1 Introduction

Suppose you are a rational, careful, and critical thinker—someone who evaluates evidence seriously and reconsiders opinion in light of compelling reason. After carefully considering relevant evidence, you become convinced that p. But then, you meet an epistemic peer whose opinion you trust in matters concerning p. He claims not-p. You have both looked at the same evidence, and you have no reason to believe that he is biased, intoxicated, or otherwise functioning below his usual competency. How should you respond? The epistemology of disagreement literature has offered various answers. At first, the debate was primarily between conciliatory views that argue to weaken your confidence and steadfast views that argue your assurance should remain unchanged.Footnote 1 Since then, many have offered positions that are more moderate. However, the debate has remained focused on peer disagreement—disagreement between those with equal or similar cognitive abilities and access to the same or similar evidence. Thomas Kelly (2010) describes a standard case: “Suppose that you and I have been exposed to the same evidence and arguments that bear on some proposition: there is no relevant consideration which is available to you but not to me, or vice versa” (p. 183).

Peer disagreement raises difficult epistemological questions. But, what about disagreement between superiors and inferiors? The literature does include some discussion. In “Reflection and Disagreement,” Adam Elga argues that his theory of peer disagreement extends to superior/inferior disagreement.Footnote 2 And, in “Discovering Disagreeing Peers and Superiors,” Bryan Frances offers a detailed account focused on the perspective of inferiors. Nonetheless, my paper aims to add to the discussion via an analysis of undertheorized aspects of superior/inferior disagreement: various types of inferiors, which of such variants bear epistemic import, and how these differing inferiors might affect appropriate superior reactions.

Prima facie, it may seem that disagreement with epistemic inferiors is trivial or at least less interesting than peer disagreement. If A claims p and find out that (1) B claims not-p and (2) B has less evidence, it seems that (2) nullifies the impact of (1). A can account for the disagreement via B’s evidence paucity and remain steadfast in her belief. The case is similar with cognitive inferiors. Maybe A is an expert on the subject and B a novice, or perhaps B is just plain stupid. In either case, it seems A has little reason to alter her belief. One aim of this paper is to show that these prima facie intuitions are often mistaken or, at least, misleading. The first step in understanding inferior disagreement is to distinguish two types of inferiors—competents and incompetents. We can then see that disagreement from competent inferiors can give superiors reason to adjust credence and reevaluate evidence, and hence that epistemic inferiority alone is insufficient grounds for dismissing opinion. Next, come more nuanced difficulties. Disagreements between superiors and incompetent inferiors should sometimes boost the superior’s justification. On the other hand, agreement from incompetent inferiors can defeat justification. Either way, inferior opinion carries epistemic weight. But, this still fails to cover all ground; uninformative inferior disagreements remain. In the final section, I describe these uninformative disagreements, how to distinguish them from the informative cases, and the proper reactions thereof.

2 Peer Relations and Epistemic Competence

2.1 What We Do When We Disagree

When met with disagreement, disagreeing parties must consider what epistemic actions are most appropriate. Two considerations are particularly important. First, subjects must consider how to adjust their belief credence or if any adjustment is needed at all. In other words, when A finds that B disagrees with him about p, A may lower, maintain, or increase belief credence. Related to yet distinct from credence adjustment is evidence reevaluation. Disagreeing parties may take steps to reconsider their evidence by engaging old evidence, searching for new, or revaluating the latter in light of the former. Sometimes an agent will adjust credence but not reevaluate evidence. This might occur, for instance, if practical considerations make reevaluation unfeasible. Alternatively, a subject may be moved to reevaluate her evidence while maintaining belief credence. Depending on the disagreement at hand, an agent may adjust credence, reevaluate evidence, do both, or neither. In Sect. 4, I argue that disagreement from certain types of epistemic inferiors can give superiors reason to both adjust their belief credence and reevaluate their evidence.

2.2 Competence

Let us consider an example similar to one offered in the peer disagreement literature by Elga (2007, p.486). A and B watch a horse race; A has great vision, a clear view of the finish line, and confidently judges that Mabel finished first. Nonetheless, B disagrees, insisting that Dixie won. Considering B her perceptual peer, A has good reason to pause, lower belief credence, and reconsider p. Perceptual disagreements like these are sometimes used to motivate a conciliatory position known as the equal weight view. Simplified, proponents argue that peers who disagree ought to give each other’s opinion equal epistemic weight, and thus that the reasonable response in the horse race example is judgment suspension.Footnote 3 This is intuitively compelling. A says one thing, B says the other, and they are equally likely to be right. But, what if A and B are not peers? Suppose that A has 20/20 vision, and B 10/20. A knows B’s vision is terrible: she has witnessed B mistake dogs for easy chairs and spouses for mothers. In short, she is skeptical because she has good reason to believe B is incompetent. It seems the reasonable response is to give B’s opinion little or no weight and for A to remain steadfast in her belief about Mabel’s triumph. In more general terms, when (1) A disagrees with B over p, and (2) A judges that B is incompetent in matters concerning p, then (2) justifies A remaining steadfast in her belief.

The above notwithstanding, disagreeing with inferiors is sometimes more complicated than the previous example suggests, because we sometimes disagree with competent inferiors. A may be B’s epistemic superior in subject S, but B might still be competent. Epistemic skill levels range from incompetence to expertise, so A can be close to the expertise end, while B just enters the competence range. Disagreement between superiors and inferiors becomes noteworthy when both parties are competent. Of course, drawing a line where incompetence ends and competence begins presents obvious difficulties, but for clarity, I define competence as follows:

Competence: A is competent in subject matter S when it is more likely that A forms true beliefs about S than false ones.

