How to Write a History of Writing Software

Isaac Asimov, John Updike, and John Hersey changed their writing habits to adapt to word processors, according to the first literary historian of the technology.

The journalist David Halberstam sits at his desk in 1993, surrounded by papers and his word processor. (Mark Lennihan / AP)

It’s hard to believe, but one of the most important changes in the way people write in the last 50 years has been largely overlooked by historians of literature. The word processor—that is, ​any computer software or hardware used for writing,  a nearly ubiquitous technology adopted by poets, novelists, graduate students, foreign correspondents, and CEOs—has never gotten its own literary history.

Perhaps it was just too much under our noses—or, I suppose, in front of them.

Now it finally has one. Five years ago, Matthew Kirschenbaum, an English professor at the University of Maryland, realized that no one seemed to know who wrote the first novel with the help of a word processor. He’s just published the fruit of his efforts: Track Changes, the first book-length story of word processing.

It is more than a history of high art. Kirschenbaum follows how writers of popular and genre fiction adopted the technology long before vaunted novelists did. He determines how their writing habits and financial powers changed once they moved from typewriter to computing. And he details the unsettled ways that the computer first entered the home. (When he first bought a computer, for example, the science-fiction legend Isaac Asimov wasn’t sure whether it should go in the living room or the study.)

His new history joins a much larger body of scholarship about other modern writing technologies—specifically, typewriters. For instance, scholars confidently believe that the first book ever written with a typewriter was Life on the Mississippi, by Mark Twain. They have conducted typographical forensics to identify precisely how T.S. Eliot’s The Wasteland was composed—which typewriters were used, and when. And they have collected certain important machines for their archives.

One day, a similarly expansive body of work may exist for writing software—and Kirschenbaum will be one of its first builders.  I spoke to him last month about his effort. This interview has been edited and condensed for the sake of clarity.

Robinson Meyer: “Who was the first author to write a novel on a word processor?” You cast that question as what drove you to write this book. Is there something close to a definitive answer for it?

Matthew Kirschenbaum: We can’t know with absolute certainty, I don’t think, but there are a couple of different answers.

If we think of a word processor or a computer as something close to what we understand today—essentially a typewriter connected to a TV set—there are a couple of contenders from the mid- to late-1970s. Notably Jerry Pournelle, who was a science fiction author. He is probably the first person to sit and compose at a “typewriter” connected to a “TV screen”—to compose there, to edit, and revise there, and then to send copy to his publisher. That was probably a novella called Spirals.

If we move back a little bit further, there’s an interesting story about a writer named John Hersey, the novelist and journalist. He did the famous book Hiroshima. He was at Yale in the early 1970s, so maybe about five years before Pournelle, and he worked on one of the mainframe systems there. He didn’t compose the draft of the novel he was working on at the keyboard, but he did edit it, and use the computer to typeset camera-ready copy.

So those are two candidates. But in many ways the most interesting, and in some sense the truest, candidate for that particular honorific is a British author named Len Deighton. He wrote espionage thrillers; I describe him as the Tom Clancy of his day, very commercially successful. He was able, in the late 1960s, to afford a piece of equipment that IBM had put on the market. It was the first product they actually sold and marketed as a word processor. It was called the Magnetic Tape/Selectric Typewriter, the MT/ST.

There was no screen on the MT/ST, but all of the character strokes and keyboard input were saved to magnetic tape. A skilled operator could then use that to revise, just as we do with a word processor. So Len Deighton, working with his secretary—a woman named Ellenor Handley—wrote a novel entitled Bomber on the MT/ST, and it was published in 1970.

That arrangement too was very typical. Often it was the secretary, who was very often female, who learned to operate the word processor. She was the one who actually processed the text, if you will.

Meyer: When someone used the MT/ST, did they have to remember what they were typing? Did it print out on a page?

Kirschenbaum: Your “screen” was the sheet of paper you had in your Selectric typewriter. You did your typing on the Selectric—which is the same typewriter, for example, we see in Mad Men; it’s a famous ’60s-era electric typewriter—and if you made mistakes, you would backspace. You would get a mess on the sheet of paper that was currently on the Selectric, but the correct sequence of character strokes was being stored on the tape. Then you would put a clean sheet of paper into the typewriter and it would automatically print out, sort of player-piano fashion, the text stored on the tape’s storage.

