Hostname: page-component-76d6cb85b7-lrvh5 Total loading time: 0 Render date: 2026-07-12T12:04:53.315Z Has data issue: false hasContentIssue false
  • English
  • Français

Still No Hope for Conciliationism

Published online by Cambridge University Press:  19 May 2026

Jonathan Dixon Open the ORCID record for Jonathan Dixon [Opens in a new window]
Jonathan Dixon*
Affiliation:
Philosophy, Wake Forest University, USA
*
Email: jdixon6611@gmail.com
Rights & Permissions [Opens in a new window]

Abstract

Conciliatory views of peer disagreement (aka Conciliationism) have repeatedly been challenged on the grounds that they are epistemically self-undermining. Recently, Dixon (2024) argues that this challenge is worse than previously thought: it is (almost certainly) permanent, and this provides strong reason to abandon Conciliationism as (almost certainly) hopeless. In response, Justin (2025) argues that the self-undermining challenge and Dixon’s arguments, at best, only apply to the belief suspension versions of Conciliationism but not to degree-of-belief (aka credal) versions of Conciliationism. Thus, not only can Conciliationists largely avoid the self-undermining challenge but, contra Dixon, this challenge also does not provide adequate reason to abandon Conciliationism as hopeless. This paper argues that not only can the self-undermining challenge be reformulated to apply to credal versions of Conciliationism but this reformulated challenge also extends and further supports Dixon’s pessimism: because there (almost certainly) is or will be peer disagreement on any way of responding to this reformulated challenge, any response (almost certainly) be defeated by the very view it is attempting to defend. Thus, credal Conciliationism is also (almost certainly) permanently self-undermining and there is (almost certainly) still no hope for Conciliationism.

Keywords

Conciliationismself-underminingdisagreementepistemology of disagreementdisagreement in philosophy

Information

Type
Article
Information
Episteme , First View , pp. 1 - 21
Check for updates
Creative Commons
Creative Common License - CCCreative Common License - BYCreative Common License - NCCreative Common License - ND
This is an Open Access article, distributed under the terms of the Creative Commons Attribution-NonCommercial-NoDerivatives licence (https://creativecommons.org/licenses/by-nc-nd/4.0/), which permits non-commercial re-use, distribution, and reproduction in any medium, provided that no alterations are made and the original article is properly cited. The written permission of Cambridge University Press or the rights holder(s) must be obtained prior to any commercial use and/or adaptation of the article.
Copyright
© The Author(s), 2026. Published by Cambridge University Press

1. Introduction

Conciliatory views of peer disagreement (aka Conciliationism) hold that in cases where acknowledged epistemic peers (i.e. those who are, or are approximately, equally well-informed and intellectually capable about p) have a recognized disagreement on p, one should significantly reduce confidence in p or suspend belief in p.Footnote 1 Such views have repeatedly been challenged on the grounds that they are epistemically self-undermining: because there is peer disagreement among philosophers on Conciliationism itself, this view dictates that these philosophical peers should significantly reduce confidence or suspend belief in Conciliationism.Footnote 2

Recently, Dixon (Reference Dixon2024) argues that this challenge is worse than previously thought: it is (almost certainly) permanent, and this provides strong reason to abandon Conciliationism as (almost certainly) hopeless. In response, Justin (Reference Justin2025) argues that the self-undermining challenge and Dixon’s arguments, at best, only apply to the belief suspension versions of Conciliationism but not to degree-of-belief (aka credal) versions of Conciliationism. Thus, not only can Conciliationists largely avoid the self-undermining challenge but, contra Dixon, this challenge also does not provide adequate reason to abandon Conciliationism as hopeless.

This paper argues, contra Justin, not only can the self-undermining challenge be reformulated to apply to credal versions of Conciliationism, this reformula challenge extends and further supports Dixon’s pessimism. In short, I argue that because there (almost certainly) is or will be peer disagreement on any way to respond to this reformulatteded challenge, any response will (almost certainly) be defeated by the very view it is attempting to defend. Thus, using Justin’s (Reference Justin2025) defense of credal Conciliationism as a foil, I establish the wider result that credal Conciliationism is also (almost certainly) permanently self-undermining and there is (almost certainly) still no hope for Conciliationism.

Importantly, I should note that, to my knowledge, no one in the peer disagreement literature has explicitly formulated the self-undermining challenge against credal versions of Conciliationism (cf. Dixon Reference Dixon2024: 3–4). Thus, this paper not only significantly contributes to this literature by precisely articulating this version of the self-undermining challenge but also argues for the substantive conclusion that credal versions of Conciliationism, like belief suspension versions of Conciliationism, also (almost certainly) succumb to the self-undermining challenge. §2 explains Justin’s (Reference Justin2025) defense of credal versions of Conciliationism from Dixon’s (Reference Dixon2024) arguments. §3 shows how the self-undermining challenge can be reformulated to apply to credal versions of Conciliationism. And §4 explains why credal Conciliationism will (almost certainly) fail, by its own lights, to avoid the reformulated self-undermining challenge.

2. Credal conciliationism vs. the self-undermining challenge

Dixon (Reference Dixon2024: 3) articulates the self-undermining challenge toward Conciliationism as:

The Self-Undermining Challenge

  1. 1. If acknowledged peers recognize they disagree on p, then [these] peers should suspend belief on p. (Conciliationism)

  2. 2. Acknowledged peers recognize they disagree on Conciliationism (Empirical premise)

  3. 3. Thus, [these] peers should suspend belief on Conciliationism.

Justin first notes that premise 1 of this argument only targets “belief suspension” versions of Conciliationism “which states that evidence of peer disagreement about p defeats the reason one has for their belief about p” (Justin Reference Justin2025: 17).Footnote 3 Dixon holds that limiting the focus of the self-undermining challenge is justified because he believes

Following (Christensen Reference Christensen, Lackey and Christensen2013, Reference Christensen2021; Frances Reference Frances2010), among others, it is not necessary to discuss more complicated versions of Conciliationism in formulating this challenge because all plausible versions of Conciliationism threaten to be self-undermining. Also, following (Christensen Reference Christensen2007; Elga Reference Elga2007; Feldman Reference Feldman2005, Reference Feldman and Hetherington2006), among others, this (or something very similar) is the most well-motivated version of Conciliationism. As such, if this version cannot adequately respond to the self-undermining challenge, then a fortiori all other plausible versions of Conciliationism are also self-undermining. (Dixon Reference Dixon2024: 3–4)

But as Justin (Reference Justin2025: 17) rightly notes, proponents of credal versions of Conciliationism might balk at this and argue that not only are credal versions of Conciliationism better motivated than belief suspension versions, they also have the resources to respond to or avoid the self-undermining challenge.

For starters, Justin argues that belief suspension versions of Conciliationism have a “limited scope” (Justin Reference Justin2025: 17) since they only apply when one peer believes p and the other believes not-p. Specifically, they do not apply in cases where one peer believes p and the other suspends belief on p: e.g. two scientists S1 and S2, who recognize each other as peers regarding treatment T, discover that S1 believes “treatment T has an effect” while S2 thinks there is not enough evidence to decide yet. Belief suspension versions are silent on how to respond to this kind of case. And since this represents a large proportion of possible peer disagreements, this limitation is a severe strike against belief suspension versions of Conciliationism.

As Justin explains, credal versions of Conciliationism do not have this limitation. To illustrate, Justin focuses on one articulation of this kind of Conciliationism that comes from Jehle and Fitelson (Reference Jehle and Fitelson2009):

Straight Averaging: disagreeing peers should resolve their disagreement by adopting an average of their prior credences as their new credences. (Justin Reference Justin2025: 16)

For example, in the case where S1 has a credence of .9 that ‘treatment T has an effect’ (i.e. they believe this proposition) and learns their acknowledged peer S2 has a credence of .5 (i.e. they suspend belief) in this proposition, Straight Averaging says that both scientists should adopt a credence of .7 in this proposition. Therefore, unlike belief suspension versions of Conciliationism, credal versions of Conciliationism can easily accommodate peer disagreements where one peer believes p and the other suspends belief on p. This provides strong reason to believe that credal versions of Conciliationism are better motivated than suspend belief versions, contra Dixon.

But more importantly, proponents of credal Conciliationism would also contest Dixon’s (Reference Dixon2024: 4) claim that “all plausible versions of Conciliationism threaten to be self-undermining.” As Justin explains, these proponents would deny premise 1 of the self-undermining challenge (i.e. if acknowledged peers recognize they disagree on p, then peers should suspend belief on p) because for them acknowledged peer disagreement does not always entail that one should suspend belief. Rather, what one should believe in cases of acknowledged peer disagreement is largely determined by three variables:

Variables of Credal Conciliationism

  1. A. The original credences of the acknowledged peers (i.e. their credences before they recognize their peer disagreement),

  2. B. The degree to which a credal version of Conciliationism says to change one’s credences after a recognized peer disagreement, and

  3. C. What the credal thresholds for rationally permissible belief and rationally required suspension of belief are.Footnote 4

To illustrate, let’s assume Straight Averaging (for B) and adopt the widely held convention (for C) that it is permissible to believe p only when one’s credence in p is greater than .5 and having a credence equal to or below .5 requires one to suspend belief on p (e.g. see Buchak (Reference Buchak2021: 211), Fleisher (Reference Fleisher2021: 9924), Jackson (Reference Jackson, Goldberg and Walkerforthcoming)). With these assumptions, in the case where (for A) acknowledged philosophical peers P1 and P2 recognize that P1 has a .9 credence, while P2 has a .5 credence, in Straight Averaging, then this view says that both epistemic peers should adopt a credence level of .7 in Straight Averaging itself. But since this is well above the threshold (.5) for belief, this peer disagreement does not require suspending belief on this version of credal Conciliationism. Indeed, contra premise 1, it is rationally permissible for P1 and P2 to believe in this version of credal Conciliationism despite their peer disagreement. Thus, contra Dixon, because plausible versions of credal Conciliationism can avoid the self-undermining challenge, not “all plausible versions of Conciliationism threaten to be self-undermining” (Reference Dixon2024: 4).

