In historical writing, declarative questions tend also to be mimetic questions. . . . If historical research were as empirical as it can be, then we might hope to see very large heuristic hypotheses put to very small controlled tests. . . . But many historians follow a different method. In common practice, a general interpretation is fashioned by an essayist not as a heuristic hypothesis but as an affirmative proposition. In the next twenty years or so, a legion of gradgrinds manufacture monographs which reify the essay, with a few nonconsequential changes. Qualifications are inserted at the end of sentences, active verbs are changed to passive, pronouns of indefinite reference are converted to proper nouns, and footnotes are added at the bottom of the page. This process continues until another essayist publishes another brilliant general interpretation, and another generation of gradgrinds are wound up like mechanical rabbits and set to running about in ever-smaller circles. The result is a dialogue between essayists and monographers which resembles the exchange between Hamlet and Polonius [at Hamlet 3.2.406 ff.]".
According to Fischer, the saying is "variously attributed to Max Beerbohm and Herbert Asquith." I am committing something like this same fallacy in not pausing at this point to track it to source myself.