Likelihood is construed modally: In 100 nearby worlds, A is competent if at least 51/100 beliefs about S are true. Likelihood is construed modally: In 100 nearby worlds, A is competent if at least 51/100 beliefs about S are true. This is not a conceptual analysis of our pre-theoretical notion of competence, but a criterion to circumscribe inferior disagreements that might provide countervailing evidence. It is important to keep this in mind, for I do not want to suggest that expertise should be understood via track records. One might be tempted to think, for instance, that if a 51 % modal track record is sufficient for competence, then a 99 % modal track record makes one an expert. But the way I am understanding competence, this is not necessarily the case. Expertise will be highly contextual and will vary from subject to subject. Competence is a notion meant simply to flag that an individual’s opinion carries epistemic merit. A 51 % track record demonstrates that the individual is more likely to be correct than incorrect. This is enough to provide some minimal degree of positive epistemic support (if one thinks 51 % is not enough for minimal support, she can increase the percentage. The exact number is not critical to my argument).

Disagreement with competent inferiors has epistemic demands importantly different from disagreement with incompetents. Knowing that competent inferiors are more likely to be right about p than wrong, superiors should recognize the inferior’s disagreeing testimony as evidence. To see why, we need only imagine a case in which a superior has no opinion on p and then learns a competent inferior believes p is true. Because competent inferiors are at least minimally reliable, their claim that p gives some reason to believe p is true. Richard Feldman has noted something similar with peer disagreement:

[L]earning what a peer thinks about evidential significance when one does not have a prior opinion does provide some evidence about the topic. But if the evidence about the other person would be epistemically significant in cases in which one did not have a view of one’s own, it is hard to see why it has no impact when one does. (2010: 309)

Feldman’s quote illustrates the arbitrariness of discounting the weight of peer opinion when and only when peers disagree. This point applies just as well to competent inferiors. Both peers and competent inferiors are reliable enough that their opinion bears some evidential significance. If we had no opinion on p, reliable testimony that p gives us some reason to believe p. But then, consistency also demands that the same testimony provides reason to believe p when we already have an opinion. In order to avoid inconsistencies, superiors should consider the evidential significance of competent inferior opinion, whether or not superiors already have an opinion of their own.

I want to briefly discuss how one’s views on the Principles of Uniqueness and Independence bear on inferior disagreement. The Principle of Uniqueness claims that given a body of evidence E, and a proposition P, there is one and only one maximally rational view to take toward P.Footnote 4 Whether or not one is inclined to accept Uniqueness, the ramifications of inferior disagreement should remain the same. Suppose, for instance, that Uniqueness is correct. Further, suppose that a superior disagrees with an inferior about p. Perhaps there is only one maximally rational attitude to hold toward p. Notwithstanding, if the inferior is competent, the superior has some reason to consider the inferior’s opinion. After all, the inferior just might have that maximally rational opinion. Now suppose Uniqueness is incorrect. Once again, if a superior and a competent inferior disagree, the inferior has some reason to consider the inferior’s opinion. Although both may have rational opinions, both may not; one person may be more rational than another. Given this, the superior has reason to take into account the inferior’s opinion (or so, I will argue).

Let us move on to the Independence Principle. Consider this formulation:

Independence: In evaluating the epistemic credentials of another person’s belief about P, to determine how (if at all) to modify one’s own belief about P, one should do so in a way that is independent of the reasoning behind one’s own initial belief about P (Christensen 2009: 758).

Those who accept this principle may be especially inclined to take inferior disagreement seriously. For suppose a superior disagrees with a competent inferior. The superior knows this inferior is more likely to be correct about p than incorrect. She is also committed to bracketing her reasoning behind her belief that p. The superior would then have good reason to consider the inferior’s opinion as evidence against her position. After all, she has come across reliable testimony that speaks against the truth of her belief. Without appealing to her own reasoning process, she ought to take the inferior’s evidence as testimony against p.

However, if one rejects Independence, then she might be less inclined to carefully consider the opinion of a competent inferior. She might, for instance, make the following sort of argument: “S might disagree with me about p, but I am justified in dismissing this disagreement. After all, when considering whether p, I rejected premises which this inferior clearly accepts. Hence, there is no reason to take her seriously.” This sort of argument, however, can be made even if the disagreeing party was a peer. Problems raised by the Independence thesis are not exclusive to inferior disagreement.

2.3 Domains of Competence

When considering whether or not a disagreeing party is competent in some domain, we should categorize subject matters broadly and intuitively. For example, A is B’s superior in math and B is A’s superior in history. Subjects can be broken down further into subcategories; A may be B’s superior in calculus, while the converse is true in geometry. This intuitive notion captures the most interesting cases of inferior disagreement: those in which most casual observers would judge B the inferior of A in S. For instance, it is interesting when we think a client has better judgment than his lawyer does; part of what makes this interesting is that most casual observers would consider lawyers superior to their clients in the domain of law.

In spite of strong intuitions about superiors and inferiors, precise subject boundaries are hard to define. Some might even argue such circumscription will be so difficult that the task is meaningless. Filling in enough details in the attorney case, for instance, might lead to the conclusion that the client is the superior of his lawyer in the domain of “suing middle class, teetotaler, retired engineers.” Along these lines, some might argue that disagreements can always be circumscribed to exclude the person in the right from inferiority. And, admittedly, in most cases, we could probably finesse just such a peerage relationship. But, this also seems like cheating. Something important occurs when casual observers would label one party the inferior, the other the superior, but nonetheless suspect that the said superior ought to respect the inferior’s opinion. This is all that is needed to recognize an interesting class of inferior disagreement. Broad and intuitive criteria capture cases in which most would label A the superior, B the inferior, and yet judge that B’s opinion bears epistemic weight for A.