However, as Justin notes, Dixon has a ready reply to this way of responding to the self-undermining challenge. Essentially, Dixon’s (Reference Dixon2024) central argument is that any way of responding to the self-undermining challenge will (almost certainly) fail, by Conciliationism’s own lights, because there is (almost certainly) peer disagreement on the reasoning or auxiliary premises of any way of responding to the self-undermining challenge. From this additional peer disagreement it follows, from (the spirit of) Conciliationism, that this way of responding to the self-undermining challenge is defeated by the very view it is attempting to defend. Dixon articulates this additional argument, which extends the original self-undermining challenge to all responsesFootnote 5 to the self-undermining challenge, as follows:

Solution Responses (almost certainly) Fail

  1. I. If argument A adequately defends [Conciliationism] from the self-undermining challenge, then epistemic peers do not disagree with the reasoning or auxiliary premises of A.

  2. II. Epistemic peers (almost certainly) disagree with the reasoning or auxiliary premises of A.

  3. III. Thus, A (almost certainly) fails to adequately defend [Conciliationism] from the self-undermining challenge. (Dixon Reference Dixon2024: 8–9).

In the present case, because epistemic peers (almost certainly) disagree about the plausibility of credal versions of Conciliationism (e.g. Straight Averaging) and/or whether they can avoid the self-undermining challenge, adopting a credal version of Conciliationism will, by Conciliationism’s own reasoning, (almost certainly) fail to avoid the self-undermining challenge.

In response, Justin argues that this additional argument fails to show that credal Conciliationism will be self-undermining because of peer disagreement on it. Specifically, contra premise I, from the fact that there (almost certainly) is peer disagreement on credal Conciliationism (e.g. Straight Averaging), it doesn’t follow that this view (almost certainly) fails to adequately defend itself from being self-undermining. For example, the case above where P1 has a credence .9 while P2 has a credence of .5 in credal Conciliationism, Straight Averaging says that once P1 and P2 recognize their peer disagreement they should adopt a credence (.7) well above the threshold for belief (.5).

But, as Justin rightly notes, “this [response] only helps in a very specific set of cases” (Justin Reference Justin2025:18), i.e. cases where the post-disagreement credence just so happens to be above .5. But there are a great many other cases where this won’t be so, e.g. P1 and P2 have respective credences of .7 and .2, Straight Averaging dictates that when they recognize their peer disagreement they should adopt a credence (.45) that is not above the threshold for belief (.5). Thus, it still seems that Dixon’s additional argument applies in many (and possibly most) cases of peer disagreement where one peer believes p and the other believes not-p.

To further contest premise I, Justin argues that credal Conciliationism has the resources to extend the range of cases where it can avoid the self-undermining challenge. For example, Justin considers a variation on Straight Averaging called

Straight Averaging*: our new post-disagreement credences should be calculated as a weighted average of peer’s original credences (Justin Reference Justin2025: 18, emphasis added).

As a credal version of Conciliationism, Straight Averaging* also applies to cases where one peer believes p and another suspends belief on p. As such, like Straight Averaging, it also resists Dixon’s underlying assumption behind premise 1 of the self-undermining challenge: that suspend belief Conciliationism is the most well-motivated version of Conciliationism. Moreover, Straight Averaging* also allows for the following additional cases where, contra premise I, it avoids being self-undermining even when peers disagree about it:

Let’s say that most of the disagreeing peers are quite positive about the [Straight Averaging*] view (although with a range of different credences). At the same time, there are only a few strong dissenters. In such a case, it is not unimaginable that after conciliating, we end up with a relatively high credence in the view. Perhaps even a credence that would warrant the belief that Straight Averaging* is the right view to have… (Justin Reference Justin2025: 18–19)

For simplicity, imagine there are six philosophical peers who have around a .8 credence in Straight Averaging* while two philosophical peers have around a .1 credence in this view. When all these two groups of peers recognize their disagreement, Straight Averaging* dictates that every peer should adopt a credence of around .625 in Straight Averaging* which is well above the (assumed) threshold for belief. In comparison, Straight Averaging dictates that every peer should adopt a credence below this threshold (.45). Thus, this case illustrates that “it is possible, pace premise [I] from above, to imagine that an argument adequately defends [credal Conciliationism], despite the fact that epistemic peers do disagree with relation to this argument” (Justin Reference Justin2025: 19).Footnote 6 In other words, even if Straight Averaging* is not ultimately the correct view of peer disagreement, “credal versions of [Conciliationism] can, in principle, walk the goldilocks line between being strong enough to give intuitive answers in [many cases of peer disagreement] while not being self-undermining” (Justin Reference Justin2025: 19).Footnote 7

In sum, Justin makes three related arguments which attempt to put the burden of proof back on Dixon (and other opponents of Conciliationism) to show that the self-undermining challenge is (almost certainly) permanent. First, credal versions are better motivated than belief suspension versions because they can accommodate a wider class of peer disagreements. Second, credal versions of Conciliationism (like Straight Averaging) can effectively avoid the self-undermining challenge and Dixon’s Solution Responses (almost certainly) Fail argument “in a very specific set of cases” (Justin Reference Justin2025: 18). Lastly, credal versions of Conciliationism (like Straight Averaging*) have the resources to extend the range of cases where they are not self-undermining and avoid (or be more resistant to) Dixon’s Solution Responses (almost certainly) Fail argument. As Justin puts it, contra Dixon, these burden-shifting arguments keep “hope alive, at least for a credential version of [Conciliationism]” (Justin Reference Justin2025: 19).

3. Reformulating the self-undermining challenge

I agree with Justin that credal versions of Conciliationism (like Straight Averaging) can avoid the self-undermining challenge as it is stated above and that there are versions of credal Conciliationism (like Straight Averaging*) that are more resilient to this challenge as it is stated above. However, these points of agreement should only provide cold comfort for proponents of credal Conciliationism. In this section, I show how the self-undermining challenge can be reformulated to apply to credal versions of Conciliationism. In the next section, I argue that Justin’s burden-shifting arguments fail to keep hope alive for credal Conciliationism.

To see how the self-undermining challenge can be reformulated, let’s first represent all credal versions of Conciliationism as having the following form:

Credal Conciliationism: In cases where acknowledged epistemic peers recognize their disagreement on p, these agents are rationally required to significantly change their original credence in p by some degree d – a significant non-zero amount between 0 and 1.Footnote 8

For example, Straight Averaging is a version of credal Conciliationism that dictates post-disagreeing peers should adopt the average of their pre-disagreement credences for d. For the time being let’s also assume, what is called, the Lockean Thesis:

Lockean Thesis: If one’s credence in p falls below some threshold t between 0 and 1, then it is irrational to continue believing p (i.e. when one’s credence falls below t, one is rationally required to suspend belief on p) (Jackson (Reference Jackson2020, Reference Jackson, Goldberg and Walkerforthcoming)).Footnote 9

As mentioned above, t is usually stipulated to be .5. These clarifications allow the self-undermining challenge to be formulated as:

The Credal Self-Undermining Challenge

  1. (i) If one’s credence in p falls below t, then one should suspend belief on p. (Lockean Thesis)

  2. (ii) If acknowledged peers recognize they disagree on p, they should change their original credence in p by d. (Credal Conciliationism)

  3. (iii) Acknowledged peers recognize they disagree on credal Conciliationism. (Empirical premise)

  4. (iv) Thus, these peers should change their original credence in credal Conciliationism by d. (ii, iii)

  5. (v) After changing one’s original credence in credal Conciliationism by d, these peers’ new credences in credal Conciliationism fall below t. (Empirical premise)

  6. (vi) Thus, acknowledged peers should suspend belief on credal Conciliationism. (i, v)

To illustrate, let’s assume Straight Averaging is plugged into premise (ii) and let’s modify the above case of peer philosophical disagreement so that P1 has a .7 credence in Straight Averaging but P2 has a .1 credence in Straight Averaging. Straight Averaging itself dictates that both peers should have a credence of .4 in this version of credal Conciliationism and, given that having a credence less than .5 requires agents to suspend belief, both peers should suspend belief on this version of credal Conciliationism.Footnote 10 Thus, this modified case illustrates that Straight Averaging does not completely avoid the self-undermining challenge since there are cases in which this version of Conciliationism applies to and defeats itself.