This unit sold in the 1960s for $10,000. That’s obviously quite a lot of money, and IBM used the term word processing as a marketing device. And Deighton was the first and one of the very few individuals who had one in his home. The real market, of course, was the business world.

Meyer: Does that mean Deighton dictated the novel?

Kirschenbaum: He had a regular Selectric that he used, then he would hand his drafts to Ellenor Handley, the secretary. She would retype them on the MT/ST, and do all of the subsequent editing, revisions, correcting there, because it made it so easy to do that revising and to produce clean copy on demand.

There are three things I really like about that story and why I feel like it’s the best candidate for quote-unquote “first.”

One, it defamiliarizes our sense of what word processing is. It’s not a typewriter connected to a TV set. The key thing turns out to be the magnetic storage layer. The other thing I like about it is—there’s a term I use in the book, “suspended encryption.” That captures that dynamic of word processing: You’re writing, but there’s a kind of suspended animation to it. The text remains in its fluid, malleable state, until such time as you commit it to hard copy.

The other thing I like about the story is that it captures that gendered dynamic, that social dimension of writing. It’s not just the author alone at his typewriter. It’s really a collaborative process, there is a gender dimension to it, and there’s something very human about it, I think.

Meyer: The Deighton example also reflects that Track Changes is a popular literary history, not just a highbrow account. And something you talk about is that everyone—the popular writers, the literary writers, and their readership—is figuring out the technology together. The writers and the readers receive, say, an Apple II at the same time. Is that a distinctive episode in history?

Kirschenbaum: One thing I’d say is that the computer is a general purpose device. A typewriter can do only one thing—it’s a typing machine, it’s a writing machine. If you were someone who got a typewriter in the late 19th, early 20th century, it was because you wanted to write. If you got a computer in the late 1980s, it wasn’t necessarily because you wanted to use it to write novels.

The other thing I would point to is that the history—this really interesting hybrid of literary and technological history—is radically compressed and foreshortened. The book has about a 20-year span that it operates in, but really, the majority of the story—if you bracket off Len Deighton and these unique circumstances—takes place from the late 1970s to around ’84 or ’85. Things were changing so fast that a writer just starting to use a word processor in 1977—or, as they were then called, a micro-computer—would have had a drastically different experience from a writer buying what would have been a top-of-the-line system come, say, 1983.

Meyer: Did they realize that was happening? Were they aware that their options were changing much faster than they could cope?

Kirschenbaum: Yes, they did. And of course there was a lot of resistance and anxiety to processing, and some of that they could ascribe to a general technophobia, an anxiety about the glowing glass screen. But a lot of it was also just writers being overwhelmed by the options.

One thing that was really different about word processing is that there were dozens and dozens of word-processing options, dozens and dozens of systems and software and formats, all of them incompatible and comparatively expensive. If you made a bad choice, that would have been a real setback for a writer. So realizing what you needed, and shopping for a computer—that in and of itself was a barrier as much as any big sense of technophobia.

Meyer: Was there a backlash to word-processing during this period? Or did you see a resumption of anxiety about the declining use of freehand?

Kirschenbaum: There were certainly writers who were the gold refuseniks. In the same sense that, a little bit later, when the internet came along and you had popular jeremiads against the internet, a number of high-profile writers complained or ranted about word processing. Gore Vidal famously claimed that word processing was “erasing literature.” Thomas Pynchon claimed that all of the writers he knew were rushing to get a word processor. Harlan Ellison swore he would never touch one. So you had these very dogmatic, colorful positions that calcified among some pretty high-profile writers.

A lot of the more interesting stories just have to do with writers who were more concerned that their editor or their agent would think that their computer had “done the work.” That was a real anxiety. One thing I did was to just go back and read popular press accounts of word processing, and a lot of them begin by assuring the reader that the computer really isn’t writing the book.

Meyer: Is that because so much of the work of writing beforehand had been so material?

