Modal Conditions: Sensitivity and Safety

Modal responses to the Gettier problem hold that knowledge requires the right counterfactual relation between belief and truth — that the belief would track the fact across nearby possibilities — and the two main proposals, sensitivity and safety, spell out that tracking in opposite directions.

Definition

Sensitivity and safety are modal conditions on knowledge: sensitivity requires that the believer would not hold the belief in the nearest worlds where it is false, while safety requires that in nearby worlds where the believer holds the belief on the same basis, it is true.

Scope

This topic covers the analysis of knowledge in terms of modal conditions: Nozick's sensitivity, on which one would not believe p if p were false, and the safety condition, on which one would believe p only if p were true. It examines how each excludes Gettier cases and lottery cases, their treatment of skeptical scenarios, and the principal objections — sensitivity's denial of closure and safety's difficulties with necessary truths and methods. Reliabilist and virtue formulations are treated separately.

Core questions

• What counterfactual relation between belief and fact is required for knowledge?
• Does sensitivity or safety better capture the kind of luck that defeats knowledge?
• Why does sensitivity seem to require denying epistemic closure?
• How should the relevant belief-forming method be held fixed across possible worlds?

Key theories

Sensitivity (the tracking theory)

Nozick analyses knowledge so that, beyond true belief, one must satisfy the subjunctive condition that were p false one would not believe p, and were p true one would believe it — the belief 'tracks' the truth.

Safety

Safety reverses the conditional: a belief is safe when in all nearby worlds where one believes p on the same basis, p is true, which preserves closure under known entailment where sensitivity does not.

History

Nozick introduced sensitivity in 1981 as a tracking analysis that elegantly handles Gettier and skeptical cases but forces the surprising denial that knowledge is closed under known entailment. Reacting to that cost, Sosa and Williamson developed safety in the late 1990s and 2000s, which keeps closure and has become the more widely favoured modal condition, though debate over methods and necessary truths continues.

Debates

Sensitivity, safety, and epistemic closure

Sensitivity invalidates closure, implying one can know ordinary propositions yet not know they entail the denial of skeptical hypotheses; safety preserves closure but must say how the belief-forming basis is held fixed, and which condition is correct remains contested.

Key figures

• Robert Nozick
• Ernest Sosa
• Timothy Williamson

Seminal works

• nozick1981
• sosa1999

Frequently asked questions

What is the difference between sensitivity and safety?

Sensitivity is the condition that you would not believe p if p were false; safety is the condition that you would believe p only if p were true. They look similar but differ in which possible worlds they hold fixed, and they diverge sharply over whether knowledge is closed under entailment.

Why does sensitivity threaten closure?

On a sensitivity account you can know an ordinary proposition like 'I have hands' yet fail to know its known consequence 'I am not a handless brain in a vat', because your belief in the latter is not sensitive. This denial of closure is widely regarded as a serious cost of the view.