1. Debunking Reasoning

Debunking reasoning plays a role in recent debates over the epistemic significance of disagreement. Everyone agrees that learning of disagreement is evidence that may undermine our views. And everyone agrees that reasonably attributing biases to disagreeing thinkers lets us resist a downward push in our confidence in our views. But conciliationists and nonconciliationists have different views about when debunking is appropriate.

For example, David Christensen, a conciliationist, thinks that recognized peer disagreement always provides a reason to reduce confidence in one's own views substantially, unless one has an independent reason to downgrade the testimonial evidence one has received. An independent reason must not depend on the original reasoning used to reach one's own view (Reference Christensen2007, Reference Christensen2011). This ‘independence’ constraint rules out reasoning of the following form: ‘Well, McCoy disagrees with me about p. But p is true. So McCoy is wrong. And so I do not need to take her disagreement as a reason to doubt my view.’ But an independent reason to think a peer is biased lets us debunk the peer's testimonial evidence, thinks Christensen (Reference Christensen and Lackey2014).

But debunking is not the exclusive practice of conciliationists. Nonconciliationists Michael Bergmann and Richard Fumerton have also suggested that we can downgrade a disagreeing peer's testimony if we have reason to regard her as tending to err. Bergmann's idea is that if a thinker has in hand an explanation for why the peer has made an error—because the peer is biased in this particular case, for instance—then the explanation can preserve rational belief in the face of disagreement (2009). And here is what Fumerton says:

By taking others to be biased, Fumerton thinks he may ‘sometimes discount to some extent the fact that well-known and respected intellectuals disagree with me’ (Reference Fumerton, Feldman and Warfield2010: 103).

On all sides, then, theorists allow reacting to disagreement by making bias attributions.Footnote ¹ But how does the debunking reasoning work? Minimally, it involves two claims:

Any instance of debunking reasoning aims to secure the Debunking Conclusion at least partly on the basis of the Bias Premise. There's no one-size-fits-all argument schema to capture every path from premise to conclusion. That is because there is no single type of inference pattern here—it may be deductive, inductive, or abductive. Does this mean that we can't begin to understand, in general, when the debunking reasoning is successful and when it's not? No. We may safely assume that we can move in appropriate steps from the Bias Premise to the Debunking Conclusion. For instance, the following offers the makings of a deductively valid route from Bias Premise to Debunking Conclusion: we have evidence for the Bias Premise; evidence that a thinker is biased is evidence that the thinker is unreliable; and evidence that the thinker is unreliable makes it rational to reduce confidence in what the thinker reports or believes. But spelling out the reasoning in detail will not answer the important question about what counts as a good reason to accept the Bias Premise. That's where the action is.

2. Debunking Strategies

So what makes it reasonable to accept the Bias Premise? As we work toward an answer, notice what the Bias Premise is. It is a claim about a thinker's disposition, in some situation, to form inaccurate or unreasonable attitudes regarding a proposition p. Biases may prevent someone from accurately forming attitudes about p or from being reasonable in responding to evidence regarding p. Being biased doesn't imply actual inaccuracy or unreasonableness, but it carries a nontrivial risk of error, due to a tendency to form attitudes or evaluate evidence in an unreliable way. Being biased means being unreliable to some degree or other, but the converse doesn't hold. Being unreliable does not mean being biased—the victim of a Cartesian demon is unreliable but not biased, for instance. Thus, affirming the Bias Premise calls for evidence that tells us about a thinker's attitude-forming tendencies.

Let us try out four potential strategies for securing the Bias Premise. I will begin with a strategy that may tempt us. Suppose we disagree with a peer and have reason to think our view is correct. But if our view is correct, then our peer's view is mistaken. It follows that we have reason to regard our peer as mistaken. Now, in peer disagreements, we can rule out certain explanations for a peer's mistake—it's not for lack of relevant evidence or intellectual abilities that a peer is wrong. But abilities are a kind of competence, and even the competent may sometimes blow it. Let's also assume, then, that we can't attribute our peer's mistake to a mere performance error. It follows that the peer's mistake is due to bias. Thus, the Bias Premise is true.

We can summarize that reasoning as follows:

On a first look, that may seem to be a promising route to the Bias Premise. But the Dogmatic Dismissal Strategy is problematic.

