Skepticism at the Service of Dogma: Descartes and the Problem of the Criterion*
In this essay, I will discuss the philosophical method used by René Descartes in his Meditations on First Philosophy. Today, it is known as “methodological skepticism”(1) — or Cartesian skepticism — a fitting description in that it was not actually Descartes’ aim to defend any form metaphysical or epistemological skepticism: skepticism about the external world, for instance, or about our ability to know anything at all, or the about the belief that we have bodies. On the contrary, Descartes was part of a long tradition of attempts to refute “the skeptical question.”(2) His skepticism was but a means to an end: to see if there are any propositions that can be known with absolute certainty and is for that reason sometimes called methodological, though I prefer the term “constructive.” Descartes exercised the (in his opinion, divine) gift of Reason, through the strategic use of doubt, for establishing indubitable premises. Is there anything, he asked, that cannot be doubted? As he answered this negatively, his more specific answers to what we would today recognize as the problem of the criterion, the problem of the external world, the problem of mind, the problem of God’s existence, etc., are best described as anti-skeptical. This essay will provide an overview of Descartes’ methodological skepticism, the anti-skeptical character of his metaphysics and epistemology, and attempt to raise one objection to this method. My focus will be on Descartes’ foundationalist approach as a potential solution to the problem of the criterion.
At the opening of the first meditation, Descartes observes that reason persuades him to withhold assent from any proposition that is not indubitably certain, in the same way that it persuades him to neglect any proposition that is demonstrably false.(3) It is crucial to take note of the very high epistemological standard for “knowledge” that Descartes has adopted for himself — for we have here no nuanced distinction between absolute certainty and degrees of certainty, and therefore no notion of probabilistic inference. In modern parlance, Descartes will only consider an article of knowledge any proposition p that has indefeasible evidence in its favor.(4) To qualify as indefeasible from the Cartesian skeptical perspective, a statement or set of statements that offer evidence in favor of some conclusion must meet two conditions: (A) the evidence must be indubitable and (B) the conclusion must follow deductively from the evidence statement(s).** Consider the following argument:
Premise 1: Either Karl*** is a four-legged hairy mammal or Karl is not a dog.
Premise 3: Karl is a four-legged mammal.
Conclusion: Karl is a dog.
The conclusion follows deductively by way of disjunctive syllogism and modus ponens and is therefore valid. On the other hand, are the premises (as statements of evidence that Karl is in fact a dog) indubitable? Hardly. Premise 1 is questionable, because it may well be the case, for instance, that Karl is a ferret or a cat. Therefore the condition of deductive validity is a necessary but insufficient condition for knowledge in the Cartesian sense. In order to establish the conclusion that Karl is a dog we must, according to Descartes, abolish all reasonable doubt, which means no more than that we should be able to rule out as false any statement incompatible with either the conclusion or the premises that attempt to prove it. There are thus two potential ways that doubt can defeat knowledge. I will refer here to Loewer’s (1980) formulation(5) of the conditions under which we cannot say that some epistemic agent A knows that p:
(A) If there is a proposition q such that A does not know that q is false and if A knows that if q is true then he does not know that p then he does not know that p.
(B) If there is a proposition q such that A does not know that q is false and q logically implies that A does not know p then he does not know that p.
In ordinary English, Loewer comprehends that Descartes has ruled out the possibility of A knowing p unless A also knows that p’s potential falsifier (q) is false(6) (one finds this, in the Stanford Encyclopedia of Philosophy online, referred to as the eliminate all doubt principle).
As an example, consider what proposition would be incompatible if it were true with our apparent knowledge that “2+2=4?” For Descartes, it is the proposition that God is a deceiver. Is it not conceivable, he asks, that God has deceived us(7) at a fundamental level — i.e., our ability to perceive clear and distinct ideas, which for Descartes is crucial to knowledge — such that we are deceived about the apparently self-evident truth of “2+2=4?” If it is conceivable then it is possible, and if it is possible, we must be able to prove that it is false to claim any justifiable knowledge about “2+2=4.” If we cannot know that “God is a deceiver” is false then we cannot claim to know that two and two add to four. Setting aside the issue of God’s sincerity (or lack thereof), the high-stakes Cartesian preconditions for knowledge are I think clearly illustrated.
