Chevalley-Eilenberg algebra in synthetic differential geometry

synthetic differential geometry


from point-set topology to differentiable manifolds

geometry of physics: coordinate systems, smooth spaces, manifolds, smooth homotopy types, supergeometry



smooth space


The magic algebraic facts




tangent cohesion

differential cohesion

graded differential cohesion

id id fermionic bosonic bosonic Rh rheonomic reduced infinitesimal infinitesimal & étale cohesive ʃ discrete discrete continuous * \array{ && id &\dashv& id \\ && \vee && \vee \\ &\stackrel{fermionic}{}& \rightrightarrows &\dashv& \rightsquigarrow & \stackrel{bosonic}{} \\ && \bot && \bot \\ &\stackrel{bosonic}{} & \rightsquigarrow &\dashv& \mathrm{R}\!\!\mathrm{h} & \stackrel{rheonomic}{} \\ && \vee && \vee \\ &\stackrel{reduced}{} & \Re &\dashv& \Im & \stackrel{infinitesimal}{} \\ && \bot && \bot \\ &\stackrel{infinitesimal}{}& \Im &\dashv& \& & \stackrel{\text{étale}}{} \\ && \vee && \vee \\ &\stackrel{cohesive}{}& ʃ &\dashv& \flat & \stackrel{discrete}{} \\ && \bot && \bot \\ &\stackrel{discrete}{}& \flat &\dashv& \sharp & \stackrel{continuous}{} \\ && \vee && \vee \\ && \emptyset &\dashv& \ast }


Lie theory, ∞-Lie theory

differential equations, variational calculus

Chern-Weil theory, ∞-Chern-Weil theory

Cartan geometry (super, higher)


Urs Schreiber. This is something I thought about in the context of my discussion at Chevalley-Eilenberg algebra.


We give a description of the Chevalley?Eilenberg algebra? of the Lie algebra of a Lie group as the ∞-quantity of functions C (BG e (1))C^\infty(\mathbf{B}G_e^{(1)}) on the simplicial space of infinitesimal neighbourhoods of the identity in the sense of synthetic differential geometry in the simplicial smooth space BG\mathbf{B}G that is the Lie ∞-groupoid incarnation of the delooping of the Lie group.

The derivation is analogous to and usefully compared with how the deRham algebra of differential forms on a manifold XX is the ∞-quantity of functions on the infinitesimal singular simplicial complex X Δ diff Π(X)X^{\Delta^\bullet_{diff}} \hookrightarrow \Pi(X) of XX, as described at differential forms in synthetic differential geometry.

We proceed entirely by using theorems and propositions from the book

in particular section 6.8 combined with section 4.3. We effectively show that these statements are precisely the ones needed to unwrap what the normalized Moore cochain complex of the cosimplicial algebra C (BG e (1))C^\infty(\mathbf{B} G_e^{(1)}) in the monoidal Dold?Kan correspondence is like.

Definitions and setup

Let GG be a Lie group (by which we mean a finite dimensional Lie group). Write BG\mathbf{B}G for the simplicial smooth space which in degree kk is the cartesian product G kG^k with the standard face and degeneray map (see the examples at nerve for details).

Let TT be some topos that models the axioms of synthetic differential geometry and which has a full and faithful embedding Diff T\hookrightarrow T.

Consider then BG\mathbf{B}G as a simplicial object in TT. As usual, we shall call objects in TT spaces in the following.

Let G e (1)GG_e^{(1)} \hookrightarrow G be the space that is the first infinitesimal neighbourhood of the neutral element ee in GG. By definition this space is ismorphic to the infinitesimal space

D(n)={(d 1,,d n)R n|i,j:d id j=0} D(n) = \{(d_1, \cdots, d_n) \in R^n | \forall i,j : d_i d_j = 0\}

for nn the dimension of DD. By the log-exp bijection in synthetic differential geometry? this space is canonically identified with the vector space g:=Lie(G)g := Lie(G) underlying the Lie algebra of GG.

Moreover, by the Kock-Lawvere axiom morphisms f:D(n)G e (1)Rf : D(n) \simeq G_e^{(1)} \to R are necessarily linear f:df 0f 1df : d \mapsto f_0 \to f_1 \cdot d, hence under the loglogexpexp bijection are nothing but elements in the dual vector space g *g^*.

Recall that the ordinary Chevalley?Eilenberg algebra? of gg is the differential graded algebra whose underlying graded-commutative algebra is the Grassmann algebra g *=g *g *g *\wedge^\bullet g^* = \mathbb{R} \oplus g^* \oplus g^* \wedge g^* \oplus \cdots.

So the subset of C (G e (1))C^\infty(G_e^{(1)}) that vanishes at 0 is naturally isomorphic to the degree-11 part of the Chevalley–Eilenberg algebra.

