The Kervaire invariant problem (Q5890860)

The purpose of this paper by Mike Hopkins is to discuss the existence and history of framed manifolds of Kervaire invariant \(1\). In particular, this is discussed in the context of the now famous and beautiful result by Hill, Hopkins, and Ravenel: Theorem. If \(M\) is a stably framed smooth, closed manifold of Kervaire invariant one, then the dimension of \(M\) is \(2,6,14,30,62\), or \(126\). Hopkins begins with the work of Pontryagin. We learn that Pontryagin, building off of the work of Brouwer and Hopf, studied maps \(f: S^{n+k}\to S^n\) by considering the inverse image of a regular value \(x\in S^n\). The pre-image (or inverse image) of such a point gives rise to a closed, smooth manifold of dimension \(k\). Given any two such points \(x\) and \(y\) which in turn give rise to smooth manifolds \(M_x\) and \(M_y\), we are interested in a third manifold \(N\) which ``connects'' \(M_x\) and \(M_y\) as boundary with \(N\) ``in between.'' This \(N\) is said to be a framed cobordism between \(M_x\) and \(M_y\). Interestingly, it can then be shown that the homotopy group \(\pi_{n+k}(S^n)\) can be described in terms of these framed cobordism classes of manifolds. Hopkins explains that ``this work established a very deep relationship between homotopy theory and geometry; one that was destined to shape the subject over the next 50 years.'' Hopkins goes on to discuss how Pontryagin used this new insight in order to compute some unknown homotopy groups of the general linear groups. The further technique of framed surgery is also discussed and illustrated. While discussing the computation, we are introduced to the Arf invariant of a certain map \(\phi\). This invariant turns out to be essential when attempting to study cobordism. The story continues with the work of Pontryagin being built upon and generalized by Kervaire and Milnor. Together, they classified smooth structures on spheres up to \(h\)-cobordism. The Kervaire invariant plays an important role in this classification theorem, and Kervaire and Milnor conjectured that the invariant of a framed manifold was zero in all dimensions other than \(2,6\), and \(14\). However, Browder showed this to be false in his 1969 paper which also showed the deep connection between differential topology and homotopy theory via the Kervaire invariant. This led to the now famous EHP sequence which has greatly aided in the computation of homotopy groups of spheres. After this brief history, Hopkins then moves on to a sketch of the proof of the statement above. It relies on several of the results he had previously stated. The paper ends with open problems and future directions.