Alice, a life-long Democrat, sits down for her nightly appointment with her favourite political commentator, Rachel Maddow. For the last several years, much of Alice's political thinking has been supported and influenced by Rachel's characteristically detailed and cheeky analysis of Washington politics. But on this particular evening, Rachel raises some concerns about a piece of legislation that Alice had previously supported. Alice finds herself in a rare moment of disagreement with Rachel.
The next evening, Alice again returns to her television for the Rachel Maddow Show. However, earlier in the day her visiting relative had been watching a programme on Fox News. Though she does not usually turn to Fox News for political insight, she finds herself curious. As the Fox News commentator is analysing the current administration's foreign policy, she notes that her thinking about foreign policy is quite different from his. Alice is not surprised to find herself in this disagreement.
Which disagreement do you think is more likely to cause Alice to revise her previous beliefs? When it comes to the beliefs of her disagreeing interlocutors, to which is Alice more likely to give weight? I suspect the answer to these questions is obvious – Alice is much more likely to give weight to the beliefs of Rachel Maddow than the Fox News commentator. But is Alice's response to the respective disagreements rational?
With regard to our beliefs about politics, it is quite easy to find someone on the other side. The same holds true in the weighty domains of religion and philosophy. Disagreements abound! And disagreement about the rational response to disagreement also abounds! In much of the philosophical literature on this topic, the focus is on the rational response to disagreement among peers. For the purposes of this article, I will use Richard Feldman and Ted Warfield's stipulated definition of intellectual peers: ‘[P]eers literally share all evidence and are equal with respect to their abilities and dispositions relevant to interpreting that evidence’ (Feldman and Warfield (Reference Feldman, Warfield, Feldman and Warfield2010), 2). Moreover, the puzzling cases of disagreement that populate the literature are not merely cases of peer disagreement but they are cases of disagreement between expert peers. When two genuine experts in a field disagree, this raises the important question: what sort of weight should each expert give to the opinion of her disagreeing peer?
Thomas Kelly suggests that there is a spectrum of positions on how to handle disagreement among expert peers. He labels one end of the spectrum ‘The Equal Weight View’. On this view, ‘In cases of peer disagreement, one should give equal weight to the opinion of a peer and to one's own opinion’ (Kelly (Reference Kelly, Feldman and Warfield2010), 112). Disagreement with a peer is taken as evidence against your belief. Kelly labels the other end of the spectrum ‘The No Independent Weight View’. On this view, ‘In at least some cases of peer disagreement, it can be perfectly reasonable to give no weight at all to the opinion of the other party’ (Kelly (Reference Kelly, Feldman and Warfield2010), 115). Catherine Elgin cleverly labels these two ends of the spectrum the ‘spineless’ and the ‘stubborn’, respectively (Elgin (Reference Elgin, Feldman and Warfield2010), 57).
In this article, I will focus on the intractable disagreements we find in politics, religion, and philosophy. I will argue that the rational response to these disagreements depends (in part) on the extent to which the disagreeing interlocutors share relevant beliefs, where the relevant beliefs are those beliefs that one might use to evaluate the proposition about which there is disagreement. If interlocutors share all of the relevant beliefs, the rational response to disagreement is spinelessness – one ought to give substantial weight to the opinion of one's disagreeing interlocutor. If interlocutors share none of the relevant beliefs, the rational response to disagreement is stubbornness – one ought to give no (or little) weight to the opinion of one's disagreeing interlocutor. Moreover, I will argue that this principle regarding the rational response to disagreement holds not only in cases of peer disagreement but also in cases in which one disagrees with a superior, as in the case of Alice and the political commentators. After defending my proposal, I will consider two objections: first, I will consider whether my thesis depends on the problematic coherence theory of justification and, second, I will consider the charge that my thesis endorses operating in an epistemic bubble. Ultimately, if my thesis is correct, it has implications for the context within which we can expect to have productive disagreements about philosophy, religion, and politics.
On weighty disagreements
We have all experienced a time when a disagreement with a peer has led us to revise a belief. David Christensen offers a prototypical case (Christensen (Reference Christensen2007), 193). You are out to dinner with friends and when the bill arrives, you decide to split the bill evenly. You calculate your portion of the bill and put cash in the middle of the table. Your friend also runs the calculation but arrives at a different number. You find yourself in a disagreement about a mathematical calculation. You have no reason to think that you are better than your friend when it comes to the basic maths required to split a bill – you are bill-splitting peers! So you recalculate and discover that in fact your friend's number was correct. One might fairly describe this as a case in which it is rational to give equal weight to your friend's calculation and your calculation when you find yourself in this disagreement. In giving equal weight to both, you will be inclined to recalculate and determine who is correct. If you and your friend are rational, one or both of you will revise your beliefs upon determining the correct split.