Intuitive domains of competence can also explain personal assessments of our own superiority. We can sometimes judge that we are generally superior in some domain, even without enough evidence to judge our superiority on an individual question. An experienced medical doctor, for instance, might be justified in believing in his superiority over medical students, in the broad category “medicinal practice.” He can make this judgment even though he cannot possibly know he is the superior of all medical students concerning all medicinal questions. This paper is concerned with the best epistemic response from someone like an experienced doctor who disagrees with someone like an inexperienced medical student. It concludes that superiors in analogous positions to the doctor often, but not always, have epistemic reason to take inferiors seriously.

3 Non-peers

3.1 Evidence

If peers are those of similar cognitive ability that hold the same or similar evidence, then non-peers are those with asymmetries in one or both of these criteria. Such asymmetries vary. Appropriate responses to inferior disagreement likewise vary, according to the type of non-peer relation. Delineating various types of non-peer relations can help us understand inferior disagreement and ideal superior response thereof. However, before analyzing non-peerage types, we should distinguish two conceptions of evidence. Some evidence is shareable and easily exchanged between epistemic parties. On the other hand, non-shareable evidence is various forms of subtle evidence not easily transferred from agent to agent; examples include unconscious beliefs, skilled interpretations, and personal memories.Footnote 5 For instance, suppose that Tony and his colleague Tim are looking for their next hire. Both have seen applicant Y’s resume, read his personal statement, and engaged Y in conversation. Tim and Tony share the same sharable evidence. However, Tony’s mafia upbringing cultivated an uncanny ability to detect dishonesty, and this leads him to doubt Y’s integrity. Tim wants to hire Y but Tony disagrees, insisting that Y is dishonest. When Tim asks why he believes this, Tony replies, “I can just tell.”

Tony does indeed have evidence of Y’s dishonesty: facial expressions and mannerisms expressed in the interview. And, in some sense, Tim at least has access to the same evidence; after all, they both witnessed the same expressions. Nonetheless, these do not serve the same evidential role in Tim’s belief forming process as in Tony’s. The latter’s special skill set enables him to interpret expressions in a way that merits evidential relevance, but this evidence is not sharable because of its dependence upon a non-transferable interpretative skill.Footnote 6 These situations are common. Persons have skill sets and histories that vary, and accordingly, so do the epistemic worth of their evidence. In this way, disagreeing parties may hold the same sharable data but retain differing non-sharable evidence.

If peerage in the actual world supervened on non-sharable evidence equality, we might never identify peers. It is rare, if not impossible, for two persons to have backgrounds and personalities so similar that their reasons for a belief include only sharable evidence.Footnote 7 Accordingly, when considering non-ideal peer disagreement, peerage should be based on sharable evidence alone. And, since it is helpful to contrast inferior and peer disagreement, these two peerage varieties should be defined according to the same conception of evidence. The best solution is to establish peer and inferior relationships based on sharable evidence.

3.2 Cognitive Ability

Cognitive ability encompasses natural skills, experiences, and physical readiness. Natural skills are raw intellectual talents and experiences the practices that develop them. Physical readiness concerns the agent’s physiological condition. For some, caffeine improves readiness while alcohol diminishes it. Extreme emotional states can also affect readiness, for instance, intellectual abilities are sometimes impaired after a death in the family, divorce, or job loss.

3.3 Superiority Relations

With that background information in place, we can define three categories of non-peer relations: strict superiority, weak evidential superiority, and weak cognitive superiority. (The reader familiar with game theory may prefer to replace “superiority” with “dominance.” I choose to use the term “superior,” because that fits with the terminology of peers in terms of “inferiors” and “superiors”).Footnote 8

Strict Superiority: Superior A is strictly superior to inferior B just in case A is relatively superior in respect to both evidence and cognitive ability. (See Model 1).

Model 1: Strict Superiority.

 

A

B

Ev

8

6

Ca

8

6

Weak Evidential Superiority: Superior A has weak evidential superiority over inferior B, just in case (1) A is relatively superior in respect to evidence, and (2) A is equal to B in respect to cognitive ability (see Model 2).

Model 2: Weak Evidential Superiority.

 

A

B

Ev

8

6

Ca

8

8

Weak Cognitive Superiority: Superior A has weak cognitive superiority over inferior B just in case (1) A is relatively superior in respect to cognitive ability, and (2) A is equal to B in respect to evidence (see Model 3).

Model 3: Weak Cognitive Superiority.

 

A

B

Ev

8

8

Ca

8

6

For simplicity, we can think of evidence and cognitive abilities on a ten-point scale: One represents complete incompetence, five, minimal competence, and ten perfect competence.

4 Defeating Disagreements

Defining competence and categories of dominance relations helps us see how inferior disagreement can have epistemic significance for superiors. In certain cases, superiors have epistemic reason to respond to inferior disagreement by lowering belief credence, reevaluating evidence, or both. This occurs with all three superiority relations. Each is discussed in turn.

4.1 Weak Evidential Superiority

Suppose A is B’s weak evidential superior as shown in Model 4.