Notice now that the multiplication on the group GG does not restrict to a multiplication on G e (1)G_e^{(1)} because the sum d 1+d 2d_1 + d_2 of two elements that each square to 0 is does in general not square to 0, – (d 1+d 2) 2=2d 1d 2(d_1 + d_2)^2 = 2 d_1 d_2 but only its cube (d 1+d 2) 3=0(d_1 + d_2)^3 = 0 does. Therefore the group multiplication induces a composition

:G e (1)×G 2 (1)G e (2). \cdot : G_e^{(1)}\times G_2^{(1)} \to G_e^{(2)} \,.

Consider therefore the space of “infinitesimal 11-cells of BG\mathbf{B}G whose composite is again an infinitesimal 11-cell”, i.e. the pullback

(G×G) e (1) G e (1)×G e (1) G e (1) G e (2). \array{ (G \times G)_e^{(1)} &\to& G_e^{(1)} \times G_e^{(1)} \\ \downarrow && \downarrow^{\cdot} \\ G_e^{(1)} &\hookrightarrow& G_e^{(2)} } \,.

By item 3) of theorem (6.8.1) this pullback is precisely the space of elements (x,y)G e (1)×G e (1)(x,y) \in G_e^{(1)} \times G_e^{(1)} such that not only xx and yy are infinitesimal neighbours of the neutral element ee, but also of each other.

By the Kock-Lawvere axiom (entirely analogous to the similar step in the derivation of simplicial differential forms in synthetic differential geometry) it should follow from this that maps (G×G) e (1)R(G \times G)_e^{(1)} \to R that vanish on degenerate elements are in bijection with antisymmetric maps that are canonically identified with elements in g * g *g^* \wedge_{\mathbb{R}} g^* (need to say this in more detail…)

Continuing in this manner (…details for higher degrees to be filled in…) we define the simplicial space

BG e (1):=((G×G) e 1G e (1)*). \mathbf{B}G_e^{(1)} := ( \cdots (G \times G)_e^{{1}} \stackrel{\to}{\stackrel{\to}{\to}} G_e^{(1)} \stackrel{\to}{\to} {*} ) \,.

The normalized Moore cochain complex N (C (BG e (1)))N^\bullet(C^\infty(\mathbf{B}G_e^{(1)})) of the cosimplicial algebra

C (BG e (1)):=(C ((G×G) e 1)C (G e (1))*) C^\infty(\mathbf{B}G_e^{(1)}) := ( \cdots C^\infty((G \times G)_e^{{1}}) \stackrel{\leftarrow}{\stackrel{\leftarrow}{\leftarrow}} C^\infty(G_e^{(1)}) \stackrel{\leftarrow}{\leftarrow} {*} )

is in degree kk given by the kernel of the joint degeneracy maps. As in the discussion at differential forms in synthetic differential geometry this picks out the functions that vanish on degenerate simplices. So from the above we get

N (C (BG e (1)))=(g *g *d= i(1) id ig *0). N^\bullet(C^\infty(\mathbf{B}G_e^{(1)})) = (\cdots g^* \wedge g^* \stackrel{d = \sum_i (-1)^i d_i}{\leftarrow} g^* \stackrel{0}{\leftarrow} \mathbb{R} ) \,.

Recall that the differential of the Chevalley?Eilenberg algebra? is on g *g^* just the dual [,] *:g *g *g *[-,-]^* : g^* \to g^* \wedge g^* of the Lie bracket [,]:ggg[-,-] : g \otimes g \to g.

We need to check that this is reproduced by the differential of the Moore cochain complex, which is the alternating sum of the face maps d= i(1) id id = \sum_i (-1)^i d_i. Let fg *f \in g^*. Then we find for all (x,y)(G×G) e (1)(x,y) \in (G \times G)_e^{(1)} that

(df)(x,y)=f(p 1(x,y))f(xy)+f(p 2(x,y))=f(x)+f(y)f(xy). (d f) (x,y) = f(p_1(x,y)) - f(x \cdot y) + f(p_2(x,y)) = f (x) + f(y) - f(x \cdot y) \,.

Now we use the crucial formula (6.8.2) from Anders Kock’s book, which says that the group product on the infinitesimal elements x,yx,y is given by

xy=x+y+12{x,y}, x \cdot y = x + y + \frac{1}{2}\{x,y\} \,,

where the last term is the group commutator

{x,y}:=xyx 1y 1. \{x,y\}:= x \cdot y \cdot x^{-1} \cdot y^{-1} \,.

So this is the term that remains in the formula for dfd f:

(df)(x,y)=12f({x,y}). (d f)(x,y) = -\frac{1}{2} f(\{x,y\}) \,.

On that we apply theorem 6.6.1 of Kock’s book, which says (in its third item) that under the loglogexpexp bijection by which we identified the infinitesimal neighbourhood G e (1)G_e^{(1)} (and functions on it) with the tangent space T e(G)T_e(G) (and linear functions on it) the group commutator maps to the Lie algebra commutator. So indeed under the identification of ff with an element in g *g^* we find

df=[,] *f. d f = [-,-]^* f \,.

This is indeed the differential of the Chevalley?Eilenberg algebra?.

(discussion needs to be completed: situation in higher degree and cup-product mapping to wedge product needs to be discussed…)

Last revised on October 30, 2009 at 16:43:01. See the history of this page for a list of all contributions to it.