Epistemology Handout 5: Thu, 9/15

( Downlaod PDF )

Inference to the Best Explanation, Russell’s Response from IBE, and Gettier Cases

1. Metaphysical backdrop to the problem of the external world

• Solipsism: the only thing that exists is my mind.

• Idealism: the only things that exist are minds. Reality is exhaustively mental.

• Phenomenalism: Reality may not be exhaustively mental. It may be the case that some things exist, yet are not identical to minds or mental states. Still, my beliefs and utterances are about my own mental states—say, about patterns in my perceptual experience. They are true or false depending on whether they accurately represent my mental states.

• Realism: Reality is not exhaustively mental. Things like mountains and planets exist, and their existence and nature does not depend on the existence of minds. With the exception of beliefs and utterances about non-existent entities (e.g., Santa Claus, phlogiston) and beliefs and utterances about my mental states or the mental states of others, my beliefs and utterances are about mind-independent entities—entities whose existence and nature does not depend on the existence of minds. These beliefs and utterances are true and false depending on whether they accurately represent the characteristics of these mind-independent entities.

• Content Internalism: The contents of a subject’s beliefs and utterances are fully determined by the state of his neural system. Thus, complete neural replicas—say, you and a particle-by-particle replica of your brain that is housed in vat and fed the same signals as your actual brain—have thoughts with the same content. Complete neural replicas represent the world in the same way.

• Content Externalism: The contents of a subject’s beliefs and utterances are not fully determined by the state of his neural system. They are in part determined by the relations that a subject bears to his actual environment. Thus, complete neural replicas need not have identical thought contents and they need not represent the world in the same way. For example, you and your corresponding envatted brain may have beliefs that differ in content.

• The problem of the external world, as it is being formulated in this course, presupposes the truth of realism and content internalism. As a consequence, though you may have true beliefs about the world that surrounds you, your envatted neural twin is systematically deceived. Aside from a priori beliefs and beliefs about his own mind, your envatted twins’ beliefs are all false.

2. The canonical argument concerning reasons for belief

• Let P be any empirical proposition.

(1) For S to have good reasons for believing that P, S must have good reasons for believing that he is not in a sceptical scenario.
(2) S cannot have good reasons for believing that he is not in a sceptical scenario.
(3) S cannot have good reasons for believing that P.

• Thus, a variant on the canonical Cartesian argument has the consequence that we cannot have good reasons for any of our empirical beliefs. In short, all of our empirical beliefs are unreasonable.

3. Inference to the best explanation and scepticism

• Inference to the best explanation:1

(IBE) If a theory of hypothesis T provides a better explanation for a body of evidence than a competing theory or hypothesis T', then one has better reasons for believing T than one has for believing T'.

• Inference to the best explanation may be even more fundamental to scientific methodology than inductive inference.

• Proponents of the IBE strategy are not merely trying to rebut some particular sceptical argument. That is, they are not trying to show that a particular sceptical argument is logically invalid or has a false premise. Rather, they are trying to refute the sceptic. Indeed, this is one of the great attractions of the IBE strategy. Proponents of it, like Russell, think they have a perfectly general argument the conclusion of which is that external world scepticism is false or (more modestly) that we have good reason to think that scepticism is false. If we can refute the sceptic, there is no need to worry about rebutting his arguments one-by-one, as we will already know in advance that they are unsound.

• Conditions of adequacy on the IBE strategy: Champions of the IBE strategy offer an argument that attempts to refute the sceptic. What conditions must such an argument A satisfy in addition to the following pair?

(1) Validity: A’s premises entail its conclusion; and
(2) Soundness: In addition to being valid, A’s premise are all true.

The following constraint is obvious:

(3) The propositions that serve as premises of A must not be empirical.

Or, what is roughly equivalent:

(3') The propositions that serve as premises of A must be ones that a subject can have reason to believe either a priori or through introspection.

For, if the argument offered by a proponent of the IBE strategy contains empirical premises, then one has reason to believe the premises of the argument only if one has reasonable empirical beliefs. But, to assume that one has reasonable empirical beliefs is to beg the question against the sceptic. It is to assume precisely what is at issue, and what the sceptic doubts or denies.