Here is a crucial step in the reasoning: our view is correct, and if we are correct, then our opponent is mistaken, and so our opponent's view is mistaken. But this reasoning, if it's any good, licenses an extremely dogmatic response to any and all opponents. Whenever others disagree with us—even recognized epistemic superiors—this strategy lets us infer that they’re mistaken by appealing to our reasons for thinking that we are correct. If we have such reasons in the first place, then we need never change our minds when encountering disagreement. Something has gone wrong.

Consider a better strategy. Suppose that we have some evidence—call it E—and it is reasonable for us to think that having E rationally compels some attitude toward a proposition. Now if we know that a thinker has E but does not hold the attitude mandated by E, then we can conclude that the thinker is in error. To illustrate this, consider an example from Peter van Inwagen:

Van Inwagen's idea is that members of the Flat Earth Society are in an evidential situation that rationally compels them to accept that the earth is spherical. It's an established fact that the world is not flat, thinks van Inwagen, because there is compelling, knockdown evidence widely available for that thesis. Since van Inwagen knows these people nevertheless believe the world is flat, he concludes that they are ‘intellectually perverse’. Let us assume that the sort of intellectual perversity at issue here involves bias. All of this suggests a second route to the Bias Premise:

In the disagreements we are party to, do we plausibly have evidence that rationally compels a particular attitude? If we do not, this strategy won't help.

Let's begin close to home and reflect for a moment on philosophical disagreements. Could we use the Unresponsiveness-to-Compelling-Evidence Strategy to debunk fellow philosophers? It's hard to say, partly because it is doubtful whether there are any philosophical arguments that rationally compel a particular attitude. Philosophers like van Inwagen (Reference Van Inwagen2009: 34, 105) and David Lewis (Reference Lewis1983: x) deny that there are conclusive, knockdown arguments in our field, at least for substantive theses. But even philosophers who fancy themselves to have discovered knockdown arguments have failed to convince everyone or even a majority of philosophers. The uncertain among us therefore may find it unclear whether or not there are any compelling philosophical arguments.Footnote ²

Even if there are compelling arguments for only a few controversial philosophical positions, it's plausible that some nonphilosophical claims are supported by compelling evidence. For instance, van Inwagen's case of the Flat Earth Society is like that. So is the case of scientists funded by the oil industry who deny certain claims about human-caused climate change. In such cases, the available evidence seems to demand a particular doxastic response. Since certain people have that evidence and yet fail to hold the mandated views, we can conclude they are biased. But in most of the disagreements we care about we won't have reason to think our evidence is so potent. Two of the cases that we began with—the ones featuring Putnam and Nozick and the smart but hotheaded talk show host—do not seem to involve compelling, knockdown evidence for a thesis. At best, the Unresponsiveness-to-Compelling-Evidence Strategy will have limited use. We can leave it to the side.

Let's turn now to a strategy for reaching the Bias Premise that promises to apply more widely, starting with an example offered by David Christensen:

Christensen thinks you can properly downgrade Aksel's father's testimony. More important, Christensen denies that anything along the lines of the Dogmatic Dismissal Strategy will explain why. According to Christensen, a good reason for reducing confidence in Aksel's father's testimony is not that you are sure that Kirsten's performance was significantly better than Aksel’s. If that were your reason, of course, you would not have an ‘independent’ reason to debunk the testimony.

What would be a better reason then? You attribute to Aksel's father a bias with respect to his judgment concerning the relative merits of the musical performances. Why think that Aksel's father is biased? Most of us can't be expected to judge our kinfolk impartially on the merits of their musical performances. Family relationships bias our judgments here as elsewhere. Thus, you can reasonably accept the Bias Premise in this case and reach the Debunking Conclusion.

But it is doubtful that all of this captures a sufficient condition for accepting the Bias Premise. Imagine a case just like Christensen’s, with one extra detail: you know a further factor holds for Aksel's father, a factor that tends to counteract or neutralize the biasing factor. More specifically, imagine that Aksel's father is a professional judge on the high-school music-contest circuit and that he has special training that helps him avoid biases toward his own students. Aksel's father may be enlightened by his music-contest-judgment expertise, and perhaps this lets him avoid or overcome the biasing influence of fatherhood. In this situation, you would not have reason to infer that Aksel's father is biased. Reason to think fatherhood biases judgments and that Aksel's father enjoys paternity, is thus not enough to accept the Bias Premise.

