Formality criteria in terms of higher Whitehead brackets (Q2329363)

In rational homotopy theory, one often considers \textit{formal} spaces, i.e. those whose rational homotopy type is determined by their cohomology algebra (equivalently, the minimal model can be constructed only knowing the cohomology of the given space, i.e. it is a ``formal'' consequence of the cohomology). As a way of showing that a space is not formal, one can hope to detect the existence of (higher) Massey products. In the present paper, the authors consider the (in a sense) dual notion of coformality. A space (we will assume simply connected throughout, as the authors do, and finite type for the most part) is \textit{coformal} if its rational homotopy type is determined by the Lie algebra structure on \(\pi_*(\Omega X) \otimes \mathbb{Q}\) given by the Samelson bracket (or equivalently, up to a degree shift, by the Whitehead bracket on \(\pi_*(X) \otimes \mathbb{Q}\)). They give a proof of the folklore result that the higher Whitehead brackets, the analogues of the higher Massey products in the previous paragraph, are obstructions to coformality. One can consider stronger versions of the above two notions, namely \textit{intrinsic} formality and coformality. A space is \textit{intrinsically formal} if it is rationally homotopy equivalent to any other space with isomorphic rational cohomology algebra; it is \textit{intrinsically coformal} if its is rationally homotopy equivalent to any other space with isomorphic rational homotopy Lie algebra. Really, these should be thought of as properties of a (Lie) algebra; intrinsic (co)formality is then rigidity in the deformation theoretic setup of Schlessinger-Stasheff. The authors give conditions under which the product of odd dimensional spheres (other than circles) is intrinsically coformal; as a special case, the product of at most four such spheres is intrinsically coformal. Next, the authors discuss a spectral sequence due to Quillen, which collapses when applied to coformal spaces. As the authors point out, it was generally accepted that the collapse of the spectral sequence does not guarantee coformality, and they provide the first explicit example of this failure of the converse; their example in fact has all higher Whitehead products vanishing. The space considered is the total space of the \(K(\mathbb{Q}, 7)\) bundle over \(K(\mathbb{Q}, 2) \times K(\mathbb{Q}, 4)\) where the degree 7 generator of the fiber kills the square of the generator of \(K(\mathbb{Q}, 2)\) times the generator of \(K(\mathbb{Q}, 4)\). This presentation gives us a minimal model of the total space, which has cubic differential. Generally, a space is coformal if its differential is purely quadratic in some minimal model. The authors show that there is no automorphism of the model which would make the differential quadratic, hence showing the considered space is not coformal. The authors then discuss the relationship between the higher terms in the differential of a dga model of a space and higher Whitehead products, extending work of Andrews-Arkowitz. The interested reader is encouraged to also see the later paper [\textit{U. Buijs} et al., Mediterr. J. Math. 17, No. 1, Paper No. 31, 15 p. (2020; Zbl 1439.55020)] which complements the present one.