, not only to my disputant but also to myself, reasons to claim that I am not the one who has unknowingly taken a drug that produces no noticeable signs.Footnote ³⁵ To have such reasons, I need to have some sort of access to the cognitive processes that generate my belief. Thus, if disagreement poses an epistemic challenge the disputants need to meet in a way they regard as responsible and nonarbitrary, then externalism does not seem to be up to the task. Note also that some externalists concede that one fails to have justification or knowledge if one has a believed undefeated defeater (e.g., Reference GoldmanGoldman 1986: 62–63, 110–112; Reference PlantingaPlantinga 1993: 40–42; Reference BergmannBergmann 2006: 153–177). Acknowledged disagreement, particularly with someone whom, prior to the disagreement, one has no reason to regard as cognitively deficient, yields a defeater. The only way to defeat that defeater is to offer compelling reasons for the belief that one has correctly evaluated the available evidence because one’s cognitive capacities are generally reliable and are functioning properly in the present circumstances. As we will see in the next section, experimental research provides us with additional defeaters inasmuch as it gives us reasons to think that cognitive capacities such as introspection and memory are much less reliable than we take them to be.

5.4 Incomplete and Misleading Information about Oneself

Over the past fifteen years, several authors have examined the philosophical implications of studies in cognitive psychology that reveal the limits of introspection and the influence of epistemically irrelevant factors on our beliefs. But few authors have examined the relevance of such studies to the question of how one should react to the discovery of disagreement (Reference BallantyneBallantyne 2019: chap. 5; Reference KellyKelly 2022: chap. 10; Reference MachucaMachuca 2022: chap. 8). By my lights, their relevance is seen particularly in relation to the strategies for resolving disagreements that were examined in Section 5.3. For those studies seem to force us to recognize, from a first-person perspective, that our self-knowledge is much more limited than we think and, hence, that it is doubtful that we can often rely on personal information to determine which of the disputants (we or our dissenter) is in a better epistemic position vis-à-vis the contested issue. In what follows, I review the results of some experimental studies on cognitive overestimation, confabulation, the bias blind spot, confirmation bias, and noise.Footnote ³⁶

5.5 Unpossessed Information about Others

When epistemic peerhood is understood as consisting in perfect cognitive and evidential parity between the disputants, the conclusions arrived at in the debate on the epistemic significance of peer disagreement cannot be fully carried over to real-world disputes.Footnote ⁵² For it is hard to believe that two flesh-and-blood individuals are both equally familiar with the same relevant evidence – particularly when the body of evidence is vast and complex – and that they both possess the very same cognitive abilities. But even if we concede that there are epistemic peers of that kind, it seems very hard to determine that any two individuals are epistemic peers so that one can legitimately talk about a peer disagreement that is acknowledged to be so by the disputants.

It seems more likely that, even though two individuals differ in their access to the relevant evidence and/or in their cognitive abilities, they are both, on the whole, equally good at evaluating the matter under consideration: one disputant’s superiority in one aspect (e.g., thoroughness or open-mindedness) may compensate for his inferiority in another (e.g., intelligence or limited background knowledge). Note, however, that even if we grant that there exist epistemic peers in this less stringent sense, it still seems quite difficult to determine, in real-life situations, that any two individuals are epistemic equals in that sense because one would have to possess a great deal of accurate information about their epistemic positions.

Given the actual conditions of real-life disputes, which are far from perfectly transparent, it may seem that suspension is the attitude to be adopted in those very few cases in which there is some sort of epistemic peerhood. However, even if we set acknowledged peer disagreement aside and focus instead on real-world disputes with all their complexities, skepticism is not out of the picture, but actually becomes more threatening. Notice that even if it were granted for the sake of argument that accurate self-assessment is possible because we have extensive self-knowledge, one often has no or partial information about one’s opponent’s evidence, the reliability (or lack thereof) of his cognitive abilities, or the functioning of these abilities in the specific circumstance of the disagreement. Such total or partial lack of information poses a serious challenge because it means that we are in the dark about information that must be taken into consideration when deciding what to believe about the disputed matter, and hence that it is no easy task to determine which, if any, of the disputants is in a better epistemic position with regard to that matter (cf. Reference KingKing 2012: 251, 267). One’s total or partial ignorance of one’s opponent’s epistemic situation should make one wonder whether he may not possess relevant evidence that one lacks, whether he may not have higher cognitive abilities, or whether he may not be employing his cognitive abilities to assess the disputed matter better than one is employing one’s own. None of these possibilities can be easily excluded in at least many cases of real-life disagreement. If so, then one should refrain from believing that, when engaged in a disagreement, one can often legitimately downgrade one’s opponent on the basis of one’s personal information despite one’s total or partial ignorance of his epistemic standing. Otherwise, one would be making a judgment on the basis of limited relevant information. In the face of real-life disagreements, one had better remind oneself of paying attention to both available and unpossessed information (cf. Reference BallantyneBallantyne 2019: chap. 7). To appreciate this, try to remember those occasions in which you downgraded an opponent because of your high degree of confidence in how smart, informed, meticulous, and objective you were in your analysis of the disputed matter, just to later realize that you were mistaken and that you should have been more open-minded and intellectually humble: your opponent turned out to be smarter or better informed or more meticulous or less biased.

