I've been reading Peirce on Retroduction; while I agree that retroduction (reasoning backwards from an idea, a consequence, a phenomena which imposes itself upon us tenaciously to a hypothesis of which it is a consequent) is certainly well illustrated in the history of science, I have to take serious exception to his belittling of the "darwinian" hypothesis that these ideas impose themselves upon us by chance. I think this explanation, when properly refined, is _just_ the one which explains many of the surprising advances in scientific theory, especially those proffered by the "genii" of science history. Peirce mentions pure play and, to be more specific, musing as the process through which we are enabled to grasp the ideas that form the fodder of retroduction. But it is in just this state that we are most open to our own internal presuppositions, those presuppositions which have been instilled in us by upbringing, environment, and accident, and give us those peculiar psychological drives which in turn influence which beliefs seem to us to be inordinately significant. By musing we do allow our 'self' to converse with 'self,' and much of the content may be novel. But the grammar of the conversation and most of the vocabulary is structured by our own previous experience. And because no two subjects will have had the same experential history, this grammar and vocabulary is going to differ on a subject to subject basis. The end result is that many of the ideas which are generated in these musings will be unique to the subject engaging in them, and of those that are held in common with other subjects, the _significance_ (the prime motivating factor in allocating mental resources to a given idea or belief) will differ. This change in signficance, besides leading to different lists of valuation, will change the character of the idea itself (because the relative importance of a belief in a system or web of ideas _is_ an aspect of its character). So while the "harmony of the spheres" was a neat idea, perhaps good for a footnote, to many of the renaissance philosophers, it was an veritable obsession with Kepler.
And it is not just conscious musing which generates the ideas which fuel retroduction; a whole host of other experiences may suggest ideas: dreams are particularly common (CF. the discovery of the shape of the Oxygen molecule after the chemist dreamt of Ouroboros). So to a very great extent it IS chance (in the sense that the psychological conditioning and personal history of any given individual is chance from a purely human perspective) which is the source of these ideas, not some kind of 'rational instinct' or 'instict towards aprehending the truth.' To apply the metaphor of biological evolution, after the chance variation (the new idea) occurs, it is then worked upon by the forces of natural selection (the methodological scrutiny which then follows). Having past the basic test of survival, the idea must then compete against rival ideas for the cognitive resources of scientists, explanatory comprehensiveness being equivalent to the ability to secure resources, parsimony being equivalent to being able to make do on little, and fruitfulness being equivalent to fecundity. And, of course, while all theories (like organisms) die, the succesful one's do leave behind something of themselves in their offspring. I think Peirce really underestimates the explanatory power of Darwinism.