Having laid down his rules of method, Descartes proceeded to deviate from them. We expect a demonstration of the truth after the principles have already been found, such as been prescribed by Plato. But we get a different way of accounting for the facts: "the way of hypothesis" proscribed by Plato. Unable to proceed further deductively, he resorted to the invention of and choice among different hypothesis, the choice being determined by crucial experiments. Thus he gave up the certainty of the a priori method in favor of the conjectural a posteriori. This meant that in his practice, the synthesis preceded analysis, for it preceded that form of it known as inductive testing.
source
Colin Murray Turbayne, The Myth of Metaphor (1962) [https://catalog.hathitrust.org/Search/Home?type%5B%5D=author&lookfor%5B%5D=Colin%20Murray%20Turbayne&page=1&pagesize=100&ft=ft] p. 38.