Putting that all together, here is what I’ll call the

Here we find an inference that's a kind of statistical syllogism. It appears to be available in a wide range of cases as there are many factors that bias thought. I’ll examine this strategy in some detail below after we’ve looked at how people form judgments about bias, but for now let's move on to a different strategy.

As I noted earlier, Richard Fumerton takes himself to have reason to regard disagreeing thinkers as suffering from biases. What is his reason? He writes:

Fumerton doesn't rely on the Biasing-Factor-Attribution Strategy. He doesn't point to some biasing factor that holds for his opponents. So how does Fumerton's reasoning deliver the Bias Premise? Doesn't it presuppose that disagreeing thinkers are biased?

Squinting a bit at what Fumerton says, his reasoning may proceed as follows. Unlike the Biasing-Factor-Attribution Strategy, which involves a kind of statistical syllogism, Fumerton's reasoning relies on an inference to the best explanation. Fumerton takes himself to know things about himself that indicate he is not biased or that at least he is relatively less biased than thinkers he disagrees with on politics and philosophy. But since these disputes (presumably) concern matters of objective fact, someone has made a mistake. Since Fumerton assumes that he and his opponents share relevant evidence and intellectual abilities, the best way to explain one side's error is to posit a bias. Given that Fumerton purportedly knows he's neutral, it's his opponents, then, who must be biased. Thus, Fumerton has a reason to accept that his opponents are biased.

We may sum up this reasoning as follows:

I will consider this strategy below, but here's one remark on (2). Many disagreements can be explained without appealing to biases: differences in evidence or intellectual abilities as well as mere performance errors are included here. Therefore, reason to accept (2) sensibly amounts to reason to think alternative, nonbias explanations for why one side has erred are not as plausible as the explanation that bias offers.Footnote ³

So far I’ve noted four reasons for accepting the Bias Premise. Since the Dogmatic Dismissal Strategy doesn't appear to offer us a good reason, we will ignore it. Although the Unresponsiveness-to-Compelling Evidence Strategy is sometimes effective, I’ll set it aside, too, because its use is rather limited. The two remaining options—the Biasing-Factor-Attribution and the Self-Exculpating strategies—look more promising. Unfortunately, as I will argue, in a great many of the cases of disagreement we take to be practically or theoretically important those two debunking strategies won't help us out. But first let us turn to a third question: how do people tend to make judgments about biases?

3. The Bias Blind Spot

Psychologists have begun to reveal how we make judgments about biases, and the lessons are fascinating. Once we understand how people tend to form judgments about biases, we’ll be able to assess critically the Biasing-Factor-Attribution and the Self-Exculpating strategies. Even if those strategies are sometimes rationally appropriate, we may often be unable to put them to use properly, given our knowledge of the psychological findings.

A growing body of work in psychology observes ‘a broad and pervasive tendency for people to see the existence and operation of bias much more in others than in themselves’ (Pronin Reference Pronin2007: 37). This is a kind of ‘bias bias’—a bias that sways judgment and reasoning about bias—and it has been called the ‘bias blind spot’. It results in the conviction that one's own judgments are less susceptible to bias than the judgments of others. Direct testing confirms the blind spot is widespread (Pronin, Lin, and Ross Reference Pronin, Lin and Ross2002). Several cognitive mechanisms have been found to generate the bias blind spot: (1) an important evidential asymmetry between judgments of self and others, (2) the tendency to assume what has been called ‘naïve realism’, and (3) the motive of self-enhancement. I’ll explain each in turn.

When people make judgments about bias in themselves, they tend to rely on introspective evidence, but when they judge bias in others they tend to rely on behavioral or extrospective evidence. People look into their own minds to judge themselves but look at outward behavior when judging others, and this evidential asymmetry shapes judgment about bias.

But a central idea in psychology is that most biases are not reliably detected by introspection (Nisbett and Wilson Reference Nisbett and Wilson1977; Wilson and Brekke Reference Wilson and Brekke1994; Kahneman Reference Kahneman2003). We typically can't figure out whether we are biased by merely gazing into our minds. Biases typically ‘leave no trace’ in consciousness. As Timothy Wilson and Nancy Brekke quip, ‘Human judgments—even very bad ones—do not smell’ (Reference Wilson and Brekke1994: 121). From the inside, biased attitudes seem just like unbiased ones. We can introspect the particular judgments our cognitive processes produce, but not the properties of the processes relevant to the judgments’ reliability. In other words, we can't normally introspect the operation of biases on our judgments even when we can detect the outputs of those biases. The machine chugs along, but we can't peek in to see whether it works properly.Footnote ⁴

