Positive definite lattices of rank at most 8. (Q1415368): Difference between revisions

The first construction of a positive definite even integral quadratic form of determinant 1 and rank 8 was given by \textit{A. Korkine} and \textit{G. Zolotareff} [Math. Ann. 6, 366--389 (1873; JFM 05.0109.01)]. It is now a well-known result in the theory of lattices that any positive definite even unimodular lattice of rank 8 is isometric to the root lattice of the \(E_8\) root system. This fact was first proved by \textit{L.-J. Mordell} [J. Math. Pures Appl. (9) 17, 41--46 (1938; Zbl 0018.29401)]. Various proofs of this uniqueness have appeared in the literature, including one by \textit{M. Kneser} [Arch. Math. 8, 241--250 (1957; Zbl 0078.03801)] that uses the method of neighbor lattices. The result can also be established by utilizing the Minkowski-Siegel mass formula or the theory of modular forms. The purpose of the present paper is to give a short, and reasonably elementary and self-contained, proof of the uniqueness of this positive definite even unimodular lattice of rank 8. The approach has similarities to that of Kneser [loc. cit.], but avoids the use of \(p\)-adic completions, and rests primarily on only an upper bound for the minimum nonzero norm of a lattice, such as the classical ones due to Hermite or Minkowski.