The reformulated self-undermining challenge makes it clear that for a proponent of credal Conciliationism to avoid this challenge they need to deny at least one of the premises which this reformulated challenge depends on, i.e. premise (i), (iii), or (v). In Justin’s case, he attempts to avoid this challenge by providing certain versions of credal Conciliationism in premise (ii) (e.g. Straight Averaging(*)) that in principle can “walk the goldilocks line” of allowing premise (iii) to be true of them but make premise (v) false in a sufficiently large number of (non-outlier) cases of peer disagreement. For example, when d in premise (ii) is Straight Averaging*, t in premise (i) is .5, and in the scenario where the original credences are sufficiently close together that they will not fall below t via d when, in premise (iii), acknowledged peers recognize their peer disagreement on Straight Averaging*.Footnote 11

One important difference between The Credal Self-Undermining Challenge and the original version (1–3) should be emphasized. By calling the former the “reformulated” challenge, I don’t mean to suggest that both versions have the same strength. Strictly speaking, The Credal Self-Undermining Challenge is weaker than the original version because the former relies on two empirical premises (i.e. premises (iii) and (v)), while the latter only relies on one (i.e. premise 2). Both versions must establish that there is some peer disagreement on some version of Conciliationism (respectively, premises (iii) and 2). But, as stated above, the reformulated challenge also has to establish, via the Lockean Thesis (premise (i)), that after peer disagreement on credal Conciliationism, these peers’ new credences fall below the threshold t for belief (i.e. premise (v)). In the next section, I show that, despite this weakening, there is sufficient peer disagreement to warrant the two empirical premises of the reformulated challenge and that credal versions of Conciliationism are (almost certainly) defeated by the very view they are attempting to defend.Footnote 12

4. Credal conciliationism is (almost certainly) self-undermining

In this section I argue, contrary to Justin’s three burden-shifting arguments, even if credal versions of Conciliationism are better motivated and (at least somewhat) more resistant to the self-undermining challenge than suspend belief versions, credal Conciliationism (almost certainly) cannot adequately respond to the credal self-undermining challenge. This is because the cases Justin provides (like the §2 disagreement between P1 and P2 on Straight Averaging) represent only possible and unlikely outlier cases. But I argue that given the actual (and likely future) amount of peer disagreement on the ways credal conciliationists might respond to the reformulated challenge, it is overwhelmingly likely that no version of credal Conciliationism will avoid the reformulated challenge. Instead, because all such responses will (almost certainly) be defeated by the very view they are attempting to defend, it is much more reasonable to believe that, by their own lights, credal versions of Conciliationism will (almost certainly) fail to adequately respond to this challenge. As such, merely possible outlier cases cannot provide an adequate response to the credal self-undermining challenge.

To appreciate this last point, notice that there are also unlikely outlier cases where suspend belief Conciliationism can also avoid the self-undermining challenge. For example, if there is no acknowledged peer disagreement on suspend belief Conciliationism, then premise 2 of the original challenge (1–3) is false and suspend belief versions of Conciliationism avoid the self-undermining challenge. But this scenario is of no consequence because, as Dixon (Reference Dixon2024) argues and Justin (Reference Justin2025) does not contest, not only does the peer disagreement literature show that there is sustained peer disagreement on suspend belief Conciliationism, the peer disagreement literature also shows there is (almost certainly) peer disagreement on any way of responding to the self-undermining challenge, per Solution Responses (almost certainly) Fail. As such, this case is possible but an unlikely outlier that does not adequately defend suspend belief versions of Conciliationism from the self-undermining challenge.

Likewise, if the cases Justin provides to illustrate how some versions of credal Conciliationism (e.g. Straight Averaging*) avoid being self-undermining are also unlikely outliers, then they cannot be used to draw the general conclusion that credal versions of Conciliationism can adequately respond to the reformulated challenge. In other words, unlikely outlier cases are insufficient to keep “hope alive” for credal Conciliationism (Justin Reference Justin2025: 19). Instead, adequately responding to the reformulated challenge requires providing good reason to believe, given the current, or likely, distributions of peer disagreements on various aspects of credal conciliationism, that some version(s) of credal Conciliationism currently, or will eventually, avoid being self-undermining in non-outlier cases. Conversely, to demonstrate that credal versions of Conciliationism are (almost certainly) permanently self-undermining requires providing good reason to believe, given the current, or likely, distributions of peer disagreements on various aspects of credal conciliationism, that credal versions of Conciliationism are or are likely to be self-undermining in non-outlier cases, in perpetuity. This is what my arguments in this section aim to accomplish: because there is (almost certainly) peer disagreement on any way of responding to this challenge, by credal conciliationists own lights, it is unreasonable to be optimistic that any version of credal Conciliationism will avoid this challenge, per Dixon’s Solution Responses (almost certainly) Fail argument.Footnote 13 Consequently, this places a heavy burden of proof on those who, like Justin (Reference Justin2025), wish to defend credal versions of Conciliationism from the self-undermining challenge. And this defends Dixon’s (Reference Dixon2024: 3–4) contention, quoted above, that “all plausible versions of Conciliationism threaten to be self-undermining.”Footnote 14

To demonstrate that Justin only provides unlikely outlier cases, in §4.1 I will temporarily assume that premise (ii) is occupied by Straight Averaging*. This assumption is justified because Justin holds this version of credal Conciliationism is more resistant (in a wider range of cases) to being self-undermining than other versions of credal Conciliationism (recall fn. 7). Moreover, by demonstrating this version of credal Conciliationism (almost certainly) cannot avoid the reformulated challenge, this assumption allows me to simultaneously accomplish two related goals: to support and defend the premises of the credal self-undermining challenge (i.e. (i), (iii), and (v)) and show that Justin’s three burden-shifting arguments fail to adequately respond to the reformulated challenge. In §4.2, I will consider other ways of formulating credal Conciliationism that might be thought of as avoiding, or as even more resistant to, the reformulated challenge. And in §4.3, I will consider the prospects of denying or revising either premise (i) (aka the Lockean Thesis) or premise (iii) to block the self-undermining challenge. In all these cases, the result is the same: because there is (almost certainly) sustained peer disagreement on these (and other) ways of responding to the reformulated challenge, they all are (almost certainly) defeated by the very view they are attempting to defend. If these arguments are sound and exhaustive, they show that any credal version of Conciliationism is also (almost certainly) permanently self-undermining and there is (almost certainly) still no hope for Conciliationism.

4.1. Outlier cases

Recall, Justin argues that when Straight Averaging* is plugged into premise (ii) it can in principle “walk the goldilocks line” of allowing premise (iii) to be true of them but make premise (v) false in a sufficiently large number of (non-outlier) cases of peer disagreement. Justin’s (Reference Justin2025: 18–19) example quoted in §2 involves a range of peers who are generally in favor of credal versions of Conciliationism vs. a few peers who are very pessimistic about credal Conciliationism. The issue now is: how likely is it that such favorable conditions to Straight Averaging* (or other versions of credal Conciliationism) obtain?

To see why such conditions are unlikely outliers, notice that Justin’s example makes the simplifying assumption that peer disagree-ers are in only two groups: those generally in favor and those generally against Straight Averaging*. But this is not representative of the peer disagreement literature. In this literature, there is a range of competing views on the rational way to respond to peer disagreement. For starters, Conciliationism is often contrasted with

Steadfastness: the family of views that says, in cases of recognized peer disagreement, it is (or can be) rationally permissible to retain one’s belief p or their original credence in p because (many, most, or all) such cases do not count as (partial or full) defeating evidence against one’s contested beliefs.Footnote 15

As defined, Conciliationism and Steadfastness are situated on the following continuum: in cases of recognized peer disagreement, one end is occupied with the extreme Conciliatory View that one should (almost) always significantly reduce their credence, while the other end is occupied with the extreme Steadfast View that one is (almost) never rationally required to reduce their credence at all. And in between these ends are intermediate views that deny, contra Steadfasters, that peer disagreement (almost) never requires a change in one’s beliefs or credences and deny, contra Conciliationists, that peer disagreement should (almost) always be given significant epistemic weight. Instead, intermediate views hold that one should sometimes reduce their credence by some degree, and other times remain steadfast.

The peer disagreement literature contains philosophical peers who occupy many positions along this spectrum. For example, an extreme Steadfast view called the Right Reasons View (defended by Kelly (Reference Kelly, Gendler and Hawthorne2005), Weatherson (Reference Weatherson, Lackey and Christensen2013), and Titelbaum (Reference Titelbaum, Gendler and Hawthorne2015)) holds that peer disagreement itself does not rationally require peers to revise their beliefs in any way. Rather, peers should revise their beliefs only in accordance with what the non-disagreement reasons and evidence actually support. In a similar vein, Enoch (Reference Enoch2010) and Lasonen-Aarnio (Reference Lasonen-Aarnio2013, Reference Lasonen-Aarnio2014) hold that, given the difficulties for Conciliatory views, there is no general requirement about how to respond to peer disagreement. Instead, how to rationally respond in cases of peer disagreement is context-dependent. In contrast, Kelly (Reference Kelly, Feldman and Warfield2010, Reference Kelly, Christensen and Lackey2013) defends an intermediate view called the Total Evidence View which says it can be permissible to remain steadfast in one’s beliefs in cases where one’s total evidence supports their beliefs. Similarly, Lackey’s (Reference Lackey, Gendler and Hawthorne2010a,b) Justificationist View holds that it can be permissible to remain steadfast in one’s beliefs in cases where one has a high degree of antecedent justification or justified confidence in one’s beliefs. And lastly, there are those who defend Conciliatory views of peer disagreement. For example, Bogardus (Reference Bogardus2009), Christensen (Reference Christensen2007), Elga (Reference Elga2007), Feldman (Reference Feldman2005, Reference Feldman and Hetherington2006, Reference Feldman and Antony2007), and Matheson (Reference Matheson2009) defend the Equal Weight View which, like Straight Averaging(*), says that we should give equal weight to the opinions of our disagreeing peers.

