I came across an interesting quote from David Lightfoot in Natural Logic and the Greek Moods:
Greek has desiderative stems (as opposed to the σείω system) which serve as presents and are bases for conjugations: ἀλέξω, νίσομαι. Equally, Latin has desiderative presents: quaeso against quaero, viso against video. However, it is generally supposed that the role of the desiderative in Greek is essentially to form the future. One sees the original desiderative value of the future in ἧλθε . . . λυσόμενος . . . θύγατρα (Iliad I, 12) and then 17 lines later τὴν δ’ἐγὼ οὐ λύσω. Many old futures have ε vocalism and these take middle endings, which might itself point to a desiderative value: πείσομαι: πασχω, χείσομαι: χανδάνω. . . .
. . .
If our interpretation is correct, it would make no sense to posit a desiderative source for the form of what we are calling the existential future. Indeed there is not much evidence for this traditional view, although I have nothing better to suggest. However, given that the future indicative and the aorist subjunctive forms are so often interchangeable, I should want to relate the sigma of the future to that of the aorist, but I can find no convincing evidence for this hypothesis either.