On the spectral characterization of manifolds (Q1946279)

The notion of a spectral triple \((\mathcal{A}, H, D)\) is invented by the author to encode the Riemannian Geometry of a generic non-commutative space using some operator algebra \(\mathcal{A}\), its representation on a Hilbert space \(H\) and its interactions with some densely defined selfadjoint operator \(D\) on \(H.\) The spectral characterisation of manifolds concerns the inverse problem to recover from a spectral triple \((\mathcal{A}, H, D)\), where \(\mathcal{A}\) is commutative, the Riemannian geometry of the manifold represented by the spectral triple. In other words, the problem is to find simple algebraic axioms on \((\mathcal{A}, H, D)\) so that there exists a unique smooth compact manifold \(X\) such that \(\mathcal{A}=C^{\infty}(X)\), the algebra of smooth functions on \(X.\) The problem was originally formulated in the author's seminal paper [Commun. Math. Phys. 182, No. 1, 155--176 (1996; Zbl 0881.58009)], in which the following five axioms were conjectured to be sufficient in recovering the smooth manifold. 1. (Dimension) The \(n\)-th characteristic value of the resolvent of \(D\) is \(O(n^{-\frac{1}{p}}).\) 2. (Universal commutation relation) \([[D, f], g]=0\) for all \(f, g\in\mathcal{A}.\) 3. (Smoothness) Let \(\delta\) be the derivation given by \(\delta(T)=[|D|, T]\) for any bounded operator \(T\) on \(H\). For any \(a\in\mathcal{A}\) and any integer \(m\), both \(a\) and \([D, a]\) belong to the domain of \(\delta^m.\) 4. (Orientability) Let \(\pi\) denote the representation of \(\mathcal{A}\) on the Hilbert space \(H\). There exists a Hochschild cycle \(c\in Z_n(\mathcal{A}, \mathcal{A})\) such that \(\pi(c)=\gamma\) when the dimension \(p\) is even and \(\pi(c)=1\) when \(p\) is odd. Here, \(\gamma\) is the grading operator in the \(\mathbb{Z}_2\)-graded \(H\) when \(p\) is even. 5. (Finiteness and absolute continuity) Viewed as an \(\mathcal{A}\)-module, the space \(\mathcal{H}_{\infty}=\cap_m\mathrm{Dom } D^m\) is finite and projective. Moreover, the following equality defines a Hermitian structure ( \(|\) ) on this module \[ \langle\xi, a\eta\rangle=\int a(\xi|\eta)|D|^{-p}. \] Here, \(\int\) denotes the noncommutative integral given by the Dixmier trace. The strategies of the proof have been sketched by the author [loc. cit.]. This paper gives a full proof of the problem of the spectral characterisation of manifolds with elegant use of (functional) analysis. The \textit{main result} of this paper is quoted as follows: Let \((\mathcal{A}, H, D)\) be a spectral triple with \(\mathcal{A}\) commutative, fulfilling the five above axioms in a slightly stronger sense where 1. The regularity holds for all \(\mathcal{A}\)-endomorphisms of \(\cap\mathrm{Dom } D^m\). 2. The Hochschild cycle \(c\) is antisymmetric. Then there is a compact oriented smooth manifold \(X\) such that \(\mathcal{A}=C^{\infty}(M).\) As a remark from the perspective of spectral geometry, the main theorem implies that a geometric space is determined by two sources of information: 1. The list of eigenvalues of \(D\). 2. The unitary isomorphism \(F\) between the Hilbert space of the canonical pair \((\mathcal{A}'', H)\) with \((H, D)\). Here, \(\mathcal{A}''\) is the von Neumann algebra weak closure of \(\mathcal{A}\) and \(F\) is analogous to the Fourier transform. These parameters for geometries have deep connections with the Yukawa coupling of the Standard Model.