All claims are not born equal

Olof Palme is talking to striking students of Stockholm Univeristy in 1968. Photo by Weine Lexius.

1. Introduction

Disagreement in the real world is a messy affair. However, the degree of messiness appears to vary from case to case. Here are two fairly realistic cases to illustrate this observation. Let us call them “Less Messy Disagreement” and “More Messy Disagreement”.

Case 1. Less Messy Disagreement
You believe that Swedish Prime Minister Olof Palme was murdered in the year of the Chernobyl disaster. I believe that he was murdered one or two years before the Chernobyl nuclear power plant went up in flames. Who is right? I take out my phone and look up “Olof Palme” and “Chernobyl” on English-language Wikipedia. To be as certain as possible, I then check the Palme and Chernobyl entries in Swedish, Ukrainian, German, French etc. I also check Encyclopaedia Britannica and the Great Russian Encyclopaedia. It says everywhere that the murder and the disaster both happened in 1986. I was wrong; you were right. Our disagreement is resolved.
Case 2. More Messy Disagreement
I believe that Olof Palme’s time in power was good for Sweden. You believe that it was bad for Sweden. Who is right? I argue that Palme’s policies strengthened the foundation of Sweden’s continued prosperity. You argue that they stifled Sweden’s full economic potential. I hold that Palme’s commitment to social justice allowed more people to overcome oppressive social structures and flourish. You hold that Palme’s obsession with social justice created a culture of resigned dependency on the welfare state. I cite statistical evidence; you cite statistical evidence. I say that Sweden compares favourably to other countries; you say that international comparisons highlight Sweden’s problems. Ten years later, we have developed somewhat more nuanced versions of our initial positions, which take into account some of the other party’s arguments. However, we are still arguing about Palme’s legacy. It looks like our disagreement may never be resolved.

Suppose further that, in the Less Messy case above, I refuse to accept as credible evidence all the multiple online sources you and I have access to. I double down on my belief that Palme was killed at least a year before the Chernobyl disaster. Any online evidence to the contrary, I claim, is a result of a wide-reaching conspiracy by some globalist Jewish elite. It seems reasonable that from now on you can dismiss me as a crackpot and disregard whatever I have to say on this matter, or indeed on most other matters. In the More Messy case, on the other hand, such a reaction doesn’t seem reasonable. To be sure, you and I might get tired of arguing about Palme’s legacy and choose to avoid the topic. We might even choose to unfollow each other on social media so as not to provoke further fruitless debate. However, neither of us could justifiably declare the other a crackpot. In fact, we might still respect each other’s opinion and expertise in matters unrelated to Swedish politics.

There appears to be an intuitively clear difference between the Less Messy case and the More Messy case. Can we account for this difference in a philosophically interesting way? I believe that we can. In this essay, I argue that a particular feature of human-made claims can be used to justify our different reactions to disagreement in such cases. This feature is the place any given claim occupies on what I will call a crispness-fuzziness spectrum. A claim is closer to the crisp end of the spectrum if it is clear that it has a definite truth value and that there is a straightforward procedure for checking its truth value. A claim is nearer the fuzzy end if it is not clear whether it has a definite truth value or how we would go about establishing it.

Here is a plan of the essay. In section 2, I argue that the bracketing of claims adopted in much of the disagreement literature is problematic. I introduce the notions of crispness and fuzziness. I then sketch out an evolutionary argument showing why our capacity to make claims with a definite truth value is likely to be unreliable, and why it is likely to be unreliable in a predictable fashion. In section 3, I consider a possible world where all claims have a definite truth value and come with a straightforward procedure for checking it. In such a world, I argue, conciliationsim would be the best attitude to any peer disagreement. In section 4, I use the findings from sections 2 and 3 as a basis for differentiated treatment of disagreements involving claims along the crispness-fuzziness spectrum. I conclude that disagreements about crisp claims usually call for conciliation and suspension of judgement. In disagreements about highly fuzzy claims, on the other hand, the parties may be justified in remaining steadfast if they need to rely on a belief as a basis for action.

2. Claims are not born equal

The central issue in much of the recent Anglophone literature on disagreement is disagreement between epistemic peers. The problem is typically set up along the following lines (e.g. Christensen 2011, 2013, 2014; Elga 2007; Elgin 2010; Kelly 2005, 2010, 2013a; Kornblith 2010; Lackey 2013; White 2005).

Suppose that two people – call them Ayna and Besha – disagree about p. Ayna believes that p on the basis of a body of evidence; Besha believes that not-p based on the same evidence. (Alternatively, Besha is agnostic about p, or they both believe that p but to different degrees of certainty.) Ayna and Besha know that they disagree. They also know they have the same evidence regarding p. Moreover, they know from past experience that they are equally “competent, intelligent, and fair-minded in their assessment” of this kind of evidence (Lackey 2013, p. 243). In other words, they know that, at least as far as p is concerned, they are “equally likely to be mistaken” (Elga 2007, p. 499).  The question is what they should do next. Should they conciliate, i.e. modify their respective beliefs in light of their disagreement? Or should they ignore the fact of their disagreement as irrelevant to their beliefs regarding p?

