The most difficult of intellectual disciplines is certainly the history of science. It requires the virtues of both the historian – exhaustive acquaintance with what can be masses of documents – and the practicing scientist – mastery of conceptually challenging subjects that more often than not involve technical ability in mathematics. You can tell when one side or the other is lacking. Without the history, scientific historians are prone to draw unfounded and often arbitrary conclusions about influence, tradition and scientific revolution (The real story of the multiple sources of seventeenth century atomic theory and its variations is a good example). Without the science, historians will as likely as not make comparisons based on accidental likenesses and differences. Their histories will read more like histories of literary themes or metaphors.
Koyré notoriously inclines to the latter fault. He rarely asks whether a scientific theory or philosophical argument from the past is valid or not, a question his chosen subject uniquely demands since, unlike the objects of a Panofskian iconology, science moves in the atmosphere of debate and verification. (Bennett, by the way, goes too far in the other direction: Not enough history, not enough concern that 17th century issues such as, for example the distinction between space and matter, though related, might have meant something different from their contemporary counterparts. He treats “early modern” texts like a paper just written by that clever Assistant Professor down the hall, ignoring significant differences in meaning, intent and context. Without the history, early science sounds puzzlingly naïve and primitive, such that one wonders how it could ever have led to clever people like ourselves.)