However, in many cases of peer disagreement, we cannot easily resolve the matter. This is especially true with respect to religious disagreement. We will not soon settle the matter of the existence of God, much less many other religious claims. Religion is not the only domain where we find intractable disagreement; the same could be said of philosophy and politics. Experts in these domains regularly disagree and a clear resolution is not forthcoming. Adam Elga describes these disagreements as ‘messy examples of real-world disagreements about hard issues’ (Elga (Reference Elga2007), 492). Let's call these important and intractable disagreements Weighty Disagreements.
Consider the following Weighty Disagreement from Peter van Inwagen (van Inwagen (Reference van Inwagen, Feldman and Warfield2010), 23–24). Van Inwagen highlights his long-standing philosophical disagreement with David Lewis about the problem of free will. Both philosophers had access to the same evidence, there is no reason to think that one of them was better at assessing the evidence – in this sense, they were genuine peers. Yet Lewis was a compatibilist and PVI is an incompatibilist. Despite conversations and correspondence, neither was able to persuade the other. Unlike the case of splitting the bill, there is no quick way to resolve this disagreement.
Note that in the case of van Inwagen and Lewis, neither gave the other's position substantial weight even in the face of this disagreement. If either gave weight to the other's view, it was not enough to move the needle of his own position on free will from belief to even suspension of belief. I suspect neither philosopher's belief was even weakened as a result of exploring this disagreement. Is this the rational way to approach a philosophical disagreement among peers? Or should the presence of disagreement cause both parties to pause and moderate their commitments to their beliefs if they are to remain rational?
Weighty Disagreements also take place in the realm of religion. Consider the disagreement between William Rowe and Alvin Plantinga on the Problem of Evil. Rowe takes the existence of apparently gratuitous suffering to be significant evidence against the existence of a theistic God (Rowe (Reference Rowe, Peterson, Hasker, Reichenbach and Basinger2014), 366–368). Plantinga takes the existence of a theistic God to be a properly basic belief (Plantinga (Reference Plantinga, Peterson, Hasker, Reichenbach and Basinger2014), 213). Thus, when presented with alleged cases of gratuitous suffering, Plantinga can reasonably perform what Rowe calls the G. E. Moore shift and argue from the existence of God to the non-existence of gratuitous suffering (Rowe (Reference Rowe, Peterson, Hasker, Reichenbach and Basinger2014), 369). Anyone who is familiar with the work of these fine philosophers will recognize that neither of them lacks access to the relevant arguments or the ability to assess them. However, they reach different conclusions from assessing the same evidence. Should Plantinga's knowledge of Rowe's alternative position cause Plantinga to be less confident in his belief in theism? And likewise, should Rowe's knowledge of Plantinga's alternative position have caused Rowe to moderate his commitment to atheism? Or is it rational for each to maintain his position even though he is aware that his peer disagrees?
On conditions for stubbornness
In his piece, ‘You Can't Trust a Philosopher’, Richard Fumerton argues that in cases of Weighty Disagreement such as those described above, there are three conditions under which one need not give any weight to the opinion of a disagreeing interlocutor. Each condition is considered individually sufficient for giving no weight. His conditions are as follows:
(1) We have good reason to believe that we have a different and better evidence base than the person with whom we disagree.
(2) We have good reason to believe that we have engaged a common evidence base more successfully that (sic) the person with whom we disagree.
(3) We have good reason to believe that the person with whom we disagree is cognitively defective. (Fumerton (Reference Fumerton, Feldman and Warfield2010), 99)
Consider again the case of splitting the bill. In this case, both parties have access to the same evidence (the bill), there is no reason to think that you split the bill better than your friend in this case, and there is no reason to think that your friend is cognitively defective with respect to the required arithmetic. Since none of these conditions is present, Fumerton suggests that the rational response to this disagreement is to give weight to the opinion of our disagreeing interlocutor.
But what about the examples of Weighty Disagreements above? As I described the cases, none of these conditions is met. Both parties have access to the same arguments. Both parties are experts in their fields and so there is no reason to think that one has engaged the evidence better than the other. Moreover, neither party appears to be cognitively defective. Yet often we do not take discovery of Weighty Disagreement to be a reason to revise or moderate our belief. Why not?