Once we accept this as a legitimate question, there seem to be three possible replies: 1) they should conciliate; 2) they should remain steadfast; 3) it depends. If “it depends” has the most intuitive appeal (as it does for me), the question is what it depends on. I want to argue that a key factor in deciding what to do in a disagreement, including but not limited to disagreement with epistemic peers, is the nature of the claim in dispute. In the literature, disputed claims are typically bracketed from the discussion as “p”, “some matter P” (White 2013, p. 312) or “some hypothesis H” (Kelly 2013b, p. 301). When they are not so bracketed, they often appear to function as mere props in scenarios that focus on other factors. As a first step in my argument, I need to present a reason why disputed claims should not receive such placeholder treatment. This is the goal of this section.

As a philosophical problem, the problem of peer disagreement involves a considerable degree of idealization. It is quite possible that genuine epistemic peers do not exist in the actual world. Perhap there is always some significant difference between the epistemic positions of the parties in any real-world disagreement. Perhaps no two actual people can be said to have exactly the same evidence or exactly the same evidence-processing capacity. None of this is to say that the problem of idealized peer disagreement has no bearing on real-world disagreement. After all, idealizations are theoretically useful precisely because they are idealizations. However, in order to remain relevant, the idealization involved in the notion of epistemic peers does need to remain firmly tied to actuality at least in one respect: the peers in question must be imperfect epistemic agents.

To explain what I mean by the notion of an imperfect epistemic agent, I will consider its opposite. A perfect epistemic agent is a being that can know everything there is to know about reality. This awesome ability involves at least three infallible skills. Firstly, such a being can infallibly establish the truth value of any claim to any degree of certainty. Secondly, it can infallibly formulate claims that have a definite truth value. Thirdly, it can infallibly recognize and dismiss claims that do not have a definite truth value.

Suppose such a being exists; call him Young Ayer. Whenever Young Ayer is confronted with a claim that p, he reacts in one of the three following ways. If the claim is true, he carries out an investigation and says, “P!” If the claim is false, he carries out an investigation and says, “Not p!” If the claim is neither true nor false, he goes, “That doesn’t even make any sense.”

Suppose now that Young Ayer has a twin, identical to him in every epistemically relevant way. Suppose also that one day Young Ayer and his twin find themselves disagreeing about a particular claim. Since both of them are perfect epistemic agents as defined above, this can mean one of three things: 1) they are in different realities; 2) there is at least one thing about reality that both is and is not at the same time; 3) there are mysteries open to perfect epistemic agents that we mere mortals can’t even coherently talk about, let alone understand.

Fascinating as these conclusions might be, at first glance they seem to have little relevance for disagreements between actual humans. It is therefore hardly surprising that epistemic gods like Young Ayer have not received much attention in the literature on peer disagreement. As Elga’s emphasis on being “equally mistaken” (rather than “equally right”) presupposes, epistemic peers worth talking about are imperfect epistemic agents. They must be similar to actual humans in that they lack all three infallible skills possessed by Young Ayer. Firstly, they can’t reliably establish the truth value of a given claim – even when it has a definite truth value. Secondly, they can’t reliably formulate claims that have a definite truth value. Thirdly, they can’t reliably recognize claims that don’t have a definite truth value. Let us call these three shortcomings the Epistemic Banes of Humankind.

Bane 1 is routinely acknowledged in the disagreement literature. Indeed, it is unclear how we could make any sense of disagreement at all if we didn’t start from the observation that humans regularly fail in their attempts to evaluate claims about the world. Banes 2 and 3, however, do not seem to receive nearly as much acknowledgement. Instead, one gets a distinct impression when reading the literature that all those sundry claims epistemic peers disagree about are somehow given. It is almost as if they came directly from on high, i. e. from some abstract realm of Platonic propositions. Some authors actually use the term “proposition” to describe a disputed claim (e.g. Christensen 2014); others avoid the p-word while still using “that p” as a placeholder for disputed claims.

To be clear, I do not wish to accuse everyone who uses “that p” of holding a Platonic view of propositions or, worse still, of granting humans epistemic access to wherever such propositions abide. My point is simply that two of the three Epistemic Banes of Humankind (our inability to reliably formulate claims with a definite truth value and our inability to reliably recognize claims that have none) do not seem to be getting as much attention as they likely deserve. Whenever humans disagree, they disagree about claims made by humans. I believe that this trivial observation has important consequences for the philosophy of disagreement.

Here is why I think so.

