Endomorphisms of \(H^*(K(V,n);\mathbb F_2)\) in the category of unstable modules (Q2498488)

At least 10 years ago Stewart Priddy posed the question of whether the Eilenberg-MacLane space \(K({\mathbb F}_2,2)\) is stably indecomposable or not. An algebraic topologist would naturally attempt to prove indecomposability algebraically, for example by considering the cohomology of this space. If the cohomology is indecomposable as a module over the Steenrod algebra then the space must be stably indecomposable. Given the close relationship between Eilenberg-MacLane spaces and the Steenrod algebra it is not surprising to find that this cohomology ring has a very intricate and beautiful structure as a module over the Steenrod algebra which, unfortunately, makes it hard to establish its indecomposability. Of course a module is indecomposable if, and only if, the endomorphism ring is trivial, and so one can more generally enquire about the endomorphism ring of \(K(V,n)\) for any elementary abelian group \(V\). One interesting case of this is known: the \textit{J. F. Adams, J. H. Gunawardena} and \textit{H. Miller} theorem [Topology 24, 435-460 (1985; Zbl 0611.55010)] asserts that the endomorphisms of \(K(V,1)\), as a module over the Steenrod algebra, are formal linear combinations of group endomorphisms of \(V\). This paper provides a huge generalization of the Adams-Gunawardena-Miller theorem by calculating the vector space of homomorphisms from \(H^*(K(V,n))\) to \(H^*(K(W,n))\) as modules over the Steenrod algebra. This space is identified with \({\mathbb F}_2[\hom(W,V)]\), just as in the Adams-Gunawardena-Miller theorem. In fact more general algebraic results are proved, calculating the module homomorphisms from \(H^*(K(V,n))\) to \(H^*(K(W,s))\) and thus giving a partial generalization of Lannes's linearization principle, cf. \textit{J. Lannes} and \textit{S. Zarati} [Math. Z. 194, 25--59 (1994; Zbl 0627.55014)]. The results are given for the prime \(2\); the odd prime case will be treated elsewhere.