Levels of Scepticism: How Even Rational People have Sceptical Blind Spots

Most of my readers doubtless recognise the importance of being skeptical about the information, arguments, and ideas that we encounter, be it dietary advice, political opinions, science news articles, or whatever else. There are, however, different levels of scepticism, corresponding to the varying degrees of sophistication we can attain in the manner in which we respond to new ideas or arguments. In this piece I wish to outline a brief topology of these different levels of scepticism. I do not pretend to offer any sort of definitive classification, nor do I claim that these levels are in any way based upon empirical psychological research. Their purpose rather is to serve as a conceptual tool to help us think about the ways in which we can improve our own thinking, and work to eliminate residual biases and blind spots that hamper our efforts to form beliefs that are best justified by strong argument and quality evidence. The hierarchy that I shall outline has four levels, ranging from least sceptical at level 0 to most sceptical at level 3. I want to emphasise that the purpose of these levels is not to create a ranking of particular people as better or worse sceptics, as most people operate at multiple different levels depending on the circumstance and the topic in question. Rather, the purpose is to rank particular types of thinking, so that we may better recognise when we are thinking in a better or worse mode of scepticism.

I will begin my discussion at the bottom of the hierarchy, level 0. When we think at this level, we do not think particularly critically or sceptically about much of anything. Though we may have opinions about various matters of political, ethical, or philosophical import, when operating at level 0 we are typically unable to clearly articulate these views to others, or explain why we hold them. Most such views are typically informed primarily by our upbringing, socialisation, and the attitudes of the people around them as they go about their lives. Many people who operate at this level have little to no ability to critically analyse evidence or analyse an abstract logical argument, having never been taught such skills or found it necessary to learn them. Even those who do have such skills, however, can sometimes be remarkably compartmentalised in the manner in which they apply them, for example being able to hold forth a detailed analytical argument about topic A, but when discussing topic B doing little more than spouting catch-phrases that resonate with them. When we operate at level 0, we tend to think that our viewpoint is ‘obvious’, and react with surprise when we find that others think differently, or that any sensible person can hold a different view. It is likely that the majority of humanity operate at level 0 most of the time, as this is the type of thinking that comes most naturally and easily to most humans. That is, we typically form beliefs about the world not on the basis of careful examination of evidence, logical analysis, or in-depth comparison of alternative perspectives, but unconsciously and reflexively as we go about our lives, drawing largely upon what we know and are familiar with. I do not want to claim that this is inappropriate in all contexts, as certainly we cannot always subject everything to detailed critical analysis. However, I do think that making a habit of thinking in this way is liable to lead us into error and confusion about a great many of our beliefs. Scepticism, logic, and science are valuable tools, and neglecting these tools leaves us intellectually impoverished and prone to biased and mistaken reasoning.

This leads me on to the next level of the scepticism hierarchy, level 1. When operating at this level, we are able to articulate clear opinions on a variety of subjects, martialling various arguments and evidences in favour of our views. We recognise the distinctiveness of different viewpoints and are able to employ the tools of scepticism and rationality to make arguments for what we regard as the correct view. However, when thinking at this level we also tend to identify strongly with one particular perspective, be it religious, political, scientific, or whatever else, and employ these sceptical tools selectively against arguments or information coming from the opposing ‘side’. We are able to spot logical fallacies, faulty reasoning, and inadequate evidence in the arguments of our ideological opponents, but are much less able to apply the same skills to arguments made by those of their own ideological persuasion. When operating at level 1, we tend to respond to new claims by ‘pattern matching’ how the claim is framed and who is making it, and on that basis classify it as ‘for’ or ‘against’ our side. We thus do not judge arguments fairly on their own merits, but subject them to an initial, largely unconscious ‘screening process’, whereby if an argument ‘sounds like’ the sort of thing someone we disagree with would say, then we subject it to closer skeptical examination. On the other hand, if it sounds like the sort of thing somebody who agrees us would say, then it typically avoids any in-depth examination. This sort of self-serving, pro in-group bias comes very naturally to humans, and thus is very difficult to overcome. It is also very difficult to notice in ourselves, because when operating at level 1 we typically are only conscious of the times when we are being skeptical and critical, not the times when we aren’t. To us it feels like we take arguments only on their merit, when in reality we are very selective about how our scepticism is applied, and make little effort to subject views that accord with our beliefs or biases to the same rigorous critical examination that we apply to those that do not. When operating at level 1 we are also liable to be misled by framing effects, slogans, buzzwords, and other irrelevancies relating not to the substance of an argument, but to how it is packaged. Selective scepticism of this sort is very common to those heavily involved in some sort of social movement or organisation, and is not always bad because it can save us time – after all, we can’t critically examine every single claim we come across. At the same time, it can become all too easy to become accustomed to operating at this level, and in doing so we fail to make proper or full use of the tools of rationality and scepticism.

