We often describe the topically or thematically linked books we publish as being in “conversation” with each other, as if our library turns tea party after hours. Of course, it’s the authors responsible for the works who actually animate those interactions, and late last year we invited two of them to engage one another more directly. In The Work of Revision, published last summer, Hannah Sullivan reveals iteration in literary production to be an inheritance bequeathed by modernism, with close and complicated ties to the advent of the typewriter. In his forthcoming Track Changes: A Literary History of Word Processing, Matthew Kirschenbaum casts a roughly similar gaze on the rise of word processing software. The two were kind enough to allow us to listen in as they discussed affinities between their work in a weeks-long email exchange; what follows is an edited and condensed version of that conversation.
Matthew Kirschenbaum: Hannah, in The Work of Revision you identify the typewriter as central to the practices (or call them poetics) of revision you read as characteristic of high literary modernism through the composition habits of figures like James, Woolf, and Eliot. For me this grounds your argument not only in the space of literary and textual criticism, but also conversations in media studies, including the technologies of writing and even what is increasingly known as media archaeology. How do you walk the line between attentiveness to the material particulars of the medium—which extend beyond the typewritten page to the ecologies and economies of inscription that surrounded it, whether James’ dictation or the stratification of the writing process into autographs, typescripts, proofs, etc.—and the slide into technological determinism?
Hannah Sullivan: Hi Matt, Yes, the typewriter is definitely a key part of my book. But I must emphasize how medially diverse the twentieth century was. The material documents I studied include many things besides typescript: notebooks, manuscripts, different stages of proof (Joyce’s proofs for the “Ithaca” episode of Ulysses are notorious), serial printings, and first and subsequent book editions. Most of the writers I study continued to draft and compose by hand, but they revised and rewrote (also usually by hand) on very different kinds of pages.
And there is quite a bit of idiosyncrasy. In the 1900s, Henry James dictated new material to his amanuensis—he had given up writing—and then corrected typescript (sometimes including aural mishearings) before seeing proof. The revision process isn’t always linear. But when he revised his earlier fiction for the New York Edition in the same decade, he was working from reproductions of his earlier published novels, pasted up with a specially wide margin for revision. Another example: in The Waste Land facsimile, we can see Eliot looping back into new composition by hand on the typescript, not a fresh sheet of paper.
The typewriter also came in at a later stage in the process than we tend to think. Media historians tend to focus more on “typing” than “typing up.” But for the high modernists (and then for their sometimes self-consciously imitative successors), the typewriter was more an instrument of re-representation, of re-presentation, than composition. Were the earliest document preserved in the story of a modernist novel’s development a typescript (even marked-up and scored with revision, and only the first of many others) the first working assumption should be that the manuscript was lost or destroyed.
This changed in the postwar period. And it doesn’t have much to do with competence at typing. Sylvia Plath was an excellent typist (and typed up Ted Hughes’s poems for him), but by writing her own poetry out in a distinctive, babyish hand, she was, I think, marking out her work as serious, modernist, and not effeminate or secretarial. (Not “written on a typewriter by a typewriter,” in Randall Jarrell’s infamous put-down of Oscar Williams’ poems.) For male writers such as Kerouac, Ginsberg, or O’Hara, the typewriter had a different symbolic meaning, untainted with the fear of secretarial work. They saw in the machine the “lyrical” possibility of spontaneous, first draft composition.
The idea that all or most first draft composition happened on the machine can lead to other determinist arguments. If something is written on a machine, does it make it mechanical? (Not for the Beats.) Isn’t that a genetic fallacy? Friedrich Kittler makes this assumption when he argues that Malling-Hansen’s writing ball made Nietzsche into a laconic. But for most of the writers I study—Henry James, T. S. Eliot, Virginia Woolf—typewriting slowed down the process of composition, introducing more stages for revision, as well as objectivity and distancing. The Waste Land facsimile, preserved forever with Pound’s cuts and jibes and “echts,” as well as Eliot’s own more tentative changes, shows how human and dialogic the typewritten page can be.
In writing the book I was interested in the story that we can tell about the medium of composition. We have exceptionally good access to the draft documents of modernism, but the writers I work on didn’t want their work to be instantiated like this. In some cases there are awkward ethical questions about whether we should even be looking at the drafts. So, unlike French genetic critics, who often take a rather hostile attitude to the published “fixed” texts, I try to emphasize the various events of publication (far-off, divine, as they may seem) as an important conceptual end goal. The documents of composition and revision, in other words, are more heterogeneous, more incoherent, of necessity, than the published books that they lead to, however sinuously.