Model 4

 

A

B

Ev

8

7

Ca

8

8

A is evidentially superior to B; however, both are competent. Suppose that A and B disagree about p. In some cases, B’s disagreement legitimizes credence revision and evidence reevaluation in spite of A’s superiority. But, at least one contingency must hold: A cannot account for the disagreement via her disagreeing party’s evidential inferiority. Subjectively, superiors have epistemic grounds to consider inferiors seriously if, from their perspective, their own superiority cannot explain the dispute. Objectively, superiors have epistemic grounds to consider inferiors seriously if, in actuality, the superior’s superiority does not explain the dispute. Sometimes superiors cannot subjectively account for disagreements in terms of their own superiority, even though this is possible objectively. For instance, from a God’s eye viewpoint, A’s superior evidential position might explain why he disagrees with B. Nonetheless, it is possible that from A’s evidential perspective, he cannot justifiably come to the same conclusion. Kelly (2010) has discussed something similar with peer disagreement:

One should give some weight to one’s peer’s opinion even when from the God’s eye point of view one has evaluated the evidence correctly and he has not. But why? Exactly because one does not occupy the God’s eye point of view with respect to the question of who has evaluated the evidence correctly and who has not (p. 57).

Because none of us is omniscient, we sometimes have superior justification without our own awareness. But, in other cases, we do have enough information to conclude that our own superiority can account for a disagreement with an inferior. Fumerton (2010) has touched on these cases, cases in which we might justifiably dismiss disagreements because we can see where our interlocutor went wrong. We might even recognize that if she could see things from our full perspective, the disagreement would vanish:

I’ve been told that our next department meeting will be this Friday at 3:00 p.m., and on the basis of this information take myself to have good reason to believe that the meeting will be at 3:00. Diane believes that the meeting is scheduled for Thursday at 7:00 a.m. (something about which she is vociferously complaining). But I also have evidence that another of my colleagues has played a practical joke on Diane and has deliberately misinformed her as to the time of the meeting. Diane and I disagree…But I have evidence that she lacks, and my additional evidence allows me to see the way in which Diane’s evidence is, in a sense, defective…Diane herself would regard her evidence as defeated should she acquire the additional information that I possess. (p. 5)

Fumerton’s example illustrates a case in which an evidential superior can justifiably dismiss an inferior simply because evidential asymmetries are both apparent and explanatory. In Fumerton’s example, the superior recognizes that the evidence his inferior lacks accounts for their disagreement. Similar situations occur in many other instances. Suppose two philosophy professors, Smith and Jones, disagree about whether to admit graduate applicant G. Smith has all the same evidence as Jones, but in addition, Smith has read a recommendation letter that reveals damaging information about G. Smith can dismiss Jones’s disagreement, for he has reason to believe that his own evidential superiority accounts for their dispute. If Jones read the additional letter, Smith believes Jones would likely change her opinion. That being said, weak evidential dominants are not always as fortunate as Smith. Sometimes evidential superiors are unable to account for the relevant disagreement via their own superiority. If they cannot account for this, they ought to lower belief credence and reevaluate evidence in response to inferior disagreement. See RECRUIT DISPUTE.

RECRUIT DISPUTE: Dixon and Spade are hockey recruiters for Port Chester University. Both have closely followed the Colombia Comets, in particular, athletes Conner and Warren. Dixon and Spade have equal cognitive abilities; however, Dixon has viewed all eight games, but Spade only seven. Dixon is Spade’s weak evidential dominant, in accordance with Model 4 (above). Port Chester can only afford one full scholarship. Dixon thinks it should go to Conner but Spade favors Warren.

We can suppose that Dixon cannot account for the disagreement via Spade’s missed game.Footnote 9 In other words, unlike with Jones and Smith, nothing stands out in Dixon’s extra evidence that can explain the disagreement. In the game that Spade missed, Warren did not make any major mistakes and Conner did not play better than usual. Because Dixon cannot account for the disagreement though his own superiority, he ought to reevaluate his evidence and lower his credence that Conner is the better recruit. Even though Dixon has more evidence overall, this evidence might be irrelevant. It would be epistemically irresponsible for Dixon to dismiss the disagreement. After all, for all Dixon knows Spade’s disagreement could be based on one of the many games both he and Spade had viewed with equal circumspect. In all situations such as these, situations in which an evidential dominant is unsure his dominance explains the relevant disagreement, superiors have epistemic reason to take inferior opinion seriously.

4.2 Weak Cognitive Superiority

The epistemic demands of weakly cognitive dominants are analogous to those of weakly evidential dominants. Consider the following dominance model:

Model 5 (as in Model 3, this is a case of weak cognitive dominance)

 

A

B

Ev

8

8

Ca

8

6

Suppose cognitive superior A disagrees with inferior B. As with evidential dominance, B’s disagreement sometimes legitimizes A lowering her credence and reevaluating evidence, but once again, an important contingency must hold: A cannot account for the disagreement via her own cognitive superiority. Like instances of evidential dominance, cognitively dominant superiors can sometimes account for the relevant disagreement in terms of their adversary’s inferiority and can then justifiably dismiss inferior opinion. For instance, suppose two English majors, Amanda and Benjie, disagree about whether a sentence demands the use of “that” or “which.” Both are competent grammarians; however, Amanda is Benjie’s superior and can immediately recognize that Benjie makes a common restrictive clause error. Because Amanda can account for the disagreement via her own superiority, she can justifiably dismiss the epistemic weight of her inferior’s disagreement. Sometimes, however, disagreements fit Model 5, yet cognitive inequalities are not explanatory. See SMART STUDENT.

SMART STUDENT: Professor Patterson composes a model essay for an introductory course. Patterson is superior to his students in both writing and grammatical skill. After class, one smart student taps him on the shoulder and says, “By the way, page two has an error; you used a comma instead of a semi-colon.” Patterson doubts the student’s claim. After all, he had just read the paper several times, leaving convinced of its grammatical perfection.

