Instructor: Eddie Cushman

Phil200 Introduction to Philosophy: Section F03 Fall 2005

• Home
• Syllabus
• Handouts
• Paper Topics
• Exam Study Sheet

Epistemology Handout 4: Tue, 9/13

( Downlaod PDF )

Descartes, Epistemic Standards, and the Canonical Argument for External World Scepticism

1. Preliminaries

• Empirical (i.e., a posteriori) propositions: propositions that depend on experience for their justification. Notice that virtually all of our beliefs about the external world—that is, our beliefs that are not about our own states of mind—are empirical beliefs.

• What are sceptical scenarios? They are logically possible scenarios where

(i) All our empirical beliefs are false; or
(ii) Even if our empirical beliefs are true, they don’t amount to knowledge.

• Typical sceptical scenarios: (i) S is dreaming; (ii) S is being systematically deceived by an omnipotent demon or god; (iii) S is a brain in a vat being systematically deceived by scientists feeding him neural signals.

• The analysis of knowledge is a difficult philosophical topic, and we’ll address it to some degree in subsequent sessions. Its leading question is the following: Under what conditions do we have knowledge? What is required for us to have knowledge? Two minimal claims should strike you as plausible:

(i) True belief is necessary for knowledge: For a S to know that P, S must have a true belief that P.
(ii) True belief is not sufficient for knowledge: For S to know that P, S needs more than simply a true belief that P.

• Why (and where) does Descartes reject the following sceptical argument, and is he correct in doing so?

(1) My senses sometimes deceive me.
(2) If my senses sometimes deceive me, then they cannot provide me with knowledge of the external world.
(3) If I can have knowledge of the external world, then it is provided by my senses.
(4) I cannot have knowledge of the external world.

2. The canonical argument for philosophical scepticism¹

• The canonical argument concerning knowledge:

Let P be any empirical proposition.
(1) For S to know that P, S must know that he is not in a sceptical scenario.
(2) S cannot know that he is not in a sceptical scenario.
(3) S cannot know that P.

• The argument appears to be valid. Thus, if we are to resist it, we must find fault with one of its premises.

• Notice that this argument is perfectly general: it applies to any subject and any empirical belief.

• Notice, as well, that the conclusion of the argument is not identical to premise (2). The conclusion does not merely affirm that we cannot know that we are not victims of a sceptical scenario. It affirms that we cannot have any knowledge of the external world. We cannot even know, say, that we have hands. (3) has the following consequence: even if all of a subject’s empirical beliefs should happen to be true, they do not amount to knowledge.

• The canonical argument concerning epistemic justification:

Again, let P be any empirical proposition.
(1) For S to be justified in believing that P, S must be justified in believing that he is not in a sceptical scenario.
(2) S cannot be justified in believing that he is not in a sceptical scenario.
(3) S cannot be justified in believing that P.

• Intuitively, a subject is justified in believing that P when she believes that P because she has good reasons for thinking P is true. Thus, a variant on the canonical Cartesian argument has the consequence that our beliefs about the external world cannot be based on good reasons. In short, they are not reasonable.

• The “straightforward strategy” (Stroud) for blocking the canonical argument grants the truth of premise (1) and attempts to contest premise (2). On pages 19-23, Stroud argues that this strategy cannot succeed. What are Stroud’s reasons for thinking that the straightforward strategy is doomed to failure, and is Stroud correct in this claim? If the straightforward strategy fails, what options are left for blocking the canonical argument?

3. Epistemic standards and the canonical argument

• If the canonical argument is sound, then (via the definition of soundness) we have to accept that its radical, sceptical conclusion is true. The argument appears to be valid. Thus, in assessing it, we should turn to the truth of its premises.

• One condition of adequacy on the canonical argument is that it does not distort or elevate our ordinary standards for possessing knowledge. If it does, it has no more force against our prethoeretic conception of ourselves as knowers than the following crackpot argument has against our conviction that there are doctors in the U.S.:

(1) To be a doctor, one must have received an M.D. on 5 continents.
(2) No one in the U.S. has received an M.D. on 5 continents.
(3) There are no doctors in the U.S.

