Ranking Theory¶
Represent plain belief and its revision through ordinal degrees of disbelief assigned to possible worlds.
Core Idea¶
Ranking theory assigns ordinal degrees of disbelief to possible worlds, with at least one zero-ranked world. A proposition receives the minimum rank among worlds where it holds. In the usual interpretation, a proposition is believed when its negation has positive rank. The retained grades also support conditional reasoning and changes under evidence.
Scope of Application¶
Spohn's ordinal conditional functions model plain but revisable belief in formal epistemology. A state can retain information about which contrary possibilities are more deeply disbelieved, even when two states have the same set of accepted propositions. The theory is used in qualitative belief change and conditional reasoning, not every task called ranking.
Clarity¶
World rank, proposition rank, and conditional rank are different operations. A zero-ranked world is merely not disbelieved, not necessarily believed. For a finite-ranked condition A, a conditional rank is calculated from κ(A and B)-κ(A); infinity-ranked evidence needs an explicit convention or different revision rule.
Manages Complexity¶
A categorical belief set loses firmness distinctions that may matter after later evidence. The ordinal ranking keeps those distinctions in a structured state, at the cost of specifying a possibility space, rank function, and update policy.
Abstract Reasoning¶
Let three worlds have ranks 0,1,3, with A true in the first two. Initially A is believed with counter-rank 3. For the finite-ranked condition E comprising the second and third worlds, subtract its minimum rank 1: Spohn's E-restricted conditional ranking gives those worlds ranks 0,2, so κ(not A|E)=2. This is a conditional result within E, not a fully specified revised state: a full update would also need the rank of the first world outside E and a firmness parameter. Retaining ranks costs extra model and update specification while preserving information for iterated revision.
Knowledge Transfer¶
Recognize the framework by normalized ordinal disbelief over possibilities, minimum proposition ranks, belief through positive counter-rank, and ranking-level evidence change. Bayesian probabilities and a bare belief-revision operator are related but do not share this whole formal identity.
Neighborhood in Abstraction Space¶
Ranking Theory sits in a sparse region of the domain-specific corpus (65th percentile for distinctiveness): few abstractions share its structure, so a faithful description tends to retrieve it precisely.
Family — Codes, Matrices & Combinatorial Problems (30 abstractions)
Nearest neighbors
- Cognitive Hierarchy Theory — 0.85
- Elementary-embedding large-cardinal schema — 0.84
- Noisy Channel Model — 0.84
- Dominant Strategy — 0.84
- Impossible World — 0.84
Computed from structural-signature embeddings · 2026-10-08