I would certainly agree that scientific method (as promulgated by philosophers of science) has 'no method intrinsic to the generation of new systems of ideas [that] guarantees their greater functionality than previously existing systems of ideas'. The standard scientific experiment, which projects the hypothesis into the world (if only the intentionally limited world of the laboratory) and if it false the world will tell you so. Then you have to generate a new hypothesis.

But I am not sure why this is so different from cognitive development. There are functional overlaps in the domains (e.g., in infancy, between visual tracking and auditory location) and also structural overlaps between the mechanisms that support these processes (e.g., control of head movement), and I would argue that it requires both these overlaps for the relevant feedback loop to have effect. But conflicts arising in these overlaps only tell the infant that there is something wrong (apologies for homuncular image - please treat as shorthand!) - not either what is wrong or what a better 'hypothesis' would be. The infant simply has to try again.

That is one reason why infants and other human beings are so _bad_ at development. Of course, as the develop, they internalise solutions to previous problems, which themselves eventually emerge as regular methods and even methodologies. So we get better at guessing what to do next. But even the cognitively most sophisticated individual (e.g., a scientist?) is not guaranteed to have any intrinsically valid insight into what to do next.

Incidentally, isn't scientific method simply one of the loftier expressions of this very fact - that we are cognitively pretty good at being methodical about our mistakes, but, lacking as we do any method for generating better hypotheses, we continue to make inspired guesses? The only difference between individual cognitive development and the development of science is that we consciously do the latter collectively.