Some arguments for skepticism stress the phenomenon of disagreement. A strength of such arguments is that the make appeal to a very plausible premise—there is disagreement, about nearly any topic one can think of. A weakness of the arguments is that they appear to conflate two very different claims. The first is descriptive facts about what people do happen to believe (or not believe). The second is a normative question about what people should believe; about what they ought to believe. This paper will argument that routes to skepticism that focus solely or primarily upon facts about disagreement are not successful. They are not successful because they attempt to use a purely descriptive premise to support a normative conclusion. Let us consider a particular example in order to focus our ideas. It is a strange fact that some people in the contemporary world continue to believe that the Earth is actually quite young, say less than 10,000 years old. (Though it is important to notice that this fact is not necessary for the argument to come. For one way of framing the conclusion of the paper is that facts about what people actually believe are irrelevant to normative questions in epistemology or metaphysics.) So one might reason as follows. People differ about the age of the Earth. Who is to say which people are right, and which people are wrong? So there is really no truth about the age of the Earth. There are only opinions.
Notice that if the argument is correct, then it generalizes. For nothing in the preceding reasoning depends upon the particular issue singled out. It is possible for people to disagree about just about anything. So the argument is a good one, it shows that there is really no truth about anything (or almost anything).
There are different ways one could criticize this reasoning. One could ask whether the premises of the argument, or its conclusion, are ‘really true’. If they are not, then it is unclear what claim they have on our attention. But there is a more fundamental problem with the reasoning, which becomes obvious when we set it out explicitly. Thus the argument can be represented like this:
(1) For almost any issue one can name, it is possible to find people who disagree about it.
(2) If people disagree about the correct verdict on a given issue, then there is no truth—no fact of the matter—concerning that issue.
(3) There is no truth—no fact of the matter—concerning most issues.
When the argument is set out explicitly in this way it becomes clear what is the problem with it. While the first premise is true, and the argument is valid, the second premise is not plausible. Why should disagreement, which is a contingent, factual matter, have consequences for what we ought to believe?
Before continuing with the critique, we notice that there are several relevant connections between the topic being pursued here. In the reading from Clifford we saw strong reasons in support of the view that one should only believe things that one has good evidence for (Clifford 1879). Evidence is a kind of normative notion. The fact that two people disagree on some matter of fact does not, taken in itself, have any normative relevance. There is also a connection here to Gettier’s demonstration that the traditional analysis of knowledge in terms of belief, truth, and justification, provides only necessary conditions for knowledge but not, even taken together, a sufficient condition (Gettier 1963). It is also worth pointing out that the ‘tripartite’ analysis of knowledge goes back at least as far as Plato, who observed that knowledge is more than true belief—it has a tendency to ‘stay put’, whereas true belief can wander away from us (Klein 1977). Notice also that Descartes, on one plausible reading, has very strict requirements upon knowledge (and perhaps even upon justified belief)—that it be infallible (Descartes 2013). This is an even stronger normative condition upon knowledge than Plato or Gettier required.
Now it is certainly true that, in the right kind of circumstances, the fact that a certain person disagrees with you on a matter of fact can have normative, and therefore epistemological, relevance. This can happen if we know that the other person is known to be reliable, or has better access to the facts than you do. But notice that here normative notions—such as reliability, and having ‘better’ access—are explicitly involved. So my criticism of the argument for relativism is not that disagreement is completely irrelevant to truth, or to knowledge. It is merely that the descriptive fact of disagreement must be supplemented in some way if it is to have epistemological relevance. And in the argument sketch set out above, it is not supplemented in any way.
This last point does perhaps show that my initial criticism of the argument for relativism was too hasty. For in the background was our knowledge (in my view) that those who believe that the Earth is much younger than scientists maintain hold their beliefs on the basis of poor reasons, or no reasons at all—depending on whether we require that reasons for belief must indicate at least some explanation for why the belief is, or is likely to be, true. Therefore a more careful analysis of the flaws with the argument for relativism (or nihilism) set out above would be this: the mere descriptive fact that some people believe that the Earth is quite young, say less than 10,000 years old, is irrelevant to the truth concerning the age of the Earth, both because it is a mere descriptive fact about what some people believe and because we have antecedent reason to believe that such people are irrational.
While this qualification does not, in my judgment, entail that the argument for skepticism is any stronger than it was initially found to be, it does point to an important fact, which ought to be explicitly acknowledged. There is no general recipe for deciding upon the relevance of descriptive facts, such as facts about disagreement, to normative issues concerning knowledge and justified belief. Each case must be examined upon its merits.
As a final matter, I should like to connect this last conclusion to Gettier’s discussion of the insufficiency of true justified belief for knowledge. There are different ways of reacting to Gettier’s arguments. One could try to provide a fourth condition which is, together with the first three, sufficient for knowledge. One could hold that Gettier showed that there is something defective about our notions of knowledge, or justified belief, and so forth. In my view, by contrast, what Gettier showed is that knowledge cannot be understood in more fundamental terms. It is a primitive notion.
It does not follow from this, of course, that we cannot isolate necessary conditions for knowledge. We can. But it does mean, if correct, that the search for the elusive ‘fourth condition’ for knowledge is a fruitless one. Just as important as this, again in my view, is that this should not be a disappointing conclusion. The history of philosophy is filled with unsuccessful attempts to find sufficient conditions for normative notions in non-normative terms. Examples are justice, moral rectitude or goodness, justified belief, and so forth. I obviously cannot prove that no such reductive understanding of any normative notion is not going to be adequate. But it does seem that it may be a mistake to be worried about this fact. So long as we have a reasonably good understanding of a notion, and can apply it correctly in a variety of cases—and this is surely the situation with knowledge—there is no obvious reason to suppose that we do not understand it: merely because past and present attempts at reduction have failed. It may be the case that the world, at least as populated by human beings, is irreducibly normative.
- Clifford, William. The ethics of belief. (1879).
- Descartes, René, and John Cottingham. René Descartes: Meditations on First Philosophy: With Selections from the Objections and Replies. Cambridge University Press (2013)..
- Gettier, Edmund L. “Is justified true belief knowledge?.” Analysis 23.6 (1963): 121-123.
- Klein, Jacob. “Plato’s Trilogy: Theaetetus, the Sophist, and the Statesman.” (1977).