Fumerton argues that in some Weighty Disagreements, we really do take our interlocutor to be cognitively defective. Thus, we rationally give no weight to the disagreement because condition (3) is satisfied. Fumerton offers his disagreement with Paul Churchland as a case in point. Fumerton describes himself as ‘a confirmed property dualist’ when it comes to the metaphysics of mental states (Fumerton (Reference Fumerton, Feldman and Warfield2010), 99). In contrast, Churchland is an eliminative materialist; he holds the view that there are no mental states. Fumerton suggests that he need not give the opinion of the eliminative materialists any weight because he thinks that they fundamentally display some sort of cognitive defect. He writes about Churchland:
[M]y first instinct is to suspect that he is not really serious – that he is just messing around a bit trying to provoke an interesting discussion. When I begin to suspect that Churchland and other eliminativists are serious, I am genuinely puzzled as to what is going on in their minds . . . I discount completely the epistemic significance of what they apparently believe. (Fumerton (Reference Fumerton, Feldman and Warfield2010), 101–102)
In this case, Fumerton seems to understand the position of his interlocutor to be so outrageous that the fact that his interlocutor holds the position is evidence of cognitive defect. The eliminativist does not meet condition (3) and so no weight need be given to his opinion.
But what about cases of Weighty Disagreements when our interlocutors do not hold views as radical as that of the eliminative materialist? Fumerton notes that to take these disagreements seriously, he would need evidence that his interlocutors are not cognitively defective (Fumerton (Reference Fumerton, Feldman and Warfield2010), 106ff.). In the case of a Weighty Disagreement about philosophy, the evidence he would need is evidence that they reliably form true philosophical beliefs. However, if we attempt to gather this evidence inductively, we will find that most philosophers must be unreliable in forming true philosophical beliefs. Given that most philosophical positions are minority views, most philosophers will hold a great many false philosophical beliefs. Fumerton thinks that the evidence supports the unreliability of philosophers in forming true philosophical beliefs and therefore the cognitive defectiveness of his interlocutors. Thus, he need not give weight to their disagreements.
Obviously, Fumerton's argument can easily be extended to religious and political disagreement. Given the wide variety of views in these subject areas, any view will be a minority view and so inductive evidence of reliability in forming religious and political opinions will be hard to come by. By the same reasoning that Fumerton employed to argue that he need not give weight to philosophical disagreements, we can also conclude that one need not give weight to religious and political disagreements.
As Fumerton notes, one disadvantage of his argument is that it also applies to him. He writes, ‘This reason for thinking that my opponents are probably cognitively defective is also a reason for thinking that I am probably cognitively defective’ (Fumerton (Reference Fumerton, Feldman and Warfield2010), 109). I would argue that the result that everyone turns out to be cognitively defective suggests that the conditions for cognitive defect are too strong. To quote the philosophically astute Dash from The Incredibles (Bird (Reference Bird2004)), ‘If everyone is special, no-one is.’ The same principle applies to cognitive defectiveness. So how should we understand the notion of cognitive defect found in condition (3)? I suspect that the intuitive force of condition (3) comes from cases in which the capabilities of the disagreeing interlocutors are obviously disparate. For example, it would seem irrational to give weight to the disagreement about splitting the bill if I am having the disagreement with a two-year-old. Likewise, it would seem irrational for a doctor to give weight to her patient's disagreement about her diagnosis; their capabilities are likely disparate. However, I don't think that appealing to condition (3) will suffice for dismissing the disagreement of an interlocutor in the Weighty Disagreements mentioned above (PVI vs Lewis, Rowe vs Plantinga, and Fumerton vs Churchland). The capabilities of these interlocutors are not disparate so accusing each other of cognitive defects seems inappropriate. Thus, if it is rational to stubbornly maintain one's original position in the face of these disagreements, I propose that we need another condition; none of Fumerton's three conditions is satisfied.
A fourth condition
In addition to Fumerton's three conditions under which we would be rational in not giving weight to the disagreement of an interlocutor, I propose a fourth:
(4) (a) We have good reason to believe that the relevant beliefs are not shared (or not shared in the same way) between us and the person with whom we disagree and (b) we have no defeaters (independent of the disagreement itself) to our relevant beliefs.
Like Fumerton's other conditions, condition (4) is individually sufficient for rationally giving no weight to the disagreement of an interlocutor.Footnote 1
My proposed condition (4) has 2 parts. Condition (4a) suggests that the rational response to disagreement depends on the degree to which relevant beliefs are shared between us and our disagreeing interlocutor. By relevant beliefs, I mean those beliefs that one might use to evaluate the proposition about which there is disagreement. When someone with whom we agree on many relevant beliefs disagrees with us about a related proposition P, it seems that the rational response to this disagreement is to be concerned that we have not seen the relationship between our beliefs and P clearly. But when someone with whom we disagree on many relevant beliefs disagrees with us about a related proposition, it is rational to disregard the disagreement. Disagreement in this case is to be expected; it can be easily explained with reference to disagreements on a more fundamental level.
