Model theory of difference fields with an additive character on the fixed field

Following a research line proposed by Hrushovski in his work on pseudofinite fields with an additive character, we investigate the theory $\mathrm{ACFA}^{+}$ which is the model companion of the theory of difference fields with an additive character on the fixed field added as a continuous logic predicate. $\mathrm{ACFA}^{+}$ is the common theory (in characteristic $0$) of the algebraic closure of finite fields with the Frobenius automorphism and the standard character on the fixed field and turns out to be a simple theory. We fully characterise 3-amalgamation and deduce that the connected component of the Kim-Pillay group (for any completion of $\mathrm{ACFA}^{+}$) is abelian as conjectured by Hrushovski. Finally, we describe a natural expansion of $\mathrm{ACFA}^{+}$ in which geometric elimination of continuous logic imaginaries holds.

💡 Research Summary

The paper introduces and studies the theory ACFA⁺, the model‑companion of the class of difference fields equipped with an additive character on the fixed field, treated as a continuous‑logic predicate. The language L⁺σ extends the usual ring language by a unary function symbol σ (the difference operator) and a continuous predicate Ψ taking values in the unit circle S¹∪{0}. In a model (K,σ,Ψ) the fixed field F = Fix(σ) carries a group homomorphism Ψ|_F : (F,+) → (S¹,·) which is required to be a standard additive character (the exponential of the trace), while Ψ vanishes on elements outside F. Moreover, for every absolutely irreducible curve C ⊂ Aⁿ defined over F and not contained in a rational hyperplane, the set Ψ(C(F)) is dense in S¹. These three axioms combine the axioms of ACFA (the model‑companion of difference fields) with those of PF⁺ (the theory of finite fields equipped with an additive character) and are expressible in continuous logic.

The first main result (Theorem A) establishes that ACFA⁺ is model‑complete, that any difference field with an additive character embeds into a model of ACFA⁺, and that the fixed field together with Ψ is stably embedded. The induced structure on F is slightly richer than the pure field structure because the restriction of σ to F names a generator of the absolute Galois group; this subtlety is handled by a careful adaptation of the quantifier‑elimination arguments for ACFA and PF⁺.

The second major achievement is a precise description of 3‑amalgamation in ACFA⁺ (Theorem C). Over a set A that is algebraically closed under σ (i.e., A = acl_σ(A)), 3‑amalgamation holds exactly when A is σ‑AS‑closed: for every a∈A there exists b∈A with σ(b)−b = a. If this condition fails, the authors construct a concrete counter‑example using the torsor Tₐ = {x | σ(x)−x = a}. When Tₐ∩A = ∅, one can freely prescribe the value of Ψ(α₁−α₂) for independent realizations α₁,α₂∈Tₐ, thereby producing an unsolvable 3‑amalgamation problem. From this analysis they deduce that ACFA⁺ is a simple theory (Corollary D).

With 3‑amalgamation under control, the paper turns to the Kim‑Pillay (KP) Galois group. In continuous logic the KP group Gal_KP(M) = Aut(M)/Aut_KP(M) carries a natural logic topology. The authors show that the connected component H_KP of this group acts transitively on the hyper‑imaginaries Tₐ/Eₐ, where Eₐ(x,y) ⇔ Ψ(x−y)=1, and that this action is essentially rotation on the unit circle. By analyzing the σ‑differences σ(a)−a, they prove (Theorem E) that

\