Questions: Saturated Models and Maximal Realization

Question 1

A model M is ω-saturated for the theory of dense linear orders without endpoints (DLO). You describe a point p as 'strictly between 1/3 and 1/2, and also between 0.4 and 0.5.' What does ω-saturation guarantee about M?

• A. M must be extended to a larger model to realize the type of p — ω-saturation only ensures types over infinite parameter sets are realized
• B. Some element of M already realizes the type of p, because ω-saturation requires every finitely-parameterized consistent type to be realized
• C. The type of p is inconsistent because the constraints conflict with one another
• D. M contains p only if the cardinality of M is large enough to accommodate the new element

Answer: (B)

ω-saturation means that every type over a *finite* parameter set that is consistent with the theory is realized in M. The description of p uses finitely many parameters (1/3, 1/2, 0.4, 0.5) and is consistent with DLO — so an ω-saturated model must contain an element realizing it. There is no need to extend M. This is the 'no missing witnesses' property: any consistent description of a hypothetical element using small enough parameters is already witnessed by an actual element in the model. The rational numbers themselves are ω-saturated for DLO precisely because no finite description of a 'missing' rational can be given that isn't already satisfied by some rational.

Question 2

What is the key structural consequence of κ-saturation that makes saturated models 'highly homogeneous'?

• A. All elements of a saturated model are definable, making the model arithmetically rigid
• B. Any two realizations of the same type over a parameter set of size < κ can be mapped to each other by an automorphism of the model
• C. Saturated models are always isomorphic to each other, regardless of cardinality
• D. Every element in a saturated model has the same type over the empty set, making all elements interchangeable

Answer: (B)

Saturation implies homogeneity: because every consistent type over a small parameter set is realized, any partial map that preserves types can be extended to an automorphism of the whole model. Two elements that satisfy the same formulas over any parameter set of size < κ are interchangeable — there is an automorphism sending one to the other. This homogeneity is why saturated models are ideal for proofs: you can pick 'representative' elements of each type, knowing they are structurally indistinguishable. Option C is close but incorrect: saturated models of different cardinalities need not be isomorphic; they are isomorphic only if they have the same cardinality and the theory is complete.

Question 3

A κ-saturated model is required to realize all types over parameter sets of size strictly less than κ, but may fail to realize types over parameter sets of exactly size κ.

• T. True
• F. False

Answer: True

This is the precise definition of κ-saturation. The threshold κ is strict: for every A ⊆ M with |A| < κ, every consistent type over A is realized in M. But types over parameter sets of size exactly κ need not be realized. This is why larger κ is harder to achieve but richer in structure: an ω-saturated model realizes all finitely-parameterized types but may omit types over countably infinite parameter sets, while a (2^ω)-saturated model handles all countably-parameterized types. Saturation is always relative to the threshold.

Question 4

Nearly every complete theory with an infinite model has a saturated model of nearly every infinite cardinality.

• T. True
• F. False

Answer: False

Every complete theory has saturated models, but not necessarily of every infinite cardinality. The existence of a κ-saturated model of cardinality κ requires κ to be large enough relative to the theory's complexity — specifically, κ must be at least 2^(|T|). Under GCH (generalized continuum hypothesis) or with sufficient set-theoretic assumptions, saturated models of every uncountable cardinality exist, but unconditionally, saturation is only guaranteed for 'sufficiently large' cardinalities. The statement that saturated models exist at *every* infinite cardinality is not a theorem of ZFC.

Question 5

Why is saturation described as the 'opposite' of the omitting types theorem, and what does it mean for a saturated model to contain 'no missing witnesses'?

Model answer: The omitting types theorem says you can build a model that deliberately leaves a type unrealized — a model with a 'gap' where certain elements could consistently exist but don't. Saturation demands the opposite: no consistent type over a small enough parameter set is omitted. Every type that could be realized — every consistent description of a hypothetical element — is already witnessed by an actual element in the model. 'No missing witnesses' means: if you can write down a consistent list of formulas that a hypothetical element would satisfy, using only a small enough set of parameters from the model, then the model already contains an element satisfying all those formulas. There is no 'ghost element' whose existence is consistent but unrealized.

This completeness of realization is what gives saturated models their homogeneity and makes them canonical representatives of their theories. In proofs, having 'no missing witnesses' means you can always find an actual element to play any role a consistent description assigns — a powerful tool for constructing elementary maps, extending partial automorphisms, and proving structural results about the theory.