How Skepticism Fails: The Dissolution of the Good/Bad Case Argument

The Skeptical Setup

The standard skeptical argument rests on a stark asymmetry:

The Good Case: Everything appears as it is. You perceive a tree. It is a tree. Your beliefs are true.

The Bad Case: A skeptical scenario (brain in vat, evil demon). Everything appears identical to the good case, but your beliefs are false.

The skeptic's conclusion: Since you cannot tell them apart, you cannot know you're not in the bad case. Therefore, you cannot know you're in the good case.

This argument has shaped epistemology for centuries. But Williamson shows it rests on a hidden assumption that is provably false.

The Hidden Assumption: Luminosity of Evidence

The skeptic's argument requires premise (1): Rational thinkers always know what their evidence is.

Why? Because the skeptic argues that if your evidence were different between good and bad cases, you could deduce that you're not in the bad case. But you can't make this deduction (it's consistent with what you know in the bad case that you could be in the good case). Therefore, your evidence must be the same in both cases.

This entire argument collapses if premise (1) is false—if rational thinkers can not always know what their evidence is.

And Williamson proves premise (1) is false using the exact same method that dissolves luminosity in Chapter 4.

The Sorites Argument Against Premise (1)

Imagine evidence gradually changing over time:

• At midnight: Freezing. Your evidence clearly involves cold temperature.
• At noon: Warm. Your evidence clearly involves warm temperature.
• Between: Temperature rises gradually, microsecond by microsecond.

If premise (1) is true (you always know your evidence), then:

• At each microsecond interval, you should know your evidence
• But each interval is indistinguishable from the previous one
• Therefore, your evidence must be identical to the previous moment
• By transitivity: evidence at midnight = evidence at noon
• Absurd

The argument proves that you don't always know what your evidence is. Just as luminosity fails at vague boundaries (feeling cold vs not feeling cold), knowledge of evidence fails at indiscriminable transitions.

The skeptic's entire reductio rests on an unprovable—and provably false—assumption.

The Phenomenal Conception Problem

If you accept the skeptic's argument that evidence must be the same in good and bad cases, you're committed to what Williamson calls the phenomenal conception of evidence: evidence is determined only by how things appear to you, independent of external facts.

But this is clearly wrong. The actual presence of the tree should matter to your evidence. Your evidence isn't just your phenomenal state—it includes facts about the world.

Williamson's response: You don't need to accept the phenomenal conception. Your evidence DOES differ between good and bad cases. But you don't always know that it differs, because knowledge of evidence isn't luminous.

The Rationality Isn't Transparent Either

If you retreat to the fallback position ("one always knows what one's evidence appears to be"), the sorites argument applies again.

The same indiscriminable transitions that prevent knowing your evidence prevent knowing how your evidence appears. Rationality itself is not transparent.

This doesn't make you irrational. It means rationality doesn't require perfect self-knowledge. You can be rational while not knowing exactly what your evidence is or how your rationality relates to it.

Skepticism Without Sameness of Evidence

Williamson goes further. Skepticism doesn't even require the good/bad case structure where false belief is possible.

Consider necessary truths (mathematical propositions). False belief in them is metaphysically impossible. Yet you can still doubt them by doubting the reliability of your methods.

Example: Someone uses a coin-flip to decide what to believe about a mathematical truth. If the coin comes up heads, they believe P; tails, they believe ~P. Even though P is necessarily true, this method is unreliable—it could easily produce false belief in a different proposition.

The skeptical structure generalizes: doubt about method reliability applies universally, not just where false belief is possible.

But the same principle applies: You don't always know whether your method is reliable or what your method even is. Therefore, the skeptical argument fails here too.

What This Proves About Consciousness

Consciousness operates through opacity, not despite it.

The skeptic's error was assuming that consciousness must be self-transparent to be rational. Williamson shows that rational consciousness is bounded:

• Evidence exists but remains partially opaque—you know some facts about your evidence, but not all
• Rationality operates through this opacity—you can be rational without perfect self-knowledge
• Authenticity emerges from this structure—not from achieving transparency but from honest inhabitation of your bounds

This is precisely what the book has been building toward since Chapter 1:

• Chapter 1: Knowing is a mental state (facticity is constitutive)
• Chapter 2-3: Knowing is broad (irreducibly coupled to world)
• Chapter 4-5: Knowing is opaque (margins prevent full access)
• Chapter 6: Margins structure rational action (even enable cooperation despite game-theoretic impossibility)
• Chapter 7: Sensitivity (perfect tracking) fails because knowledge is prime
• Chapter 8: Even rationality itself resists transparency; authenticity requires accepting this

The Liberation

This dissolution of skepticism is actually liberating. By accepting that you don't transparently know your own evidence or the reliability of your methods, you:

• Keep knowledge of the external world—nothing in skeptical arguments proves you lack it
• Maintain rationality without demanding omniscience—your opacity is structural, not a failure
• Inhabit consciousness authentically—by accepting bounds rather than seeking transcendence

The best answer to the skeptic isn't proving you know you're not in the bad case. It's showing that the skeptic's argument requires an impossible condition (luminosity of evidence) that no rational mind could satisfy.

Consciousness doesn't fail to meet skeptical standards. Skeptical standards are incoherent.

The Arc Complete: Williamson has now shown that consciousness is:

• Prime (irreducibly unified)
• Broad (necessarily world-dependent)
• Opaque (margins prevent self-access)
• Yet genuinely rational and knowing

In Part III, he will defend the equation E=K (evidence equals knowledge) and show how this unified framework resolves ongoing puzzles in epistemology. But the foundation is now set: consciousness works through its margins, not despite them. 🌑