Let us define a claim as whatever can be worded as a that-clause in a language that has something like that-clauses. For example, “Most cats have four paws”, “Die meisten Katzen haben vier Pfoten” and “La maggior parte dei gatti ha quattro zampe” all get to be claims because they can be embedded, with or without grammatical modification, in main clauses such as “It is true that…”, “Ich bezweifle, dass…” and “Sostiene che…”, respectively.


Let us further assume that some claims have a feature called “crispness”. Let us use the term “fuzziness” to describe a lack of crispness. Any claim will fall somewhere on a crispness-fuzziness spectrum. What does that mean? Roughly, a claim is crisp if it is immediately clear (a) that it has a definite truth value and (b) how we would go about checking its truth value. For example, “Most cats have four paws” is a relatively crisp claim. It is immediately clear what we need to do in order to see if it is true: we need to look at a statistically significant sample of creatures we call “cats” in different parts of the world and count their paws. Whether we can actually carry out such an investigation is not important. Even if we are stuck on a desert island with no access to cats, we still know what it takes to evaluate the claim that most of them have four paws. “Most cats have souls”, on the other hand, is a relatively fuzzy claim. It isn’t immediately clear how we would go about establishing whether any given cat had a soul. (If all this sounds suspiciously verificationist so far, please bear with me; it isn’t supposed to be.)

Is there any general explanation why some claims made by actual humans would be crisper or fuzzier than others? I believe that there is, and that it is an evolutionary one. The argument below is partly modelled on Joyce’s debunking argument for moral skepticism, as presented in Machuca (2018, p. 17). While I don’t think that evolutionary debunking succeeds as a reason for wholesale skepticism, whether moral or otherwise, I believe that at least some of the intuition behind evolutionary debunking is solid and can be put to a better use. To wit, a debunker-inspired argument can be used to show (a) that the Epistemic Banes of Humankind are inevitable and (b) that they should manifest themselves in a predictable fashion.

Here is how such an argument might go:

Premise A. Our capacity to form and evaluate claims about reality is ultimately a product of biological evolution through natural selection.
Premise B. Evolution through natural selection is not aimed at producing capacities that are indiscriminately reliable. It favours capacities that are reliable to the extent that they facilitate survival and reproduction.
Conclusion 1 (from A and B). It is unlikely that our capacity to make and evaluate claims is indiscriminately reliable.
Premise C. When it comes to survival and reproduction, our species is currently among the most successful vertebrate species on the planet.
Conclusion 2 (from A, B and C). To the extent that it facilitates survival and reproduction, our capacity to make and evaluate claims about reality is likely to be reliable.

Needless to say, the argument above is a sketch. A number of details need to be filled in. For example, “reliable” needs to be defined. The capacity to make claims needs to be defined in some neuroscience-friendly terms and tied to an evolutionary story about cognition in general and language in particular. Premise A needs to be defended against the obvious objection that the capacity to make claims about reality might not have been selected for as such; in other words, that it might piggyback on other traits, which were selected for. Premise C needs a convincing explanation of how one measures the evolutionary success of a biological species. And so on.

Fleshing out this argument would be an exciting task, but it goes beyond the scope of this paper. For my present purposes, it is enough to show that there is at least one somewhat plausible account of why our claim-making capacity would be not merely unreliable, but predictably so. On this account, human-made claims are born inherently unequal. Those that have an easily traceable link to factors important for biological survival are likely to be relatively crisp, as defined above. They will typically have a definite truth value, and they will typically come with a relatively straightforward procedure for establishing their truth value. Those that have no obvious link to biological survival are likely to be relatively fuzzy. They will often fail to have a definite truth value, and their evaluation will be far from straightforward.

In the following sections, I discuss some implications of such inherent inequality of human-made claims for the problem of disagreement.

3. What if all claims were born equal?

If some claims are inherently crisper than others, what might that mean for the problem of disagreement in general and peer disagreement in particular?

I believe that a good way to approach this question is by considering a possible world where all claims that humans ever disagree about are equally crisp. In other words, consider a world where there is only one Epistemic Bane of Humankind. In this world, humans can infallibly form claims about reality that have a definite truth value, and they can infallibly recognize a claim that has none. The only epistemic limitation humans have is that they sometimes fail to establish the truth value of a given claim correctly. Call this world Crisp World.

Crisp World is everything that it is the case. (“Die scharfe Welt ist alles, was der Fall ist”, as young Wittgenstein could have put it.) Everything that is the case in Crisp World can be stated as a unique set of impeccably crisp claims made in human language (assume that, as a matter of empirical fact rather than armchair stipulation, all human languages in Crisp World have that-clauses).

If a human Crisp-Worlder formulates a new claim about reality, another Crisp-Worlder generally responds in one of two ways. If the claim has a definite truth value, she might say something along the following lines: “Oh, how interesting. I wonder if that is true.” If the claim does not have a definite truth value, she will automatically treat it as a joke or perhaps an exercise in poetry, and she will respond accordingly.