How should Patterson respond to his student’s criticism? Immediately dismissing the opinion seems too quick, for even after triple checking, grammatical errors are easily overlooked. Moreover, recognizing the error falls within the student’s capability. Paterson’s belief in his own fallibility, combined with his belief in his student’s competence, suggests that the epistemically appropriate action is for Patterson to reduce belief credence and reassess evidence. Unlike Amanda, Patterson cannot identify where or if his inferior has gone wrong. Because he cannot account for his student’s differing opinion via an appeal to his own superiority, Patterson cannot justifiably dismiss its epistemic import.

To fully understand the epistemic lessons from SMART STUDENT, it is important to distinguish the reasonableness of Patterson’s original belief from the reasonableness of his reaction. Maybe the student was wrong and Patterson’s original belief was reasonable, but this does not make Patterson’s reaction reasonable. For instance, when discussing an example with important similarities to SMART STUDENT, Frances (2014) has noted the ambiguity of the question, “Is S rational?” He asks, “The statement ‘Bub is being rational’ is ambiguous: is it saying that Bub’s retained belief J is rational or is it saying that Bub’s retaining of that belief was rational?”

To apply Frances’s point, it may be the case that Patterson’s retained belief about the comma usage is rational.Footnote 10 Notwithstanding, Patterson’s retaining of the belief might be irrational, simply because Patterson’s own perspective gives him no reason to dismiss his student’s opinion. Because Patterson has no reason to think his own superiority explains his student’s objection, credence reduction is appropriate. Notwithstanding, it does seem reasonable that Patterson should downgrade less than he would in response to peer disagreement: Colleagues who pointed out the error should cause greater doubt than his student does. But, unlike credence adjustment, evidence reevaluation should be equally rigorous in response to this inferior disagreement as it would be in response to peer disagreement. In both cases, reconsidering consists of double-checking the purported error. Double-checking is double-checking, regardless of who motivated the subject to seek reassurance.

4.3 Strict Superiority

Finally, let us consider disagreement from a strictly dominant superior’s perspective, as in Model 6.

Model 6

 

A

B

Ev

8

5

Ca

8

5

Similar to what was said before, superior A is free to disregard B’s disagreement if she can account for it through either her cognitive or evidential dominance. However, cases will remain in which strictly superior parties cannot account for disagreements via their own superiority.

NO EXCUSES: A professional philosopher, Quinn, writes a paper on moral excuses. So far, she has devised the following scheme: Excuses are justified in situations X and Y, but not T and Z, for reasons R. She has asked several colleagues about her theory; many agreed and many disagreed. Having considered all these opinions, she is finally ready to publish. But, she first shows the paper to her sister, a non-philosopher, Schneider, who insists moral excuses are justified in situations X alone for reasons R’.

Despite Quinn’s superiority, it seems she has reason to take Schneider’s opinion seriously. As shown in Model 6, we are assuming Schneider’s competence. This may be controversial, but philosophical analysis involves, in part, coming up with reasons that explain various phenomena. Many non-philosophers are competent in this respect. This does not mean that they are likely to form true beliefs about what actually explains some phenomenon; even the best philosophers can fail to meet that standard. But, nevertheless, non-philosophers might still form true beliefs that X can explain phenomenon Y. Non-philosophers can be reliable in this sense even if less so than philosophers. So not only might Schneider’s opinion be reliable enough to carry epistemic significance, her disagreement is arguably more powerful than peer disagreement. Quinn’s peers have similar perspectives because they have had similar educational backgrounds and work in similar environments. In short, their insight brings nothing new. But, Schneider’s perspective is unlike Quinn’s usual associates and can bring to light factors Quinn overlooked. Here is another way to look at things: Schneider’s background makes it likely that she possesses some non-sharable evidence that Quinn lacks.

We come to beliefs, in part because of evidence, in part because of cognitive abilities, but also in part because of psychological particularities. Two individuals that viewed the same event might offer radically different descriptions of the occurrence, yet both may be equally accurate and someone wishing for the most complete picture will do well to consider each account. Parts of the varying stories may be mutually exclusive, other parts compatible. Explaining the differences between contrasting perspectives is difficult because complete reasons for belief are usually inaccessible. Sosa (2010) has made this point effectively: “Our basis for believing generally fails to be fully formed and operative in one fell swoop . . . A belief forms in us over time through the subtle influence of diverse sources”10 (p. 294). As Sosa suggests, it is hard to say exactly why one holds the beliefs one does, but we can reasonably assume personal perspective affects belief.Footnote 11 Regarding the disagreement between Quinn and Schneider, the former should make epistemic revisions in response to the latter because (1) Schneider is competent, and (2) Schneider’s insight is unlike Quinn’s and her peers. This unique insight can mean Schneider has reasons for her contrasting belief that Quinn cannot fully comprehend. In order to take account of these reasons, Quinn should both adjust credence and reevaluate evidence in response to her sister’s disagreement:

Credence revision: Quinn should lower her confidence that her theory best explains moral excuses, and do so even more than she would in response to a peer.

Evidene reevaluation: Quinn should reevaluate the reasons for her theoretical stance; her efforts should be greater than similar efforts motivated by peer disagreement.

