IN AN ESSAY from his 1983 collection, The Middle of My Tether, Joseph Epstein mulls over and satirizes the acknowledgments, dedications, and other secondary material found “in and around books,” particularly scholarly ones. To illustrate a typical acknowledgment of debt to an author’s wife, Epstein quotes from a diplomatically unnamed 1964 book:

For the better part of two years, she conducted research, typed manuscript, criticized drafts, ran down leads, corrected proofs, and somehow simultaneously cared for two small children and a home. Her sacrifice has improved every page of this book; the errors and shortcomings are mine.

“Ah, but yes,” Epstein puts in, “one can hear a feminine voice adding, and so was the pleasure all yours and so shall be the glory.” Milton Friedman, similarly conceding the clichés of the form, noted in the 1962 preface to Capitalism and Freedom that “the refrain, ‘But for my wife, this book would not have been written,’” had become routine, but he added that:

In this case, it happens to be the literal truth. She pieced together the scraps of the various lectures, coalesced different versions, translated lectures into something more closely approaching written English, and has throughout been the driving force in getting the book finished.

As for typing up Capitalism and Freedom’s manuscript, that was mostly handled, the preface acknowledges, by Friedman’s secretary, Muriel A. Porter.

Today, acknowledgments have escaped academic halls and routinely run on like film credits or wedding speeches in the back matter of novels. A spousal tribute is often the last word, and the expressed debt is customarily emotional, sometimes editorial, but rarely if ever secretarial. Publishing roles remain highly gendered, but if wives are still typing their husbands’ manuscripts, dudes aren’t rushing to admit it.

This change is a product not only of feminism but also of word processing. Unless one is stubbornly attached to old technology beyond the early stages of composition, one won’t need to type and retype clean copies of multiple revisions, or make sense of tousled penmanship, or master the art of smoothly obscuring liquid-paper brushstrokes. For those of us who’ve spent all or much of our writing lives in the era of personal computers, it’s sometimes hard to fathom the drudgery (and delegation) that went into the production of pre-digital books. In his new study, Track Changes: A Literary History of Word Processing, Matthew G. Kirschenbaum is especially concerned with how word processing has changed the embodied labor of writing — its actual tasks, tools, and physical demands — and with how literary writers have embraced, resisted, and interpreted that transformation.

The phrase “word processing” wasn’t coined with literary elegance in mind. It originally applied to a range of technologies and practices, including typewriters, that allowed for delayed inscription, a shift from oral to recorded dictation, copiers, and a restructuring of office domains. As Kirschenbaum explains, research and consulting outfits such as the Word Processing Institute and the American Management Association, as well as manufacturers such as IBM, promoted the adaptation of new products and operations to stanch the proliferation of paperwork and, more pressingly, to cure a “social disease” stemming from secretarial mobility and versatility. Secretaries made up a large percentage of office personnel, and their duties were a hodgepodge: taking letters, typing, filing, making coffee, scheduling meetings, answering phones, booking flights, stocking supplies, buying gifts, and so on. Accordingly, they were hardly bound to their desks, which raised inefficiency flags for management consultants. Summarizing rhetoric from business publications of the early and middle ’70s, Kirschenbaum writes of a perceived “terminal condition characterized by ceaseless circuitous movements through various arteries and corridors, impromptu meetings and interactions (obvious vectors for gossip and rumor), and other wasteful movements.”

Much of this could be eliminated, it was argued, if the “gal Friday” or “office wife” role was replaced by more specialized jobs. Experts recommended an expansion of the typing pool into word-processing centers where paperwork would be generated, often from cassette and telephonic recordings transcribed using headphones, foot pedals, and keyboards. With word processing, wrote an AMA consultant quoted by Kirschenbaum, “the random, the haphazard, the unpredictable, the largely unmeasurable can thus […] be replaced by the controlled, the organized, the automatable.” This wasn’t just a pitch to invest in the latest offering from IBM, but a call for major operational reform toward the unprecedentedly productive office of the future. “Just as magnetic tape, cartridges, and disks quickly rendered the friction of the eraser obsolete in correcting errors,” Kirschenbaum writes, “other kinds of intra-office friction — the ‘personal vagaries’ alluded to by the AMA […] — would be eliminated by making the typing secretary into a kind of removable media herself.”