This range of views shows that Justin’s simplifying assumption is false. It is not the case that there are only two groups of peer disagreement: those generally in favor and those generally not in favor of Straight Averaging*. Instead, given the range of views in the peer disagreement literature surveyed above, it is more accurate to say there are at least three general groups: Steadfast, Intermediate, and Conciliatory views. But given all the above peer disagreements, how likely is it that after taking the weighted average of all disagreeing peers in these three groups (per Straight Averaging*) that the resulting credence of these peers will be above the threshold for belief? This seems very unlikely. To illustrate, let’s consider the following scenario where philosophical peers’ credences on Straight Averaging* are as follows: seven Conciliationists with around a .7 credence, three Steadfasters with around a .1 credence, and five Intermediate defenders with around a .4 credence. Even with the favorable assumptions that Conciliationists are the largest group and those in the Intermediate group have a middling credence in Straight Averaging*, in this case Straight Averaging* still dictates that when all these acknowledged peers have a recognized disagreement on Straight Averaging*, every peer should adopt a credence of .48, which is below the assumed threshold for belief.

Moreover, even if we grant Justin’s simplifying assumption is true, since the Intermediate views deny a key aspect of Conciliatory views (i.e. that peer disagreement should (almost) always be given significant epistemic weight), it is seemingly more appropriate to group them with Steadfasters than with Conciliationists. This makes matters worse for Straight Averaging*. For example, imagine that the seven Conciliationists above with their .7 credence in Straight Averaging* now have a peer disagreement on Straight Averaging* with the now eight non-Conciliationists above. If these eight have an average credence of .25 in this view, Straight Averaging* dictates that every peer should have a credence of .46 in this view, which again is below the assumed threshold for belief. Thus, on this grouping of the range of views in the existing peer disagreement literature, Straight Averaging* does not “walk the goldilocks line” of allowing premise (iii) to be true of it while making premise (v) false. Consequently, on this grouping, Justin provides only unlikely outlier cases and his three burden-shifting arguments fail, by Straight Averaging* own lights, to adequately defend this (allegedly resistant) version of credal Conciliationism from the reformulated challenge.

To be clear, the above examples are meant to illustrate cases that are in-line with Justin’s (Reference Justin2025: 18–19) example quoted in §2 but which show when Straight Averaging* succumbs to the reformulated challenge. That said, I do not deny that there are other hypothetical cases where Straight Averaging* can avoid the reformulated challenge – i.e. where the weighted average of the disagreeing peers is above .5, contra premise (v). There are many ways to tweak the numbers to acquire this result.Footnote 16 For this reason, it is important to recall my aim in the rest of this section: to show that it is unreasonable to be optimistic that any version of credal Conciliationism will avoid this challenge because such cases are (almost certainly) unlikely outlier cases. Specifically, the rest of this section provides further empirical support that because there (almost certainly) is or will be sustained and abundant peer disagreement on all the possible ways to respond to the reformulated challenge, any way of resisting the reformulated challenged will (almost certainly) fail, by credal Conciliationism’s own lights, to adequately respond to the reformulated challenge (recall fn. 13).Footnote 17 As such, my arguments below do not crucially rely on hypothetical cases but instead place a heavy burden of proof on those who, like Justin (Reference Justin2025), wish to defend credal versions of Conciliationism from the self-undermining challenge. The rest of this section will demonstrate why this is a burden of proof that I do not believe can be met.

4.1.1. Objections

One might grant that currently there is a range of views in the peer disagreement literature and grant that this currently makes it unlikely that when the weighted average of all the above peer disagreements are considered the resulting credence in Straight Averaging* will be above the threshold for belief. But one might object that, at best, this is a temporary problem. The reformulated (and original) challenge cannot show that Straight Averaging* (or any version of credal Conciliationism) is false, but only that it currently cannot justifiably be believed. So, maybe one day in the (hopefully near) future there will be a favorable consensus on Straight Averaging* (or some credal version of Conciliationism) such that it crosses the threshold for belief (cf. Christensen (Reference Christensen2009: 763), Kornblith (Reference Kornblith and Machuca2013: 274), Littlejohn (Reference Littlejohn and Machuca2013: 178), and Matheson (Reference Matheson2015a:149)). Thus, contra Dixon (Reference Dixon2024) there is still hope for Conciliationism.

This also seems like an unlikely outlier case for two general reasons. First, according to surveys of philosophers conducted by Bourget and Chalmers (Reference Bourget and Chalmers2014, Reference Bourget and Chalmers2023), contemporary philosophers’ opinions on many current and traditional philosophical problems are fairly evenly distributed between incompatible philosophical views. Specifically, Bourget notes that the distribution of views in philosophy is particularly “messy” because, when comparing philosopher’s opinions across philosophical topics, “we don’t see large clusters or patterns emerge despite our best efforts to group similar respondents together.”Footnote 18 Instead, Bourget concludes that “everyone is their own school of thought” because “if you pick a random philosopher, you’re pretty likely to disagree with them on roughly half of the 100 questions of the survey.”Footnote 19 Couple this with the well-known general fact that philosophers rarely reach consensus on, or resolve their disagreements on, substantive philosophical propositions (cf. Kornblith Reference Kornblith, Feldman and Warfield2010, Reference Kornblith and Machuca2013), then it seems very unlikely that there will either be a significant reduction in the number of rival views in the peer disagreement literature or that the credences of future peers in Straight Averaging* will be above the threshold for belief.Footnote 20 More generally, because there is no obvious reason to think that any Conciliatory view will be exempt from these trends, there is no reason to suppose that there will be a consensus on the rational way to respond to peer disagreement, nor that this unlikely future consensus will be in Conciliationism’s favor. (More on this in §4.2).

Second, not only is there likely to be perennial peer disagreement on Straight Averaging* from a range of rival views in the peer disagreement literature but there is also (almost certainly) perennial peer disagreement on the best way to formulate Conciliationism. The most pertinent of these peer disagreements are the “family disputes” from Conciliationists themselves on the best way to formulate Conciliationism. For example, any Conciliationist who holds credal versions of Conciliationism (including Straight Averaging*) are better motivated than full belief versions, must contend with those Conciliationists who disagree (e.g. Christensen Reference Christensen2007; Elga Reference Elga2007; Feldman Reference Feldman2005, Reference Feldman and Hetherington2006).

Such peer disagreements are also compounded by the many unresolved issues regarding how to formalize Conciliatory views, like Straight Averaging*. Frances and Matheson (Reference Frances, Matheson and Zalta2024) provide the following examples regarding how to formulate the Equal Weight View which, as mentioned above, agrees with Straight Averaging(*) that we should give equal weight to the opinions of our disagreeing peers:

Fitelson and Jelhe 2009 argue that there are difficulties in making precise the Equal Weight View along Bayesian lines. In particular, they argue that the most intuitive understandings of the Equal Weight View have untenable consequences. Gardiner Reference Gardiner2014 and Wilson Reference Wilson2010 each raise an objection that Equal Weight View (at least as typically understood) violates the principle of commutativity of evidence…Lasonen-Aarnio (Reference Lasonen-Aarnio2013) sets up a trilemma for the Equal Weight View arguing that either (i) it violates intuitively correct updates, (ii) it places implausible restrictions on priors, or (iii) it is non-substantive.

These, and other,Footnote 21 additional peer disagreements further make it unlikely that there will be a consensus, among Conciliationists, on a particular version of Conciliationism that will also be able to adequately respond, by its own lights, to the self-undermining challenge.

Thus, contra Justin, all the above peer disagreements on Straight Averaging* (almost certainly) show, by this view’s own lights, that it is self-undermining in perpetuity. And again, Justin’s three burden-shifting arguments fail, by Straight Averaging* own lights, to adequately defend this (allegedly resistant) version of credal Conciliationism from the reformulated challenge. To be clear, like Dixon (Reference Dixon2024), my arguments against Straight Averaging* are not meant to decisively refute Straight Averaging*. Instead, they place a heavy burden of proof on Justin to further defend Straight Averaging* from the reformulated challenge. But, given all the above peer disagreements, this is a burden of proof that I do not believe can be met.

4.2. Other versions of credal conciliationism

Recall, to simultaneously respond to Justin’s three burden-shifting arguments, thus far I have been assuming that Straight Averaging* occupies premise (ii) of the reformulated challenge. But a proponent of credal Conciliationism might demur and argue that even if, contra Justin, Straight Averaging(*) succumbs to the self-undermining challenge, it is still possible that there are other versions of credal Conciliationism that will (hopefully soon) gain wide acceptance and walk the goldilocks line to adequately respond to the reformulated challenge. While I, for the reasons provided in the previous section, don’t think this is a plausible reply, my aim in this section is to show that all credal versions of Conciliationism (almost certainly) succumb to this challenge.