Kirschenbaum: I think so. One thing that I try to get at in the book is that there is a stock of cultural imagery around the act of writing. The hardboiled novelist pounding the keys of the manual typewriter—that’s a very iconic image. It takes a little while before similar kinds of imagery for computers enter into the literary imagination.

There’s a pretty famous R. Crumb cartoon of Charles Bukowski that I was able to include in the book. It shows Bukowski sitting in front of his Macintosh. One thing I like about the image is that it shows the way the computer had been naturalized into the stock of cultural imagery around the act of writing within a decade or so.

There was a lot of anxiety about computation and automation for a couple of decades before the personal computer arrived on the market. And in a lot of sectors of society, people had concerns—not unfounded—that they would be replaced by machines. You saw this in popular films like 2001. The computer as a threatening, otherworldly intelligence that might replace human work and creativity—that was very much an established fear and anxiety, and it fit right into writing and word processing.

Meyer: How gendered were those fears? Your description of “the hardboiled novelist, pounding at the typewriter,” that’s very violent—and you describe Deighton working with his female secretary, and the secretary actually being the one who uses the word processor. Was there some fear of emasculation?

Kirschenbaum: I think there was. Certainly any writer who bought a word processor and just looked through the instruction manual would have noticed that all of the photographs depicted women using the machine, because that was the office norm. A lot of this broke down along the distinction between speaking and writing. The assumption throughout the 1970s and the early part of the 1980s culture was that the male executive spoke—and that was dictation—and then the female secretary wrote at a keyboard, to transcribe it and turn the speech into text.

Another interesting story that’s in the book is about John Updike, who gets a Wang word processor at about the time Stephen King does, in the early 1980s. I was able to inspect the last typewriter ribbon that he used in the last typewriter he owned. A collector who had the original typewriter was kind enough to lend it to me. And you can read the text back off that typewriter ribbon—and you can’t make this stuff up, this is why it’s so wonderful to be able to write history—the last thing that Updike writes with the typewriter is a note to his secretary telling her that he won’t need her typing services because he now has a word processor.

Meyer: There is a material history you can read from a typewriter. I think you mention the example of Lawrence Rainey, a scholar of T.S. Eliot, being able to decode The Waste Land’s compositional history by looking at his typewriter. And I remember there being anxiety around writing software, and the future of that kind of scholarship. Did writing this history make you buy into the anxiety that we won’t be able to preserve contemporary literary work?

Kirschenbaum: So much of writing now, and that includes literary writing, that includes novels and poetry that will become culturally resonant and important—all of this happens now digitally. And that was something that I was interested in writing about, writing the book. What I found is that there were often very surprising examples of evidence remaining, even from these early days of word processing history.

There’s a kind of paradox at the heart of this. As you know, we’ve all lost files, or had important stuff disappear into the [digital] ether, so there’s all that volatility and fragility we associate with the computer. But it’s also a remarkably resilient medium. And there are some writers who are using the actual track-changes feature or some other kind of versioning system to preserve their own literary manuscripts literally keystroke by keystroke.

Meyer: Who are they?

Kirschenbaum: One of them is a guy named Max Barry. He’s an Australian novelist, and he uses a system that was originally developed for software engineering to record all of his drafts. You can go to his website and access them there. To go back to the Eliot example, the Wasteland manuscripts survive, and they’re a wonderful source for students and scholars, but there’s only so much depth and granularity we can see in the revisions there. What if we had had Eliot’s composition of the poem one keystroke at a time?

Meyer: Did you find that poets were typing their work?

Kirschenbaum: I did. I didn’t notice any real distinction along those kind of genre lines. There were a number of poets who were early adopters at pretty much the same time as the novel writers. What I did find was more of what you touched on earlier, which was a distinction between writers of popular fiction or genre fiction on the one hand, and highbrow or belles-lettristic writing on the other.

