The attempt to solve the central problem of philosophic method involves recourse to Newton's 'Rules of Philosophizing' rather than to Descartes' Discourse on Method, with the result that philosophy presently takes an entirely new direction. For Newton's method is not that of pure deduction, but that of analysis. He does not begin by setting up certain principles, certain general concepts and axioms, in order, by virtue of abstract inferences, to pave the way to the knowledge of the particular, the 'factual.' Newton's approach moves in just the opposite direction. His phenomena are the data of experience; his principles are the goal of his investigation. If the latter are first according to nature, then the former must always be first to us. Hence the true method of physics can never consist in proceeding from any arbitrary a priori starting-point, from a hypothesis, and in completely developing the logical conclusions implicit in it. For such hypotheses can be invented and modified as desired; logically, any one of them is as valid as any other. We can progress from this logical indifference to the truth and precision of physical science only by applying the measuring stick elsewhere. A scientific abstract or 'definition' cannot serve as a really unambiguous starting-point, for such a starting-point can only be ogtained from experience and observation. This does not mean that Newton and his disciples and followers saw a cleavage between experience and thinking, that is, between the realm of bare fact and that of pure thought. No such conflicting modes of validity, no such dualism between 'relation of ideas' and 'matters of fact' as we find in Hume's Enquiry concerning Human Understanding, is to be found among the Newtonian thinkers. For the goal and basic presupposition of Newtonian research is universal order and law in the material world. Such regularity means that facts as such are not mere matter, they are not a jumble of discrete elements; on the contrary, facts exhibit an all-pervasive form. This form appears in mathematical determinations and in arrangements according to measure and number. But such arrangements cannot be foreseen in the mere concept; they must rather be shown to exist in the facts themselves. The procedure is thus not from concepts and axioms to phenomena, but vice versa. Observation produces the datum of science; the principle and law are the object of the investigation.
This new methodological order characterizes all eighteenth century thought. The value of system, the 'esprit systematique,' is neither underestimated nor neglected; but it is sharply distinguished from the love of system for its own sake, the 'esprit de systeme.' The whole theory of knowledge of the eighteenth century strives to confirm this distinction. D'Alembert in his 'Preliminary Discourse' to the French Encyclopedia makes this distinction the central point of his argument, and Condillac in his Treatise on Systems gives it explicit form and justification. Condillac tries to subject the great systems of the seventeenth century to the test of historical criticism. He tries to show that each of them failed because, instead of sticking to the facts and developing its concepts from them, it raised some individual concept to the status of a dogma. In opposition to the 'spirit of systems' a new alliance is now called for between the 'positive' and the 'rational' spirit. The positive and the rational are never in conflict, but their true synthesis can only be achieved by the right sort of mediation. One should not seek order, law, and 'reason' as a rule that may be grasped and expressed prior to the phenomena, as their a priori; one should rather discover such regularity in the phenomena themselves, as the form of their immanent connection. Nor should on attempt to anticipate from the outset such 'reason' in the form of a closed system; one should rather permit this reason to unfold gradually, with ever increasing clarity and perfection, as knowledge of the factsa progress. The new logic that is now sought in the conviction that it is everywhere present on the path of knowledge is neither the logic of the scholastic nor of the purely mathematical concept; it is rather the 'logic of facts.' The mind must abandon itself to the abundance of phenomena and gauge itself constantly by them. For it may be that it will not get lost, but that instead it will find here its own real truth and standard. Only in this way can the genuine correlation of subject and object, of truth and reality, be achieved; only so can the correspondence between these concepts, which is the condition of all scientific knowledge, be brought about.........
http://world.std.com/~awolpert/gtr574.html