Condition (4b) specifies that the relevant beliefs cannot be held irrationally. Note that (4b) does not require the relevant beliefs to be true; surely we can rationally hold a false belief. Moreover, (4b) does not require one to have justification or evidence for the relevant beliefs, though an internalist might want to specify this stronger condition. Instead, (4b) makes the more modest claim that we must have no defeaters to our relevant beliefs. This rules out giving no weight to a disagreeing interlocutor on the basis of a difference in relevant beliefs when we have reason to believe that our relevant beliefs are false while at the same time not requiring cognitive access to the justification for our relevant beliefs. This more modest requirement for rational relevant beliefs will hopefully be acceptable to both internalist and externalist accounts of justification.
While I think condition (4) describes how we do in fact respond to disagreement, I am arguing for the further claim that this response to disagreement is rational. In the literature on disagreement, the term rational is regularly used to refer to an epistemically good response to disagreement without specifying conditions for rationality. This lack of clarity is no surprise; there is great disagreement among epistemologists on how to understand the normative concepts in epistemology of rationality, reasonability, justification, warrant, and the like. As Richard Foley points out, there is even disagreement about whether these represent distinct concepts (Foley (Reference Foley and Moser2002), 177). I will not attempt to weigh in on these debates. However, to strengthen my claim that it is rational to give no weight to the opinion of a disagreeing interlocutor under condition (4), I believe it will suffice to sketch some conditions for rational belief revision.
Under what conditions might the response of the believer to disagreement be deemed rational? When considering the spectrum of positions on the rational response to disagreement, it can seem as if the focus is on the believer's choice to either continue to believe in the face of disagreement or to abandon her belief in the face of disagreement. However, I am inclined to think that we have very little choice about what to believe; at the very least, we have very little direct control over our beliefs. I will not take the time to argue for doxastic involuntarism here but just note that this is an assumption I hold with respect to belief formation. It follows from this assumption that if I respond to disagreement by revising my original position or by holding tight to it, this response is likely to be out of my direct control.
Does doxastic involuntarism entail that we cannot assess one's response to disagreement as rational or irrational? Elgin argues no. According to Elgin, just because something is out of our control does not mean that it cannot be assessed. To adopt her example, just because I cannot control the weather does not mean that I cannot assess the coming snowstorm as bad given that it will probably cancel my classes tomorrow (Elgin (Reference Elgin, Feldman and Warfield2010), 61). Likewise, just because I cannot control my response to disagreement does not mean that I cannot assess the goodness (i.e. rationality) of that response.
So under what conditions might we describe one's response to disagreement as rational? One minimal rule of thumb for rational response would seem to be that the response should not lead to an internally inconsistent set of beliefs. Alvin Goldman states this rule for the rationality of adopting a belief as follows: ‘One should not adopt a belief in any proposition p if one already has other beliefs which, if p were adopted, would form a logically inconsistent set’ (Goldman (Reference Goldman and Moser2002), 148). Maintaining a consistent set of beliefs seems like a good indicator of a rational response to disagreement; at the very least, adopting an inconsistent set of beliefs would be rightly described as irrational. Thus, when a disagreeing interlocutor proposes a proposition P that, if adopted, would result in my constellation of beliefs being inconsistent, it seems that it would be irrational to adopt P.
But would that truly be irrational? As Goldman points out, while it would be irrational for a person to adopt an inconsistent belief constellation, this does not help us in determining which should be preferred: our original (let's suppose) consistent set of beliefs or a consistent set of beliefs which includes P and excludes the beliefs that conflict with P (Goldman (Reference Goldman and Moser2002), 148). For this, we need criteria in addition to consistency. To this end, let's consider Alvin Plantinga's description of a noetic structure.
Alvin Plantinga defines a person's noetic structure as ‘the set of propositions he believes together with certain epistemic relations that hold among him and these propositions’ (Plantinga (Reference Plantinga, Peterson, Hasker, Reichenbach and Basinger2014), 210). Plantinga highlights three main features of a noetic structure: (1) A specification of basic and non-basic beliefs; which beliefs are held on the basis of others and which beliefs are thought to be foundational. (2) An index of degree of belief; the degree of confidence with which we hold our various beliefs. (3) An index of depth of ingression; some of our beliefs are more central to our belief system – if these beliefs were to be abandoned, many other beliefs would go as well. But some of our beliefs are less central – if these less central beliefs were to be later rejected, not much would change in our noetic structure (Plantinga (Reference Plantinga, Peterson, Hasker, Reichenbach and Basinger2014), 210–211).