At the outset, I should note that this is difficult to accomplish because, as explained above, Conciliatory views exist on a spectrum – along with Intermediate and Steadfast views – of being more or less conciliatory. In particular, credal versions of Conciliationism (as defined above and per variable B) can be more or less conciliatory. This depends on how “significant” the degree d that a version of credal Conciliationism says one should change their original credence in p after recognizing peer disagreement on p. As such, where credal Conciliationism ends, these other views on the spectrum begin can be difficult to determine. The issue now is whether credal Conciliationism can adequately avoid the reformulated challenge by, for example, modifying the value for d so that peer disagreement will not require one to change their beliefs in a sufficiently large number of non-outlier cases of peer disagreement.

To overcome this difficulty, I argue other versions of credal Conciliationism will (almost certainly) fail to set an amount d so that it can walk the goldilocks line of allowing its proponents to 1). avoid the reformulated challenge in a sufficiently large number of non-outlier cases while 2). remaining a version of Conciliationism. In other words, other versions of credal Conciliationism face the following dilemma: either they will (almost certainly) remain self-undermining because the peer disagreements of the last section will still apply to these other versions of Conciliationism, or they will avoid the reformulated challenge in sufficiently large number of (now) non-outlier cases at the expense of no longer being a Conciliatory view.

To illustrate, let’s say P1 has a .6 credence in credal Conciliationism while P2 has a .3 credence in credal Conciliationism. Let’s also say this peer disagreement is over the (alleged) credal version of Conciliationism that only requires a .05 change toward their peer’s original credence. After recognizing their disagreement, the .05 change view dictates that P1 change their credence to .55 while P2 change their credence to .35 – in the .05 change view. And if t is still .5, then both peers can rationally retain their original stances: P1 can still rationally believe while P2 can rationally suspend belief on this version of credal Conciliationism. Obviously, this .05 change view allows agents to rationally retain their original stances in a larger number of cases than, for example, Straight Averaging(*). Thus, this .05 change view illustrates how a proponent of credal Conciliationism can potentially avoid the reformulated challenge and, contra Dixon (Reference Dixon2024), still (allegedly) allow hope for credal Conciliationism.

In response, while a proponent of Conciliationism can modify credal Conciliationism in a way that peer disagreement will almost never require one to change their beliefs, doing so is clearly not within the spirit of Conciliationism. For example, while this .05 change view largely avoids the credal self-undermining challenge, it does so because it is clearly a version of Steadfastness. Moreover, the prospects of avoiding the reformulated challenge by adopting other values of d that will walk the goldilocks line, while being within the spirit of Conciliationism, are dim at best. For example, let’s consider the case where P1 has a .9 credence in credal Conciliationism while P2 has a .1 credence in credal Conciliationism. Let’s also say this peer disagreement is now over the (alleged) credal version of Conciliationism that only requires a .2 change toward their peer’s original credence. After recognizing their disagreement, the .2 change view dictates that P1 change their credence to .7 while P2 change their credence to .3 – in the .2 change view itself. And like the above case, if t is still .5, then both peers can rationally retain their original stances: P1 can still rationally believe, while P2 can rationally suspend belief, on this version of Conciliationism. But while the .2 change view, like the .05 change view, largely avoids the reformulated challenge, it does so because it is either a version of Steadfastness or an Intermediate view.

Ultimately, the reason that changing value d, to allow some (alleged) version of Conciliationism to avoid the reformulated challenge in a sufficiently large number of non-outlier cases, will (almost certainly) fail is because of the general motivations for Conciliationism. Conciliationism is generally motivated by the thought that rationality requires agents to give significant, e.g. equal, weight to the opinions of their acknowledged epistemic peers because epistemic peers’ evidential situations regarding p are (roughly) symmetric. This is what makes these peers (about as) equally likely (to be in a position) to know p. For these reasons, Conciliationism normally holds that peer disagreement counts as defeating evidence that is strong enough to defeat one’s belief in p. This is why Straight Averaging(*) and the Equal Weight View are quintessential Conciliatory views. All this suggests that, for a view to be within the spirit of Conciliationism, any version of credal Conciliationism must set d to a value that roughly matches the value these quintessential versions of Conciliationism set d to, i.e. as giving significant or equal, rather than middling or low, weight to disagreements from one’s epistemic peers. But, as explained in the last section, these quintessential versions of Conciliationism (almost certainly) fail, by their own lights, to adequately respond to the reformulated challenge because there is (almost certainly) abundant and perennial peer disagreement on them.

Thus, other versions of credal Conciliationism will (almost certainly) be unable to adequately avoid the credal self-undermining challenge. To avoid this challenge, they either modify value d in ways that (almost certainly) abandon Conciliationism for a rival view or set value d in ways that align with quintessential versions of Conciliationism but are then subject to all the same peer disagreements mentioned in the last section. This further supports Dixon’s (Reference Dixon2024: 3–4) contention, quoted above, that “all plausible versions of Conciliationism threaten to be self-undermining.”

4.3. Last resorts

Recall, for a proponent of credal Conciliationism to avoid the credal self-undermining challenge, they need to deny at least one of the premises that this challenge depends on, i.e. premise (i), (iii), or (v). Thus far, I have argued, contra Justin, there is no way to plug in a version of credal Conciliationism into premise (ii) such that it allows premise (iii) to be true of it while premise (v) is false for enough non-outlier cases of peer disagreement. There are only two remaining ways to respond to the reformulated challenge. The first is to deny or neutralize premise (i), aka the Lockean Thesis. The second is to deny or neutralize premise (iii), i.e. the empirical claim that there is peer disagreement on credal versions of Conciliationism. The aim of this section is to show that both options (almost certainly) fail for the same reason as the other responses to the reformulated challenge fail: because there is (almost certainly) peer disagreement on these ways of responding to the reformulated challenge, these responses are (almost certainly) defeated by the very view they are attempting to defend, per Solution Responses (almost certainly) Fail.

Starting with premise (i), one might either outright reject this premise or revise this premise by setting the threshold t in (i) such that premise (v) is (or is often) false. Regarding the former, one might deny premise (i) outright by accepting belief-eliminativism, i.e. the view that we only have credences, not beliefs. This view is disagreed upon by many, in the literature on the relationship between beliefs and credences, who accept the Lockean Thesis (see Jackson Reference Jackson2020). This is in addition to those who also oppose belief-eliminativism by accepting credal-eliminativism, i.e. the view that we only have beliefs, not credences (e.g. Holton Reference Holton2008 and Horgan Reference Horgan2017). Alternatively, one might deny premise (i) outright because it, along with conjunction closure and belief-consistency constraints, leads to the preface paradox. This is also disagreed upon by those who defend the Lockean Thesis. For instance, to solve the preface paradox, some argue that the Lockean Thesis is true but deny either conjunction closure or belief-consistency constraints (e.g. Easwaran Reference Easwaran2016; Foley Reference Foley, Huber and Schmidt-Petri2009; Sturgeon Reference Sturgeon2008). Thus, denying premise (i) outright (almost certainly) cannot, by credal Conciliationism’s own lights, provide an adequate response to the credal self-undermining challenge.

Regarding the latter, a proponent of credal Conciliationism might accept the Lockean thesis in premise (i) but attempt to neutralize this premise by setting the threshold t such that premise (v) is (or is often) false. For example, Buchak (Reference Buchak2021) and Jackson (Reference Jackson, Goldberg and Walkerforthcoming)) accept a version of the Lockean Thesis that allows belief in p to be rational even if one’s credences are middling or low, i.e. they set the threshold t to be well below .5. This low threshold value can potentially allow for credal conciliationism to avoid the reformulated challenge in a significant number of cases of peer disagreement. For example, if philosophical peers P1 and P2 have respective credence of .7 and .1 in Straight Averaging*, upon recognizing their peer disagreement Straight Averaging* dictates that they both adopt a credence of .4 in Straight Averaging* – which is now likely above the threshold Buchak and Jackson set for permissible belief. However, this low threshold is disagreed upon by most epistemologists who only accept the Lockean Thesis when t is set at, or somewhere near, .5 (e.g. this is acknowledged by Buchak (Reference Buchak2021: 211) and Jackson (Reference Jackson, Goldberg and Walkerforthcoming), also see Fleisher (Reference Fleisher2021: 9924)). This is in addition to the disagreement with those who suggest or hold that threshold t should be set at a value significantly above .5 (e.g. Pettigrew Reference Pettigrew2015; Worsnip Reference Worsnip2015). Thus, attempts to neutralize premise (i) by lowering the threshold t also (almost certainly) cannot, by credal Conciliationism’s own lights, provide an adequate response to the reformulated challenge.

The last way for a proponent of credal Conciliationism to respond to the reformulated challenge is to deny or neutralize premise (iii), i.e. the empirical claim that there is peer disagreement on credal versions of Conciliationism. To my mind, given all the philosophical peer disagreements explained in the preceding sections (esp. from Bourget and Chalmers (Reference Bourget and Chalmers2014, Reference Bourget and Chalmers2023)), this is the least plausible way of responding to the reformulated challenge. However, some (like Cary and Matheson Reference Cary, Matheson and Machuca2013; King Reference King2012; Rotondo Reference Rotondo2015) would neutralize premise (iii) on the grounds that, because it is difficult, if not impossible, to determine whether someone is an epistemic peer, in practice there will be few, if any, acknowledged philosophical peers. Thus, one can argue that peer disagreement skepticism is rarely justified, and consequently, it is of very limited epistemological significance.