Now that the scene is set, let us take a closer look at Crisp-Worlders and disagreement. First of all, why would Crisp-Worlders disagree about anything? The answer is straightforward. Even though Crisp-Worlders are only plagued by one Epistemic Bane rather than three, they are still imperfect epistemic agents since there are claims about reality they can’t evaluate to a required degree of certainty.

Here is an illustration. Suppose that one Crisp-Worlder says, “It is the case that there are exactly 1,257,049 hedgehogs alive on Earth as we speak.” Another Crisp-Worlder might object by saying that there are exactly 1,257,051 hedgehogs. A third Crisp-Worlder, in her turn, might counter: “One of you might be right, or you might both be wrong. Unfortunately, we lack the practical means to establish the exact number of hedgehogs on the planet at any given moment. While your respective claims certainly have a definite truth value, the best we can do given our epistemic imperfection is form a more or less rough estimate and suspend our judgement as to the exact figure.”

Furthermore, not all cases of disagreement in Crisp World have to be so trivial. Consider two Crisp World experts on hedgehogs who disagree in their estimates of the global hedgehog population. One expert puts it in the ballpark of three million; the other argues that a figure of about two million is more realistic. A third expert, well versed as she is in the recent Crisp World philosophy of disagreement, might then say: “Colleagues, one of your estimates is probably closer to the truth than the other, or perhaps they are equally close to the truth, though that is less likely. There must be a difference in your assumptions, your calculations, or in the empirical data you’re relying on. Why don’t you go through your data and methodology together? Or, if that doesn’t help, why don’t you publish your results and let the academic community retrace your steps? If we join forces, we will eventually arrive at the best estimate supported by the available evidence. Until then, let us suspend judgement.”

This second case of disagreement may be less trivial, but it still sounds boringly similar to some of the less interesting disagreements that occur in our own fuzzy world. This is hardly surprising: claims about the number of hedgehogs on Earth, whether exact or approximate, are relatively crisp even when they are made in the actual world by actual humans. Even here, they are of about the same order of crispness as the claim that most cats have four paws. To bring in relief a fundamental difference between Crisp World and our world, we need to take another rigorous leap of imagination and recognize that all cases of disagreement in Crisp World are ultimately like the ones I have just described – because all claims in Crisp World are like the claim that most cats have four paws.

Demonstrating that by means of a hypothetical scenario of the kind I have already used is tricky. Recall that any claim put forward by a Crisp-Worlder (unless it is made in jest or for poetic reasons, or perhaps in a fit of madness) both has a definite truth value and is immediately recognized as having one. Moreover, for any claim made by a Crisp-Worlder, it is immediately clear what procedure could be used to establish its truth value. This is true even in those cases when following the procedure in question is not feasible. In other words, if we want to translate any given Crisp World claim into English or any other human language used in the actual world, we need somehow to approximate their original degree of crispness, so that the resulting translations do not strike us as hopelessly fuzzy right from the start. Because this is not possible (due to Epistemic Banes 2 and 3), I will instead use an asterisk to show that a claim, or a part of a claim, is in Crisp World English rather than in Actual World English.

Consider the following claims made in Actual World English:

A. God exists.
B. Most cats have souls.
C. Palme’s time in power was good for Sweden.

Now compare them with the following claims, made in Crisp World English:

A*. God exists.*
B*. Most cats have souls.*
C*. Palme’s time in power was good for Sweden.*

The two sets of claims are strikingly different. To see why, we need to look at what happens when A, B or C are uttered in our world and when A*, B* or C* are uttered in Crisp World.

Let us focus on A and A*. When we claim in public that “God exists”, a disagreement often ensues, and then a number of things tend to happen. First of all, the disagreeing parties typically start talking past each other. As a rule, each participant in the debate has her own, usually rather vague, notion of God, and attempts to agree on a common definition often lead to the participants moving the goalposts all over the logical space. To complicate matters further, there is rarely any consensus on what evidence, empirical or otherwise, can be brought to bear on the question, or if there can be, or need be, any relevant evidence in the first place. Someone may declare “God exists” to be self-evidently true; others will dismiss the statement as nonsensical; and 2400 years later we will still be debating the question.

Now consider what happens in Crisp World. Over there, when someone claims in earnest that “God exists*” and a disagreement follows, things unfold very differently. In Crisp World, the claim “God exists*” is impeccably crisp. Everyone knows exactly what God* is and what it means for God* to exist*, or what it would mean for God* to exist* in case she* doesn’t*. What is more, everyone has a fairly good idea how to check whether “God exists*” is true.

