Questions: Homogeneous and Universal Models

Question 1

In the theory of dense linear orders without endpoints (DLO), consider two finite ordered subsets of ℚ: {1, 3, 7} and {2, 5, 9}. There is an order-preserving bijection between them. Does this partial elementary map extend to an automorphism of ℚ?

• A. No — automorphisms of ℚ must fix the rationals between the mapped points, which this bijection does not guarantee
• B. Yes — homogeneity of ℚ as a model of DLO guarantees that any partial elementary map between finite subsets extends to an automorphism of the whole structure
• C. Only if the two subsets have the same sum, since automorphisms of ℚ must preserve arithmetic structure
• D. No — DLO automorphisms can only extend partial maps that include the rational 0

Answer: (B)

ℚ with the usual ordering is the unique countable homogeneous universal model of DLO. Homogeneity means exactly that any isomorphism between finite substructures — here, any order-preserving bijection between finite ordered subsets — extends to an automorphism of the whole model. The back-and-forth method constructs this automorphism: DLO guarantees that at each step, there is always a rational number in the right position to extend the map. Note that automorphisms of ℚ as a linear order need not preserve arithmetic operations — only the order relation.

Question 2

What is the key conceptual difference between a homogeneous model and a saturated model, even though the two properties often coincide?

• A. Homogeneous models are countable; saturated models are uncountable
• B. Homogeneity is about symmetry — any partial isomorphism between substructures extends to an automorphism; saturation is about realization — every type over a small parameter set is realized by some element
• C. Homogeneous models realize only complete types; saturated models realize partial types as well
• D. Saturation requires the model to be elementarily equivalent to all its elementary substructures; homogeneity does not

Answer: (B)

These are genuinely distinct concepts. Homogeneity is a symmetry condition: the automorphism group acts transitively on finite substructures of the same isomorphism type — the model 'looks the same' from every finite vantage point. Saturation is a completeness condition on type realization: every consistent type over a parameter set smaller than |M| is realized in M. Saturated models are homogeneous (and universal), but one can have homogeneous models that are not fully saturated, and the definitions target different aspects of the model's richness.

Question 3

In a homogeneous model, if two elements realize the same complete type over a finite parameter set, there is an automorphism of the model sending one element to the other.

• T. True
• F. False

Answer: True

True — this is one of the most important consequences of homogeneity, and it gives types a geometric meaning. In a homogeneous model, realizing the same type is equivalent to lying in the same automorphism orbit. Since types encode all first-order properties an element has relative to a parameter set, two elements with identical types are structurally indistinguishable — and homogeneity ensures this indistinguishability translates into an actual symmetry of the model. This is why stability theory can treat 'same type' as a meaningful equivalence relation with algebraic and geometric content.

Question 4

Universality and homogeneity are the same property stated in different terms: a model that is universal for its theory should also be homogeneous, and vice versa.

• T. True
• F. False

Answer: False

False — they are distinct properties. A universal model must contain an isomorphic copy of every model of the theory of appropriate cardinality. A homogeneous model must have every partial isomorphism between finite subsets extend to a full automorphism. These conditions are logically independent: a model can be universal without being homogeneous (it contains all models but has few automorphisms), or homogeneous without being universal (it has rich symmetry but doesn't embed all models). Saturated models achieve both simultaneously, which is why they are the canonical choice for stability-theoretic analysis.

Question 5

Why do model theorists work with homogeneous universal 'monster models' rather than reasoning directly about the class of all models of a theory?

Model answer: A monster model provides a single ambient structure in which all models of the theory appear as elementary substructures, and whose rich automorphism group makes algebraic arguments available. Instead of quantifying over many different models, you reason about types and definable sets within the monster, where the automorphism group acts transitively on tuples realizing the same type. This transforms questions about what is true in 'some model' or 'all models' into questions about orbits and definable sets in one canonical structure.

The practical power is significant. Many stability-theoretic arguments that would require quantification over all models of a theory — asking whether something is consistent, whether a type is definable, whether an independence relation holds — become local questions inside the monster model. The homogeneity guarantee ensures that orbits under automorphisms correspond to types, making the automorphism group a concrete tool rather than an abstract symmetry group. The monster model is not a real object that 'exists' in any ordinary sense; it is a convenient fiction whose existence follows from compactness and whose usefulness comes from centralizing all the models of a theory in one place.