5 Objections

5.1 Owing to Inferiority

Some might agree that Quinn should consider Schneider’s opinion seriously, but nonetheless argue that this has nothing to do with her inferiority. In other words, it is not in virtue of Schneider’s inferiority that Quinn should take Schneider seriously. For example, one might think it is in virtue of Quinn’s creativity and non-conformity that Schneider has reason to take her opinion seriously. There is a sense in which this is correct; it is plausible to suppose these qualities play an important role in my example. However, recall that we are thinking of the inferior/superior divide as contextualized to a subject matter. It might be the case that an individual who is considered “superior” according to common standards is less creative than an individual considered “inferior.” A first-year philosophy major, for instance, may be more creative than her philosophy professor. I do not think this is enough to say the student is a superior philosopher. I do think it is enough to give the professor reason to take the opinion of the student seriously, which is preciously my point. Although many people would consider the philosophy professor the superior, she has reason to take the philosophical opinion of her inferior (the student) seriously.

Now, of course there is a way we might construe the above case so that the student is not inferior after all. I do not think it is wrong to do that, but I think it misses something important. Most people consider a philosophy professor the superior of her student, in the domain of philosophy. Nonetheless, the student’s opinion is epistemically relevant for the professor. So hence, there are cases where, under one intuitive understanding of the superior/inferior relation, superiors have reason to take their inferior seriously.

Another point to consider is that if it were not for Schneider’s inferiority, she would lack the relevant insight. If Schneider had attended graduate school and worked in academia, her valuable perspective would likely be lost. Schneider’s opinion is noteworthy because her education and cognitive abilities contribute to a psychological perspective with less expertise yet notable insight. In other words, the epistemic value of the disagreeing party’s opinion specially connects to her inferiority. For even if we were to say it is ultimately creativity and non-conformity that contribute to Schneider’s difference of opinion, it is likely that her life outside of academia contributed to these character traits. And, since the status of “non-academic” is what leads many people to consider Schneider an inferior, these qualities are specially related to aspects of her inferiority.

5.2 Peer Relations

Another objection concerns categorizing peer relations. Some might say that SMART STUDENT and NO EXCUSES suggest that inferior opinion is relevant according to broadly defined superiority relations. But, we should define things more narrowly, and when we do, all of the previous examples are instances of peer disagreement, not inferior disagreement. For instance, one might argue that in SMART STUDENT, the student is the professor’s peer in recognizing semi-colon mistakes on papers written by the professor. Because we are all prone to overlook our own mistakes, the professor’s usual competence level should be adjusted down to match his student. Similarly, when the other examples are properly circumscribed, the party in the right is most aptly labeled the superior.

We can call the above the “particularizing peerhood” objection, for its main force comes from the suggestion that we should particularize judgments of peerhood to particular questions. Many difficulties arise with such conception. First, particularizing peerhood would lead us to dismiss both interesting cases of what we would intuitively call peer disagreement and cases we would intuitively judge as inferior disagreement. Concerning the former, particularized peerage would result in very few cases of peer disagreement at all. Two persons are rarely of equal abilities regarding an individual question, for one party’s history or cognitive skills will have advantages over the other. Although some peer disagreement may remain, the set would be significantly smaller than intuitions suggest. Minor evidential or cognitive advantages would nearly always result in considering one party the superior over the other.

In addition to unreasonably limiting the scope of peerhood, particularizing peerage would lead us to forgo many common sense judgments of superior/inferior relationships. For instance, we would lose ground to call history professors the superiors of their students in the domain of history. If superiority relations must be defined according to individual questions, it would make no sense to say something like, “History professors are in a better epistemic position regarding history than their students are.” After all, it is surely possible that some students happen to be superior to their professor regarding some particular historical inquiry. We would then have to define peer relations in terms of each individual history question, rather than the in terms of the much more general domain of what we might call “historical studies.”

It seems that particularizing peerhood throws away too much. Everyday categories of superiority relations help us navigate the practical world. In order to increase justified beliefs, we seek expert testimony and avoid advice from amateurs. Laypersons, for instance, cannot know if every lawyer is their superior in every legal nuance. Nonetheless, non-lawyers seek advice from lawyers because the latter are the general superior in law. If there were no general superiors, there would be no reason for non-lawyers to seek professional legal advice. But, it seems that there is such reason, which explains why we should understand superiority relations in terms of subject matter rather than individual questions.

It is through the lens of such general peerhood categories that we ought to look at SMART STUDENT and NO EXCUSES. These examples show that even in cases where disagreeing parties are aptly categorized inferior, their opinion might be epistemically relevant. We can even place the chosen examples aside. What matters is that it is conceptually possible for superiors and inferiors to disagree when the superior’s opinion is mistaken. If this much is possible, you can disagree that the given examples are inferior disagreements while still agree with the more important claim: at times, superior parties have epistemic cause to adjust credence and reevaluate evidence in response to inferior disagreement.

5.3 Epistemic Pushovers

Those sympathetic-to-steadfast sentiments might argue that what was said so far glorifies epistemic pushovers. If we are required to lower belief credence when probed by inferiors, we might never hold any beliefs at all. However, this is a misguided suggestion. Nothing said thus far requires epistemic superiors cower to inferior opinion, because generally, taking inferiors seriously does not require taking them as seriously as peers. Indeed, most of the time, inferior disagreement should be taken less seriously. This was the case in NO EXCUSES and RECRUIT DISPUTE. Professor Patterson ought to have taken his student’s opinion seriously, but less seriously than a peer. Similarly, Coach Dixon had reason to take Spade’s opinion seriously, although his consideration might had been greater if Spade were his evidential equal. Admittedly, in some special cases, contextual considerations can merit taking inferior opinion as seriously as or even more seriously than peer opinion (NO EXCUSES). However, such instances are the exception, not the rule. Furthermore, all disagreements discussed thus far were between superiors and competent inferiors. But, when disagreeing interlocutors are more likely to get things wrong than right, when inferiors are incompetent, inferior disagreement is not defeating and superiors are anything but pushovers: In these cases, superiors can indeed simply dismiss the inferior’s opinion.

