Motivation

• All theories discussed so far are internalist theories.

• Doxastic theories: justifiability of a belief depends exclusively on what else one believes.

• Internalist theories more generally: justifiability depends on one’s internal states, including beliefs, perceptual states, memory states, etc.

• Internal states are vaguely characterized as those we have “direct access” to; clarification to come in later chapters.

• Externalist theories loosen justifiability requirements further, allowing factors beyond internal states to influence justification (e.g., reliabilism bases justification on reliability of cognitive processes).

• Main motivation for externalism has two stages:

  1. Rejection of all doxastic theories, which fail to handle perceptual input (a nondoxastic but rationally evaluable process).

  2. Intuition that we want our beliefs to be probable, i.e., should not hold beliefs unless they are probably true. Probability and reliability are external considerations, not directly accessible to the believer.

• Externalism evaluates justification partly on these external probabilistic considerations.

• Secondary motivation: internalist theories propose complex, ad hoc, and incomplete arrays of epistemic rules without systematic justification or illuminating analysis.

• Internalist theories take epistemic justification as primitive and unanalyzed, building rules from intuitions about justification but lacking grounds to justify those intuitions.

• Some argue internalist criteria of justifiedness are analyses of justification, but even if correct, they do not illuminate what epistemic justification is about or why we should care.

• Internalist theories produce complicated, ad hoc criteria that give correct answers but no deep explanation or principled grounding for epistemic norms.

• Most doxastic theorists have little to say about analysis of epistemic justification due to lack of ideas; it is a difficult problem.

• Externalist theories have an advantage: allowing justification to depend on external factors makes it feasible to analyze justification in terms of probability.

• Lehrer (1974) criticizes foundationalism for postulating without justification, noting the underlying justification is that beliefs are constructed to have a reasonable probability of truth.

• However, Lehrer’s own theory diverges from this basic intuition and becomes as complex as internalist theories, merely replacing one complicated structure with another, thus still postulating epistemic rules ad hoc.

• The hope that epistemic justification can be simply analyzed in terms of probability has driven interest in externalism, but if the externalist’s analysis is also complicated, it is no better than internalism and remains ad hoc.

• The primary virtue of any such theory would be to correctly identify which beliefs are justified, which is no small feat.

• Summary of externalism’s appeal:

  • Captures intuition that justification is connected to probability.

  • Offers hope for an analysis that explains what justification is rather than merely providing criteria.

• To evaluate externalism’s promise, we must examine actual externalist theories.

• Two main types of externalist theories:

  • Probabilism: justifiability of a belief based on its probability and the probabilities of related beliefs.

  • Reliabilism: justifiability based on the reliability of cognitive processes producing the belief.

• Reliabilism is more familiar among epistemologists; probabilism will be discussed first, followed by reliabilism.

• Some groundwork must be laid before discussing these theories.

Varieties of Probability

• Philosophers often use the term probability too loosely, assuming a single unified concept when there are actually multiple kinds.

• A reasoned assessment of externalist theories requires distinguishing between different types of probability.

• One key distinction is between physical probability and epistemic probability:

  • Physical probability pertains to the physical world’s structure, independent of knowledge or opinion (e.g., laws of quantum mechanics).

  • Physical probabilities are discovered by observing relative frequencies and are central to statistics.

  • Physical probability is the stereotype most philosophers associate with probability.

• Epistemic probability relates to the degree of justification or belief about a proposition and is concerned with knowledge and opinion rather than physical facts.

  • Epistemic probability is relative to a person and a time.

  • It is an open question whether epistemic probabilities can be assigned numerical values and whether these values conform to the standard probability calculus used for physical probabilities.

• Besides physical and epistemic probabilities, there exist mixed physical/epistemic probabilities used in contexts like decision theory and weather forecasting.

  • These involve both physical facts about the world and knowledge of present circumstances.

• A second important distinction is between definite probabilities and indefinite probabilities:

  • Definite probabilities concern particular propositions or specific states of affairs (e.g., the probability that a coin lands heads).

  • Indefinite probabilities relate to concepts, classes, or properties rather than particular propositions (e.g., probability that a smoker contracts lung cancer, referring to the class of smokers).

• Some probability theories consider definite probabilities as basic; others start with indefinite probabilities.

• Epistemic probabilities are always definite probabilities since they reflect belief in a specific proposition.

• Physical probabilities may be either definite or indefinite:

  • Theories linking physical probability to relative frequencies treat them as indefinite probabilities since frequencies concern classes/properties, not individuals.

  • Propensity theories regard physical probabilities as applying to individual objects, making them definite probabilities (e.g., the probability a specific coin lands heads next toss).

• With these distinctions in mind, there are three broad categories of probability theories in the literature, any of which externalists could potentially use.

• Next, a brief sketch of each kind of probability theory will be presented.