Although introspection does not reliably detect the operation of biases, that isn't to say people don't still try. But using introspection to attempt to discover bias has an important upshot: people may end up with the impression they have acted in spite of their own expectations and interests rather than because of them. As Joyce Ehrlinger, Thomas Gilovich, and Lee Ross point out, ‘one's conscious efforts to have avoided bias, in fact to have “bent over backwards” to do so, are likely to be highly salient’ in one's thinking about whether one is biased (Reference Ehrlinger, Gilovich and Ross2005: 686). The feeling that we’ve done our level best to be unbiased will encourage us to think we are unbiased, but that feeling should not be trusted.

Different judgments arise from differences in introspective access—no news there. But it gets worse. It's not just that people actually rely asymmetrically on introspective and behavioral evidence. They also think they should place less weight on behavioral evidence in their own case and more weight on their own introspective evidence. In fact, researchers have noted that subjects sometimes show surprising disregard for the subjects’ own actions even when those actions become salient. Subjects insist their behavior isn't relevant to decide whether they are biased. This is striking. So often in life we are judged by our actions, not our intentions or hopes or feelings. But in self-judgment about bias, we overlook our actions and instead cling to how things feel on the inside (Pronin and Kugler Reference Pronin and Kugler2007).

The evidential asymmetry driving the blind spot seems to prove the old wisdom: nobody should be a judge in his own case (nemo iudex in causa sua).Footnote ⁵ But what if we break down the evidential asymmetry and inform people about others’ thoughts and feelings? Will the blind spot vanish? Regrettably not. Even after people are given others’ introspective reports, the blind spot tends to persist. In some experiments, psychologists provided observers with a sample of actors’ introspections, but observers continued to impute more bias to others than to themselves even when the observers had read and listened to detailed reports from actors and believed those reports correctly reflected the actors’ thoughts (Pronin and Kugler Reference Pronin and Kugler2007).

A second source of the blind spot is our tendency to presume what psychologists have called ‘naïve realism’—the idea that our experience of the world, others, and ourselves is veridical. We normally assume that our experience gives us a more or less ‘unmediated’ picture of how things really are.

How does naïve realism influence judgment about biases? It turns out that when we discover that others disagree, we often attribute biases to them. In one experiment, American and Canadian undergraduate students who disagreed with the US president's decision to invade Iraq attributed a greater degree of self-interest bias to the president than did students who agreed with him (Reeder et al. Reference Reeder, Pryor, Wohl and Griswell2005). Those in the grips of naïve realism believe objective thinkers will agree with them. When others disagree with us, we are prompted to ask whether they’ve missed relevant evidence. If we think they are well-informed, naïve realism leads us to conclude they are biased (Pronin Reference Pronin2007: 39–40). Our tendency to try to resolve cognitive dissonance plausibly explains this sort of effect sometimes. When we find that people disagree, we treat this as prima facie evidence against our views. Then we resolve the dissonance in favor of our own views, often by way of bias attributions.

Evidential asymmetry and naïve realism are two sources of the bias blind spot: the motive of self-enhancement is a third. Research has established that we see ourselves in an overly positive light.Footnote ⁶ For valuable or desirable traits, we tend to overrate ourselves, even when the evidence suggests otherwise. In a classic study that should be close to every college teacher's heart, 94 percent of college teachers rated themselves as doing above-average work (Cross Reference Cross1977). And when people lack a talent or positive trait, they are sometimes oblivious. These sorts of effects stem from powerful ‘ego-protective’ biases. We think well of ourselves, objective evidence be damned, but most people rarely notice this. In fact, self-enhancement bias is regarded as a key element of health. People with major depression tend to have more accurate self-assessments and fewer illusions about themselves, but at the cost of healthy functioning (Taylor and Brown Reference Taylor and Brown1988).

The self-enhancement motive gives us no protective illusions concerning other people, naturally enough (leaving aside, for now, the interesting case of family and friends). We attribute biases to others with ease. We expect them to make judgments that serve their self-interest and to make overly positive self-assessments. As a result, judgments of bias in self and others will differ. The motive to self-enhance thus makes us less likely to find bias in ourselves than in others (Pronin Reference Pronin2007: 37).