6 Defeating Agreement

6.1 The Problem of Inferior Agreement

Although inferior disagreement sometimes gives reason to lower credence or reevaluate evidence, many times it does not. When we disagree with those who are incompetent, those more likely to get things wrong than right, we have no reason to doubt our belief nor to reexamine evidence. Since incompetent inferiors are unreliable, their disagreement, if anything, might justifiably boost our belief credence. Conversely, agreement from incompetent inferiors can be epistemically worrisome. In The Critic as Artist, Oscar Wilde’s (2000) Gilbert exclaims, “Ah! Don’t say that you agree with me. When people agree with me I always feel I must be wrong” (p. 292). Gilbert may take things too far, but most of us have felt similar sentiments. The uneasy feeling when someone who we believe to be our inferior enthusiastically asserts his agreement. We find ourselves thinking, “Oh no, if that person thinks I am right, then surely I have gone wrong!”

Suppose that A believes p. P falls under domain S, and A is competent in S. A then learns that B agrees; he also believes p. The problem for A is that B is incompetent. In the past, he has consistently arrived at false beliefs about S. It seems that B’s unreliable record renders his agreement positive evidence against p and that A might lower his credence in response. This is reasonable. If an agreeing party is incompetent, his agreement can be considered higher-order evidence, evidence about one’s competence in interpreting evidence.Footnote 12 Agreement from incompetent inferiors is higher-order evidence that our belief-forming process is less than optimal, that somehow or someway we have gone wrong in our assessment. This gives us reason to lower our belief credence and reevaluate our evidence.

When considering incompetent inferiors, however, we must draw certain boundaries, lest we end up similar to Wilde’s Gilbert and take every such instance as defeating evidence. This would be a mistake, because upon reflection, not every case of inferior agreement is higher-order evidence or any evidence at all. If an inferior’s incompetence is arbitrary, inferior agreement speaks neither for nor against the superior’s relevant belief or belief forming process. On the other hand, incompetent inferiors who consistently generate false beliefs might offer defeating evidence. We can refer to the latter as epistemic maliciousness, the former epistemic inadequacy, and explain their differences probabilistically. Suppose someone is presented with two propositions, p1, which is true, and p2, which is false. Given only these choices, epistemically inadequate agents are just as likely to believe p1 as p2; inadequate inferiors track neither truth nor falsity. Their opinion is completely arbitrary. On the other hand, epistemically malicious agents are more likely to believe p2 (the false belief) than p1 (the true belief). Unlike the epistemically inadequate, the epistemically malicious do not have arbitrary opinions: Malicious opinion tracks falsity.

6.2 Inadequate Inferiors

Inadequate inferiors can be divided into two classes that circumscribe two different ways in which disagreeing parties might be arbitrarily wrong. Some inadequate inferiors are exhausting personalities who talk just to talk, and occasionally say something true. In other terms, these inadequate inferiors spew bullshit. According to Frankfurt and Wilson (2005), the essence of bullshit is a “lack of connection to a concern with truth… [an] indifference to how things really are…” (p.33). When bullshitters say something false, it is not because they try to deceive but that they are disinterested in truth. Bullshitters sometimes speak the truth, but they can be just as likely to speak falsely. If an epistemic superior believes p and a bullshitting inferior claims not-p, the superior has little reason to adjust credence either one way or the other. As Frankfurt and Wilson (2005) reminds us, “The bullshitter is faking things. But this does not mean that he necessarily gets them wrong” (p. 48). For all the superior knows, the inferior bullshitter could be right, but he could just as easily be wrong.

Ignorant inadequates, unlike bullshitting inadequates, do have a concern for truth. However, their terrible ignorance of the relevant subject makes any opinion a wild guess. For instance, suppose you ask your friend about the medical severity of your scratchy throat. However, your friend is not only your inferior but incompetent.Footnote 13 Because he is not a medical doctor, but a meta-ethicist, his confidence in your health should offer no assurance. However, neither is his confidence reason for concern; it is simply uninformative. What ignorant and bullshitting inadequates have in common is opinions so arbitrary they are irrelevant.

At times it may be unclear whether an inadequate inferior is a bullshitter or simply ignorant. Consider political pundits, the ones often featured on afternoon US cable television shows. These pundits are often asked to predict political events. While extremely confident in their assertions, many such pundits have terrible track records. I tend to think that most are not lying or bullshitting, rather, they honestly believe what they say, but are simply incompetent. Regardless of whether these pundits are ignorant or bullshitting, what they have in common, and what is common to all inadequates, is that their opinion is analogous to flipping a coin, shaking a magic eight ball, or consulting a psychic. My magic eight ball’s testimony that p gives no epistemic reason to so believe, but neither does it give reason for disbelief. Likewise, agreement from inadequate inferiors gives us no reason for nor against belief.

Everything said about agreeing inadequates, mutatis mutandis, applies to disagreement. The arbitrariness of inadequate inferiors’ opinion means that neither their agreement nor disagreement justifies credence revision or evidential reassessment. We cannot, however, say the same about malicious inferiors.

6.3 Malicious Inferiors

Malicious inferiors track not truth but falsity. This falsity tracking might prove epistemically helpful when correctly identified. If A is aware of B’s record of falsity tracking, A can take B’s opinion that p as evidence that not-p. On the other hand, B’s opinion is misleading if A mistakenly considers B competent. In such epistemically dangerous situations, A may consider B’s opinion evidence for p when he ought to consider it the opposite. In this way, malicious agents contribute to the acquisition of false beliefs or unjustified credence revision.

