12 Examples

Theorem 12.1. ACF⁡ has quantifier elimination.

Proof. We will show that condition Theorem 11.4(iii) holds.

So we need to show that for any finitely generated A, T∪D(A) is complete.

Fix a finitely generated L-structure A and show ACF⁡∪D(A) is complete. Use Los-Vaught test. Fix K1,K2⊨ACF⁡∪D(A) uncountable, wiht |K1|=|K2|. A is a finitely generated integral domain contained in K1,K2.

So since A contains 1, it determines the characteristic. So char⁡(K1)=char⁡(K2). So K1≅K2 (in Lrings).

Need an LA-isomorphism, i.e. an isomorphism Φ:K1→K2 preserving A. Consider Fi the fraction field of A in Ki. The field of fractions of an integral domain is unique up to isomorphism, i.e. ∃⁡τ:F1→F2 preserves A pointwise.

A is finitely generated (hence finite trdeg⁡), so trdeg⁡(K1∕F1)=trdeg⁡(K2∕F2). Therefore τ extends to τ∗:K1→K2 fixing A pointwise. So K1≅LAK2. □

Definition 12.2 (Constructible). Let F be a field. We say that X⊆Fn is constructible if it is a boolean combination of subsets of Fn defined by p(x1,…,xn)=0 for p(x¯)∈F[x¯].

Corollary 12.3 (Chevalley’s Theorem). Assuming that:

  • K⊨ACF⁡

  • X⊆Kn a constructible set

Then the projection
Y={(a1,…,an−1)∈Kn−1:(a¯,b)∈X for some b∈K}

of X is also constructible.

Definition 12.4 (Rado graph). Let L={E}, with E a binary relation. A Rado or Random graph is a graph (V,E) such that V≠∅ and for any finite disjoint X,Y⊆V there is some v∈V such that:

  • E(v,x) for all x∈X

  • ¬⁡(v,y) for all y∈Y

We denote the theory of Rado graphs as RC. It consists of

  • Graph axioms (irreflexive and symmetric).

  • For any k≥1 and l≥1,

    ∀⁡x1,…,xk,∀⁡y1,…,yl(∧i,jxi≠yj→∃⁡v(∧i,jE(v,xi)∧¬⁡E(v,yj))).

Facts:

Theorem 12.5. RG has quantifier elimination.

Proof. Option 1: Show RG∪D(A), with A a finite graph is complete.

Option 2: Use (ii) of Theorem 11.4. Fix M,N⊨RG, A⊆M∩N. Fix a quantifier-free formula φ(x1,…,xn,y), a¯∈An. Assume that there exists b∈M, M⊨φ(a¯,b). Want to show ∃⁡c∈N such that N⊨φ(a¯,c).

Write φ(x¯,y) in disjunction normal form:

∨s=1k∧t=1ls𝜃s,t(x¯,y)

where 𝜃s,t(x¯,y) is atomic or negated atomic. There is some s⊆k such that M⊨∧⁡t=1ls𝜃s,t(a¯,b).

Each of 𝜃s,t(x¯,y) is one of xi=xj, xi=y, E(xi,xj), E(xi,y) and negations.

If xi=y appears then b=ai∈A⊆N and so N⊨φ(a¯,b).

We may assume xi=y does not appear in 𝜃s,t. Let

X={ai:M⊨E(ai,b)}⊆a¯Y={ai:M⊨¬⁡E(ai,b)}⊆a¯

Then X and Y are finite and disjoint. So we have c∈N such that

N⊨E(ai,c)∀⁡ai∈XN⊨¬⁡E(ai,c)∀⁡ai∈Yc∉{ai,…,an}

So N⊨∧⁡t=1ls𝜃s,t(a¯,c) and thus N⊨φ(a¯,c). □