One of the genres that I look at pretty closely is science fiction. It shouldn’t come as a surprise that the science-fiction writers collectively, as a community, were amongst the early adopters. But as it turns out, it’s not for the reason you might expect—like, oh, science fiction, of course these are really tech-y people, they’re all going to get a computer and love it as soon as they can. Rather, the reason why science-fiction writers were among the first to come to word processing is that it had such an impact on the material work of writing a novel that writers who used a word processor found that they were more productive. And as several of my interviewees told me, to make a living as a science fiction novelist in the late 1970s, early ’80s, you needed to be able to do two or three books a year. That was sort of the threshold for being able to do it for a living. I suspect there were similar dynamics in romance fiction and other popular genres.

Meyer: You talk a little bit about looking at different paths for word processing after Word. You go into “austerityware,” which is your phrase for software like WriteRoom, which tries to cut down on distractions. Is there any prognosticating you feel like you could do about what’s catching on next?

Kirschenbaum: I do think we’re seeing this interesting return to what instructors of writing for a long time have called free writing, which is just about the uninhibited process of getting stuff out there, doing that sort of initial quick and dirty draft. What’s interesting to me is that there are now particular tools and platforms that are emerging with that precise model of writing in mind.

The one that’s gotten the most attention is the one I write about at the end of the book. At the time I was writing, it was called the Hemingwrite, but now it’s called Freewrite. It’s essentially a very lightweight, very portable keyboard, with a small screen and portable memory. It reminds me of the way a lot of writers talk about their fountain pens—these exquisitely crafted and engineers fine instruments for writing. The Freewrite aspires to bring that same level of craft and deliberation to the fabrication of a purpose-built writing instrument.

So, you know, in a sense, I think we’re going to see more and more of those special-purpose writing platforms. I think writing might move away from the general-purpose computer—we’ll still do lots of writing of all sorts at our regular laptop, but it might be your email, your social media. For dedicated long-form writing, I think there may be more and more alternatives.

Meyer: Do you think there will be a bifurcation between tools for writing for pleasure or edification, and tools for writing for utility?

Kirschenbaum: I don’t know if it will be a bifurcation per se. I think there are the realities of different markets. Certain kinds of writing associated with the workplace is in turn going to be bound up with the way that much larger question of how it gets disseminated throughout the workplace environment. At work, a device like the Freewrite probably isn’t going to have much of a footprint. But I do think for individuals, there will be more choices at our disposal.

In some ways, I think our situation might end up resembling those of writers back in the early 1980s who had dozens and dozens of different options at their disposal, to the level of producing some anxiety, and who really had to think through what they wanted out of their word processing software. Maybe, hopefully, without some of that attendant anxiety.

We will really be able to think through what kind of writing instrument, writing platform, we want to use, and whether it should be the one software package that’s dominated the market for the last couple decades.

Meyer: How did you write this book?

Kirschenbaum: I used that one piece of software that’s dominated the market for the last couple of decades. [laughs] There’s nothing interesting about how I wrote the book—I did it with Word and a Thinkpad.

Meyer: In the book, you mention how Isaac Asimov didn’t know where the computer should go in the house. Could you retell that story?

Kirschenbaum: Isaac Asimov, of course, is one of the grand patriarchs of science fiction. By the time he gets his first computer, he’s close to about 70 years of age. It’s 1983—which is actually a little bit late already in the adoption curve. He gets an offer from Popular Computing magazine to write a series of articles documenting his coming to grips with the word processor.

And what’s interesting is that, as a result of those stories that he wrote, we have quite a lot of particular detail about how that unfolded. What Asimov describes, for example—and I do think this is revelatory and resonant—when the computer, which is a TRS-80, is originally delivered to his apartment, his immediate instinct is to put it in the back bedroom, sort of out of the way. But his wife insists that it goes into the living room, where it’s unavoidable.

At the time, computers often demanded their own space in the household. The family computer room was its own phenomenon. One of the interesting things once Asimov overcomes his reluctance to use the computer is that he discovers he’s sitting in the living room working at it, he can actually listen to and watch TV, because he can hear the TV over the keys. That’s something that wouldn’t have been possible with a manual typewriter, because it was so noisy.

I single that detail out because it’s one of those tiny, resonant details that captures the kinds of disturbances in a domestic space that introducing a new kind of media into that space resulted in. The computer reshaped the environment around it and other kinds of habits that we wouldn’t associate with the act of writing.