Thus, if one Crisp-Worlder says “God exists*” and another replies “No, she doesn’t*”, a third Crisp-Worlder will react in one of the following ways. If establishing whether “God exists*” is about as straightforward as checking if it is raining outside, she might say: “Why don’t you both suspend judgement until we can go out and check?” If it is more complicated, she might reply: “It takes a lot of energy to check if that’s true. However, I’ve read that China is building a Large Godhead Diviner for that very purpose. Soon we will all know if God* exists*. Why don’t you suspend judgement until then?” Finally, checking if “God exists*” might not be feasible for practical reasons; perhaps it requires more energy than Crisp World humanity can hope to produce in the foreseeable future. In that case, the third Crisp-Worlder might say: “Unfortunately, we can’t generate enough energy to check whether God* exists*. We all just have to remain agnostic on this one, so why don’t you two stop being silly.”

Cases of Crisp World disagreement regarding B* or C* would follow the same script. Everyone in Crisp World knows what it entails for a cat* to have* a soul*, or how exactly someone’s time* in power* can be* good* for Sweden*, and everyone knows how to go about checking claims that involve those notions, even if an actual investigation is not always possible for practical reasons. Thus, a Crisp World Wikipedia entry on Palme* might state as a matter of empirical fact that “Palme’s time in power was good for Sweden*”; or that “Palme’s time in power was not good for Sweden*”; or that there is currently no way to establish whether “Palme’s time in power was good for Sweden*”. It is perhaps worth noting that this last agnostic statement would not mean (as similar statements often do in our world) that the relevant expert community in Crisp World was somehow divided on the issue, with some arguing that p, others defending not-p, and still others sitting on the fence. Instead, it would refer to a genuine agnostic consensus that there was no way to tell either way (perhaps because all the relevant historical data had been destroyed by the Freemasons*).

In other words, Crisp World is a world where conciliationism reigns supreme when it comes to disagreement between epistemic peers. As we have seen, if you and your epistemic peer disagree about a claim in Crisp World, the only rational reaction is to modify your belief in light of the disagreement and suspend judgement until you can look and see who is right – or suspend judgement forever, if checking who is right is not possible for practical reasons.

The conclusion I am led to draw after considering disagreement in Crisp World is this: if all claims were born equally crisp, indiscriminate conciliationism would be the most appropriate normative attitude to peer disagreement. However, as I have argued earlier, claims made by humans in the actual world are not indiscriminately crisp. For evolutionary reasons, some of them are likely to have a definite, straightforwardly knowable, truth value; others are likely to have none; still others are likely to languish somewhere in the middle, all fuzzy around the edges but with something like a crisp core inside.

To what extent does conciliationism still hold in our world of widespread fuzziness? This is the subject of the next section.

4. Conciliatory crispness and steadfast fuzziness

In this section, I argue that we should adopt a differentiated approach to disagreement (including but not necessarily limited to peer disagreement) based on the position of a given claim on the crispness-fuzziness spectrum.

Here is a short summary of the reasons for such an approach that I have offered so far. I began by introducing the notion of crisp and fuzzy claims. Crisp claims have a definite truth value and come with a straightforward procedure for establishing their truth value. Fuzzy claims may or may not have a definite truth value, and it is not clear how they should be evaluated, or whether they can be evaluated in the first place. We have evolutionary reasons to believe, I then argued, that some human claims are inherently crisper than others, and that this variation in crispness is predictable. Following that, I discussed a possible world with no such variation, where all claims made and evaluated by humans are crisp. I concluded that in such a world the best reaction to any peer disagreement is to conciliate by suspending judgement, either until the disagreement can be resolved by looking at the relevant evidence, or forever if looking at the evidence is not possible.

Given the discussion so far, what follows for disagreement in the actual world? The first thing to note is that some actual-world disagreements are about relatively crisp claims. It seems reasonable to suggest that the correct attitude to disagreement about such claims should be thoroughly conciliationist. If a claim appears to have a definite truth value, and it is fairly clear how that value could be established, we should suspend judgement until we can perform the necessary check. Once we have carried out the check, we can either resolve the disagreement or dismiss the other party, just like a Crisp-Worlder would automatically dismiss a nonsensical claim as a joke, a piece of art, or a sign of madness.

It is important to note that the duration of the check required can be as short as a split second. For example, if you believe that there is no bear standing right in front of you in a brightly lit room, and someone claims that there is, it is usually enough to look right in front of you to dismiss the claim as a silly joke or a sign of delirium tremens, provided that you are in full possession of your eyesight and know yourself to be sober and sane. Compare this with a disagreement about a different relatively crisp claim: that there was a bear exactly on the spot of the Carolina Rediviva library in Uppsala on the day Socrates drank hemlock. Suppose that your statistical calculations and your estimate of the bear population density in the Uppsala area around 399 B.C.E., based on the available evidence, lead you to believe that the claim is likely true. A friend, who is generally just as good at statistics and at estimating past bear populations, takes the same evidence to indicate that the claim is likely false. In this case, it might take a while, if not forever, to unearth any conclusive evidence in favour of either view, and the suspension of judgement may well need to be permanent.


