Scribd
Upload
40 views4 pages

A M N A F M N F F Am Af M M A A M M A M F M N A A A F A M N A F F F A F M F F F A M N F M N M F F A A A N N N N N N N ' ' M F A M ' M M F M ' A A M M M M M ' A A F A M M F M F M

The document discusses two perspectives on working with super objects: 1) The "even rules" perspective treats super objects in terms of their even parts and works functorially by considering how objects change under extensions of scalars. 2) The "odd rules" perspective (in a minority report) expresses the notion of a super Lie algebra in terms of a quadratic map from the object to its even part, functorial under extensions of scalars. This de fines a bracket satisfying certain properties.

Uploaded by

luisdaniel
Copyright
© Attribution Non-Commercial (BY-NC)
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PS, PDF, TXT or read online on Scribd
40 views4 pages

A M N A F M N F F Am Af M M A A M M A M F M N A A A F A M N A F F F A F M F F F A M N F M N M F F A A A N N N N N N N ' ' M F A M ' M M F M ' A A M M M M M ' A A F A M M F M F M

The document discusses two perspectives on working with super objects: 1) The "even rules" perspective treats super objects in terms of their even parts and works functorially by considering how objects change under extensions of scalars. 2) The "odd rules" perspective (in a minority report) expresses the notion of a super Lie algebra in terms of a quadratic map from the object to its even part, functorial under extensions of scalars. This de fines a bracket satisfying certain properties.

Uploaded by

luisdaniel
Copyright
© Attribution Non-Commercial (BY-NC)
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PS, PDF, TXT or read online on Scribd
You are on page 1/ 4

Appendix: Even rules

P. Deligne

Let A be a commutative super ring and M , N be A-modules. A morphism f : M ! N


is a parity respecting map f such that f (am) = af (m).
The \even rules" point of view is that
(a) one should think to M in terms of the functor which to any extension of scalars A ! A0
attach the even part (M 0 )+ of M 0 := A0
A M ;
(b) it amounts to the same to give a morphism of modules f : M ! N or to give for each
commutative super A-algebra A0 a (A0 )+ -linear morphism F [A0 ] from (M 0 )+ to (N 0 )+,
functorial in A0 .
More precisely, the map f 7! F :

F [A0 ] = f 0 jM 0 + ;
with f 0 deduced from f by extension of scalars, is bijective.
Proof. To a functorial F [A0 ]: (M 0 )+ ! (N 0 )+ one attaches a morphism f : M ! N as
follows. On M + , f := F [A]. Take now A0 = A[] with  odd. An element n of N 0 + can be
written uniquely in the form n0 + n1 with n0 2 N + and n1 2 N . De ne '0 and '1 on
M by
F [A0 ](m) = '0(m) + 1 (m);
and de ne f on M to be '1.
Apply functoriality to A0 ! A:  7! 0. The corresponding map from (M 0 )+ to M +
maps m (m 2 M ) to zero, and it follows that '0 = 0: for A0 = A[], we have

F (A0 )(m0 + m1 ) = f (m0 ) + f (m1 )


1
for m0 in M + and m1 in M .
We now check that f is A-linear. The A+-linearity is clear. If a is inA and m in M + ,

f (am) = F (A0 )(am) = aF (A0 )(m) = af (m) :

For m in M , applying functoriality to A0 ! A:  ! a, we get f (am) = af (m) as well.


If F is obtained from a morphism f , one has for A0 = A[]

f 0 (m0 + m1 ) = f (m0 ) + f (m1 )

and it follows that this construction gives back f .


Conversely, the f deduced from F gives back F by extensions of scalars: for a0 2 A0
and m 2 M even,
F (a0 m) = a0 F (m) = a0 f (m)
by A ! A0 functoriality. For a0 and m odd, functoriality for A[] ! A0 :  7! a0 gives

F (a0 m) = a0 f (m)

as well.
A similar, more complicated argument, gives the
Theorem. Let Mi be a nite family of A-modules. The data of a morphism
f :
Mi ! N

amounts to the data of a A0 +-multilinear map, functorial in A0 , of the (Mi0 )+ to (N 0 )+ .


The proof is left to the reader.
Corollary. Let B be an A-module. To give on B a structure of A algebra amounts to give
on (B0 )+ a structure of (A0 )+ algebra, functorially in A0 . The algebra B will be associative
(resp. commutative, resp. Lie, : : : ) if and only if the algebras (B0 )+ are.
Note that the (B0 )+ are purely even objects: at the cost of having to think functorially,
which one wants to do anyway, this removes signs from the de nitions.
2
Example. Let X be a supermanifold and L be an (even) function on X . The locus where
L is stationary is the subspace S of X (a sub-supermanifold in good cases) de ned by the
vanishing of the section dL of the cotangent bundle of X .
When thinking functorially, one thinks to X as determined by the functor which to
any supermanifold B attaches the set B-points of X , where a B-point is a morphism
x: B ! X . Intuitively: a family of points of X parametrized by B. In this language,
describing S means describing which B-points x: B ! X map to S , i.e. factors through
S ,! X . Such B-points will be called stationary.
By de nition, a B-point x is stationary if for any even or odd section s on B of x TX
(the \ ber of TX at x"), hdL; si = 0 on B. If s is odd, let B0 = B  R0;1: we add an
odd parameter . Then, s is even and hdL; si = 0 on B if and only if hdL; si = 0 on B0 .
We see that a B-point x is stationary if and only if after any change of basis B0 ! B,
the B0 -point B0 ! B ! X , still noted x, is such that for any even section s of x TM ,
hdL; si = 0 on B0 .
This can be rephrased as follows: after any change of basis B0 ! B, one wants that
for any family of points x(t) ( " < t < "): B0 ] "; "[! X , one has
d
dt L(x) = 0 at t=0
or, more cumbersomely written: the pull back of L to B0 ] "; "[ is such that @L
@t = 0 at
t = 0.
This formulation is convenient when X is \in nite dimensional", and when what makes
good sense is not the space X and the function L, but the notion of map x: B ! X and
of the pull back by x of L.

Minority report: Odd rules.


The notion of super Lie algebra can be expressed nicely as follows:
Data: after any extension of scalars A ! A0 , a quadratic map x2 from L to L+, functorial
in A0 .
Applying the theorem to the associated bilinear form, from (L)+
(L)+ to L+,
we see that x 7! x2 de nes a bracket [x; y]: L
L ! L such that for x and y in L ,
[x; y] = (x + y)2 x2 y2 . In particular, for x in L , [x; x] = 2x2 .
3
Axiom: after any extension of scalars A ! A0 , for any x in L , one has

[x; x2 ] = 0 :

Provided 6 is invertible in A, this de nition is equivalent to the one in term of the


Jacobi identity (which is a polarized version of the axiom).

You might also like