Questions: Morley's Theorem on Uncountable Categoricity

Question 1

A countable theory T is found to be categorical in ℵ₂ (it has exactly one model of cardinality ℵ₂ up to isomorphism). What does Morley's theorem immediately allow you to conclude?

• A. T is also categorical in ℵ₀, because uncountable categoricity implies countable categoricity
• B. T is categorical in every uncountable cardinality — ℵ₁, ℵ₂, ℵ₃, and all larger cardinals
• C. T is categorical in ℵ₁ and ℵ₂ but may fail to be categorical in ℵ₃ and above
• D. T is ω-stable but may still have multiple non-isomorphic uncountable models in some cardinalities

Answer: (B)

Morley's theorem states that uncountable categoricity is all-or-nothing for countable theories: if T is categorical in *any* uncountable cardinality, then T is categorical in *every* uncountable cardinality. Finding categoricity in ℵ₂ is sufficient to conclude categoricity in all of ℵ₁, ℵ₃, ℵ₄, and every larger uncountable cardinal. Option A is wrong: Morley's theorem says nothing about ℵ₀-categoricity, which is an independent phenomenon. DLO is ℵ₀-categorical but not uncountably categorical; ACF_p is uncountably categorical but not ℵ₀-categorical.

Question 2

The theory of dense linear orders without endpoints (DLO) is ℵ₀-categorical. What does Morley's theorem predict about its behavior in uncountable cardinalities?

• A. DLO must also be categorical in all uncountable cardinalities, by Morley's theorem
• B. Morley's theorem says nothing about DLO in uncountable cardinalities — the theorem concerns what uncountable categoricity implies, not what ℵ₀-categoricity implies
• C. DLO cannot be categorical in any uncountable cardinality because ℵ₀-categorical theories are never uncountably categorical
• D. DLO is categorical in ℵ₁ but not in larger uncountable cardinalities

Answer: (B)

Morley's theorem has a specific direction: it says that uncountable categoricity at any cardinality implies uncountable categoricity at all cardinalities. It does not say anything about the relationship between ℵ₀-categoricity and uncountable categoricity — those are independent phenomena. In fact, DLO is ℵ₀-categorical (by back-and-forth) but has many non-isomorphic models of every uncountable cardinality. Option C might sound plausible but is wrong: the theorem does not establish any such rule. The correct statement is that the two categoricity regimes (countable and uncountable) are simply independent of each other for countable theories.

Question 3

A countable theory that is categorical in some uncountable cardinality is expected to be ω-categorical (categorical in ℵ₀).

• T. True
• F. False

Answer: False

This is false and illustrates the independence of the two categoricity regimes. Morley's theorem shows that uncountable categoricity forces ω-*stability* (a constraint on the number of types over countable parameter sets), but ω-stability is weaker than ω-categoricity. The theory of algebraically closed fields of characteristic p (ACF_p) is the canonical example: it is categorical in every uncountable cardinality (uncountably categorical) but has many non-isomorphic countable models (not ℵ₀-categorical). Uncountable categoricity and ℵ₀-categoricity constrain structure in different ways and do not imply each other.

Question 4

Morley's theorem implies that for a countable theory, the question 'in which uncountable cardinalities is it categorical?' has only two possible answers: either none or all.

• T. True
• F. False

Answer: True

This is the precise content of Morley's theorem. The set of uncountable cardinalities in which a countable theory is categorical is either empty (the theory has multiple non-isomorphic models of every uncountable cardinality) or all uncountable cardinalities simultaneously. There is no intermediate case where a countable theory is categorical in some uncountable cardinalities but not others. This is what makes the theorem so striking — it might have been expected that categoricity could hold at some uncountable cardinalities but fail at others, but the actual situation is maximally clean.

Question 5

What role does ω-stability play in the proof of Morley's theorem, and why does constraining the type space force categoricity in uncountable cardinalities?

Model answer: Morley showed that uncountable categoricity implies ω-stability: for every countable set of parameters A, the space of complete types over A is countable. This constrains how 'complex' the theory can be — there are not too many distinct types, so models cannot differ in too many ways over countable sets. The key ingredient is Morley rank, an ordinal-valued dimension assigned to definable sets in an ω-stable theory. Two models of the same uncountable cardinality in a Morley-categorical theory have the same Morley rank everywhere, and a back-and-forth construction using this rank produces an isomorphism between them. Without ω-stability, the type space could be uncountable, preventing the rank from being well-defined and breaking the back-and-forth argument.

The connection to classical algebra is illuminating: in ACF_p, Morley rank coincides with the Krull dimension of algebraic varieties. The abstract model-theoretic concept of 'how many types are there over this set?' turns out to measure the same structural rigidity as classical geometric dimension. This is why Morley's theorem launched modern stability theory: it showed that counting types was the right tool for classifying the complexity of first-order theories.