It can be objected here that the notion of a “relatively crisp claim” in the context of actual-world disagreement is circular because it seems to boil down to “a claim that doesn’t really cause disagreement”. One might argue, in other words, that advocating suspension of judgement in disagreements about crisp claims is like advocating suspension of judgement in disagreements that require suspension of judgement.

This objection, however, does not fully appreciate the evolutionary origin of crispness in human claims. As I argued in section 2, the reliability of our capacity to make claims about reality is likely to be linked to survival and reproduction. It is worth stressing here that by the reliability of our capacity to make claims about reality I don’t mean a capacity to make true claims about reality. Rather, I mean a capacity to make claims about reality, i.e. claims that have a definite truth value. In other words, what I have in mind is crispness, as defined above, and what I am arguing is that the inherent crispness of our claims is ultimately linked to our success as a species. By “ultimately”, I mean that the link goes through an intermediary. To the extent that the reliability of our claim-making capacity can be linked to survival and reproduction, it can be so linked by means of the concepts that we use to make claims.

Before I go on, I probably need to clarify what I mean by “concept”. For the purpose of this essay, a concept should be understood as a functional kind: concepts are whatever mental representations we use in categorizing and organizing our past experience so that it can serve as a basis for future action (this definition is based on the discussion in Lalumera 2009, pp. 11-15). I take it for granted that at least some lexemes in a given human language are used to signify concepts thus understood (though this fact by no means exhausts the question of what else these lexemes mean as part of their linguistic system or when used in a particular speech act). For example, the words medved’, ours and bear all signify a particular mental representation in my brain that can roughly (i.e. non-academically) be described as a large animal that lives in the forest, has brown fur, eats berries, wild honey and pretty much everything else, sleeps in a den during winter, features prominently in Russian folk tales, and so on. It is perhaps worth stressing that concepts, understood functionally, don’t have to be linked to any linguistic labels. Pre-linguistic children and non-human animals alike have mental representations that they use to categorize and organize their experience. However, since making claims (as defined in section 2) is a decidedly linguistic capacity, my primary interest is in concepts that do have linguistic labels. In what follows, when I use a word such as “bear” in quotation marks, I simultaneously refer to the mental representation (if any) that the word signifies.

What does it mean for a concept to be linked to survival and reproduction? To stick with the same example, let us consider the concept “bear”. I want to argue that concepts such as “bear” can be linked to our success as a species. I don’t mean that as a caricature of an argument from evolution, whereby the concept “bear” is especially cognitively salient because prehistoric hominids who failed to perceive bears got eaten and eliminated from the gene pool. What I mean instead is that bears are typical robust medium-sized, medium-speed phenomena of the kind that have surrounded us and our ancestors on this planet for as long as we have been able to form mental representations. For millions of years, our ability to survive and reproduce must have depended on our ability to deal with robust medium-sized, medium-speed (RMS2 for short) phenomena, for the obvious reason that we are a phenomenon of the same scale.

Dealing with RMS2 phenomena involves skills such as distinguishing their presence from their absence in space and time, establishing their potential causal relation to oneself (e.g. Can I eat it? Can it hurt me? Can I mate with it? Can I hide in it?), predicting their motion and their reaction to stimuli, and so on. The fact that our species has been able to survive, multiply and kill off much of the competition strongly suggests that we have mastered such skills to a sufficiently high degree. Put differently, it suggests that we can reliably use concepts to represent reality at least to the extent necessary for certain kinds of interactions with RMS2 phenomena such as ourselves. This conclusion, so far as I can tell, doesn’t hinge on any particular theoretical commitment regarding the ultimate ontology of RMS2 phenomena. Whether RMS2 phenomena are ultimately fields or quantum gravity loops or internal fluctuations of one big timeless thing akin to Spinoza’s god is not important. The only ontological assumption we do need to make is that these phenomena are at least as real as we are; that they are a part of reality at least to the same extent as we are (regardless of our ultimate ontology). For reasons that I can’t go into here, this assumption strikes me as both modest and sufficient to warrant talk of reality in conjunction with our capacity to make claims.

To put all this in terms of my earlier discussion, claims involving concepts that categorize and organize RMS2 phenomena stand a better chance of being crisp than claims that do not involve, or only partially involve, such phenomena. As an illustration, let us briefly revisit some of the claims considered in this essay.

“Most cats have four paws.”

This claim is highly crisp as it only involves RMS2-related concepts. Cats and their paws are typical RMS2 phenomena. Part-whole relations of the kind implicit in the relevant use of “have” are ubiquitous throughout the RMS2 domain, including our own bodies. Quantification (“most”, “four”) is arguably a straightforward extension of the capacity to distinguish absence from presence, as evidenced by studies of rudimentary numerical cognition in non-human species, including non-primates (Agrillo 2016; Beran et al. 2016).

“Most cats have souls.”