With these features of a noetic structure in mind, we can begin to sketch when it would be rational to prefer our original consistent set of beliefs and when it would be rational to prefer a consistent set of beliefs which includes P and excludes previously held beliefs that conflict with P. Consider the following example. Suppose that Jack determines that proposition P is inconsistent with other beliefs in his noetic structure that are either basic, or deeply ingressed, or about which Jack has a high degree of confidence. Suppose also that Jack has no reason to think the relevant beliefs in his noetic structure are false. If Jack were to believe P and maintain a consistent constellation of beliefs, he would have to abandon a certain number of beliefs that are well-established in his noetic structure in favour of P – a proposition that (let's suppose) would be on the periphery of Jack's noetic structure and about which he has only moderate confidence. It seems that abandoning these basic or deeply ingressed or highly justified beliefs in order to accept P, which lacks these features, would rightly be deemed irrational.
By way of contrast, suppose that Jill does not share the beliefs that led Jack to reject P or Jill holds those beliefs but with much less confidence or those beliefs are not as deeply ingressed in Jill's noetic structure. Like Jack, Jill has no reason to think that the relevant beliefs that she does have are false. Would Jill's original consistent constellation of beliefs be preferable to a constellation of beliefs in which Jill adopts P and revises other parts of her noetic structure as needed? The answer is not so obvious. Maybe the only revision required would be for Jill to reject ~P, a belief she held on the periphery of her noetic structure with very little confidence. In this case, the new constellation of beliefs may very well be preferable and thus Jill's response to her encounter with P would be assessed as rational.
All of this suggests the following principle: assuming one has no defeaters for the relevant beliefs in one's original constellation of beliefs (independent of the disagreement of his interlocutor), the original consistent constellation of beliefs should be preferred to a revised consistent constellation of beliefs that accepts P only if accepting P does not require abandoning (1) properly basic beliefs, (2) beliefs that are more deeply ingressed than P, or (3) beliefs held with a greater degree of confidence than P. Thus, Jack is rational in preferring his original constellation of beliefs to the revised consistent constellation of beliefs that accepts P. However, it is quite possible that Jill is rational in revising her original consistent constellation of beliefs in order to adopt P. This rough principle for rational belief revision supports condition (4). When we believe that our disagreeing interlocutor does not share our undefeated relevant beliefs or they are not shared in the same way (i.e. basically, deeply ingressed, or believed with confidence), then we can rationally give no weight to their disagreement.
Why should we add condition (4)? Condition (4) allows us to explain the rationality of dismissing disagreement in some cases without accusing a disagreeing interlocutor of cognitive defect or intellectual vice. Consider again Fumerton's disagreement with the eliminative materialist. Like Fumerton, I am inclined to disagree with eliminative materialism. The eliminative materialist is committed to a materialist view of persons; for many eliminative materialists, this belief in materialism more broadly is held with a high degree of confidence and it is deeply ingressed. I do not share their commitment to materialism and so it would be irrational for me to accept the costly position that there are no mental states. However, I do not take the eliminativist to be cognitively defective. In fact, I think the eliminativist takes a materialist view of persons to its logical and simplest conclusion. We share an appreciation for the difficulty of reducing mental states to brain states. But it would seem irrational for me to give weight to the beliefs of the eliminative materialists because I do not share their deeply ingressed belief in a materialist view on the nature of persons. I see how they came to their view on the non-existence of mental states but I do not give weight to their opinion about mental states because it stems from a more fundamental disagreement.
A similar sort of assessment will apply to Rowe and Plantinga's disagreement about the problem of evil. As Plantinga has argued, belief in God should be considered properly basic (Plantinga (Reference Plantinga, Peterson, Hasker, Reichenbach and Basinger2014), 213). Given his commitment to God's existence (in particular, an omniscient, omnibenevolent, omnipotent God), it will follow that there are not instances of gratuitous suffering. All the suffering that we experience is suffering that must have a purpose. Rowe did not share Plantinga's belief in God, much less would he have taken that belief to be properly basic. So when Rowe assesses an instance of suffering for which it is difficult to see a purpose, he concludes that it is reasonable to see this as an instance of gratuitous suffering. Without a basic (or at least deeply ingressed) belief in the existence of a God with the above omni attributes, it seems reasonable for Rowe to draw such a conclusion.
Given the differences in their relevant beliefs, it is rational for both to give no weight to the disagreement of the other. Though both philosophers are able to understand and assess the cases that support the evidential problem of evil, they will probably come to different conclusions about it because of their different relevant beliefs. I take it that this is what Rowe is suggesting when he defines ‘Friendly Atheism’ (Rowe (Reference Rowe, Peterson, Hasker, Reichenbach and Basinger2014), 370). Rowe suggests that while he is convinced (has a high degree of confidence) that atheism is true on the basis of the evidential problem of evil, he maintains that the theist who is acquainted with the arguments for and against belief in God can nonetheless be rational in her belief in theism. Rowe's ‘Friendly Atheism’ seems to be a precursor to the ‘No Independent Weight’ view on disagreement. He acknowledges the rationality of the theist, all the while not considering her opinion to move him from his atheism.
