Williamson's argument against luminosity relies on which key feature of mental states?
AMental states are completely inaccessible to introspection in all cases
BMental states change gradually, so at any boundary point nearby situations include states where the belief is wrong
CMental states are too fast-changing for the brain to track accurately
DIntrospection requires language, but mental states are pre-linguistic
Williamson's margins-for-error argument targets the gradual nature of mental state change. Consider warmth: as you cool imperceptibly, at every point your current feeling is so similar to adjacent moments that knowing you feel warm would require being reliably right in nearby situations — but in nearby situations you no longer feel warm. Knowledge requires safety (being right not just now but in nearby cases), and gradual change guarantees nearby cases where your belief would be wrong. This argument applies to any mental state that can change gradually.
Question 2 Multiple Choice
You are experiencing a slowly decreasing level of warmth, imperceptible from moment to moment. At the threshold between 'warm' and 'not warm,' Williamson's argument says you cannot know you feel warm. The best explanation for this is:
AYou are temporarily unconscious at the threshold and cannot form beliefs
BThe concept of 'warm' is too vague to have a determinate truth condition
CKnowledge requires a margin for error — nearby situations are ones where you're no longer warm but still believe you are
DAt the threshold, the feeling itself does not exist and so cannot be known
Williamson's key move is the margin-for-error requirement: you know P only if, in situations relevantly similar to the current one, you would still believe P and it would still be true. At the threshold of warmth, the immediately adjacent moments are ones where you are no longer warm — but by hypothesis, you cannot detect the difference, so you still believe you are warm. This means your belief is unsafe: nearby it is false while you still hold it. Therefore it doesn't qualify as knowledge, even though in this very moment it happens to be true.
Question 3 True / False
Rejecting luminosity entails that introspection is mostly unreliable and provides no epistemic access to one's own mental states.
TTrue
FFalse
Answer: False
This is an overreaction that Williamson explicitly avoids. Rejecting luminosity means only that introspection is not *infallible* or *perfectly privileged* — it is subject to the same kinds of margins-for-error and reliability constraints as perception. Introspection can still be a generally reliable guide to one's mental states without being an authoritative, error-proof source. The result is that self-knowledge is more like perceptual knowledge than traditional epistemology assumed, not that it is useless.
Question 4 True / False
The KK principle states: if you know that P, then you know that you know that P.
TTrue
FFalse
Answer: True
This is the correct statement of the KK principle. It says epistemic states are transparent: knowing P automatically puts you in a position to know your own knowledge state. Williamson's argument against KK parallels his luminosity argument — knowing that you know requires satisfying the conditions for knowledge at a higher order, and the margin-for-error argument shows those conditions may fail even when first-order knowledge obtains. The KK principle fails for the same reason luminosity fails: safety conditions at the second order are not automatically satisfied by safety conditions at the first.
Question 5 Short Answer
Why does Williamson use a case of gradually changing mental states (like slowly cooling down) to argue against luminosity, rather than focusing on cases of dramatic mental state change?
Think about your answer, then reveal below.
Model answer: Gradual change is what generates the margin-for-error problem. If states changed abruptly, there would be a clear before-and-after with no ambiguous intermediate cases. But with imperceptible gradual change, at every moment your current state is extremely similar to the adjacent moments — some of which have crossed the threshold. Knowledge requires safety (being right in nearby situations), and gradual change ensures that nearby situations include ones where the state has changed but your belief hasn't. Cases of dramatic change wouldn't create this structural problem.
The force of the argument depends on the inability to detect fine-grained differences. The gradual-change scenario creates a series of adjacent cases across which the mental state shifts while the introspective belief stays constant — and this is exactly what knowledge requires you not to do. Williamson's genius is recognizing that this pattern is not a special edge case but a structural feature of any mental state that admits of degrees, which is most of them.