At least as from Edgeworth and Pareto we think about utility in mathematical terms. My question twofold
(i) about the start of the usage of binary relations to model preferences in economics, and hence the start of the axiomatic utility representation exercise.
(ii) about the start of usage of the symbol $\succsim$ to represent weak preference.
About (i) the earlier reference I could find is Wold's (1943): "A synthesis of pure demand analysis I-III" pioneering work on utility representation; and about (ii) is Herstein and Milnor (1953): "An axiomatic approach to measurable utility" work on expected utility.
Does anyone know about earlier uses??
Ant further comments, references or ideas are most welcome.