Suppose I am correct that two peers need not give weight to a disagreement if they have different (undefeated) relevant beliefs or the shared beliefs play different roles in their noetic structures. Does this mean that they are not peers? This is the position of both Elga (Reference Elga2007) and Regina Riti (Reference Riti2017). Elga argues that a person who is otherwise as ‘quick-witted, well-informed, intellectually honest, and thorough’ as you does not count as your peer in assessing political claims if they do not share your political framework (Elga (Reference Elga2007), 493). Riti argues that it is reasonable (in some cases) to trust the political testimony of those who share our partisan affiliations because they share our value commitments. For Riti, sharing value commitments is a necessary condition for being epistemic peers (Riti (Reference Riti2017), E-51).
While I largely agree with Elga and Riti – both seem to support a version of my thesis that the weight we give to an interlocutor will depend on shared relevant beliefs – I want to resist adding this requirement to peerhood. If genuine peerhood requires the peers to share exactly the same noetic structure (with respect to basicality, confidence, and ingression) or at least exactly the same noetic structure with respect to any of the relevant beliefs, then the notion of a peer Weighty Disagreement is a genuine fiction. There would be no peers. But it is precisely Weighty Disagreements among peers with different noetic structures like PVI and Lewis, Plantinga and Rowe, even Fumerton and Churchland, that give rise to the predicament of how to weigh such disagreement. When the disagreement is a ‘clean, pure’ disagreement, like that of splitting the bill, the rational response is much less puzzling (Elga (Reference Elga2007), 492). Thus, to require such strict standards for peerhood avoids the problem of peer disagreement only by defining it out of existence.
Moreover, if peer disagreement does require identical noetic structures, disagreement should always carry weight. If we have no disagreement with an interlocutor about the order and confidence of other relevant beliefs, there seems to be no rational basis upon which we would reach different conclusions about the same evidence. So instead of construing condition (4) as defining a peer, I think it should be understood as describing a correlation between the similarity of the noetic structures of peers and the degree to which disagreement between these peers affects justification. The more similar their relevant beliefs and the roles they play in their respective noetic structures, the more weight should be given to disagreement.
Peers and superiors
I have focused thus far on cases of Weighty Disagreement between expert peers, the kind of disagreement you find in the case of van Inwagen and Lewis on free will, Rowe and Plantinga on the existence of God, and Fumerton and Churchland on the nature of mental states. I have argued that there are individually sufficient conditions for giving no weight to the opinion of a disagreeing peer – in addition to Fumerton's three conditions, I have proposed a fourth. Moreover, I have suggested that when none of these conditions is met (as in the case of splitting the bill), we ought to give weight to the opinion of a disagreeing interlocutor. My condition (4) suggests that the amount of weight that we give to the opinion of the disagreeing interlocutor should be proportional to the degree to which we share the relevant beliefs in the same way (i.e. the relative role they play in our noetic structures with respect to basicality, depth of ingression, and confidence).
I now would like to return to the case of Alice and her disagreements with Rachel Maddow and the Fox News commentator. Let's suppose that both of these disagreements are not peer disagreements; let's suppose that both Rachel Maddow and the Fox News commentator are Alice's superiors when it comes to matters of politics. What is the rational response of Alice to her disagreement with her superiors?
Surely the rational response to all disagreements with superiors cannot be to revise one's beliefs to conform to the beliefs of the superior. As our discussion thus far has amply illustrated, there are Weighty Disagreements between peer experts who are surely our superiors about their expertise. If the rational response to Weighty Disagreement with a superior is revision, then which superior's beliefs should we accept? Or must we continually revise as we encounter yet another superior's opinion on the matter? Intuitively, this sort of arbitrary picking and choosing among expert opinions does not characterize rational belief formation.
I want to suggest that my proposed condition (4) offers guidance on rational belief revision whether or not the interlocutors are peers or one is superior. Condition (4) does not specify the expertise of the interlocutor. It is not based on which of two constellations of beliefs is better justified, closer to the truth, etc. It is simply based on the similarity of two constellations with respect to relevant beliefs and the role they play in the believer's noetic structure. Thus it seems that the same principle of proportioning the weight one gives to the beliefs of a disagreeing interlocutor to this relative similarity will hold in disagreements with a superior. The main difference is that when disagreeing with a superior (at least a recognized superior), we are more likely to evaluate their opinion on, say P, as being more justified than our opinion on P. Thus, when evaluating our original constellation of beliefs versus the revised constellation that includes P, it would be rational to weigh the latter differently when a superior is advancing P than when a peer is advancing P. But it still may be rational to maintain our original constellation of beliefs if adopting P requires significant revision to it.
