nonabelian de Rham cohomology > history

differential cohomology in an (∞,1)-topos -- survey
**structures in an (∞,1)-topos**
* **shape**
* **cohomology**
* cocycle/characteristic class
* twisted cohomology
* principal ∞-bundle
* ∞-vector bundle
* **homotopy**
* covering ∞-bundles
* Postnikov system
* path ∞-groupoid
* geometric realization
* Galois theory
* internal homotopy ∞-groupoid?
* Whitehead system
* **rational homotopy**
* ∞-Lie algebroid
* ordinary rational homotopy
* internal rational homotopy
* Chern-character
* **differential cohomology**
* flat differential cohomology
* de Rham cohomology
* de Rham theorem
* **relative theory over a base**
* relative homotopy theory
* Lie theory
## Examples
(...)
## Applications
* Background fields in twisted differential nonabelian cohomology
* Differential twisted String and Fivebrane structures
* D'Auria-Fre formulation of supergravity

In a locally contractible (∞,1)-topos $\mathbf{H}$ a cocycle in *(nonabelian) de Rham cohomology* is a cocycle $\Pi(X) \to A$ in flat differential cohomology whose underlying cocycle $X \hookrightarrow \Pi(X) \to A$ in (nonabelian) cohomology is trivial: it encodes a trivial principal ∞-bundle with possibly nontrivial but *flat* connection.

If $\mathbf{H}$ is a smooth (∞,1)-topos, then nonabelian deRham cocycles are represented by flat? ∞-Lie algebroid valued differential forms $\omega$:

$\array{
\Pi^{inf}(X) &\stackrel{\omega}{\to}& \mathfrak{a}
\\
\downarrow && \downarrow
\\
\Pi(X) &\to& A
}
\,.$

If $A = \mathbf{B}^n R/Z$ then $\omega$ is an ordinary closed n-form.

A differential 1-form $A \in \Omega^1(X)$ on a smooth manifold $X$ may be thought of as a connection on the trivial $U(1)$- or $\mathbb{R}$-principal bundle on $X$.

Similarly a differential 2-form $B \in \Omega^2(X)$ on a manifold $X$ may be thought of as a connection on the trivial $U(1)$-bundle gerbe on $X$; or on the trivial $\mathbf{B}(1)$-principal 2-bundle.

This pattern continues: a differential $n$-form is the same as a connection on a trivial $\mathbf{B}^n U(1)$-principal ∞-bundle.

Moreover this pattern generalizes to $G$-principal bundles for nonabelian groups $G$:

for $\mathfrak{g}$ the Lie algebra of a Lie group $G$ – possibly nonabelian – a Lie-algebra valued 1-form $A \in \Omega^1(X,g)$ may be thought of as a connection on the trivial $G$-principal bundle on $X$.

While it may seem that the notion of differential form is more fundamental than that of a connection, in the context of differential nonabelian cohomology in an arbitrary path-structured (∞,1)-topos? the most fundamental notion of a differential cocycle is that of a *flat connection on a principal ∞-bundle* : on an space $X$ this is simply given by a morphism $\Pi(X) \to A$ from the path ∞-groupoid to the given coefficient object $A$.

The underlying principal ∞-bundle is that characterized by the cocycle that is given by the composite morphism $X \to \Pi(X) \to A$.

We may therefore characterize flat connections on trivial $A$-principal ∞-bundles as those morphisms $\Pi(X) \to A$ for which the composite $X \to \Pi(X) \to A$ trivializes. This way we characterize $A$-valued deRham cohomology in the (∞,1)-topos $\mathbf{H}$.

Fix a model for the (∞,1)-topos $\mathbf{H}$ in terms of the local model structure on simplicial presheaves $SPSh(C)^{loc}$ as described at path ∞-groupoid.

For $A \in SPSh(C)$ a pointed object with point $pt_A : {*} \to A$ define $A_{dR} \in SPSh(C)$ by

$A_{dR} : U \mapsto [I,SPSh(C)]
\left(
\array{
U
\\
\downarrow
\\
\Pi(U)
}
\,,\;
\array{
{*}
\\
\downarrow^{\mathrlap{pt_A}}
\\
A
}
\right)
\,.$

This we call the **de Rham differential refinement** of $A$.

The cohomology with coefficients in $A_{dR}$

$H_{dR}(X,A) := \pi_0 \mathbf{H}(X,A_{dR})$

we call **$A$-valued de Rham cohomology**

**(de Rham cohomology in terms of differential forms)**

The definition does not actually presuppose that the ambient (∞,1)-topos is a smooth (∞,1)-topos in which a concrete notion of ∞-Lie algebroid valued differential forms exists. It defines a notion of “de Rham cohomology” even in the absence of an ordinary notion of differential forms.

But if $\mathbf{H}$ does happen to be a smooth (∞,1)-topos then both notions are compatible.

Created on February 22, 2010 17:38:38
by Urs Schreiber
(131.211.234.184)