Noncommutative geometry and lower dimensional volumes in Riemannian geometry (Q2476619)

The author aims to represent the volume of any compact Riemannian manifold by means of ``lower dimensional'' ones. This nonsense program in the framework of classical differential geometry, acquires an unsuspected real meaning when one passes across Connes' noncommutative geometry. In fact, the author uses a Connes' idea who interprets the Einstein-Hilbert action by means of the Dirac operator Dir on a spin manifold \(M\), and the identification of the noncommutative residue \(\oint \text{Dir}^{-2}\) as the area in Planck's units of the manifold \(M\). (The author refers to the noncommutative residue of \textit{M. Wodzicki} [Invent. Math. 75, 143--178 (1984; Zbl 0538.58038), Lect. Notes Math. 1289, 320--399 (1987; Zbl 0649.58033)] and \textit{V. Guillemin} [Adv. Math. 55, 131--160 (1985; Zbl 0559.58025)].) Thus he pushes further Connes' idea defining \(\text{Dir}^{-n}\) as the \textit{noncommutative volume element} of \(M\), and \(ds\simeq \kappa | \text{Dir}| ^{-1}\) as the \textit{noncommutative length element} of \(M\), where \(\kappa\) is a constant. And he defines the \textit{\(k\)th dimensional volume} of \((M,g)\), \(\text{vol}_g^{(k)}=\oint ds^k\), \(k=1,\cdots,n\), where the integral is calculated as noncommutative residue in the sense of Wodzicki and Guillemin. Then he defines the lower dimensional volume \(vol_g^{(k)}\) in a purely differential geometric way, by means of some results by Atiyah-Bott-Patodi, Gilkey and others. In this way the lower dimensional volumes are directly calculated by means of integrals of local Riemannian invariants, forgetting the noncommutative road passing across the Dirac operator and spin structure. Thereby the author is able to give sense to questions like: ``what is the length of an \(n>1\)-dimensional Riemannian manifold ?'', or ``what is the area of an \(n>2\) Riemannian manifold?'' \dots Of course these questions, are nonsense in classical differential geometry!!! Reviewer's remark: This intriguing paper put really a rebus to the reader: \textit{How is it possible to give sense to a nonsense question, by simply passing to a non commutative geometric framework, when the final result does not depend from the noncommutative machinery used?} The answer to this rebus, must be recognized in the same type of noncommutative geometry considered. In fact, the noncommutative integration to which the author refers, is non-completely noncommutative, in the sense that this integration takes value in commutative numbers. (Instead a full noncommutative integration must take value in a noncommutative algebra. See the works on quantum manifolds by the reviewer of this paper.) Really the spectral characterization used by Connes and coworkers, of spin manifolds by means of the Dirac operator, uses the classic spectrum of this differential operator. Thus this noncommutative geometry can be considered a spectral representation of commutative geometry. Then the above rebus disappears, because the spectral representations of manifolds is likely to put these into a larger (noncommutative) framework, where it is posssible to recognize properties of these classic objects which (i.e., in the commutative framework) was impossible to see before. In some sense this paper is another beautiful example that confirms Gödel's theory on the \textit{``undecidable propositions''}, applied to Riemannian manifolds.