To sum up: the bias blind spot stems from an important evidential asymmetry, the assumption of naïve realism, and self-enhancement bias. These psychological mechanisms lead to the conviction that our judgments are less susceptible to bias than the judgments of others.

Here is an important part of the story we shouldn't ignore. To judge bias in others we focus on behavioral evidence, but that evidence by itself does not disclose others’ inner dispositions. The evidence needs interpretation. We use ‘abstract theories’ of bias to decipher it. Let me sketch the content of our abstract theories (cf. Ross, Ehrlinger, and Gilovich, forthcoming). We know people are motivated to seek pleasure and avoid pain. They give heavy weight to their own wants and needs. They are sheltered from reality by various psychological defense mechanisms. People sometimes view issues through the lens of their ideology and social position. And they’re capable of (self-)deception concerning the influence of nonrational factors on their judgments. These sorts of thoughts comprise our abstract theories. Our theories tell us when motives, needs, expectations, and context invite bias, allowing us to regard others’ behavior as indicating the presence or risk of bias. These theories are rough guidelines. As a matter of fact, they are imperfect guides. For instance, experiments show that people sometimes cynically expect others to be more biased than they in fact are (Kruger and Gilovich Reference Kruger and Gilovich1999).

Thus concludes my review of psychological research on the blind spot. Let's see how this psychological picture might guide us in the business of debunking.

4. Two Debunking Strategies and the Psychological Picture

Here is my fourth and final question. How often do we have good reasons to debunk testimony from those who disagree with us?

Recall that debunking reasoning moves from the Bias Premise to the Debunking Conclusion, and we earlier looked at four strategies to get the Bias Premise in hand. The Dogmatic Dismissal Strategy is a nonstarter. The Unresponsiveness-to-Compelling Evidence Strategy won't apply widely enough to situations where we hope to resist the downward push of testimony from disagreeing thinkers. But the Biasing-Factor-Attribution and the Self-Exculpating strategies show more promise. Let us consider, then, how often we can reasonably follow these two routes to the Bias Premise.

Awareness of the psychological picture brings trouble for anyone who hopes to deploy either the Biasing-Factor-Attribution or the Self-Exculpating strategies in a wide range of important cases. The trouble is that those debunking strategies require particular judgments about biases, but reflection on the psychological picture raises doubts about whether we reliably make those judgments. Accordingly, we should often doubt that it's reasonable for us to deploy these strategies.Footnote ⁷

Let us begin with the

As already noted, (2) requires reason to think alternative, nonbias explanations for why one side has erred are less plausible than the explanation bias provides. Let's grant for now that (2) can be reasonably accepted. I will argue that our inclination to accept (1) should be curbed by what we know about the bias blind spot.

Richard Fumerton appears to endorse the Self-Exculpating Strategy. Consider what he says about his reasoning: ‘When I argue this way, I again risk sounding like a bit of a jerk. Do I really suppose that I am justified in thinking that there is an asymmetry between myself and others when it comes to various epistemic defects? Am I any less likely to be blinded to what is reasonable to believe by antecedent views or desires? Well, to be honest I suppose that I think I am’ (Reference Fumerton, Feldman and Warfield2010: 105). How should we react to this disclosure? You may decide that Fumerton is a poster child for a public service ad campaign by psychologists to raise public awareness of the bias blind spot. (‘The Bias Blind Spot: It Makes You Sound like a Jerk’.) But perhaps Fumerton's honesty is to be admired. The gist of his reasoning is pretty much standard operating procedure for us human beings, but none of this is normally expressed so candidly.

The blind spot neatly explains why Fumerton judges that (1) is true. He doesn't tell us his reasons for (1)—he just says he thinks he is justified to think it is true in some disagreements. Plausibly, Fumerton introspects to check for bias in himself and relies on behavioral evidence, guided by abstract theories of bias, to check for bias in others. Of course, introspective evidence is not a reliable way to recognize subconscious biases in ourselves and may even leave us with the feeling of having successfully overcome judgment-distorting influences after ‘bending over backward’ to be neutral.