This is a fairly fuzzy claim. Though the “most cats” part is anchored in the RMS2 domain as per the analysis above, we quickly run into trouble at “soul”, and the trouble with “soul” affects the interpretation of “have”. Unless we define “having a soul” as another way of saying “being alive” or “being clingy and easily scared”, to name just two such options, the concept “soul” doesn’t seem to capture any easily recognizable RMS2 phenomenon and lends itself to a number of competing interpretations. Even if we agree on a particular non-RMS2-based definition (e.g. “A soul is an immaterial substance that departs the body once the cat is dead” or “A soul is an effervescing spark of the world spirit”), we might simply end up with a different fuzzy claim on our hands and with just as little idea of how to check if it’s true.

“God exists.”

This is a fuzzy claim. Regarding “God”, see the discussion of “soul” above. “Exist” may look like an RMS2-friendly concept if we define it as “there is” and apply it to a cat or an island in a particular location, such as “your flat” or “Lake Baikal”, but it is not immediately clear what “there is” means when applied to an entity such as the God of the Gospel of John residing in reality as a whole.

“Olof Palme was murdered in the year of the Chernobyl disaster.”

This is a relatively crisp claim. Olof Palme, the Chernobyl disaster and the action of murdering someone are all RMS2 phenomena, easily localizable in spacetime thanks to our practice of copious record-keeping. “In the year of” is likely to be referring to a particular system of keeping track of time, based on counting day-and-night cycles. Once we have clarified which particular time-tracking system we are talking about (say, the Gregorian calendar), establishing the truth value of the claim is quite straightforward.

“Palme’s time in power was good for Sweden.”

This is a relatively fuzzy claim. “Palme”, “time” and the genitive link between them are firmly RMS2-anchored and can be used to make relatively crisp claims such as “Palme’s time in the laundry room is over” or “It was Palme’s time to bake cinnamon buns”. “Time in power”, like “time in the laundry room”, refers to an explicit social convention, in this case of making someone a leader based on a sequence of highly ritualized interactions. Social interactions, including those that form power structures, are squarely in the realm of RMS2 phenomena that we have evolved to track. In other words, even a claim such as “Palme spent some time in power” would still be relatively crisp. What makes “Palme’s time in power was good for Sweden” hopelessly fuzzy is the use of “good” in conjunction with “for Sweden” and someone’s time in power. The use of “good” in this claim parallels its use in claims like “Fresh air is good for your health” or “Her time in St. Petersburg was good for her Russian”, where “good” indicates that something helps someone achieve a desired goal, namely becoming healthier or improving one’s fluency in Russian. However, these last two claims are firmly rooted in RMS2 concepts: each of them states that a particular RMS2 phenomenon (breathing fresh air and spending some time in a community of speakers of a language, respectively) conferred a particular benefit on a particular person. This is manifestly not the case in “Palme’s time in power was good for Sweden”. Sweden is not a robust medium-sized, medium-speed phenomenon of the kind we have been dealing with for millions of years. Whether we define it as an “imagined community” (Anderson 1991) or a set of millions of actual people living on an actual piece of the Earth’s surface, it is not immediately clear what desired goals it should have – or indeed could have, considering that it is not clear how such a thing can desire anything in the first place.

Needless to say, those were five rough applications of a rough idea. My hope is that they can at least serve to clarify how a differentiated claim-based approach to disagreement might work. In particular, the fuzzy examples aim to show that, as we move towards the fuzzy end of the crispness-fuzziness spectrum, purely epistemic considerations might eventually give way to practical ones. As noted by Catherine Elgin, to suspend judgement regarding a claim because of peer disagreement often means “refraining from using either p or not-p as a premise in assertoric inferences or as a basis for action”. This, Elgin observes, is a “cognitively impoverished stance” (Elgin 2010, p. 65).

Consider again what happens when we disagree about crisp claims. In such cases, we can be relatively certain that our respective stances are either right or wrong as stated; that there is a relatively straightforward procedure for establishing who is right; and that we can resolve our disagreement either by using that procedure ourselves or by deferring to those who can do it for us. If the disputed claim is crisp, the distinction between peer disagreement and non-peer disagreement can be very important. We have every reason to expect an expert to know more than a layperson does about the truth value of crisp claims, or about the procedures required to establish their truth. To use Elgin’s expression, we run a high risk of cognitive impoverishment if we are not prepared to conciliate (and suspend judgement until we know better) in a disagreement about a crisp claim.

