What are anticommuting spinor parameters $\zeta^\alpha$? I'm reading Martin F.Sohnius, Introducing supersymmetry, page 82. It is the first time he introduces the anticommuting spinor parameters $\zeta^\alpha$ to calculate the supersymmetry variations of a field. At this point, I think he still stays in the ordinary quantum field theory (no superspace, no superfield yet).
So my question is, what exactly are these spinor parameters? For me, it is very reasonable to see them as some anticommuting operators. (which may satisfy $\eta\zeta=\eta^\alpha\zeta_\alpha=c\in\mathbb{C}$...)
 A: They shouldn't be thought of as operators i.e. $q$-numbers; instead, they should be thought of as $c$-numbers. They're mutually anticommuting but otherwise they play exactly the same role as $\Delta x^\mu$ for translations or angles $\varphi$ for rotations.
They're spinor variables which means that under a Lorentz transformation $\Lambda\in SO(3,1)$, they transform as 2-spinors
$$\zeta^{\prime\alpha} = \Lambda^\alpha{}_{\beta} \zeta^\beta $$
where $\Lambda^\alpha{}_\beta$ is the $SL(2,C)$ matrix equivalent to the $SO(3,1)$ transformations.
The variables $\zeta^\alpha$ anticommute with all other Grassmann-odd $c$-numbers, e.g. with each other, 
$$ \zeta^\alpha \zeta^\beta = -\zeta^\beta \zeta^\alpha$$
etc. and with all Grassmann-odd $q$-numbers (i.e. fermionic operators) but they commute with everything else.
These variables carry the units of $[\zeta^\alpha]={\rm length}^{1/2}$ which may be said to be a geometric average of the units of rotations or Lorentz transformations (those are dimensionless) and the translations (${\rm length}$). This fact is obviously equivalent to saying that the generators $Q_\alpha$ they are multiplied by (and contacted with) have the units of ${\rm mass}^{1/2}$.
Because the Grassmann numbers are anticommuting, they can't take any particular nonzero "ordinary" real values. But one should still imagine that there is an "undefined" set of values they may take – and these values may be integrated over via the Grassmann (Berezin) integrals.
See e.g.

http://motls.blogspot.com/2011/11/celebrating-grassmann-numbers.html?m=1

"Celebrating Grassmann numbers" for more details and comments.
