Questions: Possible Worlds Semantics for Knowledge

Question 1

In the possible worlds framework, what is the condition for an agent to count as knowing that p?

• A. p is true in the actual world and the agent believes it with high confidence
• B. p is true in at least one world within the agent's epistemic range
• C. p is true in every world within the agent's epistemic range
• D. p is necessarily true — true in all possible worlds without restriction

Answer: (C)

Knowledge requires p to be true in ALL accessible worlds — every world the agent cannot rule out given her evidence. Belief only requires p to be true in SOME accessible worlds. This difference in quantifier (all vs. some) is the structural gap between knowledge and belief in the possible worlds framework. Option A describes a common intuitive view that conflates confidence with the modal structure; option D confuses knowledge with logical necessity.

Question 2

Amara is looking at a real barn in good lighting. But she is in Fake Barn County, where most roadside barn-like structures are facades indistinguishable from real barns. Does Amara know there is a barn in front of her?

• A. Yes — her belief is true, she has perceptual justification, and she is looking at a genuine barn
• B. Yes — knowledge only requires truth in the actual world, and the actual world contains a real barn
• C. No — her epistemic range includes accessible worlds (compatible with her visual evidence) where she is facing a facade, so she has not ruled out all error-possibilities
• D. It depends on whether Amara is aware that she is in Fake Barn County

Answer: (C)

In the possible worlds model, Amara's visual evidence does not distinguish between facing a real barn and facing a perfect facade. So her epistemic range — the worlds compatible with her evidence — includes worlds where p is false. Since not all accessible worlds are p-worlds, she does not know p, even though her belief is true and her perception is functioning normally. This is the Gettier-style problem made geometrically precise: truth plus justification is insufficient if nearby accessible worlds contain error.

Question 3

In the possible worlds framework, the difference between believing p and knowing p is a matter of how confident the agent is in p.

• T. True
• F. False

Answer: False

The difference is structural, not a matter of degree. A believer has accessible worlds where p is false mixed in with worlds where p is true — she has not eliminated all error-possibilities. A knower's entire epistemic range consists of p-worlds — she has ruled out every accessible world where p fails. Confidence (a psychological intensity) is irrelevant to this modal structure. An agent could be supremely confident in a false belief or quietly certain of a known truth; the framework tracks the worlds, not the inner states.

Question 4

If the accessibility relation in an epistemic logic is reflexive, then knowledge is veridical — an agent cannot know a false proposition.

• T. True
• F. False

Answer: True

Reflexivity means every world is accessible to itself. If the agent knows p (p holds in all accessible worlds), and the actual world is accessible to itself, then p must hold in the actual world. This is axiom T: Kp → p. It captures the basic constraint that knowledge requires truth — you cannot know something false. Without reflexivity, the formal system would allow 'knowledge' of falsehoods, which most epistemologists reject as a constraint on any adequate analysis of knowledge.

Question 5

How does the structure of the accessibility relation in possible worlds semantics determine which epistemic axioms hold? Give one concrete example.

Model answer: Epistemic axioms correspond directly to geometric constraints on the accessibility relation. Each structural property of the relation validates a specific axiom. For example: reflexivity (every world accesses itself) validates axiom T (Kp → p — knowledge is veridical). Transitivity (if w₁ accesses w₂ and w₂ accesses w₃, then w₁ accesses w₃) validates axiom 4 (Kp → KKp — knowing implies knowing that you know). Adding symmetry yields axiom B and the S5 system. The power of the framework is that debates about epistemic principles become tractable questions about relational geometry.

This connection between syntax (axioms) and semantics (relational properties) is what makes the possible worlds framework so productive in formal epistemology. Instead of arguing verbally about whether 'knowing implies knowing that you know,' you ask: does the accessibility relation in your model have the transitivity property? If yes, axiom 4 is valid; if not, it fails. This makes epistemological commitments visible and testable in a way that purely verbal formulations often cannot achieve.