When operating at the next level up in the hierarchy, level 2, we are able to apply critical thinking skills and skeptical analysis consistently and fairly both to arguments that we find agreeable and those that we find disagreeable. We allow the arguments and evidence to be persuasive in their own right, with minimal influence based on who has made them, or how they have been formulated. We consciously recognise our tendencies to favour ‘our side’ over the ‘other side’, and make efforts to circumvent this by deliberately taking time to critique arguments made by those who agree with us, and likewise by finding the strongest, most able defenders of ideas we disagree with. This, of course, is not easy to do, and requires careful attention and genuine effort to fairly engage with different perspectives and ideas. There is, however, one significant failing that we still commonly experience when operating at level 2. Namely, we instinctively and reflexively retain an unreasonable overconfidence in our own reasoning abilities. We tend to believe that our perspectives or conclusions on some issue are the ‘right’ ones, and everyone else has got it ‘wrong’. Taken to extremes, this type of thinking can lead to habitual contrarianism and even conspiratorial thinking. In such cases, we may think that both sides of some major dispute have it wrong, and we are the ‘lone genius’ able to see the correct answer. While most people do not reach such extremes, what those operating at level 2 have in common is their inability or unwillingness to apply the same sceptical attitude and critical examination to their own thought processes that they do to the arguments of others. We thus do not properly appreciate the many limitations of memory, rationality, and knowledge that we ourselves are subject to, and which hamper our efforts to draw correct conclusions. We are skeptical of everyone else, but not sufficiently skeptical of themselves, of our own biases and limitations.

The highest level of my hierarchy is level 3, and it is the level I believe we should all aspire to use as regularly as possible. When operating at level 3, we properly apply scepticism and critical analysis not only to everyone else, but also to ourselves and our own beliefs, preconceptions, and thought processes. We are often hesitant to attach strong credence to the conclusions we reach, because we know that our rationality is grossly imperfect and our knowledge and perspectives sorely limited. This of course should not lead us to radical scepticism or keep us from forming opinions about anything, but it should temper our confidence considerably and keep us from becoming dogmatically attached to our conclusions and perspectives. In level 3 we are also much more self-critical, actively setting out to uncover our own biases and doing our best to compensate for them, and not just criticising the biases and errors of others. Likewise, we actively seek out the viewpoints of other informed persons to critique our opinions and point out our cognitive ‘blind spots’, helping us to apply scepticism to our own thought processes and reasoning. Level 3 is often an uncomfortable state to operate in, for it robs us of the overconfidence in our beliefs that is reassuring to most people, and also requires a degree of active self-criticism which is unnatural and effortful to maintain. We also must also make an effort to find the right balance between appropriate self-criticism and scepticism on the one hand, and paralysing self-doubt, apathy, or total mistrust of reason on the other. Operating at level 3 is neither easy nor natural, but I do believe it is the highest form of ‘true scepticism’, and the ideal to which we should all aspire. Operating in this level may not always be possible, but nevertheless is worth striving for since it allows us to take the fullest advantage of the tools afforded by logic, rationality, and scepticism, thereby providing us maximum chances for ultimately forming accurate beliefs free from error, bias, and distortion.


