Defining transcendentals in function fields (Q4779644)

There are two notions of ``relative algebraic closure'' of a field \(K\) in a field extension \(F/K\): the field-theoretic relative algebraic closure \[ \overline{K} \cap F = \{x\in F \mid f(x)=0\text{ for some } f\in K[T] \} \] and the model-theoretic relative algebraic closure \[ \text{acl}_{F}(K):=\{ x\in F \mid x\in A \text{ for some finite }K\text{-definable }A\subseteq F \} \] (where ``\(K\)-definable'' means definable by a first-order formula in the language of fields with parameters from \(K\)). Clearly, \(\overline{K}\cap F\subseteq \text{acl}_F(K)\); the present paper provides examples of field extensions \(F/K\) where this is a proper inclusion (i.e., where the two notions of relative algebraic closure of \(K\) in \(F\) do not coincide, or, equivalently, where \(F\) contains \(K\)-definable transcendentals over \(K\)).