Meyer: I’d bet that the role of a computer in the home still feels up in the air and kind of permanently undecided to a lot of people. I mean, there’s people who just carry their laptops around their house. It’s what they work on, but it’s also what they watch Netflix on.

Kirschenbaum: There’s a picture in the book I really like of Eve Kosofsky Sedgwick, the famous scholar, who has at the time what were known as a “luggable” computer—as opposed to a “portable” computer. These were machines like the Osborne and Kaypro. They weighed in at around 25 pounds. The picture shows her sitting in a recliner, with the chair tilted back. She has the keyboard on her lap, and the rest of the unit is on a small table next to the recliner. It captures the way that the places in which one could write with a computer were already changing—we were already moving away from the formal posture of the computer at the desk, and us sitting toward it. Of course, nowadays we contort ourselves into all kinds of positions and postures as we sit with our tablet and our laptop.

Meyer: You don’t advance many, or any, theories about how the word processor led to any kind of stylistic change. I wanted to give you the opportunity to let you go further than that. Was there any change you think you observed?

Kirschenbaum: Well, I think that work could be done, but it wasn’t my work. I imagine somebody who was interested—particularly nowadays with what we popularly call digital humanities—somebody who’s interested in a particular writer, assembling their corpus. We have great text analysis tools that allow you to do all kinds of stylistic analysis, and someone could certainly do it.

My real question would be, not so much could you do it, but: What would it tell you? If you were able to ascertain that so-and-so’s use of adjectives either proliferated or diminished after their word processor adoption, well then what? What comes after that?

There was a recent study in The Atlantic, actually, where a couple of my colleagues, lit professors elsewhere, did a study of people who had been through MFA programs. They wanted to know if the MFA made a stylistic difference, and the upshot was no, or at least not in any way they could isolate. And the conclusion they came to, rightly, was that the MFA was much more about a process of professional socialization than it was about actually encroaching on any stylistic tendencies they had. And that seems right to me.

My view on this is that writing is a very multi-dimensional, complex kind of activity. There are lots of different things that account for what we abstract as a writer’s style. A computer, or whatever the instrument of composition is, can certainly be one of them, but so can all kinds of other things going on in a writer’s life, the writer’s biography, his or her physical environment, the circumstances under which the book is being written, the audience for which it’s going to be written—I don’t believe in the project of isolating any one variable, and certainly not a technological variable, as the causal agent in a process that’s as rich and complex as that one is.

Meyer: Where would someone who wanted to research this even start?

Kirschenbaum: You would want a writer precisely like Asimov, where there’s sufficient documentation that we can really—I mean, in Asimov’s case, it’s quite literally that we can pinpoint to the day when he gets his computer. Then you would assemble your corpus. You would want digitized copies of a lot of the books before and a lot of the books after, and then you would see what you could find with your algorithms.

I think much more interesting than doing that for one individual writer would be the prospect for doing it for a community of writers. So, science fiction—which we were talking about earlier—was a relatively small community at that time. There were a few hundred people probably making their living actively writing science fiction in the early 1980s, and a lot of them knew one another. They were friends, they were rivals, but they all sort of knew each other and knew what they were up to. Reading other people’s work, collaborating—when you have a community with those coherent social relations, introducing the question of technological disruption, as we now say, of course that would be interesting to look at.

Meyer: One thing I love about the book are all the office pictures—the pictures from ’80s offices, especially. There is a sense looking at the images that the desks are retrofitted writers’s desks, rather than the kind of generic surface-with-a-laptop setup that I think a lot of people work at now.

Kirschenbaum: The visual history of all of this is really interesting. One of the hard things was trying to figure out, what is a literary history of word processing, how do you go about researching it? Maybe by going to the archives, but you also do it by looking at the way in which computers really were represented in the kind of imagery I was looking at earlier. You look at the old office photographs. You see a picture of Amy Tan sitting with a laptop and you try to figure out what kind of laptop it is, and lastly you do it by talking to people. It was the oral histories I did that were the best research for the book.