In response, nearly everyone in the peer disagreement literature (including Conciliationists like Justin (Reference Justin2025)) hold that philosophers can, and often do, accurately acknowledge each other as epistemic peers (cf. Christensen Reference Christensen, Lackey and Christensen2013, Reference Christensen2021; Feldman Reference Feldman and Antony2007; Kornblith Reference Kornblith, Feldman and Warfield2010, Reference Kornblith and Machuca2013; Machery Reference Machery2017: 131). It is generally thought that perfect epistemic peerhood is unlikely or impossible, but this is not required to accurately regard others as epistemic peers. Instead, one can accurately regard another as an epistemic peer on p if one believes they are roughly in the same epistemic position to know p, or lack any good reason to believe either is in a significantly epistemically superior position to know p (cf. Frances and Matheson Reference Frances, Matheson and Zalta2024). But more importantly, denying premise (iii) on these grounds is not a viable way for a Conciliationist to avoid the self-undermining challenge. If peer disagreement skepticism is impossible or rarely justified, then Conciliationism is a view that does not or rarely applies to agents in the world. As such, it is of little consolation to the Conciliationist that they can avoid the self-undermining challenge by accepting that their view is a concept without an extension. Conciliationists are likely to view this cure as worse than the disease.Footnote 22

In sum, because there is ample peer disagreement on denying or neutralizing the Lockean Thesis in premise (i) and the empirical claim in premise (iii), these ways of blocking the credal self-undermining challenge will (almost certainly) fail, by credal Conciliationism’s own lights, to save credal versions of Conciliationism from the credal self-undermining challenge. Moreover, it is worth reiterating that the failure of these ways to undermine premise (i) and (iii) are also supported by the facts, in §4.1., that philosophers’ opinions on current and traditional philosophical problems are fairly evenly distributed across a range of incompatible philosophical views and the well-known general fact that philosophers rarely resolve their disagreements or come to a consensus on substantive philosophical theses. Because there is no obvious reason to think that disagreements on the Lockean Thesis in premise (i) and the empirical claim in premise (iii) will be exempt from these trends, there is no reason to suppose that there ever will be a consensus on these claims, nor that this unlikely consensus will be in credal Conciliationism’s favor.

5. Conclusion

To be clear, my arguments above are not meant to decisively establish that all possible credal versions of Conciliationism succumb to the self-undermining challenge. Instead, my argument places a heavy burden of proof on those who, like Justin (Reference Justin2025), wish to defend credal versions of Conciliationism from the self-undermining challenge. But, given the empirical facts about the pervasiveness and persistence of disagreement among philosophers about substantive philosophical propositions and the fact that the peer disagreement literature itself contains epistemic peers who disagree both about various aspects of credal versions of Conciliationism and the ways one might deny the premises in the credal self-undermining challenge, this is a burden of proof that I do not believe can be met. If these arguments are sound and exhaustive, they show that credal Conciliationism is also (almost certainly) permanently self-undermining. Dixon (Reference Dixon2024) is correct that credal versions of Conciliationism are not better able to avoid being self-undermining than full belief versions of Conciliationism, and there is (almost certainly) still no hope for Conciliationism.Footnote 23

Competing interests

The author(s) declare none.

Declarations, permissions, funding, and AI statement

There are no declarations or permissions to declare, I received no funding for this project, and AI was not used in any way to write this manuscript.

Footnotes

1 Conciliationism is defended by Bogardus (Reference Bogardus2009), Christensen (Reference Christensen2007, Reference Christensen2009, Reference Christensen, Lackey and Christensen2013, Reference Christensen2021), Elga (Reference Elga, Feldman and Warfield2010), Feldman (Reference Feldman2005, Reference Feldman and Hetherington2006, Reference Feldman and Antony2007), Littlejohn (Reference Littlejohn and Machuca2013), Matheson (Reference Matheson2009, Reference Matheson2015a,b), and Vavova (Reference Vavova2014a,b). See Frances and Matheson (Reference Frances, Matheson and Zalta2024) for an overview of the peer disagreement literature.

2 The self-undermining challenge is discussed by Bogardus (Reference Bogardus2009), Christensen (Reference Christensen2009, Reference Christensen, Lackey and Christensen2013, Reference Christensen2021), Decker (Reference Decker2014), Dixon (Reference Dixon2024), Elga (Reference Elga, Feldman and Warfield2010), Fleisher (Reference Fleisher2021), Littlejohn (Reference Littlejohn and Machuca2013, Reference Littlejohn2014 Reference Littlejohn2020), Matheson (Reference Matheson2015a,b), Pittard (Reference Pittard2015), Walker (Reference Walker2024), and Weatherson (Reference Weatherson, Lackey and Christensen2013).

3 Justin makes a more specific claim that this version of the self-undermining challenge only targets a specific interpretation of, what he calls, “first-order import,” which is a necessary feature of Conciliatory views. I set this complication aside since it has no bearing on my arguments below.

4 To be clear, while Justin does not explicitly mention these factors, it is uncontroversial that these factors are a part of credal Conciliationism. Furthermore, it will be helpful for explaining my arguments below to explicitly explicate them.

5 Dixon says that this argument only applies to what he calls “Solution Responses” to the self-undermining challenge. He provides another argument that challenges, what he calls, “Skeptical Responses” to this challenge, and together these arguments are supposed to challenge all responses to the self-undermining challenge. I set this complication aside since it has no bearing on my arguments below.

6 To be clear, this example is not Justin’s but my way of illustrating how Straight Averaging and Straight Averaging* (allegedly) diverge. To accomplish this, this example assumes that Straight Averaging holds that the credences of these two groups of peer disagree-ers (regardless of size) be averaged. But a proponent of Straight Averaging might deny this assumption. Instead, they might plausibly hold that this view only applies in one-on-one peer disagreements; and in cases where there are more than two peer disagree-ers, we should implement a related view: take the weighted average, not of the disagreeing groups but, of the individual credences, i.e. do what Straight Averaging* dictates. (Assuming these individuals formed their credences independently of one another (see Lackey (Reference Lackey, Christensen and Lackey2013)). On this interpretation of Straight Averaging, my example above can also make it clear how this view differs from Straight Averaging*: only the latter applies (but see the next fn.).

7 However, while Justin (Reference Justin2025) treats Straight Averaging and Straight Averaging* as distinct views, it is not clear (to me) how exactly these views differ or are related to another. A proponent of Straight Averaging might plausibly hold that these views are not distinct, but that Straight Averaging is just a version of Straight Averaging* that only concerns one-on-one peer disagreements. I believe the unclarity here is due in large part to some commentators failing to be clear on how versions of Conciliationism, like Straight Averaging, should be extended past one-on-one peer disagreements (e.g. Jehle and Fitelson (Reference Jehle and Fitelson2009) only define Straight Averaging for one-on-one peer disagreements). While Justin’s (Reference Justin2025) definitions of these views also doesn’t make it clear how these views differ from one another, a close textual reading shows that he restricts the use of these views in the following way: Straight Averaging only concerns one-on-one peer disagreement while Straight Averaging* concerns more than two peer disagree-ers. In the rest of the paper, I follow Justin’s (Reference Justin2025) use of these views and assume these views are distinct. Thanks to an anonymous referee for pushing me to address these issues.

8 How to interpret “significant” will be discussed in §4.2 when discussing alternative versions of credal Conciliationism than Straight Averaging(*). Also, “change” one’s credence is needed rather than “decrease” one’s credence. If one with a low credence in p encounters a peer with a high credence in p, the spirit of Conciliationism dictates that each peer should change their credence to be closer to the other, i.e. the first peer should raise, while the second peer should lower, their credence in p.

9 In full, the Lockean Thesis states: “S rationally believes p iff S has a rational credence in p above some (probabilistic) threshold” (Jackson (Reference Jackson, Goldberg and Walkerforthcoming)). Only the left-right direction of this biconditional is relevant for my purposes.

10 In terms of variables A–C above, this case results in suspending belief on Straight Averaging because for (B) t = .5, for (C) d = Straight Averaging, and for (A) the original credence’s are sufficiently far apart that they fall below t via d when acknowledge peers recognize their peer disagreement.

11 In terms of variables A–C, Justin attempts to avoid this challenge by plugging Straight Averaging = d into (C), t = .5 for (B), and for (A) the original credence’s are sufficiently close together that they do not fall below t via d when acknowledge peers recognize their peer disagreement.

12 Thanks to an anonymous reviewer for pushing me to make these points clearer.

13 Indeed, given the contingent nature of the self-undermining challenge (i.e. it is contingent on the current, or likely, distributions of peer disagreements on credal Conciliationism and responses to the reformulated challenge) it is likely impossible to decisively establish (i.e. with certainty) that all possible credal versions of Conciliationism succumb to this challenge. But, because it is unreasonable to require decisively establishing a contingent proposition, one cannot object to my arguments on the grounds that they are not decisive. Hence, it is appropriate that my arguments only aim to establish that, given that the actual and likely distributions of peer disagreements, we have sufficient, but not decisive, reason to believe it is unreasonable to remain hopeful that credal Conciliationism will ever avoid this challenge. Thanks to Luis Oliveira for making me realize the importance of this point.

14 Dixon (Reference Dixon2024: fn. 7, 8, and 41) mentions this way of extending the self-undermining challenge to credal Conciliationism. But his remarks are brief and need to be further explained to be convincing. The aim of this section is to accomplish this.

