
Definition of 
Supercommutative algebra and its dual coalgebra
Case of 
Definition of 
Lie superalgebra structure on 
Differential
of 
Representation as vector fields
Simpler notations
Ghost number
Are
Faddev-Popov ghosts?
Definition of 
As far as we know,
was introduced in Alekseev:2010gr, in the context of current algebras.
We will now present it in the language of quadratic algebras.
Supercommutative algebra and its dual coalgebra
Let

be a super linear space, and

be the tensor algebra generated by

:
Here the upper index

implies that we consider tensors as forming an
algebra.
Let

be the free super-commutative algebra generated by

. We define it as the subspace
of

consisting of symmetric tensors.
(To define the structure of algebra, we notice that this subspace is isomorphic to the
factorspace by
quadratiic relations of

, which are antisymmetric tensors in

.)
Then, the dual coalgebra

, by definition, the subspace of
symmetric tensors in the tensor
coalgebra:
(Here the upper index

means that we consider it as a coalgebra.)
For example, let

be even and

be odd.
The following tensors belong to

:
The bar construction is:
The subspace

is annihilated by
because of relations of

.
In particular, the following are elements of

:
These are
symmetric tensors in

.
To summarize, if
is the algebra of symmetric tensors in
,
then
is the coalgebra of symmetric tensors in
.
Case of 
Let
be a Lie superalgebra. We consider it as a graded Lie superalgebra, with all elements
having grade zero. Let us apply the construction of
〚Supercommutative algebra and its dual coalgebra〛
to
.
Let

be the free commutative superalgebra generated by

, and

its Koszul dual
coalgebra:
At this point, we consider

only as a linear superspace.
The Lie algebra structure on

is forgotten.
Let us consider the cobar construction of

:
Now the index

in

means that we consider tensors as forming an algebra, the free algebra.
The overline over

means that we remove
the counit, see
LodayVallette for precise definitions.
We actually only need a subspace:
which is generated by commutators. This is a free Lie superalgebra.
Consider the natural twisting morphism
(which is denoted

in Chapter 2 of
LodayVallette,
but we reserve

for contraction of a vector field into a form).
Its image belongs to

. It satisfies
the Maurer-Cartan equation;
using the notations of Chapter 2 of
LodayVallette:
where

is the differential on

induced by the coalgebra structure on

.
Since

is a supercommutative algebra,

actually
belongs to the subspace:
This implies that

preserves the subspace:
(Indeed, the

of the cobar complex is

, and

is supercommutative;

“kills the commutator”.)
This means that we may write

instead of

.
To summarize,
with
is a differential graded Lie superalgebra.
Definition of 
Let us consider a larger space:
Here

stands for semidirect sum of Lie superalgebras, with arrow pointing towards the ideal.
The embedding of

into

as the first summand will be denoted

:
Lie superalgebra structure on 
We define the commutator of two elements of
as follows:
Differential
of 
There is a natural projection:
annihilating all tensors with rank

(
i.e. 
).
We define a differential

on

, in the following way.
We postulate that the action of

on

(the first term in Eq. (
6)) be zero:
It is enough to define the action of

on

.
We put:
We then extend Eqs. (
8) and (
9) to the differential

of

.
We consider

with the differential

which will be called

:
The nilpotence of

follows from the fact that

anticommutes with

:
Representation as vector fields
Consider the cone of our free Lie algebra:
and its universal enveloping algebra

. The Maurer-Cartan Eq. (
5) implies:
This is an equation in the completion of

.
Let

be some supermanifold, and

the algebra of vector fields on it.
Suppose that we are given a map of linear spaces:
Such a map defines a representation of

in the space of pseudo-differential forms (PDFs) on

.
We want to project Eq. (
11) on the space of PDFs on

. It is not possible to do directly,
because we do not require that

act on PDFs.
Instead, we will use the following version of Eq. (
11):
for all

.
Simpler notations
For us

, and

is the coalgebra
of symmetric tensors in

. Therefore, the space

can be thought of as the space of formal Taylor series of functions on the superspace

with values in a linear superspace

.
The subspace

,
where

consists of rank

tensors, is the space
of

-th coefficients of the Taylor series. In particular, we interpret

as the space of functions on the supermanifold

:
where

means that we are not being rigorous. We ignore the question of which functions are allowed,
i.e. do not explain the precise meaning of

.
Let

denote some basis in

, and

the dual basis in the space

of linear functions on

:
In this language:
for

any tensor symmetric in

.
To agree with
Alekseev:2010gr, we will denote:
Here the notation

agrees with Eq. (
7).
Ghost number
In our notations, if

has ghost number

then

has ghost number

.
In other words,
lowers ghost number. In particular, the cone of the Lie superalgebra

is:
The generator

has ghost number

.
For example, in
bosonic string theory,

corresponds to

where

is the BV antifield for ghost and

the odd Poisson bracket. As a mnemonic rule,

has the same ghost number as

. Elements of

(the first summand in

) all have ghost number zero.
Are
Faddev-Popov ghosts?
As an ideal in
, the
is Lie superalgebra with zero commutator. We can think of
as Faddeev-Popov ghosts of the BRST complex of
. Since the commutator is zero,
the BRST differential annihilates
.
But when we consider

and not

, the commutator

is nonzero;
we consider the free Lie superalgebra generated by

. We might have introduced new
Faddeev-Popov ghost for nested commutators, and the corresponding BRST differential.
But this is not how the construction goes.
Instead, we introduce new generators, such as

, and the differential

such that
etc. This differential

acts on elements
of the Lie superalgebra, not on Faddeev-Popov ghosts.
This is
not the Faddeev-Popov construction. Notice that

gets contracted
with products of

. In this sense, we may say that we replace the Faddeev-Popov

with nonlinear functions of

. We replace:
with: