Author Information: Chris Lepock, Athabasca University, email@example.com
Lepock, Chris. “Norms and the Temptations of Relativism: A Reply to Sankey.” Social Epistemology Review and Reply Collective 4, no. 5 (2015): 37-42.
The PDF of the article gives specific page numbers. Shortlink: http://wp.me/p1Bfg0-25N
Please refer to:
- Sankey, Howard. “Scepticism, Relativism and a Naturalistic Particularism.” Social Epistemology (2015): 1-18. doi: 10.1080/02691728.2014.971909.
Image credit: Dermot O’Halloran, via flickr
Epistemic relativism is the view that justification and rationality are relative to sets of epistemic norms. We can examine whether a claim and its evidence meets or fails a set of norms, but there is no meaningful answer to the question of what set of norms are correct. All norms have equivalent status; none can claim to be more rational than any other.
Howard Sankey (2015) argues that this sort of relativism is motivated by the problem of the criterion, which runs like this. If you are to know that p or be rational in believing it, you need some criterion or principle that identifies p as likely to be true. But if you have such a criterion, you need some reason for thinking it is accurate. That requires either a meta-criterion that identifies accurate criteria, which would have to be supported by a meta-meta-criterion, etc. ad infinitum. Or it would require a pre-existing stock of beliefs known to be rational that you could test the criterion against – but those beliefs would need to be vindicated by criteria, putting us right back on the merry-go-round. Either way, the argument goes, you can’t show that any proposition is more credible than any other.
Relativists are not skeptics. Relativists allow that claims can be justified relative to arbitrarily chosen norms, whereas skeptics think that an unsupported norm justifies nothing. That is, relativists limit their skepticism to the question of how epistemic norms are justified; they do not let the unjustifiedness of the norms infect beliefs. Nonetheless, Sankey thinks the link between skepticism and relativism is strong enough that he can answer the relativist with an antiskeptical strategy.
He advocates a form of naturalism, according to which the only standards of rationality we must meet are the ones used in ordinary scientific investigation. Skeptical questions about how we know that this method is reliable are set aside. So we can study epistemic norms empirically by looking at whether investigations using them yield the right results or not. For instance, take an epistemic norm of the type “believe everything very old book V says.” We can test this norm by testing what V says. If too much comes out false, we can infer that the V-norm is not reliable and thus inferior to the scientific method. So contra relativism, we have a method for evaluating epistemic norms, and we know that they are not all equivalent.
Now the relativist has an obvious reply here. All we are doing is judging one norm (the V-norm) by whether it conforms to another set of norms (standard scientific method). So the argument for the V-norm’s inferiority is circular: it presumes that scientific norms are correct, and then rejects the V-norm for disagreeing. But if we flipped it around and judged the scientific method by the V-norm, then we would “discover” that science was unreliable. (Or maybe large chunks of science are only metaphorically true.) Since we can use either norm to judge the other, their status must be equivalent. Epistemic relativism is not only not refuted, but affirmed.
Sankey argues that we can legitimately use scientific methods to evaluate other norms. As for why, at times he seems to endorse two different (though related) positions. In section 5, most of 6, and footnote 18 he seems to argue for Roderick Chisholm’s particularism; in the last paragraph of 6 and section 7, he seems to advocate reliabilism.  I’ll briefly discuss each interpretation and why neither one makes for much of an argument against relativism.
Chisholm (1973) argues that there is no way out of the problem of the criterion without begging the question. If you do not reject the skeptic’s demands, you have to accept the skeptic’s conclusion. He categorizes positions based on which of two questions they beg.  Methodists beg the question by postulating criteria for knowledge – say, that everything coming from the senses or the meanings of words is known, or everything clearly and distinctly perceived is known. From this the methodists try to figure out what they know. Chisholm’s favoured position, which he traces back to William Reid and G. E. Moore, is particularism. This begs the question by postulating there are things we obviously know – I have a hand, 2+2=4, pure snow is white, etc. Then by examining where these obvious cases of knowledge come from, we can determine what the criteria of rationality are. 
So if Sankey is embracing particularism, then he’s not begging the question against relativism by assuming scientific norms are correct. He’s assuming we can evaluate norms by comparing them to what we know. If examining chicken entrails predicts there will be a great storm tomorrow, and the local meteorologist says it will be sunny, then we can see what happens and determine which norm worked better: “trust your meteorologist” or “trust your (chicken’s) gut”. We’re not preferring one set of norms over another because we’re evaluating them all by the same standard: whether they are reliable given what we know.
But the relativist would be leaning over my shoulder to blue-pencil that into “what we think we know”. Very much hinges on what instances of knowledge we start with. If we say it is obvious that the earth goes round the sun, that humans are a type of great ape, or that Jupiter has moons, we will quickly find that scientific norms are generally reliable. But if we started off by assuming that enough of very old book V was obviously true, we’d get very different results.
Maybe what is fair is to restrict the obvious cases of knowledge to the sorts of thing any minimally rational person of any background could agree on – something like the intersection of all human cultures’ sets of putatively obvious truths. But then we face the oft-noted problem that much scientific evidence is too theory-dependent for there to be any clear way of showing it true using only obvious commonsense knowledge as a starting point. (See, e.g., Kuhn 1996.) If all we will accept as obvious is that looking down a certain tube on a clear night lets you see a round thing and some littler ones, there is still lots of work that needs to be done to establish that you can see Jupiter’s moons. And it’s a very hard slog from claims every minimally rational person would endorse to the result that certain numbers on a certain screen at CERN are evidence of the Higgs boson.
Relativists can say that this situation illustrates exactly what they’re trying to get at. Anyone can call their most confidently held beliefs “knowledge” and test norms by whether they agree with the “knowledge”. If the relativists want to be extra-cheeky, they’ll advocate relativist particularism, the view that a norm is justified iff it is reliable over some set of beliefs deemed as obvious instances of knowledge, and any set of obvious instances is as good as any other.
So let’s look at the reliabilist response and how it fares. On this account, the skeptic’s error is insisting that if you do not know that a reliable criterion vindicates your belief, you do not know that belief. The belief only needs to be reliable – produced and sustained in a way that is sufficiently likely to yield truth. Key here is that the believer need not know the belief is reliable. Not being able to give an adequate account of the reasons for your beliefs does not mean you lack good reasons for them. 
If our picture of the world is generally right, then our usual belief-forming methods (empirical science, perception, etc.) are reliable. So we can use them to evaluate methods we’re not sure about (like checking chicken entrails). Note that the problem of the criterion doesn’t claim that our picture of the world is not right, only that we cannot give adequate reasons for thinking that it is. Without positive evidence that human reasoning or the scientific method is unreliable,  we should continue using our usual belief-forming strategies without worrying about skeptical doubts.
As we saw with particularism, others can use this strategy too. Those who follow a V-norm can say that if V is a God-given source of truths (as they believe), then they know what it tells them even if they have no corroborating evidence. If consulting chicken entrails is reliable, as the augurers believe, then we know those prognostications.  The externalist does not take this to show that these norms are on a par with ours because the arbiter is actual, not perceived, reliability. From our point of view (so to speak) we may seem in the same boat as V-normers and entrail inspectors, but somebody’s methods are not actually reliable, and that somebody is wrong. An unreliable method is not on a par with a reliable one just because its users cannot tell it is unreliable. Error can run so deep it cannot recognize itself as such.
However, the relativist can still ask: what reason do we have for thinking that our methods are the ones that reliably uncover the truth, so that using our methods to evaluate epistemic norms can tell us whether those norms are reliable? After all, the naturalist has to choose whether to start off with scientific inquiry or a V-norm, and if the only answer is “well, I wasn’t brought up to believe V,” the relativist is going to pounce.
In section 7 (in his response to the second and third objections), Sankey seems to give an answer. Every culture needs to acquire true beliefs about the world and make accurate predictions in order to survive. Thus anyone has an interest in using reliable norms and eschewing unreliable ones, and no one can ignore empirical evidence about the reliability of their norms.
We may want to qualify this: no one can ignore what they consider to be empirical evidence. We would ignore a believer in Azande poison oracles who tested our psychology and found that it is just awful at identifying witches. But oraculists could similarly reject plenty of the data we offer for failing to meet their standards of evidence. Besides, there’s no shortage of creative ways to accommodate or discount evidence of norm unreliability. (Big Pharma manipulated the results so they could sell more vaccines. My divination method only failed your test because the gods find your lack of faith disturbing.) There can even be epistemic norms that command this. (The gods sent empirical psychologists to tempt us into error.)
I think Sankey’s idea is that no one should reject reliable data about epistemic norms, regardless of what their norms say about it. If so, I agree heartily. Medical knowledge cannot just be relative to cultural beliefs and practices because it is literally a matter of life and death. In saying this, Sankey and I are supposing that reliability is a criterion of rationality regardless of what epistemic norms a person endorses. A belief can be irrational because unreliable, even if the believer’s norms vindicate it.
Now Sankey assumes that his opponent also agrees. More than that, he assumes that his opponent agrees that epistemic norms must tend to lead to truth and that truth is objective (14). Mind you, if that’s the view Sankey wants to refute, a much briefer argument will do it: there is no way that all epistemic norms could be on a par because there is no way all epistemic norms are equally conducive to reaching a single objective truth.  Take these three candidates for norms of rationality:
Believe the simplest explanation of your sensory experience.
Believe everything Hesiod says.
Believe everything The Silmarillion says.
No matter what the truth is, at least two of those putative norms are not truth-conducive. Sankey doesn’t say what sort of jiggery-pokery he thinks his opponent is using to keep this view halfway coherent.
But jiggery-pokery abounds in the literature, and much of it has the potential to defuse Sankey’s argument from reliabilism. One possibility is to reject either the objectivity of truth or truth-conduciveness of norms. Either of these might be considered too revisionary, to put it politely. A more interesting idea is to accept both postulates but take them as platitudes, more or less: objective reliability, construed as something independent of our conception of the world, can have no explanatory power. There is no “God’s-Eye View” from which we can examine different norms and see which yields truth most often. The only test of truth is measuring beliefs and their etiologies by the standards we believe to be most truth-conducive. So truth, inasmuch as it is a notion we can use, is closely connected with rational acceptability—and measuring rational acceptability by likelihood of attaining truth becomes a pretty tight circle. 
Richard Rorty – not a relativist per se, but a “pragmatist” who doth protest too much – extends this latter line of reasoning. All that can meaningfully be said about truth is how the predicate “is true” is used (see, e.g., Rorty 1986). Absolute truth is not a standard for inquiry, because “there are no constraints on inquiry save conversational ones — no wholesale constraints derived from the nature of the objects, or of the mind, or of language, but only those retail constraints provided by the remarks of our fellow-inquirers” (1980, 726).
On this account, empirical studies of epistemic norms can at best be among the “retail constraints” for those inquirers who think them relevant. Sankey’s argument needs reliability to be something more than a certain move in a certain “conversation”, and Rorty doesn’t admit a conception of truth with enough substance for that.
So, there are many ways for the relativistically inclined to respond to this line of argument. Sankey might nonetheless claim to have refuted a certain brand of relativism. But crude relativism needs refutation about as much as the divine right of kings. It’s never been easy to pull off an absolute claim that all claims are relative. The interesting positions are the ones like Rorty’s that are not bluntly relativist, but still have the implications that epistemic norms are historically contingent, that there is no neutral ground from which to evaluate alternative views, and that there is minimal difference between what a community calls “right” and what is right for them.
There must be some dark temptation that leads people to such views. We should try to understand what it is, and how to show it wrong, rather than argue against simplified versions of the tempters.
BonJour, Laurence, and Ernest Sosa. Epistemic Justification: Internalism vs. Externalism, Foundations vs. Virtues. Malden, MA: Blackwell, 2003.
Chisholm, Roderick. The problem of the Criterion. Milwaukee, WI: Marquette University Press, 1973.