4. A lame response to the external world sceptic (p. 21)

• The Real World Hypothesis: 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 of other people are similar to mine.
(2) The Real World Hypothesis provides a better explanation for this evidence than the hypothesis that I am the victim of a sceptical scenario.
(3) If a theory of hypothesis T provides a better explanation for a body of evidence than a competing theory or hypothesis T', then one has better reasons for believing T than one has for believing T'.
(4) I have better reasons for believing the Real World Hypothesis than I have for believing the hypothesis that I am the victim of a sceptical scenario. [From (1), (2), (3)]

• Russell claims that these considerations “beg the very question at issue” (p.21)? What does he mean by this? What, exactly, is wrong with this response to the sceptic.

5. Russell’s response from inference to the best explanation (pp. 22-3)

(1) My perceptual experiences are patterned and organized.

(2) The Real World Hypothesis provides a better explanation for this evidence than the hypothesis that I am the victim of a sceptical scenario.

(3) If a theory of hypothesis T provides a better explanation for a body of evidence than a competing theory or hypothesis T', then one has better reasons for believing T than one has for believing T'.

(4) I have better reasons for believing the Real World Hypothesis than I have for believing the hypothesis that I am the victim of a sceptical scenario. [From (1), (2), (3)]

• Questions:

1. Is Russell correct in affirming that the Real World Hypothesis provides a simpler explanation for the patterned and organized character of my perceptual experiences than some rival sceptical hypothesis? What makes one explanation simpler than another?

2. Does Russell’s response from IBE “beg the question” as well? That is, does it presuppose that the subject in question already has reasonable empirical beliefs? In particular, does Premise (3) meet condition of adequacy (3)? Or, does our justification for premise (3) depend on empirical evidence?

6. Russell’s response from epistemic conservatism

Epistemic conservatism: One has reason to abandon a belief only if one has evidence against it.

(1) I (“instinctively”) believe in the Real World Hypothesis.

(2) Though it is possible that I am the victim of a sceptical scenario, I have no evidence supporting this hypothesis.

(3) If I have no evidence supporting the hypothesis that I am the victim of a sceptical scenario, I have no evidence against the Real World Hypothesis.

(4) I have no evidence against the Real World Hypothesis. [From (2), (3)]

(5) I have reason to abandon a belief only if I have evidence against it.

(6) I have no reason to abandon my belief in the Real World Hypothesis [From (1), (4), (5)]

• None of the premises of this argument appear contentious aside from premise (5)—that is, the assumption of epistemic conservatism. Is the assumption of epistemic conservatism compelling?

7. The traditional analysis of knowledge

• The traditional analysis (i.e., the ‘TJB’ account): S knows that P iff:

(1) P is true; and
(2) S believes that P; and
(3) S is justified in believing that P.

• Questions:

1. Why is true belief not sufficient for knowledge?
2. Intuitively, what do we mean by ‘S is justified in believing that P’?

8. Gettier cases

• Gettier’s two principles:

(1) S can be justified in believing that P even if P is false.
(2) If (i) S is justified in believing that P; and (ii) P entails Q; and (iii) S deduces Q from P; and (iv) S believes that Q as a result of this deduction, then S is justified in believing that Q.

• Four Gettier cases:

1. Professor Spliff.
2. Jones’ job. (Gettier)
3. Jones’ Ford. (Gettier)
4. Holographs.

• Gettier’s diagnosis: in both of these cases, the connection between Smith’s belief and the fact that makes it true is accidental.

9. The no false lemmas response to Gettier

• The no false lemmas principle:

(NFL) For a subject to be justified in believing that P on the basis of his justified belief that Q, Q must be true.

• Questions:

1. Which of Gettier’s principles is being attacked?
2. Does (NFL) succeed in ruling out all of the above Gettier cases?
3. Is (NFL) plausible?

1 ‘Inference to the best explanation’ describes an inference pattern that was first systematically analyzed by Charles Sanders Peirce in the late 19th century. Peirce’s term for this inference form was ‘abduction’. H introduced the term to distinguish this type of inference from inductive inference. Contemporary philosophers use ‘inference to the best explanation’ and ‘abduction’ interchangeably. The relations between IBE and induction are very complex. Some philosophers argue that all inductive inferences reduce to inferences to the best explanation, whereas the converse of this claim is false. Thus, on this line of thought, IBE is a more fundamental form of reasoning than induction. Other philosophers deny that induction reduces to IBE, and thus see it as an autonomous form of reasoning.