However, the conciliation pendulum may well swing the other way the fuzzier things become. To begin with, in a disagreement about highly fuzzy claims it is not even always clear if there is a relevant epistemic hierarchy to defer to. What does it mean to be in the same epistemic position with regard to claims such as “Most cats have souls”, “God exists” or “Palme’s time in power was good for Sweden”? Even more intriguingly, what does it mean to be in a superior epistemic position vis-à-vis such claims? Conceivably, someone might suggest that we defer to vets on the issue of feline souls and to priests, or to Richard Dawkins, on the issue of God’s existence. Up to a point (say, the point of secondary school) this might even make sense. There is, however, an unmistakable whiff of absurdity about the suggestion that most adults should simply accept expert opinion on such matters. The absurdity becomes even more pronounced when we turn to politics. Should a left-wing political scientist and a right-wing political scientist conciliate regarding Palme’s legacy? Come to think of it, should a miner who had never finished high school necessarily defer to a political scientist on this matter?

The questions are only partly rhetorical. A considerable number of authors would be quick to say “Yes, they should” and “Yes, he should”. However, my attempt to treat the nature of disputed claims as a key consideration in approaching disagreement has, at least for now, led me to believe otherwise. When it comes to fuzzy claims with clear practical implications, our goal of avoiding false beliefs about reality may sometimes become less pressing than our need to have a basis for further action (including further theorizing in philosophy). In such cases, if we become convinced that some p is more likely that non-p and then discover that some very intelligent person thinks otherwise, we may well be justified in remaining steadfast.

In case anyone cares to cite this essay

Zarubin, K. (2018). All claims are not born equal. Unpublished manuscript. Retrieved from


Agrillo, C. (2016). Numerical and Arithmetic Abilities in Non-primate Species. In R. Kadosh & A. Dowker (Eds.), Oxford Handbook of Numerical Cognition (pp. 214–236). Oxford: Oxford University Press.

Anderson, B. (1991). Imagined Communities. London: Verso.

Beran, M. J., Perdue, B. M., & Evans, T. A. (2016). Monkey Mathematical Abilities. In R. Kadosh & A. Dowker (Eds.), Oxford Handbook of Numerical Cognition (pp. 237–257). Oxford: Oxford University Press.

Christensen, D. (2011). Disagreement, Question-Begging, and Epistemic Self-Criticism. Philosophers’ Imprint, 11(6).

Christensen, D. (2013). Epistemic Modesty Defended. In D. Christensen & J. Lackey (Eds.), The Epistemology of Disagreement. New Essays (pp. 77–97). Oxford: Oxford University Press.

Christensen, D. (2014). Conciliation, Uniqueness and Rational Toxicity. Noûs, 50(3), 584–603.

Elga, A. (2007). Reflection and Disagreement. Noûs, 41(3), 478–502.

Elgin, C. Z. (2010). Persistent Disagreement. In R. Feldman & T. A. Warfield (Eds.), Disagreement (pp. 53–68). Oxford: Oxford University Press.

Kelly, T. (2005). The epistemic significance of disagreement. In J. Hawthorne & T. Gendler (Eds.), Oxford Studies in Epistemology, Volume 1 (pp. 167–196). Oxford: Oxford University Press.

Kelly, T. (2010). Peer Disagreement and Higher-Order Evidence. In R. Feldman & T. A. Warfield (Eds.), Disagreement (pp. 111–174). Oxford: Oxford University Press.

Kelly, T. (2013a). Disagreement and the Burdens of Judgement. In D. Christensen & J. Lackey (Eds.), The Epistemology of Disagreement. New Essays (pp. 31–53). Oxford: Oxford University Press.

Kelly, T. (2013b). Evidence Can Be Permissive. In M. Steup, J. Turri, & E. Sosa (Eds.), Contemporary Debates in Epistemology (2nd ed., pp. 298–311). Wiley-Blackwell.

Kornblith, H. (2010). Belief in the Face of Controversy. In R. Feldman & T. A. Warfield (Eds.), Disagreement (pp. 29–52). Oxford: Oxford University Press.

Lackey, J. (2013). Disagreement and Belief Dependence: Why Numbers Matter. In D. Christensen & J. Lackey (Eds.), The Epistemology of Disagreement: New Essays (pp. 243–268). Oxford: Oxford University Press.

Lalumera, E. (2009). Cosa sono i concetti. Roma-Bari: Editori Laterza.

Machuca, D. E. (2018). Moral Skepticism: An Introduction and Overview. In D. E. Machuca (Ed.), Moral Skepticism: New Essays (pp. 1–31). New York: Routledge.

White, R. (2005). Epistemic Permissiveness. Epistemology, 19(1), 445–459.

White, R. (2013). Evidence Cannot Be Permissive. In M. Steup, J. Turri, & E. Sosa (Eds.), Contemporary Debates in Epistemology (2nd ed., pp. 312–323). Wiley-Blackwell.


Один комментарий на «»All claims are not born equal»»

Добавить комментарий

Заполните поля или щелкните по значку, чтобы оставить свой комментарий:


Для комментария используется ваша учётная запись Выход /  Изменить )

Фотография Facebook

Для комментария используется ваша учётная запись Facebook. Выход /  Изменить )

Connecting to %s

Создайте сайт или блог на

%d такие блоггеры, как: