Questions: Multi-Case Analysis and Knowledge Conditions

Question 1

An epistemologist proposes that sensitivity (if P were false, the agent wouldn't believe P) is necessary for knowledge. A critic constructs a case where sensitivity holds but knowledge clearly doesn't, then another where sensitivity fails but knowledge seems present. What do these two cases together establish?

• A. That sensitivity is neither necessary nor sufficient for knowledge
• B. That intuitions about knowledge are too unreliable to use as evidence
• C. That sensitivity is sufficient but not necessary for knowledge
• D. That the Gettier problem has been definitively resolved by the sensitivity condition

Answer: (A)

Two cases are needed to establish two distinct failures. A case where sensitivity holds but knowledge doesn't shows sensitivity is not sufficient (you can have sensitivity without knowledge). A case where sensitivity fails but knowledge seems present shows sensitivity is not necessary (you can have knowledge without sensitivity). Together, they establish that sensitivity neither suffices nor is required — it is neither necessary nor sufficient. This is controlled variation in action: each case changes exactly the variable being tested.

Question 2

What is the primary purpose of controlled variation in multi-case analysis?

• A. To generate as many cases as possible so that no critic can address them all
• B. To change exactly one feature between a pair of cases so that any change in intuitions can be attributed specifically to that feature
• C. To confirm that philosophical intuitions are consistent across all cultures and individuals
• D. To show that a single well-constructed case is sufficient to establish a philosophical conclusion

Answer: (B)

Controlled variation is a methodological discipline: hold everything constant except the one feature being tested. If intuitions change between the two cases, the changed feature is responsible. If they don't change, the feature doesn't affect knowledge attribution. This allows epistemologists to systematically test proposed conditions rather than generating cases haphazardly. It is the same logic as a controlled experiment: isolate the variable of interest by keeping everything else fixed.

Question 3

The evidential strength of multi-case analysis comes from the convergence of intuitions across many carefully varied cases, not from any single case being conclusive.

• T. True
• F. False

Answer: True

This is the methodological core of the approach. A single case can always be dismissed as an intuition pump, a misleading scenario, or a confusion about what the case stipulates. But when dozens of varied cases — manipulation cases, fake-barn cases, doxastic incontinence cases — all point in the same direction, that convergence constitutes evidence that is much harder to dismiss. The methodology treats intuitions as fallible data points; accumulation and convergence is what builds philosophical confidence.

Question 4

If a proposed knowledge condition fails to correctly categorize most test cases, the entire methodology of case analysis is undermined.

• T. True
• F. False

Answer: False

A condition failing for some cases is evidence about the condition, not evidence against the methodology. The methodology is designed to detect exactly this: if a condition doesn't work for all cases, that tells you the condition needs refinement. Cases are data; the methodology is the process of collecting and interpreting them. A failed condition prompts a revised condition, which gets tested against further cases. The methodology is self-correcting, not invalidated by any single failure.

Question 5

Why does a large collection of carefully varied cases provide stronger evidence about knowledge conditions than a single counterexample like the original Gettier case?

Model answer: A single case can be dismissed as a misleading intuition pump, an unusual exception, or a scenario that trades on ambiguity. Many carefully varied cases that all point in the same direction are much harder to dismiss — if different scenarios with different details all produce the same pattern of intuitions, that convergence suggests the pattern reflects something real about the structure of our knowledge concept rather than an artifact of one particular case's framing.

The Gettier cases were powerful enough to overturn a millennia-old analysis — but they were still debated. Subsequent case analysis built up an entire landscape: fake-barn cases, barn-façade cases, manipulation cases, each varying different features. When the same problems recur across all of them, the evidence is cumulative. This is why the post-Gettier literature generated hundreds of cases rather than just arguing from the original two — cases are the medium of evidence in analytic epistemology.