Normally a Boolean function is self dual if:
$$f(x_1, x_2, ..., x_n) = \lnot f(\lnot x_1, \lnot x_2, ..., \lnot x_n)$$
For example the "not" function is self-dual:
$$\lnot x_1 = \lnot (\lnot (\lnot x_1)) = \lnot x_1$$
But the "and" function is not:
$$x_1 \land x_2 \neq \lnot(\lnot x_1 \land \lnot x_2) = x_1 \lor x_2$$
But how do we model something like true $1$ or false $0$? Are they self-dual? I don't know how to model them as functions since they don't take arguments.
For example if $f = \text{True}$ then I don't even know how we'd write a self-dual when there are no function arguments to negate.
They are actually each other's dual, because you can see the True as a conjunction of $0$ conjuncts (it is trivially true that 'all' $0$ conjuncts are true) and the False as a disjunction of $0$ disjuncts (if there are no disjuncts, then there is not at least one true disjunct) and the conjunction and disjunction are each other's dual.