Let us turn, then, to the question of what role skepticism has up to now played in Descartes’ method. First of all, Descartes is not a skeptic. Rather, his “skepticism” is merely an instrument for combating skepticism, which is why it is properly described as “constructive.” As Scruton astutely observes:
Scepticism begins by identifying some set of beliefs which are basic to our view of the world, and whose truth we do not question. It then identifies all the grounds for those beliefs: not the actual grounds that this or that person may have, but all the possible ground [e.g., potential falsifiers]. It proceeds to show that those grounds do not justify the beliefs. Mild skepticism argues that they do not prove the beliefs conclusively; radical skepticism argues that they offer no reason for believing at all.(8)
Descartes’ mission is to discover whether or not there are any absolutely certain propositions, of which the cogito is the prime affirmative example. Since, for Descartes, it cannot possibly be doubted that “I am a thing that thinks,”(9) we have in this proposition absolute and certain knowledge from which other absolutely certain propositions, such as the existence of God,(10) can be deduced. The profoundly anti-skeptical character of Descartes’ epistemology is thus evident. The only kind of knowledge in the Cartesian sense is certain knowledge, and it begins with the cogito.
What are we to make of Descartes’ assertion that the cogito is true and certain, and of his methodological skepticism in general? According to recent scholarship, Descartes was part of a new tradition of rationalism and “constructive skepticism” that attempted to use skeptical arguments to address the problem of the criterion and refute the Pyrrhonian skepticism of Sextus Empiricus.(11) The question at issue, then, is whether Descartes has satisfactorily resolved the problem of the criterion. I believe he has not. It is to the problem of the criterion we turn next.
The philosophical skepticism of Sextus Empiricus saw a revival in the Renaissance, and it thereafter became a major task of Renaissance intellectuals to address it. R.H. Popkin argues that, “through Montaigne, Renaissance scepticism became crucial in the formation of modern philosophy, contrary to the view that it was only a transitional moment in the history of thought.”(12) Popkin sheds light for us on the problem of the criterion by situating it in its historical context:
With the rediscovery in the fifteenth and sixteenth centuries of writings of the Greek Pyrrhonist Sextus Empiricus, the arguments and views of the Greek skeptics became part of the philosophical core of the religious struggles then taking place [i.e., the Protestant Reformation]. The problem of finding a criterion of truth, first raised in theological disputes, was then later raised with regard to natural knowledge, leading to la crise pyrrhonienne of the early sixteenth century.(13)
The problem has not by any means gone away. Roderick Chisholm called it the “one of the most important and one of the most difficult of all the problems of philosophy,” and added, “I am tempted to say that one has not begun to philosophize until one has faced this problem and has recognized how unappealing, in the end, each of the possible solutions is….”(14) It is the problem of the vicious circle (the “diallelus”) and it is, I argue, no less vicious in ensnaring the epistemological foundationalism of Descartes. It seems a challenge to find a better formulation of the problem of the criterion than Chisholm’s so it is to his I turn. He distinguishes two sets of questions related to the criterion. They are (A): “What do we know?” and (B): “How are we to decide whether we know? What are the criteria of knowledge?”(15) (p. 2) The skeptic, Chisholm says, will argue:
You cannot answer question A until you have answered question B. And you cannot answer question B until you have answered question A. Therefore you cannot answer either question. You cannot know what, if anything, you know, and there is no possible way for you to decide in any particular case.