Even so, we don't know that Fumerton is actually subject to the bias blind spot. All we know is what the psychological evidence says: humans in general tend to make judgments like his. An important issue, then, concerns the connection between the psychological evidence and a thinker's reasons to accept (1). Here's what I propose. Evidence about how people judge bias is relevant for assessing the premise that some person is not biased. The psychological evidence is the kind of evidence that tells us about how effectively we tend to assess evidence about bias. This is what has been called ‘higher-order evidence’—the kind of evidence that tells us about our evaluation of, or responsiveness to, a body of first-order evidence (see Feldman Reference Feldman2005; Christensen Reference Christensen2010). Learning of the psychological evidence, then, tends to undermine belief in (1).Footnote ⁸ If we ourselves wish to accept (1), in full awareness of the psychological picture, we need to have a reason to think that we are not subject to the blind spot. Insofar as we lack reason to think we’re not subject to it, we have reason to doubt (1) is true, and accordingly the Self-Exculpating Strategy will fail to deliver the Bias Premise.

An analogy will amplify the reasoning I have proposed (see Elga ms.). Imagine that Earhart is piloting a small aircraft above 10,000 feet. Earhart knows that people in small aircraft flying at high altitudes often suffer from hypoxia—a condition where the brain is oxygen-deprived—and, as a result, their judgments become unreliable. Once hypoxia has taken effect, it will typically seem to the hypoxic subject that her reasoning is perfectly good even when it is bad. Hypoxia has an effect on thinking without ‘leaving a trace’ in consciousness.Footnote ⁹ Since Earhart recognizes she may be hypoxic at this high altitude, she has reason to invest much less confidence in her calculations about her remaining fuel. To restore confidence in her calculations, she needs reason to think that she is not hypoxic.Footnote ¹⁰

If her belief is reasonable, she needs independent means to determine that she's not hypoxic, such as an O₂ detector. And our situation with respect to bias is not unlike that of the high-flying pilot. Since we know about the bias blind spot, we should reduce confidence in (1), the premise that we are not biased, unless we have good reason to think that we are not suffering from the blind spot.

What counts as good reason to think we’ve overcome the blind spot? Back in the hypoxia example, once Earhart suspects she's hypoxic, a reason to regard herself as not hypoxic must arise from a belief-forming method she reasonably thinks is unaffected by high altitude. Likewise, once we suspect we may be blinkered by the blind spot, any reason to think that we are unbiased must trace back to a method we reasonably think is not biased. That's why introspection alone won't normally dispel doubts about the blind spot—we know introspection is prejudiced in our favor.

If introspection is off-limits, where shall we turn? To avoid biases, thinkers sometimes try to debias—to identify and avoid biases or adjust their judgments to counteract the negative effects of biases (Wilson and Brekke Reference Wilson and Brekke1994; Wilson, Centerbar, and Brekke Reference Wilson, Centerbar, Brekke, Gilovich, Griffin and Kahneman2002; Larrick Reference Larrick, Koehler and Harvey2004). Let's say we know of a reliable method to judge our susceptibility to the blind spot. If we reasonably think that our judgment that we aren't afflicted by the blind spot traces back to a reliable debiasing method, we get reason for thinking that we have avoided or overcome that bias. Thus, recognizing we’ve debiased would permit us to accept premise (1), our awareness of the blind spot notwithstanding.

As I’ll now argue, reason to think that we’ve successfully debiased is going to be uncommon. That's because debiasing is extraordinarily hard. For a start, we naturally rely on abstract theories of bias to debias, but our theories may lead us astray. For instance, the feeling that we can debias by carefully thinking things over is tempting but mistaken (cf. Kenyon Reference Kenyon2014: sec. 2). And so we’ll want to take our cues from research on debiasing. But even when we do, debiasing in real life demands extensive knowledge of ourselves and the nature of our biases.

To see why, let's consider Timothy Wilson and Nancy Brekke's prominent account of debiasing (Reference Wilson and Brekke1994 and Wilson, Centerbar, and Brekke Reference Wilson, Centerbar, Brekke, Gilovich, Griffin and Kahneman2002). In brief, Wilson and Brekke see a thinker's inability to debias as stemming from a number of common sources: (1) that thinker's lack of awareness of his or her mental processes (e.g., the extent to which the thinker's positive evaluation was due to a self-enhancement motive), (2) lack of control over mental processes (e.g., the thinker's inability to prevent the fact that his or her status is at issue from influencing self-judgment), (3) inaccurate theories about biasing influences on judgment (e.g., the thinker's failure to appreciate how his or her own status could nonconsciously influence self-judgment), and (4) inadequate motivation to correct for bias (e.g., an insufficient desire to avoid a self-enhancing judgment) (cf. Reference Wilson, Centerbar, Brekke, Gilovich, Griffin and Kahneman2002: 187). Any one of (1) through (4) will prevent successful debiasing. In light of how commonly people find themselves in those conditions, Wilson and Brekke are ‘rather pessimistic’ about our ability to debias—and their pessimism even extends to trained psychologists who are familiar with the literature on biases (Reference Wilson and Brekke1994: 120; Wilson, Centerbar, and Brekke Reference Wilson, Centerbar, Brekke, Gilovich, Griffin and Kahneman2002: 190–91, 200). This bleak assessment confirms what other psychologists have found: debiasing demands knowledge individuals often lack (Kahneman Reference Kahneman2003).