Kuhn, Thomas S. The Structure of Scientific Revolutions. 3rd ed. Chicago: University of Chicago Press, 1996.
Nisbett, Richard, and Lee Ross. Human Inference: Strategies and Shortcomings of Social Judgment. Englewood Cliffs, N.J.: Prentice-Hall, 1980.
Putnam, Hilary. Realism with a Human Face. Cambridge, MA: Harvard University Press, 1990.
Rorty, Richard. “Pragmatism, Relativism, and Irrationalism.” Proceedings and Addresses of the American Philosophical Association 53 (1980): 719-38.
Rorty, Richard. “Pragmatism, Davidson and Truth.” In Truth and interpretation: Perspectives on the philosophy of Donald Davidson, edited by Ernest LePore, 333-55. Oxford: Blackwell, 1986.
Russell, Bertrand. The Problems of Philosophy. New York: H. Holt, 1912.
Sankey, Howard. “Scepticism, Relativism and a Naturalistic Particularism.” Social Epistemology (2015): 1-18. doi: 10.1080/02691728.2014.971909.
Siegel, Harvey. “Epistemological Relativism: Pro and Con.” In A Companion to Relativism, edited by Steven D. Hales, 201-18. Oxford: Wiley-Blackwell, 2011.
Sosa, Ernest. “The Raft and the Pyramid: Coherence Versus Foundations in the Theory of Knowledge.” Midwest Studies in Philosophy 5 (1980): 3-25.
 The relation between these is that Sosa (1980) argues that reliabilism has the antiskeptical merits of particularism, but gives a more comprehensive theory of knowledge.
This is not an exhaustive categorization – it misses, for one, coherentism, which would have us start off with a hypothesis about criteria and a hypothesis about items of knowledge and then make mutual adjustments until it all makes sense. The externalist response is similar to particularism but tries to beg the question less bluntly – it tries to redefine the standard for knowledge in a way that takes away the need for criteria. I’ll discuss that shortly.
 It’s maybe worth noting that Sankey omits Chisholm’s argument for particularism, only giving the preliminary step that to answer the skeptic we must beg the question. Chisholm’s not arguing that question-begging is underrated, but that against the skeptic it is inevitable. So the mere fact that a position begs the question against the skeptic is not reason to reject it; we need to look at the position’s merits. Chisholm’s argument for particularism is at root that some instances of knowledge are more obvious than any philosophical theory could be: “[i]f you come across some philosophical theory that implies that you and I cannot know that this is a hand, then so much the worse for the theory” (1973 §9). So we should not doubt obvious cases of knowledge on philosophical grounds, nor start with a theory of knowledge and then see if it gives the result that I know I have a hand.
Chisholm rejects this move on the grounds that any criterion of knowledge must be internal: the mind must find adequate grounds for belief within itself (1973 §3). Particularism satisfies this because we can tell if an item of knowledge is internal. An excellent discussion of the reliabilist response to skepticism, and internalist criticisms of it, is in BonJour & Sosa 2003.
There is of course evidence that human reasoning is often unreliable – the locus classicus is Nisbett & Ross 1980. But that’s a different problem.
 Most reliabilist epistemologists think that reliability alone is not sufficient for knowledge – there are other conditions as well, and methods like those two may fail to satisfy the extra conditions. That can help us get a theory of knowledge that conforms to our intuitions about what is known, but it won’t get around the problem I’m laying out here. If we add too many requirements for knowledge, the reliabilist antiskeptical strategy fails. For it to work, it has to be sometimes rational to believe in the absence of internally available evidence; whenever that is possible, you get the problem discussed above.
 This is the root of the plurality objection against coherentism (the view that justification is a matter of how well the agent’s system of beliefs hangs together, justifying and explaining each of its parts). See Russell 1912 ch. 12 for a quick statement of the objection.
 Possibly the best statement of this idea is in Putnam 1990. Siegel calls it the “no-transcendence argument” (2011, 209-12).
Categories: Critical Replies
Leave a Reply