Let $(M,g)$ be a spin manifold. We define the spinors and the spinorial Clifford algebra:
$$\psi . \psi' + \psi' . \psi = 2 < \bar \psi, \psi'>$$
with $\psi,\psi'$ two spinors and
$$(X.\psi).\psi' =\psi .(X. \psi')$$
with $X$ a vector.
$$\psi^{-1}= \frac {\bar \psi}{<\psi,\psi>}$$
We then can define the action of a spinor $\psi$ over a function $f$:
$$\psi (f)=df^* . \psi$$
and a spinorial connection:
$$\nabla_{\psi} (fs)= \psi (f).s + f \nabla_{\psi} (s)$$
with $s$ a section of a module over the spinorial Clifford algebra.
The spinorial Dirac operator is then:
$${\cal D}_{\psi}= \sum_{i \in I} \psi_i . \nabla_{\psi_i}$$
with $(\psi_i)_{i\in I}$ an hermitian basis of the space of spinors.
Has such a construction a physical meaning?