i finally found the link for my presentation on reductionism v. complexity theory -- a 1938 book by a french epistemologist.

The scientific mind is first seen clearly and incontrovertibly when it makes representation geometrical, that is to say when it delineates phenomena and puts an experience's decisive events into an ordered series. This is indeed how we arrive at figured quantity, halfway betweenthe concrete and the abstract in an intermediate zone where the mind aspires to reconcile mathematics and experience, laws and facts. Yet while such geometrisation seemed to have often been achieved... in the end it is always shown to be in some way deficient.

- gaston bachelard, formation of the scientific mind

bachelard's "stages" of scientific development: observation, generalisation of observation into some form of abstraction, and complete abstraction, are more appealing than thomas kuhn's "normal" vs. "paradigm-shifting" science in that these three stages seem to describe how an individual thinker can engage the process of scientific discovery. as i remember it, kuhn's descriptions of scientific revolutions seemed to start out with "and then, this guy suddenly had an idea..."