Psychologists say that our best shot at successful debiasing lies in debiasing techniques. One self-administered technique is to ‘consider the opposite’, to argue against one's own initial judgment.Footnote ¹¹ This debiasing advice comes with a warning attached: this technique may backfire. ‘Ironically, the more people try to consider the opposite,’ Norbert Schwartz and colleagues observe, ‘the more they often convince themselves that their initial judgment was right on target’ (Reference Schwartz, Sanna, Skurnik and Yoon2007: 128). In general, psychologists are fond of pointing out the inbuilt fallibility of debiasing techniques. Such techniques may be our best shot to debias, but that doesn't mean they’re a good shot.

Remember, we wanted to know whether debiasing could let us reasonably accept (1), the premise that we are not biased, in spite of our awareness of the blind spot. Would-be debiasers face obstacles, as noted. Of course, some of us will still regard ourselves as having successfully debiased. But without instruction in debiasing and practice in implementing the techniques, aren't we just fooling ourselves? True, philosophers are trained to think carefully about arguments, but we are not trained how to debias for the blind spot. Again, rechecking our arguments and finding they still look good to us is not enough. At the very least, we should think it is unclear whether or not we’ve effectively debiased. The take-home message is that good reason for regarding ourselves as having debiased for the blind spot, and thus good reason for (1), is uncommon in the kind of disagreements we care about.

In the end, we might expect a person unknowingly subject to the bias blind spot to use the Self-Exculpating Strategy. But forewarned against the blind spot, we shouldn't deploy that strategy unless, again, we have good reason to think our self-judgment about bias is reliable.

In hopes of finding a better way to debunk our dissenters, let's turn to the

How does this strategy look against the backdrop of the psychological picture? Here's a general worry to start with. One precondition for successful use of this strategy is that we lack a reason to accept that a biasing factor like F holds for us. Otherwise, the strategy will explode in our hands—it will debunk our dissenters and ourselves and fail to preserve our rational belief. But now we have reason to doubt that our perspective on ourselves is objective, given our inclination to cling to (unreliable) introspective evidence and disregard behavioral evidence concerning ourselves, for instance. Thus, we should sometimes suspect that the biases we attribute to others apply to our own judgment, too. Even if we set that point aside, both (1) and (2) are problematic for other reasons. I’ll consider two.

First, our abstract theories of bias aren't always well-attuned to reality. We occasionally cynically overestimate bias in others (Kruger and Gilovich Reference Kruger and Gilovich1999). Let's suppose our theories say some factor F biases certain kinds of judgments. Then we should ask: on reflection, do we have good reason to think that our abstract theories of bias are right? Suppose we lack such reason. Then it would be strange for us to accept (1): if we are unsure whether our abstract theories are correct, we should also be unsure about whether F really biases. This suggests that, once we reflect, reason to accept (1) amounts to reason to think that our abstract theories are reliable guides to bias, but we may often lack any reason to accept this.

Second, and shifting to (2), we appeal to behavioral evidence when determining whether some thinker is biased, but our uptake of that evidence may be influenced by unreliable evidence-gathering methods. Psychologists have noted that observers often place heavy weight on others’ characters to explain their behavior in a particular situation, rather than thinking about how their behavior may have been shaped by the situation (Jones and Harris Reference Jones and Harris1967). This has been called the ‘fundamental attribution error’ because it is arguably a central engine of social judgment. We explain how others act primarily by appeal to their characters, not to the situations they’re in. When a bicyclist crashes, for example, we are apt to judge he's an unskilled or reckless rider. But that judgment is an error when the crash is explained instead by some part of the rider's situation. Maybe he's a good rider, but he was late for an appointment and knew he had to ride a bit dangerously around a corner to arrive on time. (Unsurprisingly, we readily regard situational factors as influencing our own behavior.) We should ask: on reflection, are the methods we use for collecting behavioral evidence concerning our disagreeing peers any good? If we lack reason to think those methods are good, it would be strange to accept (2): if we are unsure whether we can competently gather evidence relevant to factor F concerning a disagreeing thinker, we should also be unsure about whether F holds for that thinker's judgment. The idea here is that reason to accept (2) amounts to reason to think that our techniques for gathering behavioral evidence are reliable. But if we take the psychological research seriously, many of us often lack reason to think this.