This was all very well for the business world, but the potentially alienating effects of such efforts ran counter to the Romantic ideas of composition to which writers still pledged nominal allegiance. “For the literati,” Kirschenbaum writes, “word processing was invariably burdened (not buoyed) by the associations carried by the gerund in that compound term, which could function as a foil for nothing less than humanity itself.” Few of the first literary writers to try word processing trumpeted their use of the technology, lest their work be lowered by the taint of automation. Kirschenbaum devotes a good portion of his book to early adopters who, sometimes through institutional associations, wrote or revised novels on cumbersome first-wave word processors during the ’60s and ’70s. Kirschenbaum presents Len Deighton’s Bomber as a strong candidate for the first novel significantly created on a word processor. A multiperspective war novel published in 1970, Bomber was drafted on a state-of-the-art typewriter, the famous IBM Selectric, but then retyped (and processed), by Deighton’s literary secretary, Ellenor Handley, on IBM’s MT/ST: a Selectric equipped with a magnetic tape storage unit whose cartridges held about 24,000 characters, remembered revisions, and could print and reprint text at 170 words per minute.

By the late 1970s, a few novelists, mostly genre practitioners like the science-fiction writers Jerry Pournelle and Larry Niven, were working with word processors in the narrower, more modern sense: that is, using an application to compose, edit, and format text on microcomputers with keyboards, monitors, and improved storage capacity. Less pioneering but still ahead-of-the-curve writers who made the switch in the early ’80s often did so at considerable expense. Peter Straub paid $14,000 for his IBM Displaywriter 6580, on which he co-wrote The Talisman in mostly remote, modem-enabled collaboration with Stephen King, who used a Wang System 5 Model 3. By the time that novel was published in 1984, according to data compiled by the Association of American Publishers, approximately half of literary authors were using word processors.

Culling from specialized publications, mainstream journalism, and author interviews, Kirschenbaum recaptures the excitement and optimism writers often felt in the face of this magical new technology. To many, word processing seemed to promise a new possibility for aesthetic perfection. “Perfect” was the leading marketing keyword, found in ad copy and in product names such as WordPerfect, Letter Perfect, and Perfect Writer, and more than a few novelists greeted the mantra as something more than hype. If, in one traditional view, literary perfection was either illusory or the province of poems and other short works, now, it seemed, even a long novel could be refined to an apotheosis of unalterable integrity. The modularity of word-processed text made major structural reorganization a matter of a few clicks (well, you’d probably need to switch back and forth between several floppy disks). You could tinker endlessly with sentences: transposing phrases, deleting a comma, replacing an adjective, restoring the comma. You could search out and decimate pet words and phrases. Hannah Sullivan, a scholar quoted by Kirschenbaum, wrote in 2013 that, with word processing, “the cost of revision” had “fallen almost to zero.” Kirschenbaum quotes a 1988 interview with Anne Rice in which she held that, with word processing, “there’s really no excuse for not writing the perfect book.”

Given the vaunted word-processing breakthroughs of the 1980s, it’s unsurprising that some contemporary writers remain loyal to their original systems. George R. R. Martin, for example, still uses a DOS-type computer on which he runs the long-superannuated WordStar (which would doubtless be an intolerable editorial hassle if his books didn’t make so much money). Edward Mendelson, as he explained in a New York Review of Books essay, is a WordPerfect holdout, as was Philip Roth until his retirement.