15 Defined this way, proponents of Steadfastness include Enoch (2010), Kelly (Reference Kelly, Gendler and Hawthorne2005, Reference Kelly, Feldman and Warfield2010, Reference Kelly, Christensen and Lackey2013), Lackey (Reference Lackey, Gendler and Hawthorne2010a,b), Lasonen-Aarnio (Reference Lasonen-Aarnio2013, Reference Lasonen-Aarnio2014), Titelbaum (Reference Titelbaum, Gendler and Hawthorne2015), and Weatherson (Reference Weatherson, Lackey and Christensen2013).

16 For instance, if the seven Conciliationists in the previous example were more optimistic and had a credence of .8 in Straight Averaging*, the weighted average of these fifteen philosophers would be .52 – i.e. above the assumed threshold for belief, contra premise (v). Alternatively, one might contest the above grouping of Intermediate and Steadfast views and argue that it is more appropriate to group Intermediate views with Conciliationism, than with Steadfastness. On this grouping, it is more likely that the resulting credence on Straight Averaging* could be above the threshold for belief, e.g., if twelve peers have a credence of .6, while only three have a credence of .1, in Straight Averaging*, the weighted average of these peers equals a .5 credence in Straight Averaging*, enough to believe Straight Averaging*.

17 To illustrate, the empirical support in the rest of this section shows that, contra the examples in the previous footnote, because there is (almost certainly) sufficient peer disagreement on credal Conciliationism and anyway of grouping Conciliationism vs. non-Conciliationism, it is (almost certain) that such favorable distributions of credences and groupings are unlikely outlier cases. Consequently, such cases (almost certainly) fail, by credal Conciliationism’s own lights, to save such views from the reformulated challenge. Thanks again to an anonymous referee for pushing me to be clearer on these points.

18 See https://survey2020.philpeople.org/survey/results/additional-figures (accessed Feb 13, 2026).

19 See Bourget’s Philpeople.org post on 2/22/23, https://philpeople.org/profiles/david-bourget/news (accessed Feb 13, 2026).

20 This assumes that only actual, and not potential, peer disagreements have defeating power. But this is also a point on which there is peer disagreement. See Barnett and Li (Reference Barnett and Li2016), Cary (Reference Cary2011), Cary and Matheson (Reference Cary, Matheson and Machuca2013), Christensen (Reference Christensen2007), Kelly (Reference Kelly, Gendler and Hawthorne2005), and Kornblith (Reference Kornblith, Feldman and Warfield2010).

21 For instance, views like Straight Averaging* are motivated by the thought that only those peer disagreements that arise from agents who formed their beliefs independently from each other can have epistemic weight. Lackey (Reference Lackey, Christensen and Lackey2013) argues that this is mistaken because the plausible ways of formulating belief-independence are false. Moreover, there are more general peer disagreements in the distinct literature about higher-order evidence that bear directly on all versions of Conciliationism. See Horowitz (Reference Horowitz, Zalta and Nodelman2025) and Whiting (Reference Whiting2021).

22 Alternatively, one might deny premise (iii) on the grounds that alleged peer disagreements on credal Conciliationism, and disagreements in philosophy more generally, are not genuine disagreements. My evidence for various peer disagreements above derives from journal articles published by professional philosophers who defend opposing views. But one might hold that this is a poor indicator of what authors truly believe, rather than what one merely endorses, accepts, etc. Moreover, one might apply this reasoning to discount the Bourget and Chalmer’s (Reference Bourget and Chalmers2014, Reference Bourget and Chalmers2023) surveys of philosophers’ differing views as evidence of widespread “genuine” disagreement amongst philosophers. In response, while I grant that at least some of these (alleged) disagreements between philosophers are not genuine in this sense, but to my knowledge, no one has defended this view in print (see Plakias (Reference Plakias2019) and Fleisher (Reference Fleisher2020)). And, for what it is worth, my anecdotal evidence of interacting with other philosophers makes me believe that, by and large, philosophers do have genuine disagreements. But more importantly, given the diverse views philosophers have (Bourget and Chalmers Reference Bourget and Chalmers2014, Reference Bourget and Chalmers2023), there will (almost certainly) be enough genuine peer disagreement on this way of denying premise (iii) that it will (almost certainly) fail, by credal Conciliationism’s own lights, to adequately respond to the reformulated challenge.

23 Thanks to Sophie Horowitz, Hilary Kornblith, and Christopher Meacham for providing me with written comments on an earlier version of this paper. Thanks to Matt Sheilds, Luis Oliveira, Jonathan Matheson, and an anonymous reviewer at this journal for providing encouraging and helpful written feedback on more recent versions of this paper. Thank you to the philosophy faculty at Wake Forest University for giving me valuable feedback on this paper at a work in progress meeting. And thanks to Elizabeth Hupfer for discussing several different versions of this paper and for making this project possible. All the above made this paper much better than it otherwise would have been.

