On \(\vartriangleleft^{*}\)-maximality. (Q1428039)

This article belongs to the group of papers about the classification of unstable theories without strict order property (SOP) by Shelah and others. The authors compare the model-theoretic notion \(\vartriangleleft^*\) and the syntactic \(\text{SOP}_i\), \(i= 1,2,3\), and they establish the implications: \[ \text{SOP}_3\Rightarrow\;\vartriangleleft^*\!\text{-maximality}\Rightarrow \text{SOP}_2\Rightarrow \text{SOP}_1\Rightarrow\text{not simple}. \] Not all implications are reversible, but which one is or is not is a ``burning question'', they say. Here are some indications of notions involved. Between two theories \(T_0\) and \(T_1\), \(T_0\vartriangleleft^* T_1\) holds if there is a third theory \(T\) into which \(T_0\) and \(T_1\) are interpretable, and in each \({\mathfrak M}\models T\) when the inner model of \(T_1\) in \({\mathfrak M}\) is \(\lambda\)-saturated, so is that of \(T_0\) for all large enough \(\lambda\). The notions \(\text{SOP}_i\) for \(i\geq 3\) were used previously. The new notions \(\text{SOP}_2\) and \(\text{SOP}_1\) are derived from a reformulation of the old \(\text{SOP}_3\). \(\text{SOP}_2\) declares the existence of a formula \(\varphi(x,y)\) and elements \(a_\eta\) (in the `monster' model) indexed by all the finite sequences of \(0\) and \(1\) such that although \(\{\varphi(x, a_\eta), \varphi(x, a_\nu)\}\) is inconsistent if \(\eta\) and \(\nu\) are incompatible, the set \(\{\varphi(x, a_{\rho| n})/n\in \omega\}\) is consistent for any \(\rho: \omega\to \{0,1\}\). \(\text{SOP}_1\) is seemingly weaker but whether it is, is an open question. On the way to obtaining these results, the authors build the theory \(T_{\text{feq}}\) which is not simple and is not SOP\(_3\) (hence irreversibility of the above implications somewhere), and is strictly below the theory of dense linear order with respect to \(\vartriangleleft^*\). They also give detailed studies of SOP's -- equivalent conditions, comparisons, etc. This paper is well organized and pleasantly written. Many definitions, constructions, and proofs are long and complicated. But the authors break them up into stages with good explanations of motivations and strategies. Remark: Mathematics is not literature. Neverthless, it is delightful to come across ``Before laying down the organisation of the paper\dots'', and the like.