The School of Athens, Fresco, Raphael, 1509-11
Suppose we assume, roughly, that epistemic peerhood is a symmetrical relation between any persons A and B such that, given each person’s education, experience, intelligence, background information, intellectual honesty, fairmindedness, due diligence, critical examination, general willingness to follow the evidence where it leads, etc., with respect to some topic t, A’s intellectual reliability concerning the truth and rationality of t is commensurate with B’s and vice versa. [1]
Suppose that A and B are epistemic peers concerning proposition p and yet disagree about the truth of p. A affirms (i.e., believes) that p and B denies (i.e., disbelieves) that p. [2] What should A and B do?
The literature on this issue typically recognizes (at least) two options: conciliation and steadfastness. [3] In situations of peer disagreement, the thoroughgoing conciliationist abandons his position on p and perhaps remains uncommitted about whether or not p. (There are less extreme versions of conciliationism that permit one to keep one’s belief while decreasing one’s confidence in it. More on this option below.) Moreover, he holds that this is what epistemically ought to be done in such circumstances, given the peer disagreement. In contrast, the total steadfaster retains his position on p despite the disagreement, perhaps downgrading his estimate of his peer’s epistemic reliability – which raises the question of whether or not A and B are epistemic peers in the first place. Anyhow, the steadfaster avers that peers ought to retain their beliefs in such cases. Conciliationism might be motivated primarily by intellectual humility and steadfastness by self-confidence (proper or improper) and courage of conviction.
The dichotomy between thoroughgoing conciliation and total steadfastness appears to be a false dilemma. There is a via media between the two sides – or so it seems to me. I will articulate this middle path below. But first, let’s identify some important prefatory matters.
Plausibly, what one epistemically ought to do in cases of peer disagreement depends on several factors such as (i) the context of the disagreement about p, (ii) the kind of proposition p is [4], and (iii) the goal of affirming or denying p. Suppose in Case 1 that the context of disagreement about p is a matter of high epistemic importance, p is the kind of proposition which is not knowable with epistemic certainty by human beings, and the goal of affirming or denying p is the acquisition of a reasonable belief (assume that p is some difficult matter of metaphysics – say, that human beings have libertarian free will (LFW) – and that the proposition has a considerable degree of practical encroachment on one’s life). Now suppose in Case 2 that the context of disagreement about p is a matter of low epistemic importance, p is the kind of proposition that is knowable with epistemic certainty by human beings, and the goal of affirming or denying p is the acquisition of certain knowledge (say p is some proposition of concerning long division in arithmetic and p has relatively low practical encroachment on one’s life, such as a precisely accurate conclusion about the tip one desires to leave for a server after a meal at a local eatery.) It is intuitively plausible to hold that what one epistemically ought to do in Case 1 might well differ from what one should do in Case 2.
Now, it seems to me that there is a tertium quid, that is, a middle course between thoroughgoing conciliation and total steadfastness. In some cases, one can lower one’s credence (i.e., degree of (subjective) confidence) with respect to one’s belief that p or that ~p and yet retain one’s belief. [5] The disagreement of one’s epistemic peer is, in a sense, a bit of counter-evidence to one’s position.
In what cases might the middle way be fitting? Take Case 1. Here, it seems reasonable to say that if there is a disagreement between epistemic peers concerning whether or not humans have LFW, the peers don’t need wholly to conciliate, nor should they remain absolutely steadfast. Rather, each might reduce his level of confidence that p or that ~p. Suppose that, prior to disagreement, A affirms p with a credence of .8, and B denies p with a credence of .8. The middle path suggests that both A and B should reduce their credence levels, perhaps to something around, say, .6. In this case, they keep the beliefs that they had before the disagreement, thus respecting themselves as reliable thinkers, and yet they recognize and respect each other as dependable as well, which is demonstrated in their intellectual humility and mutual reductions of confidence levels.
Cases of epistemic certainty might require a different approach. Take Case 2. Suppose that A and B disagree. A checks his long division by hand twice and then confirms a third time with a pocket calculator. It seems that A should remain steadfast in his belief about the correct answer to the division problem even if B disagrees. A might consider reasonable explanations for why it seems that B is wrong, but arguably A is justified in continuing to believe with high confidence that his answer is the correct one.
