Moretti’s Methodology: The Cure for Symptomatic Reading?

Moretti’s Methodology: The Cure for Symptomatic Reading?

There are four competing versions for the “original” publication of The Portrait of a Lady: two were simultaneously serialized in England and the United States, and each was bound into book form shortly thereafter. I had assumed that perusal of the textual variants might tell us something about the ways in which English usage in England was different from that in the United States. Using Moretti’s system, I confirmed the existence of textual variation, but found also that I had to take care that the reading of these findings was not too symptomatic, as per Rooney’s discussion. Implementing his system myself at once transformed my symptomatic reading into a surface reading and made me desire more context for Moretti’s investigation.

I looked at various versions of the text of The Portrait of a Lady to see how many dashes had been printed; I chose the dash for the unscientific reason of its seeming to me an interesting piece of punctuation in that it has no fast rules governing it; it is a piece of punctuation that is open to interpretation -- thus, one can imagine how my symptomatic reading of the dash might easily lead me to some conclusions re: the temper of the English reader in coping with the mystery of the dash versus that of the American reader; who would be more willing to engage with this nebulus? I assumed we could learn something about the transatlantic readership in the early 1880s: a natural time constraint inheres in the fact that the versions I looked at did not appear over decades like some revisions do, as the artist matures and looks back or as conventions of language change; instead, these four versions of The Portrait of a Lady are issued within the space of only two and a half years.

First, I looked at the first serial installation of the text as it appeared in the English Macmillan’s Magazine in October 1880; next, at the same installation as it appeared in the American Atlantic Monthly in November 1881. After that, I examined the same portion of the work when it was published by Macmillan in book form in 1882; and finally, I looked at the same portion as published in the United States by Houghton Mifflin in 1881. To approximate Moretti’s method, I counted the dashes in each version by downloading an editable version of each and manually using the Find + Replace function (same results, more laborious process).

In Volume I, the Atlantic stripped the Macmillan version of the text of 28% of its dashes, retaining only 80 of the original 110. Although the English Macmillan’s Magazine serial publication began before the American Atlantic Monthly started running The Portrait of a Lady, an American publisher, Houghton Mifflin, bound it as a book first. In doing so, they reinserted (or left intact) 108 of the original 110 dashes. And then Macmillan did the same thing -- a curious choice, since Macmillan already had an edited version of the text. Because both the American and English versions of the bound text appear to be the same, my original hypothesis -- being able to “read” the readerships on either side of the Atlantic through their relationship to the dash -- was undone. The editorial choices made in England were thought fit to travel without doing any real violence to an American set of readers and, in this instance, the dash tells us nothing about the mind of the English reader versus that of the American.

Fortunately, the history of The Portrait of a Lady’s publication is available to further inquiry: the plates for the book were originally fabricated by Macmillan in England then sent over to the United States, where they were first published by Houghton Mifflin and later used by Macmillan to print the English version. This information confirmed the unraveling of my symptomatic reading I would have done. It is, however, not the case that history would yield the 7,000 novels investigated by Moretti to further, perhaps contradictory, investigation. That is, we can never know if the shortening of novels’ titles is a response to the exigencies of advertising unless we are able to confirm this with supporting evidence (i.e. a control group of long titles juxtaposed with shorter ones). So, while his method yields data that would otherwise be too burdensome to calculate and might, as in my case, thwart a desired symptomatic reading, I am not sure that the data does not simply produce another “text” that is equally open to interpretation. The shortening of titles over time may just as well have informed us as to the price of ink and paper as to the necessity of advertising in circulating libraries’ catalogs. While his set of data may offer another data point in the reading of a body of texts, it may be just another data point.