References

Barnett, Z. and Li, H. (2016). ‘Conciliationism and Merely Possible Disagreement,’ Synthese 193(9), 2973–85.10.1007/s11229-015-0898-7CrossRefGoogle Scholar
Bogardus, T. (2009). ‘A Vindication of the Equal Weight View,’ Episteme 6(3), 324–35.10.3366/E1742360009000744CrossRefGoogle Scholar
Bourget, D. and Chalmers, D. (2014). ‘What Do Philosophers Believe?’ Philosophical Studies 170(3), 465–500.10.1007/s11098-013-0259-7CrossRefGoogle Scholar
Bourget, D. and Chalmers, D. (2023). ‘Philosophers on Philosophy: The 2020 PhilPapers Survey,’ Philosophers’ Imprint 23(1), 1–53.Google Scholar
Buchak, L. (2021). ‘A Faithful Response to Disagreement,’ The Philosophical Review 130(2), 191–226.10.1215/00318108-8809893CrossRefGoogle Scholar
Cary, B. (2011). ‘Possible Disagreements and Defeat,’ Philosophical Studies 155(3), 371–81.10.1007/s11098-010-9581-5CrossRefGoogle Scholar
Cary, B. and Matheson, J. (2013). ‘How Skeptical is the Equal Weight View.’ In Machuca, D. (ed.), Disagreement and Skepticism, pp. 131–149. New York: Routledge.Google Scholar
Christensen, D. (2007). ‘Epistemology of Disagreement: The Good News,’ Philosophical Review 116, 187–218.10.1215/00318108-2006-035CrossRefGoogle Scholar
Christensen, D. (2009). ‘Disagreement as Evidence: The Epistemology of Controversy,’ Philosophy Compass 4(5), 756–67.10.1111/j.1747-9991.2009.00237.xCrossRefGoogle Scholar
Christensen, D. (2013). ‘Epistemic Modesty Defended.’ In Lackey, J. and Christensen, D. (eds), The Epistemology of Disagreement: New Essays, pp. 77–97. Oxford: Oxford University Press.10.1093/acprof:oso/9780199698370.001.0001CrossRefGoogle Scholar
Christensen, D. (2021). ‘Akratic (Epistemic) Modesty,’ Philosophical Studies 178(7), 2191–214.10.1007/s11098-020-01536-6CrossRefGoogle Scholar
Decker, J. (2014). ‘Conciliationism and Self-Incrimination,’ Erkenntnis 79, 1099–134.10.1007/s10670-014-9599-8CrossRefGoogle Scholar
Dixon, J. (2024). ‘No Hope for Conciliationism,’ Synthese 203(148), 1–30.10.1007/s11229-024-04590-3CrossRefGoogle Scholar
Easwaran, K. (2016). ‘Dr. Truthlove or: How I Learned to Stop Worrying and Love Bayesian Probabilities,’ Noûs 50(4), 816–53.10.1111/nous.12099CrossRefGoogle Scholar
Elga, A. (2007). ‘Reflection and Disagreement,’ Noûs 41, 478–502.10.1111/j.1468-0068.2007.00656.xCrossRefGoogle Scholar
Elga, A. (2010). ‘How to Disagree About How to Disagree.’ In Feldman, R. and Warfield, T. (eds), Disagreement, pp.175–186. New York: Oxford University Press.10.1093/acprof:oso/9780199226078.003.0008CrossRefGoogle Scholar
Enoch, D. (2010). “Not Just a Truthometer: Taking Oneself Seriously (but not Too Seriously) in Cases of Peer Disagreement”, Mind 119(476):953–97.10.1093/mind/fzq070CrossRefGoogle Scholar
Feldman, R. (2005). ‘Respecting the Evidence,’ Philosophical Perspectives 19, 95–119.10.1111/j.1520-8583.2005.00055.xCrossRefGoogle Scholar
Feldman, R. (2007). ‘Reasonable Religious Disagreements.’ In Antony, L. (ed), Philosophers without Gods: Meditations on Atheism and the Secular Life, pp. 194–214. New York: Oxford University Press.10.1093/oso/9780195173079.003.0016CrossRefGoogle Scholar
Feldman, R. (2006). ‘Epistemological Puzzles about Disagreement.’ In Hetherington, S. (ed), Epistemic Futures, pp. 216–236. New York: Oxford University Press.Google Scholar
Fleisher, W. (2020). ‘Publishing without (Some) Belief,’ Thought: A Journal of Philosophy 9(4), 237–46.10.1002/tht3.466CrossRefGoogle Scholar
Fleisher, W. (2021). ‘How to Endorse Conciliationism,’ Synthese 198(10), 9913–39.10.1007/s11229-020-02695-zCrossRefGoogle Scholar
Foley, R. (2009). ‘Beliefs, Degrees of Belief, and the Lockean Thesis.’ In Huber, F. and Schmidt-Petri, C. (eds), Degrees of Belief, pp. 37–47. London: Springer.10.1007/978-1-4020-9198-8_2CrossRefGoogle Scholar
Frances, B. (2010). ‘The Reflective Epistemic Renegade,’ Philosophy and Phenomenological Research 81(2), 419–63.10.1111/j.1933-1592.2010.00372.xCrossRefGoogle Scholar
Frances, B. and Matheson, J. (2024). ‘Disagreement.’ In Zalta, Edward N. (ed), The Stanford Encyclopedia of Philosophy, pp. 1–37. (Spring 2024 Edition). https://plato.stanford.edu/archives/win2024/entries/disagreement/ Google Scholar
Gardiner, G. (2014). ‘The Commutativity of Evidence: A Problem for Conciliatory Views of Disagreement,’ Episteme 1(1), 83–95.10.1017/epi.2013.42CrossRefGoogle Scholar
Holton, R. (2008). ‘Partial Belief, Partial Intention,’ Mind 117(465), 27–58.10.1093/mind/fzn002CrossRefGoogle Scholar
Horgan, T. (2017). ‘Troubles for Bayesian Formal Epistemology,’ Res Philosophica 94(2), 233–55.10.11612/resphil.1535CrossRefGoogle Scholar
Horowitz, S. (2025). ‘Higher-Order Evidence.’ In Zalta, Edward N. & Nodelman, Uri (eds), The Stanford Encyclopedia of Philosophy, pp. 1–23. (Spring 2025 Edition).Google Scholar
Jackson, E. (2020). ‘The Relationship between Belief and Credence,’ Philosophy Compass 15(6), 1–13.Google Scholar
Jackson, E. (forthcoming) ‘How Low Can You Go? A Defense of Believing Philosophical Theories.’ In Goldberg, S. and Walker, M. (eds), Attitude in Philosophy. Oxford. Oxford University Press.Google Scholar
Jehle, D. and Fitelson, B. (2009). ‘What is the “Equal Weight View”?’ Episteme 6, 280–93.10.3366/E1742360009000719CrossRefGoogle Scholar
Justin, M. (2025). ‘More Hope for Conciliationism,’ Episteme 1–21. https://doi.org/10.1017/epi.2025.22 CrossRefGoogle Scholar
Kelly, T. (2005). ‘The Epistemic Significance of Disagreement.’ In Gendler, T. and Hawthorne, J. (eds), Oxford Studies in Epistemology, vol. 1, pp. 167–196. Oxford: Oxford University Press.10.1093/oso/9780199285891.003.0007CrossRefGoogle Scholar
Kelly, T. (2010). ‘Peer Disagreement and Higher Order Evidence.’ In Feldman, R. and Warfield, T. (eds), Disagreement, pp. 111–174. New York: Oxford University Press.10.1093/acprof:oso/9780199226078.003.0007CrossRefGoogle Scholar
Kelly, T. (2013). ‘Disagreement and the Burdens of Judgment.’ In Christensen, D. and Lackey, J. (eds), The Epistemology of Disagreement: New Essays, pp. 31–53. Oxford: Oxford University Press.10.1093/acprof:oso/9780199698370.003.0003CrossRefGoogle Scholar
King, N. (2012). ‘Disagreement: What’s the Problem? Or a Good Peer is Hard to Find,’ Philosophy and Phenomenological Research 85(2), 249–72.10.1111/j.1933-1592.2010.00441.xCrossRefGoogle Scholar
Kornblith, H. (2010). ‘Belief in the Face of Controversy.’ In Feldman, R. and Warfield, T. (eds), Disagreement, pp. 29–52. New York: Oxford University Press.10.1093/acprof:oso/9780199226078.003.0003CrossRefGoogle Scholar
Kornblith, H. (2013). ‘Is Philosophical Knowledge Possible?’ In Machuca, D. (ed), Disagreement and Skepticism, pp. 260–276. New York: Routledge.Google Scholar
Lackey, J. (2010a). ‘What Should We Do When We Disagree?’ In Gendler, T.S. and Hawthorne, J. (eds), Oxford Studies in Epistemology, Oxford: Oxford University Press.Google Scholar
Lackey, J. (2010b). ‘A Justificationalist View of Disagreement’s Epistemic Significance.’ In Haddock, A., Millar, A. and Pritchard, D. (eds), Social Epistemology, pp. 298–325. Oxford: Oxford University Press.10.1093/acprof:oso/9780199577477.003.0015CrossRefGoogle Scholar
Lackey, J. (2013). ‘Disagreement and Belief Dependence: Why Numbers Matter.’ In Christensen, D. and Lackey, J. (eds), The Epistemology of Disagreement: New Essays, pp. 243–268. Oxford: Oxford University Press.10.1093/acprof:oso/9780199698370.003.0012CrossRefGoogle Scholar
Lasonen-Aarnio, M. (2013). ‘Disagreement and Evidential Attenuation,’ Nous 47(4), 767–94.10.1111/nous.12050CrossRefGoogle Scholar
Lasonen-Aarnio, M. (2014). ‘Higher-Order Evidence and the Limits of Defeat,’ Philosophy and Phenomenological Research 88(2), 314–45.10.1111/phpr.12090CrossRefGoogle Scholar
Littlejohn, C. (2014). ‘A Note Concerning Conciliationism and Self-Defeat: A Reply to Matheson,’ Social Epistemology Review and Reply Collective 3(12), 104–12.Google Scholar
Littlejohn, C. (2020). ‘Should we be Dogmatically Conciliatory?’ Philosophical Studies 177(5), 1381–98.10.1007/s11098-019-01258-4CrossRefGoogle Scholar
Littlejohn, C. (2013). ‘Disagreement and Defeat.’ In Machuca, D. (ed), Disagreement and Skepticism, pp. 169–192. New York: Routledge.Google Scholar
Machery, E. (2017). Philosophy Within its Proper Bounds. Oxford, Oxford University Press.10.1093/oso/9780198807520.001.0001CrossRefGoogle Scholar
Matheson, J. (2009). ‘Conciliatory Views of Disagreement and Higher-Order Evidence,’ Episteme 6(3), 269–79.10.3366/E1742360009000707CrossRefGoogle Scholar
Matheson, J. (2015a). ‘Are Conciliatory Views of Disagreement Self-Defeating?’ Social Epistemology 29(2), 145–59.10.1080/02691728.2014.907833CrossRefGoogle Scholar
Matheson, J. (2015b). ‘Epistemic Norms and Self-Defeat: A Reply to Littlejohn.’ Social Epistemology Review and Reply Collective 4(2), 26–32.Google Scholar
Pettigrew, R. (2015). ‘Accuracy and the Credence-Belief Connection,’ Philosopher’s Imprint 15(16), 1–20.Google Scholar
Pittard, J. (2015). ‘Resolute Conciliationism,’ Philosophical Quarterly 65(260), 442–63.10.1093/pq/pqv022CrossRefGoogle Scholar
Plakias, A. (2019). ‘Publishing Without Belief,’ Analysis 79(4), 638–46.10.1093/analys/anz040CrossRefGoogle Scholar
Rotondo, A. (2015). ‘Disagreement and Intellectual Skepticism,’ Australasian Journal of Philosophy 93(2), 251–71.10.1080/00048402.2014.986746CrossRefGoogle Scholar
Sturgeon, S. (2008). ‘Reason and the Grain of Belief,’ Noûs 42(1), 139–65.10.1111/j.1468-0068.2007.00676.xCrossRefGoogle Scholar
Titelbaum, M. (2015). ‘Rationality’s Fixed Point (Or: In Defense of Right Reason).’ In Gendler, T. and Hawthorne, J. (eds), Oxford Studies in Epistemology, vol. 5, pp. 253–294. Oxford: Oxford University Press.10.1093/acprof:oso/9780198722762.003.0009CrossRefGoogle Scholar
Vavova, K. (2014a). ‘Moral Disagreement and Moral Skepticism,’ Philosophical Perspectives 28(1): 302–33.10.1111/phpe.12049CrossRefGoogle Scholar
Vavova, K. (2014b). ‘Confidence, Evidence, and Disagreement,’ Erkenntnis 79(1), 173–83.10.1007/s10670-013-9451-6CrossRefGoogle Scholar
Walker, M. (2024). Outlines of Skeptical-Dogmatism: On Disbelieving Our Philosophical Views. Lanham: Lexington Books.Google Scholar
Weatherson, B. (2013). ‘Disagreements, Philosophical and Otherwise.’ In Lackey, J. and Christensen, D. (eds), The Epistemology of Disagreement: New Essays, pp. 54–73. Oxford University Press.10.1093/acprof:oso/9780199698370.003.0004CrossRefGoogle Scholar
Whiting, D. (2021). ‘Higher-Order Evidence,’ Analysis 80(4), 789–807.10.1093/analys/anaa056CrossRefGoogle Scholar
Wilson, A. (2010). ‘Disagreement, Equal Weight and Commutativity,’ Philosophical Studies 149(3): 321–6.10.1007/s11098-009-9362-1CrossRefGoogle Scholar
Worsnip, A. (2015). ‘Belief, Credence, and the Preface Paradox,’ Australiasian Journal of Philosophy 94(6), 549–62.10.1080/00048402.2015.1084343CrossRefGoogle Scholar