Kirschenbaum: There is so much to engage here! Part of what interests me in these conversations are the iconic dimensions of different writing technologies, their place in both the popular and the authorial imagination; you touch on this above with your remarks about the different and _gendered_ dimensions that the typewriter assumed as an instrument of composition. I’m struck by the popular online lists that circulate nowadays of writers and their writing tools, like this one for example: http://flavorwire.com/410384/the-writing-tools-of-20-famous-authors/. It’s the typewriter, along with the fountain pen, that are celebrated—not to be too Derridean about it, but the fact is the computer marks the entire list precisely through its absence.
I mention this because after talking with and researching dozens of authors and their particular writing habits for Track Changes, I’ve discovered that very few only ever use any one single “tool.” Michael Ondaatje, who I had the privilege of interviewing, is famous for writing drafts over and over again in longhand. But at some point in the process he also dictates (making additional changes by voice) and then an assistant types the spoken and recorded text into a Word file, which Ondaatje will revise and edit further. (Incidentally, dictation was once explicitly part of how word processing was defined, one stage in a complete and routinized system for the production of written documents.)
I would like to ask you now about the archival dimension of your research, particularly in relation to your remarks about publication as an “end goal” above. Did these writers imagine or anticipate an archival identity for their work at all, and can we think of the availability of pre-publication materials in archives (however formidable the practical barriers to access may be in individual cases) as marking an equally significant event horizon, in the sense that given our habits and predilections as textual scholars we are as compelled by the archival record as we are the finished text? Of course sometimes the archive inverts itself, or like a Klein bottle it _becomes_ the publication: I’m thinking, of course, of Valerie Eliot’s Waste Land edition here, among others. Has your immersion in the archives of modernists suggested that we perhaps approach, reconcile, _read_ the modernist archive as though it itself were a modernist event?
Sullivan: Your point about writers’ tools and the heterogeneity of materials is fascinating. I’m struck by the fetishistic quality of some of the material descriptions in the “writing tools” piece, and remember feeling the same way when I was beginning this book and heard Claire Messud talk about the fact she could only write in a particular brand of French notebook (graph paper; Clairefontaine). The audience admired The Emperor’s Children, so we all took this down, as a kind of tip. Magical thinking, no?
A kind of aura attaches to the HB pencil, the cheap biro, and the Mont Blanc equally—and to the typewriter too, as the prices in vintage and antique shops show. Do you think this will also be the case in time for word processors and then computers? The difficulty seems to be keeping the machines in working order—enabling them to be long-lived writing tools, rather than simply curiosities. And I am certain that we need to question the idea that the relationship between tool and product is iconic. We understand that the biro doesn’t produce “cheaper” or quicker prose than the Mont Blanc. So why should a machine produce literature that is mechanical? Why couldn’t avant-garde poetry be written in bog-standard MS Word? My work on the laborious, individual, highly artisanal, slow typescripts of modernism suggests (contra Kittler) that the typewriter doesn’t make everyone a laconic.
I don’t think it’s until the postwar period that writers started building the possibility of future manuscript and typescript sales into their compositional horizons. This is a period of rapid growth in funding for US universities, and it had already become clear that private individuals were willing to pay large amounts for writers’ manuscripts. (The Berg Collection was donated in 1940; the Ransom Center was building up its collections in the 1920s and 1930s). And of course the timing of this is not unconnected to the history of modernism. The war was better for American universities than for the high modernists!
The boundary between public and private documents alters when one knows that even unpublished documents might be read by scholars, and, yes, absolutely, the sense of publication as a single, dramatic, estranging “event” must alter. Archives are places where reading and interpretation happen quite slowly, I think, piecemeal, on different visits, and often over many years. But the modernists I discuss were, I think, relatively naive about the prospect of institutional collection and examination of their papers. In some ways, this makes their drafts more interesting to look at—the fantasy of peering over someone’s shoulder is easier to sustain when we imagine they didn’t want us to do it. In this sense, the jumble of documents that are gathered together in the facsimile of The Waste Land, which Eliot took remarkably little care over in the period post-1922, are exemplary for being rather uncurated! James wanted everything destroyed.