We have already covered, to an extent, Descartes’ criteria of knowledge (though we added a modern spin to them in order to give Descartes the strongest possible case), beginning with indefeasible evidence and deductive validity. Naturally, the question must arise: what qualifies as indefeasible evidence? In the Meditations such evidence takes the form of the cogito. By way of the cogito, Descartes finds that the criterion of its truth and “knowability” (if the reader will forgive the term) is also revealed:
I am certain that I am a thing which thinks; but do I not then likewise know what is requisite to render me certain of a truth? Certainly in this first knowledge there is nothing that assures me of its truth, excepting the clear and distinct perception of that which I state, which would not indeed suffice to assure me that what I say is true, if it could ever happen that a thing which I conceived so clearly and distinctly could be false; and accordingly it seems to me that already I can establish as a general rule that all things which I perceive very clearly and very distinctly are true.(16)
The criterion of clarity and distinctness, if I may, completes our account of the Cartesian criteria for any A knowing any p. We now turn our gaze directly to the abyss that is the problem of the criterion, with which I conclude this paper and for which the method of Descartes offers no satisfactory response. Descartes’ criterion of clarity and distinctness clearly underpins the other two criteria: evidence is indefeasible if and only if it can be so perceived. Correlatively, a conclusion reached deductively, if based on indefeasible evidence, is indubitable.
Either it is true or it is false that, “anything I perceive very clearly and very distinctly is true.” If it is true, then Descartes’ Meditations provide a powerful guide as to how certain knowledge can be discovered. If it is false, then mere clarity and distinctness of an idea or thought must be insufficient to assure us of its truth. How are we to determine the truth-value of the criterion? We clearly cannot do so without reference to a new criterion, and so on ad infinitum, nor can we say that the criterion of clarity is true because it can be perceived clearly and distinctly for that way the vicious circle lurks, and the edifice of Cartesian knowledge cannot be founded on specious reasoning.
In short, to the first part of the problem of the criterion — “What do I know?” — Descartes answers: “I am a thinking thing.” To (B) — “How do I know it?” — he answers: “Because I clearly and distinctly perceive it.” But is it true that he clearly and distinctly perceives it? If so, then this fact can be counted with the cogito amongst the things that he knows — and yet, who would be willing to admit that Descartes knows he perceives clearly and distinctly because he perceives that he perceives clearly and distinctly? I am certain, at least, that I would not. That I am a thinking thing I likewise cannot claim to know in the Cartesian sense just because I so perceive it, especially when Descartes’ criterion by which to judge beliefs as knowledge is itself dubious.
(1)Scruton, Roger. Modern Philosophy: An Introduction and Survey (Penguin: New York, 1996)
(2)Ibid, p. 16
(3)Descartes, René. “Meditations on First Philosophy” in The Philosophical Works of Descartes, trans. by Elizabeth S. Haldane (Cambridge University Press, 1911), p. 6
(4)Appiah, Kwame. Thinking It Through (Oxford University Press, 2003), pp. 43–45
(5)Loewer, Barry. “Descartes’ Skeptical and Anti-Skeptical Arguments” in Philosophical Studies 39 (Reidel Publishing: Boston, 1981), pp. 165–166
(7)Descartes, Meditation III
(8)Scruton, p. 17
(9)Descartes, Meditation II
(10)Descartes, Meditation III
(11)Popkin, R.H. The History of Scepticism from Savonarola to Bayle (Oxford University Press, 2003), p. 18
(12)ibid, p. 43
(13)ibid, p. 3
(14)Chisholm, Roderick. “The Problem of the Criterion” in The Foundation of Knowing (Harvester Press: Sussex, 1982), pp. 61–75
(16)Descartes, p. 13
* The following is a paper I wrote while still an undergraduate some time around 2015 or 2016.
** Evidence given in support of inductive conclusions is technically always defeasible, regardless of the alleged strength of the probabilistic support lent to the conclusion, due to the problem of underdetermination. If a conclusion follows with absolute certainty from a set of premises, the inference must therefore be deductive.
*** I have a dog named Karl, after Sir Karl Popper.