• Thus, in assessing the canonical argument, a central question is whether it distorts our ordinary standards for knowledge. A deeper question is methodological: How do we go about determining what our ordinary standards for knowledge amount to?

4.  Some salient epistemic standards

• The Simple Closure Principle:²

(SC) If S knows that P and if P entails Q, then S knows that Q.

Equivalently: For S to know that P, S must know -Q for all propositions Q that are incompatible with P.

Intuitive idea: For S to know that P, S must be able to rule out all alternatives to P.

• The Goldfinch Principle:³

(GP) If S knows that P and if S knows that P entails Q, then S knows that Q.

Equivalently: For S to know that P, S must know -Q for all propositions Q that S knows to be incompatible with P.

Intuitive idea: For S to know that P, S must be able to rule out all alternatives that he knows to be incompatible with P.

• The Witness Principle:

(WP) If S knows that P and if S’s knowing that P entails Q, then S knows that Q.

Equivalently: For S to know that P, S must know -Q for all propositions Q that are incompatible with S’s knowing that P.

Intuitive idea: For S to know that P, S must be able to rule out all alternatives that are incompatible with his knowing that P.

5. Subsidiary arguments for premise (1) of the canonical argument

• Subsidiary argument 1:

(1) (SC) If S knows that P and if P entails Q, then S must know that Q.
(2) P entails that S is not in a sceptical scenario.
(3) If S knows that P, then S must know that he is not in a sceptical scenario.

• Subsidiary argument 2:

(1') (GP) If S knows that P and if S knows that P entails Q, then S must know that Q.
(2') S knows that: P entails that S is not in a sceptical scenario.
(3) If S knows that P, then S must know that he is not in a sceptical scenario.

• Subsidiary argument 3:

(1") (WP) If S knows that P and if S’s knowing that P entails Q, then S knows that Q.
(2") S’s knowing that P entails that S is not in a sceptical scenario.
(3) If S knows that P, then S must know that he is not in a sceptical scenario.

Notice that (2") simply follows from the definition of a sceptical scenario.

• Questions about these subsidiary arguments:

1. Why does Stroud think Premise (1) of subsidiary argument 1 is false? Why does it have the consequence that number theory is either trivial or impossible?

2. Is the Goldfinch Principle true? Is the Witness Principle true? Or, due both of these principles involve some sort of distortion of our ordinary standards for knowledge? How does Stroud try to motivate these principles, and is he successful?

3. The Witness Principle has the following underlying rationale: for a subject to have knowledge that is based on a certain source (e.g., the testimony of witnesses, the testimony of his senses), the subject must have reason to think that the source is reliable. Further, a subject cannot have reason to think that a source is reliable unless he has reason to think that it is not systematically misleading. Thus, a subject cannot have knowledge based on his sensory experience unless he has reason to believe that he is not in a sceptical scenario. How compelling is this rationale?

6. A subsidiary argument for premise (2) of the canonical argument

• Let the Real World Hypothesis be the claim that our beliefs about the external world are largely true—that is, we are surrounded by objects, events, and people that for most part possess the properties we believe them to have.

(1) The perceptual experiences S has cannot provide him with better reason to believe the Real World Hypothesis than sceptical hypotheses.

(2) If the perceptual experiences S has cannot provide him with better reason to believe the Real World Hypothesis than sceptical hypotheses, then S cannot be justified in believing the Real World Hypothesis.

(3) S cannot be justified in believing the Real World Hypothesis. [From (1), (2)]

(4) For S to know that P, S must be justified in believing that P.

(5) S cannot know that the Real World Hypothesis is true. [From (3), (4)]

(6) If S cannot know that the Real World Hypothesis is true, S cannot know that he is not in a sceptical scenario.

(7) S cannot know that he is not in a sceptical scenario. [From (5), (6)]

• This argument does not figure explicitly in Descartes First Meditation, though it is strongly suggested by his discussion.

---

¹ This argument is often referred to as ‘The Argument From Ignorance’ or ‘The Transmission Argument.’

² The principle is known as a ‘closure’ or ‘transmission’ principle because it affirms that knowledge is closed under, or transmitted across, the entailment relation.

³ Also known as a ‘closure’ or ‘transmission’ principle because it affirms that knowledge is closed under, or transmitted across, the known entailment relation.