Is Conciliationism Self-defeating?


In this piece I argue that conciliationism, the position that in cases of peer disagreement we ought to moderate our beliefs between the extreme positions, is not self-defeating, or at least is not self-defeating in any way which undermines the argument for conciliationism. I provide three related arguments in support of this contention. First, I argue that the ‘self-defeating objection’ can be applied to the ‘self-defeating objection’ itself, such that if conciliation is self-defeating, then so is this critique of it. Second, I argue that many apparently very reasonable epistemic standards also can be potentially self-undermining in some circumstances, thus illustrating that this problem is a general one not specific to conciliationism. Third, I will argue that there are good reasons to think that difficulties arise generally from the attempt to recursively apply epistemic principles to themselves, and therefore treating such self-referential cases as special is not arbitrary, but perfectly justified.

Introduction to the Positions

Conciliationism is the position that, when faced with disagreement between two epistemic peers (persons of roughly equal knowledge, intelligence, free of bias, etc), the most rational response is to conciliate: that is, either suspend judgement, or otherwise adopt some sort of compromise position between the two extremes. The idea is that when there exists disagreement between epistemic peers, there exists no rational reason to prefer one position over the other, and hence the most justifiable response is to conciliate.

This position has been attacked has being self-undermining. The idea is that there exists peer disagreement about the topic of peer disagreement itself – some philosophers advocate conciliationism, whilst others advocate steadfastness. It would seem, therefore, that the conciliationist position would in this instance advocate suspension of judgement, or some sort of compromise between these two extremes. Thus the conciliationist’s own position leads them to adopt a less conciliatory position. In this way, so the argument goes, conciliationism is self-undermining.

The ‘Self-defeating Objection’ is Self-defeating

Let us suppose that we accept the critique as outlined above, and conclude that conciliationism is self-defeating. How should someone who was antecedently convinced of the superiority of the conciliationist view respond? On the one hand, it seems that since this critique undermines conciliationism, existing conciliationists should abandon the view, or at least substantially reduce the credence they place in it – that is, they should adopt or move toward steadfastness (or something similar). On the other hand, what would it mean for such persons to adopt steadfastness? In this context, it would mean nothing other than sticking with their original position, namely conciliationism.

We thus arrive at a contradiction: it seems that if we accept the ‘self-defeating’ argument, it follows that existing conciliationists should abandon conciliationism, and simultaneously continue to uphold it. By applying the ‘self-defeating’ refutation to itself, we thus find that the ‘self-defeating rebuttal’ is, by its own logic, self-defeating.

As I will argue later, I do not actually think this is a problem, because a great many epistemic positions encounter difficulties when applied to themselves. As such, I am not arguing that the refutation fails because it is self-defeating, since I do not think being self-defeating in this manner is necessarily a problem. Rather, I am arguing that because it is not necessarily a problem for an epistemic principle to be self-defeating, the ‘self-defeating refutation’ simply has no purchase in the first place – the fact that conciliationism is (in this sense) self-defeating does not, by itself, constitute a good argument against it.

Many Epistemic Standards are Potentially Self-defeating

It is relatively easy to construct examples of highly plausible epistemic principles which are nevertheless self-undermining in at least some circumstances. For example, consider the principle “don’t place too much confidence in any new idea that you come up with at 4am in the morning after having drunk six beers”. This principle, subject perhaps to minor caveats or rewording, surely seems quite reasonable. However, as the astute reader will immediately notice, it is also potentially self-undermining – what if someone came up with this very idea at 4am in the morning after having drunk six beers? It seems that under these circumstances, the principle says not to place much confidence in itself, and so is self-undermining.

Consider another (I think) very plausible epistemic principle: “don’t believe anything important said by a pathological liar and known con-artist”. But what if a pathological liar and con-artist were to tell you this? Again, it seems that in this case the principle would assert disbelief in itself, and would therefore be self-undermining. These cases (and many more like them which can easily be constructed), I think allow us to see that merely because an epistemic principle is sometimes self-undermining, it does not follow that the principle is invalid.

