The distinction between "scientist" and "scholar" has been increasingly blurred, especially in the social sciences (owing no doubt to a certain blurriness around the notion of "science") by the image of a "researcher". Heidegger was to my knowledge the first to point out that "modern" science really meant the disappearance of "the scholar" and the emergence of the "researcher". (See this post.) For the purpose of the comparison I want to make in this post, i.e., of the social sciences to the humanities, I will need to reconstruct this distinction. While it has become common, especially in the administrative sciences to talk about oneself as a "scholar" and one's work as "scholarship", this general sense is not my meaning here. Following Heidegger, I will mean by "scholar" someone whose "learning" is rooted in erudition, while a "scientist" is one who is engaged in developing a "theory of the real".
The scholar, unlike the scientist, has no method and no theory, only a style. While scientists can claim to have done something very specific (method) and seen the world from a particular perspective (theory), scholars working in the traditional humanistic disciplines can only claim to approach their material with a sort of "attitude". We might say that in so far as the scholar has a "theory" and a "method" it is only in the rhetorical senses that I encourage social scientists to adopt when writing. A theory is just a system of expectations; a method is just a source of credibility. Scholars working in the humanities can make specific efforts in their writing to arouse the reader's expectations and to win the reader's trust, but this will not happen by appeal to some shared set of "categories of observation" (concepts) or to some procedure by which to establish the "given in experience" (data). (All "data" is of course relative to method, i.e, data is a "methodological" issue, just as all concepts are "theoretical".) Rather, the scholar must cultivate a distinct, yet somehow recognizably "academic" style.
While such a style does not have to reach the level of high literature, Proust's famous definition* can help us to understand what is at stake:
What we call reality is a certain relationship between sensations and memories which surround us at the same time, the only true relationship, which the writer must recapture so that he may for ever link together in his phrase its two distinct elements. One may list in an interminable description the objects that figured in the place described, but truth will begin only when the writer takes two different objects, establishes their relationship, and encloses them in the necessary rings of his style (art)...
What the scientist calls "reality" is of course something a bit different, or is at least a relation between sensation and memory somewhat differently construed. Memory is brought to bear upon the scientist's experience through the intermediary of theoretically formed concepts, which summarize the result of past observations and experiments. Sensation is allowed only in the form of carefully collected data, derived from the flux of experience by a set of increasingly refined operations that allow them to make observations that are not colored by memories; or at least not colored in a way that is not controlled by the concepts that the theory makes explicit. We might say that where scientists take great pains to establish their objectivity with respect to their object, scholars cultivate a studied subjectivity about their subject. This may be what Norman Mailer meant when he thanked Diana Trilling for reading him—indeed, misreading him—with her "full and specific sympathy".
It's always my aim to be practical. So I can't leave this post at a mere theoretical distinction between "scholarly" and "scientific" writing. How, then, we may ask, do scholars "enclose [things] in the necessary rings of [their] style" if not, like scientists, by framing them with theory and probing them with method? The sense in which style can do double-duty for theory and method should become clear once we realize that the only thing that informs writing of scholars, the only thing that shapes their specifically scholarly sympathies, is their reading. Scholarly inquiry is simply reading enclosed by rings of reading. (This is not true of novelists, mind you. Ideally, their work is about life enclosed by living.) But we can distinguish between different kinds of reading, and we can distinguish between different reading materials.
In our writing as scholars we are telling our reader what we have read and how we have read it. Since the first is likely to arouse particular expectations in a reader that shares our frame of reference, the "what" of our reading serves a purpose similar to theory in social science. The "how" of our reading, meanwhile, goes along way toward establishing our credibility, so there is a direct analogy to the methods section of a social science paper. Now, it can be useful to distinguish between our primary and our secondary sources, between, for example, works by Kafka and works about Kafka, but it's important to keep in mind that in both cases the readers's expectations of our analysis will be aroused by what we've read, whether by or about Kafka, and the reader's trust will be won by how we have read these works, again regardless of whether they are by or about the author under study.
For every paragraph you write in a humanities paper, then, ask yourself whether it tells the reader primarily (1) what you have read by the author your have studied; or (2) what you have read about the author you have studied; or (3) how you have read the author you have studied; or (4) how you have read the work of your scholarly peers. This will tell you what "ring" of your style you are at this moment, i.e., during the 27 minutes you have devoted to writing this paragraph, trying to enclose your subject in.
I hope that helps. I have a feeling I need to say more about this.
*Update: It occurs to me that this isn't really a definition of "style" but of "reality". I would argue, however, that he's saying precisely that reality is a stylistic construction, and that style is simply a bringing together of sensation and memory.