Le carre de l'egalite (Q4934044)

In their classical paper on superstable groups, Berline and Lascar proved that every superstable field has Lascar rank \(\omega^{\alpha} \cdot n\) and conjectured that \(n\) must be 1. Here the author gives a negative answer to this question, shows how to construct some (expansions of) algebraically closed fields having Morley rank \(2\) and \(\omega \cdot 2\) and suggests how to generalize this construction in order to produce examples of rank \(\omega^{\alpha} \cdot n\) for every \(\alpha\) and \(n \geq 2\). The construction draws inspiration from Hrushovski's ideas producing a strongly minimal non-trivial structure interpreting no infinite group, or a stable unsuperstable \(\aleph_0\)-categorical theory. Here the basic structures are pairs \((K,N)\) where \(K\) is a field and \(N\) is just a subset of \(K\). Provide each finite \(A \subseteq K\) with a dimension \(\delta (A)\): this is twice the transcendence degree of \(A\) minus the number of points in \(A \cap N\). Call \((K,N)\) a bicolour field when \(K\) is algebraically closed and \(\delta (A) \geq 0\) for every \(A\); let \(T_0\) denote the first-order theory of these structures. For every finite \(A\), put \(d(A) = \min \{ \delta (B) : A \subseteq B \subseteq K \}\). One can see that there is a minimal algebraically closed subfield \(k\) of \(K\) such that \(A \subseteq k\) and \(d(A) = \delta (k)\); \(k\) is called the self-sufficient closure of \(A\). This notion can be extended in a natural way to arbitrary subsets of \(K\). At this point, the author defines, for \(k\) a bicolour field and \(K\), \(L\) two self-sufficient extensions of \(k\), a free amalgamation of \(K\) and \(L\) over \(k\) and shows a key lemma saying that this is still a bicolour field. This amalgamation result yields a homogeneous universal bicolour field \(K_{\omega}\). \(K_{\omega}\) is the unique countable bicolour field (in its characteristic) satisfying the following condition: for every algebraically closed self-sufficient bicolour subfield \(k\) of \(K_{\omega}\) and for every self-sufficient extension \(K\) of \(k\) such that \(K\) (and \(k\)) have a finite transcendence degree, there is a self-sufficient embedding of \(K\) in \(K_{\omega}\) fixing \(k\) pointwise. In general, a bicolour field with this property is called rich. Two rich bicolour fields in the same characteristic are elementarily equivalent, and it is shown that they are just the \(\omega\)-saturated models of their first-order theory \(T_{\omega}\). But the main point is that rich bicolour fields have Morley rank \(\omega \cdot 2\). Moreover, rich bicoulour fields eliminate imaginaries, and a group definable in one of them is definably isomorphic to a group definable in the pure underlying field, hence to an algebraic group. At this point the previous construction is rearranged in order to build a field of Morley rank 2. Now one considers bicolour fields \((K,N)\) of characteristic \(\not = 2\) such that each point can be expressed in at most one way as a sum of two elements in \(N\). A suitable amalgamation lemma again yields a homogeneous universal model \(M\) in this setting. The saturation of \(M\) is discussed in another paper of Baldwin and Holland. Here it is shown that, when \(M\) is \(\omega\)-saturated, then the theory of \(M\) is \(\omega\)-stable of Morley rank 2. Moreover it eliminates imaginaries, as in the rank \(\omega \cdot 2\) case. The analysis of definable groups is left as an open question.