To return to Alice, given that she takes the Fox News commentator to be her superior, she may very well believe that his opinion on foreign policy is more justified than hers. However, if she finds that accepting his views on foreign policy would require significant revision to her undefeated relevant beliefs and she evaluates these beliefs in her noetic structure to be properly basic or deeply ingressed or beliefs about which Alice has great confidence, it is rational for her to give no weight to the disagreement. To adopt his (let's suppose) superior opinion in this one area at significant cost to her constellation of beliefs does not seem to be a rational response.
But condition (4) suggests that the rational response to disagreement with Rachel Maddow is for Alice to give weight to Maddow's opinion. Given the similarity of their relevant beliefs and noetic structure, belief revision would presumably be very low cost for Alice. Paying a low epistemic cost for accepting the belief of a superior with whom you share other relevant beliefs seems to characterize rationality. Moreover, it would be rational for Alice to be concerned that she has misevaluated the relationships between the beliefs she shares with Rachel Maddow and the one about which they disagree. Thus, I argue that condition (4) can be used to evaluate the rational response to disagreement with a superior as well as disagreement between peers.
Objection 1: the coherence theory of justification
I have proposed that we need not give weight to the opinion of a disagreeing interlocutor if we have good reason to believe that the relevant beliefs are not shared between us and the person with whom we disagree. One worry about this proposal is that it is based on a coherence theory of justification. According to the coherence theory of justification, there are no basic beliefs. Beliefs are justified simply in terms of their internal relationships with other beliefs. I am justified in believing P as long as P coheres with my other beliefs. If P does not cohere with my interlocutor's other beliefs, it is not justified for her. However, the lack of justification for P for my interlocutor does not affect my justification since, according to the coherence theory of justification, all of my justification for P comes from the relationship between P and the other beliefs in my noetic structure.
As Fumerton notes, the coherence theory of justification is certainly at home with the notion that we need not give weight to the opinion of a disagreeing peer. He writes,
As long as we view the justification of a person's belief as a function of that belief's coherence with the rest of the person's beliefs, it should become immediately obvious that the existence of another (perhaps rational) person with whom I disagree is no real threat to the justification I possess for my beliefs. (Fumerton (Reference Fumerton, Feldman and Warfield2010), 104)
The coherence theory of justification, if true, would support my claim that I need not give weight to a disagreeing peer with different relevant beliefs. It would also explain why I do give weight to a disagreeing peer with similar relevant beliefs – if we share the beliefs that are relevant to justifying P and justification is derived solely from P's relationship to these beliefs, then we should not have a disagreement about P. If we do have a disagreement about P, I should take that disagreement seriously and seek to understand and resolve it.
The coherence theory of justification has well-known problems.Footnote 2 According to the alternative systems objection, there could be competing coherent belief systems. It is not difficult to imagine two noetic structures, each of which is internally consistent, and yet these noetic structures contain conflicting beliefs. The coherence theory of justification would suggest that both of these conflicting beliefs are justified. Moreover, because beliefs are justified simply in terms of their internal relationships with other beliefs on the coherence theory of justification, it is possible to have a set of justified beliefs that cohere with one another but are utterly isolated from reality. Proponents of this isolation objection to coherence theory are concerned about an account of justification that does not require any input from external reality to justify a belief.
Does my proposal depend on the problematic coherence theory of justification? Even if a disagreement about P is rational due to a difference in relevant beliefs, those beliefs are not necessarily in turn justified solely by their internal relationships with my other beliefs. The isolation objection to a coherence theory of justification suggests that any rational noetic structure will need input from external reality. The relevant supporting beliefs for P may be basic beliefs or they may be justified by our experience of external reality. For that matter, P itself may be justified in part from external input in addition to other beliefs in my noetic structure. In other words, my proposal is consistent with a view of justification that does not merely define justification as an internal relationship among beliefs.
To illustrate my proposal's independence from the coherence theory of justification, consider again the disagreement between Rowe and Plantinga on the existence of gratuitous evil. I propose that they need not give weight to the disagreement about gratuitous evil because their disagreement stems from a disagreement about other relevant beliefs. However, both Plantinga and Rowe offer justification for these relevant beliefs that does not depend on other beliefs in their noetic structures. For Plantinga, his belief that there is no gratuitous evil depends on his basic belief that God exists. For Rowe, his belief that there is gratuitous evil depends on a plausible assessment of particular instances of evil. He appeals to his experience of the world; he considers cases like that of a fawn burning for days in a forest fire without anyone around. So even if I am correct that a disagreement on relevant beliefs is a reason not to give weight to the opinion of a disagreeing interlocutor, this does not entail that the only source of justification is the internal relationship among a person's beliefs. My proposal is consistent with the existence of other sources of justification, such as those appealed to by Plantinga and Rowe. Thus, my proposal does not depend on the coherence theory of justification.
