How Should You Use Generalizations? Or Should You Avoid Them?

When you are studying a subject or person -- especially a person -- in depth, you start to notice things that make you say, "oh, he was always doing that sort of thing." I do it, for instance, when I say such things, as I did in an earlier post about Douglass, as  "He writes like this a lot to people." That was in a blog post, however; and, if I were to put something like that in my manuscript, I would add a couple of examples in either the text or a discursive end note. Am I being to picky or overly cautious about such things?

I am thinking about this because of some of the generalizations made in the book I love to hate to love to Zapruder, Diedrich's Love Across Color Lines. In one instance, she writes, "Several manuscripts in Assing’s handwriting in the Douglass Papers show that she sometimes served as his secretary, and it is possible that she even drafted letters, speeches, and editorials for him." [p. 193] Since Diedrich also makes the claim that Assing wrote Douglass's editorials in the New National Era, a claim I find troubling and that is based only on a letter from Assing to her sister (an interaction that renders the claim unreliable), that I cannot check right now because it is written in German 19th century script, and that I have no idea of verifying independent of Assing, I thought I should at least check similar claims that might strengthen or weaken that assertion. This seemed to be one.

Fortunately, there is a footnote for this sentence that refers readers to "Drafts of letters to George T. Downing on Ebenezer Bassett’s appointment as minister to Haiti (FD to G.T. Downing. Undated but 1869. FDP. LC) and to Senator Henry Wilson of Massachusetts (FD to H. Wilson. 12 September 1866. FDP. LC) are in Assing’s handwriting.”

You can see both of those letters online, as well as a sample of Assing's handwriting. They match.  So, yes, Assing seems to have sometimes written drafts of letters for Douglass. At least, she did so on two occasions during a three year period. Does the rest of Diedrich's assertion, then, follow from this citation? Did Assing also write "speeches and editorials for him"? Does this mean that she "served as his secretary"? Do these two examples -- and I have found no others in the two years in question, and I am looking for other examples, including among the other types of documents in the Douglass Papers at the Library of Congress -- even merit the adjective "several"?

Am I being too strict here? How do you convey something common that you have observed in your research that might earn "several" or "usual" or any such other sort of descriptor without entering every single instance in your notes? Should you enter every single instance, to CYA? Would a better note in this particular example have given citations of not only letters, but speeches and editorials, and given citations over a broader period of time? Would a better way to have handled this have been to write the sentence differently to narrow down the period of time or type of document? Was this an awkward use of these two letters, which may have served another argument better, if at all?

I'm also resisting the urge to say that Diedrich does this frequently. I feel as if I should give more examples to prove "frequently." Perhaps this should serve as another. The example in that post says that "Douglass hinted at marital problems in letters to friends, describing himself in 1848, for example, as a 'most unhappy man.'" In that post, I pointed out the problems with that citation, which I think might disqualify it from being an example for that particular point. Yet, no other examples appear in that citation. You can find evidence for marital problems -- this example just is not among them -- and you can find at least one example from Douglass himself from a rather shocking letter from 1857. Can, however, this be described  in the way that Diedrich does here? Is this just an example of a poorly argued point that might have been rewritten with better use of the sources? Should you avoid the use of generalizations at all?

This troubles me because the framing of the story in this way -- that this or that sort of thing was constantly happening, without defining when and under what conditions and without enough evidence to back up the assertion of continual behavior -- is the way that dubious information enters into the message creep syndrome. I have, in the past six months, heard repeated both the "fact" that Anna Murray was pregnant before she got married to Douglass and that the scene in Douglass's autobiographies in which a black man in Manhattan helped him find David Ruggles was a gay pickup. The first "fact" comes from William McFeely's biography, in which he asserts his claim in the face of evidence to the contrary, and the second comes from John Stauffer's book Giants, in which he throws in that interpretation into his narrative -- absent any queer theory or secondary literature about the history of homosexuality -- in order to have some sort of parallel with the oft-questioned belief that Lincoln was gay. In both instances, the ideas are intriguing, but the evidence weak if non-existant. Yet, the ideas are being repeated without examination of the source.

In other words, this troubles me because this is the way that myth gets made, not history.