Systematic and consistently misleading testimony of malicious inferiors is often due to bias or false higher-order beliefs. Because of such mistakes and biases, malicious inferiors can be even less reliable than the ignorant. Consider someone who rejects evolutionary theory and accepts creationism; this creationist may be more likely to make false claims about human development than someone who has hardly reflected on the matter at all, for the ignorant might entertain the truth of claims that the creationist necessarily rejects. Because of his open-mindedness, the person completely ignorant of scientific theories might hold a few true beliefs via sheer chance. Hence, the creationist claim that p can speak more strongly against p’s truth than the claim of someone more ignorant but less biased: biases can generate reliability records worse than chance.

We can reach similar conclusions about malicious agents who have systematic biases against certain races, classes, or ideologies. Prejudice begets malicious inadequacy. Imagine superior A disagrees with inferior B over the merit of initiative I. Moreover, B is a Klansman. A could learn more from this disagreement than he could from inadequate inferiors. Because B’s prejudice leads him to consistently support initiatives A opposes and oppose initiatives A supports, when A learns of B’s dissent, A should be reassured that he is taking the correct stance. A might justifiably think to himself, “I am glad that Klansman disagrees. I must be on the right side of things.” On the other hand, A would learn little from inadequate inferiors. For instance, disagreement from low-information voters with arbitrary political positions would speak neither for nor against the merit of A’s stance on I. Footnote 14

Rather than a magic eight ball or a coin flip, eliciting agreement from malicious epistemic agents is like consulting the satirical newspaper The Onion for information on current events.Footnote 15 Suppose you believe that rabid zebras killed two zookeepers, but you are unsure where you heard the story. You then pick up The Onion and find that very zebra story on the front page. This provides reason to doubt your original belief: The Onion reporting story X is evidence that not-X. Since satirical news publications consistently and as a matter of practice publish false stories, an article on event E in such publication is evidence E never happened. In the same way, we must be wary of agreement from inferiors whose epistemic merits are comparable to The Onion.

Reading the satirical report on killer zebras is higher-order evidence that your belief-forming process has gone awry; this gives reason to adjust credence and reassess evidence. Likewise, agreeing inferiors who have a consistent record of inaccuracy can give us reason to doubt and reevaluate. If an evolutionary theorist learns that a creationist agrees with his latest hypothesis, he might have reason for worry. This applies to many similar instances when we know agreeing parties are biased or misinformed. If I find out that my latest moral theory receives widespread support from the morally depraved, I would be wise to reevaluate.

7 Epistemic Opportunity Costs

If we were only concerned with disagreement in an ideal world with unlimited time and resources, it would make little difference whether disagreeing parties are peers, superiors, or inferiors. At least it would make little difference for evidence reevaluation. In this ideal world, we should reinvestigate every belief held with less than absolute certainty for there would be nothing to lose pursing not-p possibilities ad infinitum. But, in the actual world, disagreements are between those with limited time, money, and intellectual energy. This scarcity of epistemic resources gives superiors strong reason to consider epistemic opportunity cost, defined as followsFootnote 16:

Epistemic Opportunity Cost: epistemic goods sacrificed while pursuing other epistemic goods.

Reconsidering any belief is senseless unless the epistemic benefit of holding the true belief is greater than the epistemic cost of reconsideration. Sometimes acquiring a true belief is just too costly: Resources are better used on other epistemic activities. For instance, suppose A and B disagree about the proper amount for a tip. A is sure it is $47, but B, her competent epistemic inferior, disagrees arguing it should be $48. Recognizing the reasonable possibility that B is right, this would often be enough to motivate A to recalculate. However, suppose that as soon as A completed her calculation, she began reading the Treaty of Versailles. A should lower her belief credence that the proper tip amount is $47, but pass on reconsideration because recalculating would cut into her historical scholarship. Things also work the other way around. Sometimes we have little faith in our disagreeing party’s competence, but consideration costs are cheap enough to make epistemic sense. B may believe it is unlikely that his epistemic inferior, C, is correct. But, when the cost is low, reconsidering might still be worthwhile. B can be confident that the sun is shining; nonetheless, he may glance out the window when C disagrees. Reconsideration costs are so low he has nothing to lose.

8 Conclusion

The epistemology of disagreement literature has been overwhelmingly focused on peer disagreement. This is not without cause. Intuitively, disagreement from those of similar epistemic merit is especially puzzling; it seems hard to admit that others are just as qualified as we are and simultaneously ignore their dissenting viewpoints. Nevertheless, instances of inferior disagreement have an epistemic significance of their own.

Dismissing disagreement from epistemic inferiors might seem prima facie justified, but this is incomplete and misleading. Disagreement from inferiors who are competent does have epistemic import. And, sometimes, even an incompetent inferior’s opinion is epistemically significant. Disagreement from malicious inferiors, inferiors that make systematic mistakes, may justify credence boosting and agreement credence reduction. On the other hand, inadequate inferiors have an arbitrary track record and so their agreement or disagreement is irrelevant. Assessing the appropriate epistemic response when you suspect you disagree with an inferior requires asking first: “Is the inferior competent?” If yes, we should prepare to lower credence or reevaluate evidence. If no, we must determine whether the disagreeing party is malicious or inadequate. Agreement from malicious inferiors should reduce confidence, and disagreement should have the opposite effect. But, if an inferior is merely inadequate, we can dismiss the opinion without further evaluation.