Objection 2: a vicious epistemic bubble?
According to Fumerton's condition (3), if we believe our interlocutor has a cognitive defect, we need not give weight to their disagreeing opinion. He then offers inductive support for the claim that his philosophical interlocutors do in fact have cognitive defects – given that most philosophical positions are minority positions, it follows that most philosophers are unreliable in forming true philosophical beliefs. I proposed condition (4) to explain why one can rationally give no weight to (some) disagreeing interlocutors without accusing them of a cognitive defect. But I wonder if condition (4) has a similar result. Though I do not think that my disagreeing interlocutors are cognitively defective, I have argued that I need not give weight to their opinion if they do not share my undefeated relevant beliefs. Does this mean that I give a privileged position to my own beliefs? Does my proposed condition in effect recommend that I operate within an epistemic bubble, only taking seriously the beliefs of those who share the vast majority of my relevant beliefs, and organize those beliefs in a way that is similar to my noetic structure?Footnote 3
As for the first question, I think the answer is no. I do not give a privileged position to my own beliefs. In order to give no weight to your opinion about P, I do not have to believe that my opinion about P is true and yours is false. I give no weight to your disagreeing opinion because I recognize that you do not share the beliefs that I hold that are relevant to my assessment of P. If the beliefs that are relevant to my assessment of P are otherwise undefeated and have a well-established place in my noetic structure, then it seems rational for me to reject P. So the fact that we disagree about P does not mean that I have access to better arguments than you or that I have cognitive superiority in my ability to assess these arguments. The disagreement stems from other differences in our noetic structure. Moreover, the situation is symmetrical; I take you to be rational in disregarding my disagreeing opinion for exactly the same reasons.
However, for the second question, I do think that my proposal suggests some previously unrecognized advantages of an epistemic bubble. Thi Nguyen defines an epistemic bubble as ‘a social epistemic structure which has inadequate coverage through a process of exclusion by omission’ (Nguyen (Reference Nguyen2020), 143). Epistemic bubbles form somewhat unintentionally due to the way we naturally obtain our information from those with whom we share certain affinities. In doing so, our information coverage is inadequate; certain sources will inevitably be left out. We usually deride the notion of an epistemic bubble because surrounding ourselves with only those with whom we largely agree has the effect of reinforcing our biases. I do not deny this concern; there is certainly value to exposing oneself to a broad spectrum of information sources. However, my proposal suggests that it is within the context of a community with shared relevant beliefs (an epistemic bubble?) that we do and should take disagreement seriously. Thus, a non-vicious employment of an epistemic bubble would be to help us refine and tweak our beliefs. In fact, I would suggest that it is this sort of incremental development of one's noetic structure that is characteristic of rational intellectual growth (as opposed to the rarely seen radical conversion of one's noetic structure). To return to the case of Alice, listening to Rachel Maddow may place Alice in an epistemic bubble but given the similarity of Alice's noetic structure to others in the bubble, this is the context within which disagreement could inspire rational belief revision. Disagreements well outside the bubble (as in Alice's disagreement with the Fox News commentator) can be rationally disregarded. This suggests that the environment within which disagreement is most likely to produce rational, incremental change in our religious, philosophical, and political beliefs is a community whose members have relevantly similar noetic structures (i.e. an epistemic bubble).
Conclusion
Throughout this discussion, I have focused on Weighty Disagreements – disagreements about religion, philosophy, and politics that are not easily resolved and have substantial consequences for how one lives. Fumerton offers three conditions under which we need not give weight to the opinion of a disagreeing peer; I propose a fourth. My goal in proposing condition (4) is to give a fuller explanation for why it is rational to give no weight to disagreement when our interlocutor has a relevantly dissimilar noetic structure.
As I have argued above, my proposal does not require one to take one's own beliefs to be superior to those of someone wildly different from oneself; this is not a claim to cognitive superiority. It is merely the claim that the rational response to disagreement depends on the degree to which relevant beliefs are shared. So built into this proposal is both a sort of philosophical, religious, and political tolerance – I understand why you came to your view in light of your beliefs and experiences – and an explanation for why a rational person need not always re-evaluate belief in the face of disagreement across philosophy, religion, and politics. As we navigate an ideologically diverse society, this insight could provide guidance on the context and conditions for productive disagreement.
Acknowledgements
Special thanks to the Philosophy of Religion Reading Group at the United States Air Force Academy for discussing an earlier draft of this article and offering valuable insights and criticisms. I greatly benefited from their challenging questions, especially regarding the statement of the fourth condition. I would also like to thank EJ Coffman, Allison Postell, and Mark Jensen for helpful comments and encouragement on an early draft of the article. Lastly, I am indebted to the journal reviewers for thorough comments on my first submission.