The psychological picture raises doubts about whether we have good reason to accept (1) and (2), which sometimes calls into doubt our use of the Biasing-Factor-Attribution Strategy (cf. Christensen Reference Christensen and Lackey2014: 160–61). At the same time, (3) is problematic, too. Recall that (3) is satisfied when we’ve reason to think that no enlightening factor—one that would counteract a biasing factor—holds for a disagreeing thinker. But this condition is too easy to meet. We could satisfy it, without fail, by remaining oblivious to the presence of potential enlightening factors. As a matter of fact, it's doubtful whether we are normally ‘on the lookout’ for enlightening factors operating in others who disagree with us. Given naïve realism—that is, the presumption that our views are objective—whenever others disagree with us, we tend to be on the lookout for biasing factors. The engine driving this may be confirmation bias. In general, we search for evidence to confirm our hypothesis that we are objective and not for evidence to disconfirm it; it's thus unsurprising that we would search for biasing factors, not for enlightening ones, because we expect that our dissenters are biased. As a result, we may take ourselves to accept (3) reasonably just because we’ve failed to search adequately for enlightening factors.

This point about (3) suggests a change to the Biasing-Factor-Attribution Strategy. Suppose that we reasonably accept premises (1), (2), and (3). Imagine further, as the psychological picture recommends, we reasonably believe we are not always disposed to search for factors that may enlighten disagreeing thinkers, even when such factors are present. But this perhaps reveals that the strategy at issue doesn't set down conditions sufficient for us to have reason to accept the Bias Premise. If there's a gap, we may fill it with a further premise, namely, that we have reason to accept premise (4), that we have adequately searched for potential enlightening factors that hold for the disagreeing thinker's judgment about proposition p. Once again, the psychological picture should give us pause. Since it's doubtful we are in the habit of searching for factors that may enlighten our disputants, we have some reason to doubt whether we’d satisfy the condition in (4).

My suggestion is not that the Self-Exculpating Strategy and the Biasing-Factor-Attribution Strategy can't ever help us debunk testimony from those who disagree with us. The idea is that, in light of the psychological picture, we often have reason to doubt that these strategies do help. To make use of them in view of that picture, we must face up to the following questions. What is our reason to think that our self-judgments about biases are reliably formed? Why do we think our abstract theories of bias are any good? Why do we think our methods of collecting behavioral evidence about our dissenters are reliable? Why do we think we have adequately searched for factors that may enlighten disagreeing thinkers? To answer well, we must know how to make principled judgments about bias.

To be sure, sometimes we can answer those questions with relative ease and reasonably debunk dissenters. For instance, when we imagine Christensen's high-school music contest, it's natural to assume (1) that you’ve got good reasons to think your abstract theory of bias that implies Aksel's father is biased is in fact reliable, (2) that your gathering of behavioral evidence concerning Aksel's father's judgment has been reliable, and (3) that you’ve properly searched for factors that might enlighten him. Thus, you can launch the Biasing-Factor-Attribution Strategy and debunk Aksel's father. The same plausibly goes for the case of the climate scientists sponsored by the oil industry. But these are easy cases, and we are here concerned with less straightforward ones. In more difficult cases, we will on reflection have reason to doubt that these strategies are appropriate. While we may be tempted to see our intelligent and informed opponents as unable to hear the good sense we’re talking, because of their social position or intellectual commitments, we now know that is too convenient. We may be inclined to treat ourselves as the unbiased side in disputes, to see ourselves as judge and jury, prosecution and defense, in our own case. But can't we do better? Ordinarily, we do not demand good reasons for our own judgments concerning biases even when we hold others to tougher standards. The psychological picture pushes us to resist the temptation to intellectual hypocrisy.Footnote ¹²