This middle path might go by either of two names depending on one’s emphasis. We might dub the path moderate conciliationism, according to which each peer reduces his confidence in p but may continue to believe that p if the total evidence seems to him sufficient. Regarding this name, the emphasis is placed on reducing one’s credence level. Alternatively, we might call this path moderate steadfastism, according to which each peer reduces his confidence in p but may continue to believe it if the total evidence seems sufficient. Here, the emphasis is placed on the peer’s continuing to believe that p.
Question: have I forged one middle path or two? In other words, is there a clear difference between moderate conciliationism and moderate steadfastism?
Another Question: suppose that before the disagreement, you estimate your view and your peer’s view as roughly equally reliable. But suppose also that you justifiably place greater confidence in your estimate of your own reliability and less confidence in your estimate of your peer’s reliability. After all, you know yourself better than you know your peer. How does this factor influence our reflection on what one ought to do in cases of peer disagreement?
A third set of questions: Are there precise epistemic peers? Is epistemic peerhood an imprecise relation? If it’s an imprecise relation, how relatively precise must the relation be to count as a peerhood relation? Is there a sorites problem lurking here? (These are ontological questions.) If peerhood is a precise relation, are we capable of recognizing such relations? What are the standards for achieving such recognition? (These are epistemological questions.)
There is much more to say about this topic, but I must leave the dialectic here for now.
[1] Peerhood can extend beyond two persons, but for the sake of time and space, I’ll limit the discussion to two. Moreover, peers can be competent to some degree, but also can be incompetent to some degree. We might think of a range of competence between wise and foolish. Peerhood can be present at any point on the spectrum.
[2] It seems that there are three doxastic attitudes one can take to a proposition: one either believes it (i.e., takes it to be true), disbelieves it (takes it to be false), or remains uncommitted to its truth value. Maybe there is a fourth option, which is something like practical acceptance: one doesn’t quite believe that p, and yet one doesn’t disbelieve that p, but one is not neutral either. Instead, for practical purposes, one accepts that p (i.e., for practical purposes, acts as if p is the case though one is not prepared to claim that p is the case).
[3] Sometimes, the dichotomy is said to be between equal-weight views and steadfast views such that the equal-weight approach is conciliatory while the steadfast view is not conciliatory.
[4] For example: What is the total evidence for and against p? Is p prima facie plausible? Implausible? Can p be known with epistemic certainty? Is it merely a matter of probable belief?
[5] In epistemology, credence is a particular degree of confidence on the [0,1] scale of subjective probability one places in one’s belief that p. For example, take the proposition: “The Lakers will win the game tomorrow.” Jones might believe that proposition, but only with a degree of confidence at .65. In this post, I set aside ontological questions about whether both beliefs and credences exist.
There are at least three conceptions of probability to consider here: (a) epistemic probability, (b) objective probability, and (c) subjective probability. To elaborate: (a) is a matter of p’s being probable to some degree on the [0,1] scale of non-subjective probability, given the support for p of relevant evidence e. And (b) is, like (a), an aspect of reality which holds regardless of what anyone thinks about the matter; in other words, the objective probability of an event occurring is what it is independently of what we might believe, know, think about, or desire. The concept of objective probability is sometimes described in terms of the frequency with which an event occurs. (This might be called the frequentist theory of objective probability.) Alternatively, objective probability might be characterized as a matter of a physical object’s mind-independent tendencies or propensities to behave in particular ways under specific conditions over time. (This might be called the propensity theory of objective probability.) For example, a baseball player’s batting average or a basketball player’s free-throw shooting percentage might be construed as matters of objective probability in the frequentist sense. In contrast, a fair coin toss might be construed as a matter of objective probability in the sense of propensities; i.e., the physical characteristics of a genuine coin are such that if the coin is tossed fairly, it will land according to a probability fixed by its physical characteristics. In addition to objective probability and epistemic probability, there is another type of probability which epistemologists sometimes call “credence,” “subjective probability,” or “degree of confidence.” This type of probability is subjective since it depends on the subject’s mind-dependent level of confidence in a proposition that the subject believes. Such confidence is relative to the person who has it and thus can vary from person to person. Ideally, one’s degree of subjective probability or credence aligns with one’s degree of epistemic probability.