Treating Self-Reference Differently is not Special Pleading

Is it not arbitrary special-pleading to say that “we should conciliate about all beliefs except for conciliationism”? Is this a sensible position to take? Though it may seem so at first, I do not think this position is arbitrary special-pleading. Rather, as I have mentioned previously, I believe there is something intrinsically difficult about applying epistemic (or semantic) principles to themselves, and as such applying conciliationism to itself is marked out as being ‘special’, and hence treating it differently to other cases of disagreement is justified.

To see why this is the case, imagine a consumer reports magazine, which conducts product reviews of a wide range of consumer goods and makes recommendations to potential buyers. Imagine that our hypothetical magazine has developed over many years a strong reputation for impartiality and delivering careful, critical reviews of the consumer goods they examine. Now suppose that our magazine wanted to undertake a comparison and review of consumer reports magazines themselves. Suppose further that our magazine already knows a great deal about its competitor magazines and the sort of product reviews they write, so there is no chance of our magazine uncovering new information in the course of its review or of realizing that it had been ‘wrong’ about any of its past product recommendations.

In such a situation where our magazine wishes to stand by all their past product recommendations, the only reasonable outcome is for them to rank themselves as the top consumer reports magazine. To do otherwise would be to contradict themselves by asserting that they stand by all the particular recommendations and rankings they had made in the past, and simultaneously asserting that they think the rankings and recommendations of some other magazine are superior to their own.

The point to be made here, is that it does not make any sense for us, as loyal readers, to angrily demand that our magazine provide a more neutral, unbiased analysis of the best consumer affairs magazine, just as they do for all other products. For if they wish to stand by their own previous decisions (which we presume they do, for they believe they are justified), there is only one possible consistent magazine recommendation for them to make. This is not a case of special pleading on behalf of the magazine; the decision stems from the very nature and logic of applying criteria such as this to themselves.

Though the magazine case is only an analogy, I think it helps to illustrate the point that self-reference is an intrinsically tricky problem. It is not arbitrary or special-pleading to declare that certain philosophical principles or ideas may work differently when applied recursively to themselves. We can see another example of this broad point in the form of Tarski’s undefinability theorem, which (loosely speaking) says that arithmetic truth cannot be defined within the language of arithmetic, precisely because of this problem of the ability to formulate recursive self-contradictory statements such as “this statement is false” in the language of arithmetic.

I am not arguing that Tarski or the magazine case are exactly the same as what is happening in the case of conciliationism. Rather, I am using these cases as illustrative of the broader point, which is that strange things can easily happen when we apply certain epistemic or semantic ideas recursively to themselves, and that as such it is not arbitrary or special-pleading for a conciliationist to say something like “we should conciliate about every position except for conciliationism”.

Conclusion and Caveats

In this piece I have argued that conciliationism survives the ‘self-defeating’ critique. The key reason why this critique does not hold is because applying epistemic or semantic theories to themselves quite often leads to problematic or potentially self-undermining consequences. I illustrated this broad problem by a number of examples, including the very ‘self-defeating critique’ itself. The problem lies not with conciliationism as a position; the problem lies with self-reference much more broadly. As a result of these considerations, the self-defeating critique of conciliationism fails.

Notwithstanding my arguments here, I do actually think that the presence of peer disagreement about the question of peer disagreement should make both conciliationists and steadfasters less confident in their respective positions – peer disagreement does still matter when applied to itself. I do not, however, think the degree of conciliation need or ought be as great as I would ordinarily advocate for other questions, because as I have argued, things get ‘tricky’ when applying such concepts to themselves. Since this principle applies broadly across many epistemic theories, I do not think it is arbitrary or special pleading for a conciliationist not to conciliate about conciliationism as much as they would ordinarily conciliate on other questions.