Questions: Epistemic Accessibility Relations

Question 1

In the possible-worlds framework, why does reflexivity of the accessibility relation R (every world accesses itself, wRw) correspond to the axiom 'Kp → p' (knowledge implies truth)?

• A. Because reflexive relations generate complete and consistent modal theories by construction
• B. Because if the actual world always accesses itself, anything true in all accessible worlds must be true in the actual world — so what is known must actually be true
• C. Because reflexive relations produce Euclidean accessibility structures, which independently enforce that knowledge requires truth
• D. Because reflexivity makes the relation symmetric, which links knowledge to mutual belief

Answer: (B)

The semantics of 'Kp' says: p is true in ALL worlds accessible from the actual world w. If wRw (the actual world accesses itself), then w is among the worlds where p must hold. So Kp in w implies p is true in w — knowledge entails truth. Without reflexivity, an agent could 'know' p while p is false in the actual world, which is the defining difference between knowledge and mere belief. This is the T axiom: Kp → p.

Question 2

An agent has justified belief in proposition p, but p is actually false. In the possible-worlds framework, this scenario is correctly modeled by:

• A. A reflexive accessibility relation where the actual world does not access itself
• B. A non-reflexive accessibility relation — the actual world is not included in the agent's epistemic range
• C. A symmetric but non-transitive accessibility relation, allowing belief without introspection
• D. An accessibility relation satisfying the T axiom but not the 4 axiom

Answer: (B)

Justified false belief is modeled by non-reflexivity: the agent's accessible worlds (those they cannot rule out) do not include the actual world w. The believed worlds all make p true — that is why the agent believes p — but the actual world is not among them, so p is false there. This is precisely why you cannot KNOW a falsehood but can BELIEVE one: knowledge requires reflexivity (the actual world must be accessible), belief does not. Option A is contradictory: reflexivity is the property wRw, so it cannot simultaneously hold and fail.

Question 3

If an agent's epistemic accessibility relation satisfies transitivity, the agent achieves omniscience — they know everything that is true in most logically possible worlds.

• T. True
• F. False

Answer: False

Transitivity gives the 4 axiom: Kp → KKp (positive introspection — if you know p, you know that you know p). It does NOT give omniscience. The agent still only knows what holds in all their accessible worlds; transitivity just says that what is accessible from accessible worlds is also directly accessible. The agent remains ignorant of facts outside their epistemic range. Omniscience would require a universal accessibility relation (every world accesses every other), which is a far stronger property.

Question 4

The philosophical debate about whether knowledge requires 'negative introspection' (¬Kp → K¬Kp — if you don't know p, you know that you don't know it) can be precisely restated as a question about whether the epistemic accessibility relation is Euclidean.

• T. True
• F. False

Answer: True

This is exactly the formal translation. The Euclidean property says: if wRw' and wRw'', then w'Rw''. This corresponds to the 5 axiom (¬Kp → K¬Kp). The philosophical intuition is: if there are two worlds w' and w'' both accessible from w (both epistemically possible for the agent), then from w's perspective, w'' must also be accessible from w' — the agent 'knows what they don't know.' Whether this property should be required for knowledge is a substantive philosophical question that the framework translates into a precise structural one.

Question 5

Explain why reflexivity of the accessibility relation is equivalent to the T axiom (knowledge implies truth). What would a non-reflexive accessibility relation for knowledge allow, and why is that philosophically unacceptable?

Model answer: Reflexivity says the actual world w always accesses itself (wRw). The truth conditions for Kp in w are: p is true in ALL worlds accessible from w. Since w is one of those worlds (by reflexivity), Kp at w requires p to be true at w — hence Kp → p. Without reflexivity, the actual world might not be in the agent's epistemic range. Then Kp could hold at w (p is true in all worlds the agent considers possible) even if p is false at w — the agent 'knows' something false. This violates the basic intuition that knowledge, unlike belief, is factive: you cannot know what is not so. A non-reflexive accessibility relation for knowledge would conflate knowledge with (possibly false) justified belief.

The T axiom is often described as distinguishing knowledge from mere belief: it is the one axiom that captures the factivity of knowledge. Formally, removing reflexivity moves from the S5 or S4 system (knowledge) to weaker systems like K or KD (often used for belief), where the accessibility relation need not include the actual world.