As Kirschenbaum notes, “our instruments of composition, be they a Remington or a Macintosh, all serve to focalize and amplify our imagination of what writing is.” And writers, he finds, have traced the word processor’s essential ancestry to about every conceivable spot on the compositional spectrum. To Jorie Graham, it’s just “a fancy typewriter.” For Kamau Brathwaite, who used his Macintosh to craft unique typographies, “the computer is getting as close as you can to the spoken word.” To others, word processing’s advantages had to do precisely with its return, after the relentlessly linear typewriter, to the flexibility of longhand: writing on a computer felt something like flipping through a notebook, scrawling a bent arrow from one paragraph to the next, filling a page with proofreaders’ marks without actually making the marks. If you write and edit in a single ever-changing document, the (elusive) path toward perfection is one in which the finished work, as Kirschenbaum writes, “bears no visible trace of its prior history; indeed, it is as though the document did not have a history, but rather emerged, fully formed in its first and final iteration, from the mind of the author.”

In the era of word processing, the ideal of multiple drafts, once so important to the literary writer’s amour propre, has for many of us become something of an abstraction. Several years ago, I attended a presentation by a well-known writer who boasted of the 30-odd drafts that went into her books. She challenged aspirants to follow suit. I’d just read a short story by this writer in which more than a few clichés had survived those pains, so I wondered if the trouble was entirely necessary. But, more than that, I didn’t know how or why I would properly enumerate my drafts. In writing novels, essays, and stories, I’ve often enough changed a document’s title, printed it out, created a new document of excised material, or otherwise left a draft-like trail — and there are notebooks, index cards, and other artifacts awaiting a water-heater mishap in my basement. But, for the most part, writing on a word processor is a constant braid of composition, revision, and erasure. The text looks instantly professional, feels infinitely provisional.

On a word processor I often revise compulsively and carelessly: it’s diverting and self-flattering to make endless little supposed improvements to sentences, but if I make the wrong choice, there’s always, it seems, another chance to get it right. (Until there isn’t.) Word processing, Kirschenbaum argues,

emerges as a combination of the indefinite suspension of inscription and the allure of real-time editorial intervention […]. In effect, the writing surface becomes a Möbius strip, with the writer both writing and not-writing at the same time — which is to say, writing in multiple locations simultaneously, one text made of light and another stored indefinitely prior to printing onto yet another (even more durable) surface.

Could be; I certainly spend a lot of time at my desk both writing and not-writing, though the latter might involve checking Facebook or conducting consumer research on sweatshirts.

Beyond the ken of Kirschenbaum’s study are questions of how word processing affected the style of individual writers who transitioned from earlier methods of composing and revising, or how it has affected literary style more generally. He cites the work of scholars who’ve done important work in that dauntingly broad area, but, on this matter and others, he’s chary about sweeping summation and loose speculation. In the concluding chapter, he writes, “Every impulse that I had to generalize about word processing — that it made books longer, that it made sentences shorter, that it made sentences longer, that it made authors more prolific — was seemingly countered by some equally compelling exemplar suggesting otherwise.” This seems right: by some measure, for example, we’re in an extended period of over-fastidious attention to sentence construction that must have something to do with computers; by another, we’re in the ungainly middle of an anti-style period that must have something to do with computers.

To further complicate such questions, most writers, maybe more than ever, employ catch-as-catch-can methods and technology. They do so even for relatively undemanding projects. To write this review, for instance, I took notes in the book itself, composed a few early sentences in my head while chopping a salad, drafted roughly half the piece longhand on a legal pad, wrote and revised the rest on Microsoft Word, checked my email for the 17th time that day, printed the more or less finished draft in order to note further revisions by hand, entered those changes while making a few more, repeated these last two steps, and sent the article off to an editor, who might follow a similarly mixed approach.

Of course my hopes for perfection are as slim as they are for book-length projects. Spurred by Kirschenbaum to try a new application, I fed the review through a word-frequency text analyzer, which led me to delete all uses of “perhaps,” one of my tics. I should have done more, but it seemed like it was time to think about lunch. For his part, Kirschenbaum uses “opined” twice on one page, but I suppose I only noticed because I’m averse to even one use of “opined.” Well, there’s only so much we can do. Word processors give us endless tools, but not endless energy. They can’t, alas, prompt us with satisfyingly incisive endings. Someday, perhaps.


Dylan Hicks is a writer, musician, and the author of the novels Amateurs and Boarded Windows.