Singular cohomology from supersymmetric field theories (Q822717)

There is a beautiful interpretation, due in large part to Kontsevich, of the de Rham complex of a manifold in terms on supergeometry. Let \(M\) be a finite-dimensional smooth manifold. Write \(\mathbb{R}^{0|1}\) for the supermanifold called variously the ``superpoint'' or the ``odd line''; its superalgebra of smooth functions is \(\mathbb{R} \oplus \mathbb{R}x\), where \(x\) is odd and \(x^2=0\). Then there is a finite-dimensional (!) smooth supermanifold \(\pi\mathrm{T}M := \underline{\mathrm{sMan}}(\mathbb{R}^{0|1}, M)\) of smooth functions \(\mathbb{R}^{0|1} \to M\), which naturally carries an action of the supermonoid \(\underline{\mathrm{sMan}}(\mathbb{R}^{0|1}, \mathbb{R}^{0|1})\) of endomorphisms of \(\mathbb{R}^{0|1}\). The superalgebra of functions on \(\pi\mathrm{T}M\) turns out to be isomorphic to the total de Rham complex \(\Omega_{\mathrm{dR}}(M)\) with its wedge product. The grading and differential on \(\Omega_{\mathrm{dR}}(M)\) turn out to exactly encode the action of \(\underline{\mathrm{sMan}}(\mathbb{R}^{0|1}, \mathbb{R}^{0|1})\). This beautiful interpretation of the de Rham complex has a further reformulation, due to Stolz and Teichner, in terms of ``supersymmetric quantum field theory in zero spacetime dimensions''. Let's say that a field theory is ``over \(M\)'' if it couples to a (background, a.k.a. nondynamical) scalar field (a.k.a. function) valued in \(M\). Such a field theory will have a ``partition function'' which is a function on the space of spacetimes equipped with such a field. In general a quantum field theory contains more data than just its partition function: it contains the information of how the partition function can be computed ``locally''. But in zero spacetime dimensions, the partition function is the complete data of the quantum field theory. For supersymmetric field theories over \(M\), the partition function is a function on \(\pi\mathrm{T}M\), i.e. a de Rham form. Requiring that the field theory be topological restricts to closed de Rham forms. The de Rham cohomology of \(M\) can then be recovered as the set of equivalence classes of field theories over \(M\) modulo ``concordance'', i.e. interpolations by field theories over \(M \times \mathbb{R}\). This interpretation of de Rham cohomology is a test case for the proposal by Segal, Stolz, and Teichner to construct a geometric cocycle model for Topological Modular Forms in terms of 2-spacetime-dimensional supersymmetric field theory. The present paper describes a wide-reaching generalization of this construction. The authors introduce a new notion of ``superalgebraic cartesian sets'' over any commutative base ring \(R\). Among these are, on the one hand all simplicial sets and on the other hand an algebraic version \(\mathbb{A}^{0|1}\) of \(\mathbb{R}^{0|1}\). It contains all simplicial sets. Moreover, the category of superalgebraic cartesian sets is a presheaf topos, and so enjoys very good categorical properties. In particular, for any simplicial set \(X\) there is an internal mapping space \(\pi\mathrm{T}X := \underline{\mathrm{sCart}}(\mathbb{A}^{0|1},X)\). The first result of the paper is that, when \(R \supset \mathbb{Q}\), the functions on \(\pi\mathrm{T}X\) are precisely Sullivan's rational differential forms, and an version of the story above about topological supersymmetric quantum field theories recovers the singular cohomology of \(X\) with coefficients in \(R\). For base rings \(S\) not containing \(\mathbb{Q}\), a version of the construction working with certain divided power algebras recovers \(\mathrm{H}^\bullet(X, S)\) linearly, and recovers a portion of the multiplicative structure.