With regard to one’s lack of (full) access to one’s dissenter’s (total) evidence, the case of religious experience – a topic touched on in Section 4 – is particularly relevant. Consider the claim made by many religious believers to the effect that they have had visions and auditions (through their eyes and ears or in a nonphysical way), or ineffable mystical insights, or a direct awareness of the presence of God. Suppose that you are engaged in a disagreement about the epistemic justification of religious belief with someone who makes such a claim. How should you react? Should you explain it away on naturalistic grounds, arguing that he is merely delusional? Or should you rather entertain the possibility that your dissenter has had a veridical experience of a kind that is (so far at least) entirely foreign to you? It is possible that he is lying and that you have no hint that he is doing so, in which case you would still lack crucial information, namely, information to the effect that his report should be dismissed out of hand. But, as noted in Section 4, one need not question that people in general have religious experiences; the key issue concerns the veridicality of such experiences. If you have never had such an experience (as in my case), then you lack a kind of experience that may be key in determining the justification of religious belief inasmuch as there might be relevant information that can only be obtained by having that experience – full disclosure does not seem feasible in at least most instances of religious experience (cf. Reference PittardPittard 2019: 183). Or if you have had an experience of that kind but have dismissed it as non-veridical, perhaps your dissenter’s specific religious experience is different from your own in a crucial way or perhaps your dissenter is open enough to appreciate the full epistemic significance of religious experience. Imagine that it is Alson or Plantinga or some other Christian philosopher whose positive epistemic credentials you recognize who claims to have had a religious experience he takes to be veridical. One can assume that philosophers of such caliber have carefully considered the possibility that they may be delusional and have come to the conclusion that they are not. Of course, they may well be wrong because they may have fallen prey to certain cognitive or motivational biases despite their general competence and best efforts. The problem is that it is hard to determine whether or not that is the case. Hence, can you rule out the possibility that they have relevant evidence that you lack?

With respect to one’s lack of (full) knowledge of one’s dissenter’s (full range of) cognitive capacities and their reliability, another religious example may be illuminating. How should you react if your rival in the disagreement about the epistemic justification of religious belief claims that he forms his theistic beliefs by means of a special cognitive faculty – such as the sensus divinitatis – that God have implanted in us? Should you affirm that he is merely delusional in believing that such a capacity exists, that he possesses it, and that it is reliable? Or should you rather entertain the possibility that he may have a different cognitive capacity by means of which he comes to a different view on the subject of the disagreement, or that you both have the same capacity but that yours malfunctions? As far as I am concerned, I feel the pressure to take that possibility seriously, and hence to suspend judgment, when it is Reference PlantingaPlantinga (2000) and other seemingly competent philosophers who defend the existence of such a cognitive capacity as the sensus divinitatis.

In sum, even though I myself am not a religious person, and so even though Reference AlstonAlston’s (1991) view of the experiential awareness of God and Reference PlantingaPlantinga’s (2000) view of the sensus divinitatis are entirely foreign to me, the fact that competent philosophers defend the rationality of religious belief puts pressure on me not to rule out the possibility that they have access to relevant evidence that I do not have, or that they possess a religious belief-producing capacity that I lack, or that we all have such a capacity but that in some of us it malfunctions. This is not to say that very smart people (and people much smarter than I) whose views I do not share are shielded from cognitive and motivational biases and other epistemically contaminating factors. (As we saw in Section 5.4, that is not at all the case.) Rather, it is to recognize that the range of my experiences may be significantly limited and that others may have cognitive abilities that I lack. It is also to recognize that I myself am not shielded from various epistemically contaminating factors either. Given that, at present at least, I have no access to information that is relevant to establishing what the epistemic position of believers is with regard to religious matters, I cannot rule out the possibility that their epistemic position is superior to mine. Maybe people who claim to have certain religious experiences are lying, maybe they do have the experiences they report but are delusional, or maybe their experiences are veridical. In many real-life cases at least, it is extremely hard to determine which is the case.

In sum, we are at least quite often in the dark about relevant information concerning the epistemic position of our dissenters. And if we accept – as argued in Section 5.4 – that people know much less about their evidence, cognitive capacities, and performance than they believe, then there is even more information about which we are in the dark. It is therefore extremely difficult to establish with the required precision what my epistemic situation is in relation to that of my rival – and it is equally difficult for him to do so. If I cannot justifiably affirm that I am epistemically superior to my opponent with regard to the disputed matter, and vice versa, then it appears that for both of us suspension is called for. If so, then it is not the case that a broad skepticism threatens only if one restricts oneself to acknowledged peer disagreements. What we do not know both about ourselves and about others poses a serious challenge to the epistemic credentials of our beliefs about controversial matters.Footnote ⁵³

5.6 Skeptical Dogmatism

In present-day epistemology, when awareness of disagreement is taken to lead to a skeptical stance, the skepticism in question is typically agnostic. A different form of disagreement-based skepticism has recently been proposed by Mark Reference WalkerWalker (2023), who defends a view he calls “skeptical dogmatism.” The skeptical dogmatist maintains that, when confronted with many multi-proposition philosophical disagreements (i.e., philosophical disagreements where there are more than two contrary views), the correct doxastic attitude to adopt for the disputants is disbelief (or a degree of confidence below 0.5) rather than suspension inasmuch as all the conflicting views are probably false. For example, if there is a philosophical disagreement between four views that are mutually exclusive, jointly exhaustive, and equally credible, then there is only a 0.25 chance that any one of them is true. Skeptical dogmatism is also to be adopted in the case of a multi-proposition philosophical disagreement in which one or more views are deemed to be more credible than the others but not to the point of being more likely true than not. Thus, in philosophical disagreements such as the two just described, each of the parties should believe that their preferred philosophical view is probably false.

It might be argued that suspension is still called for in the cases considered by Walker. If asked which of the mutually exclusive and jointly exhaustive philosophical views on p is the correct one, the skeptical dogmatist will reply that he does not know or that he cannot decide. Even if, with regard to each individual view, the skeptical dogmatist feels rationally compelled to claim that it is probably false, with regard to the whole set of competing views, he is undecided about which one is correct. Suspension about whether p can be conceived of as a state of indecision reached after having evaluated the competing views on whether p.