Kirschenbaum: To your point about whether the relationship between tool and product will persist into the age of computers, laptops, and word processing: there are some tantalizing examples. George R. R. Martin (of Game of Thrones fame) reportedly still writes with WordStar on an MS-DOS machine he maintains for that purpose. (Sites like eBay and Craigslist make the acquisition of much vintage computing equipment quite trivial, but this too of course is a phenomenon with a half-life.)
I’m also struck by Charles Stross’s recent self-described “rant” against MS Word, not only the virulence (and characteristic eloquence) of it, but also by the glee which accompanied its widespread circulation. Some of that, of course, is basic anti-Microsoft sentiment, but I think it’s also about bearing witness to the spectacle of a writer so passionately invested in his or her tools. WordStar and Scrivener (the latter is Stross’s software of choice) have very different affordances from Word of course, but the fetishistic desire remains as well: I think about Peter Carey’s laptop on display under glass at the State Library of Victoria in Melbourne, or the justly lauded emulations of Salman Rushdie’s personal computers available to patrons of the rare books reading room at Emory. Derrida has written a bit about his belief that an “aura” will come to surround famous computers in the same way as famous manuscripts. And William Gibson’s throwaway line about “Stephen King’s Wang” in his novel Pattern Recognition was my point of departure for Track Changes.
There is one final question I wanted to ask you, which is about the role of digital humanities (so-called) in your research. One is struck by the screenshots from the Juxta tool that populate your book, for example; can you talk about digital tools as part of your research method and the extent of your interest in using them to extend your further textual investigations?
Sullivan: I think these questions about tools are absolutely fascinating, and you bring up terrific examples. In my own case, all the same, they’re mostly interesting along the way—the end goal is always analysis of what they produce, the final text. And I do take seriously the idea that writing tools might not be tremendously determinative; this is the Diving Bell and the Butterfly argument, and a good humanist point of view.
The emulations of Rushdie’s computers seemed bizarre to me when I first heard about them. Rather brilliant, but also rather pointless. For Benjamin, the aura doesn’t attach to the tool (chisel, pen, paintbrush) that produces an object, but to the object that the tool makes. Derrida might be right to say that, in the future, we’ll find writers’ computers inordinately interesting, but I don’t think there is a very good parallel for this in the past; we’re interested in the blood, sweat, tears, and paralipomena on the page because they are hermeneutically interesting, not in the pen itself.
A lot of my work is geared towards putting the author—and his or her rational, conscious, executable intention—back in the picture. To me, it’s exciting to realise that (at least sometimes) individuals can write and revise in conscious pursuit and realisation of a sophisticated intention. (It so often doesn’t happen, for me, for my students, for many people.) But I’m like almost all other modern literary critics in caring primarily about the work, and the traces through which it comes to be, rather than the person who makes it, or the tools they use. e.g., I would definitely rather know what Michael Pietsch did in editing The Pale King on his own computer than have the computer that David Foster Wallace wrote on, as an object in itself.
On tools: I think Juxta is so very useful. I had far more Juxta screenshots in my original manuscript, and I think this was partly because it seemed to me an intellectually honest and straightforward way of presenting variation, acknowledging (often) that two states being compared were not binary alternatives, but two of many (the drop-down menu offering more). And it was convenient. Some of these screenshots made it through into the book, but a simple parallel text would have done as well. Probably this will look very dated soon!
The real beauty of Juxta, which I use quite a bit with students, is being able to produce a list of post-publication revisions very quickly. When I taught “The Middle Years” at Stanford, I was amazed that students could notice in ten seconds the things that had taken me weeks to unearth. On the other hand, I’ve tended to find that the human eye is pretty good at discovering semantically significant variation (“lived” to “dived,” for example)—it’s the commas, the small adjective shifts, that we tend not to notice...
I am working on a new project now on free verse. There I am trying to do a formalist, non-semantic version of what Franco Moretti has called “distant reading.” The Work of Revision isn’t really about (nor does it benefit from) digital humanities, in any significant sense. But I do think the idea of using computers to notice formal patterns (revision; line length) rather than/as well as semantic meaning (e.g. the collocation of two nouns, like “love” and “marriage”) is important, and something that has been weirdly neglected. I think old-fashioned formalists, true close readers, could both gain from, and give much, to digital humanities. Learning about and summarizing the content of texts—the plots of millions of 20c novels, say—doesn’t seem to me to be the distinctive ability of a computer over a human. But who wants to count the average number of syllables per line in mediocre verse of the 1860s?