Computing, Humanism, and the Coming Age of Print

Stuart Moulthrop
School of Communications Design
University of Baltimore
Seminar: Is Humanities Computing an Academic Discipline?
The University of Virginia
December 3, 1999

On the assumption that no academic enterprise can call itself a discipline until it learns to value inquiry above certainty, I wish to meet a question with a question. This seminar began by asking: Is humanities computing an academic discipline? To which I can only respond: What else could it be? Though this reply may seem suspiciously ironic, I mean it in simple sincerity, not as wisecrack but as the dialectical opposite, an (apparently) foolish question.

The preceding papers in this seminar, which I have been privileged to read before writing, argue compellingly for the growing importance of humanities computing within the academy. Yet our (federated) (inter)discipline -- humanities computing, multimedia, humanistic informatics, or whatever we choose to call it -- makes its appearance in interesting times. As Willard McCarty notes, universities "are now reconfiguring themselves to adapt to changing conditions of societies that are no longer certain of what they're for" ("Interdiscipline") -- an uncertainty that perhaps applies as much to societies as to universities. Conditions of textual production and reception are changing in profound ways throughout the culture, not just within the scholastic world, and while the concept of academic discipline confers the benefit of focus it may also mask important patterns of connection and consequence. To ask a less foolish question: what might we mean by humanities computing if we looked oustide the narrowly institutional context? Though I do not propose that we shed our academic identity, I suggest we think about points of contact between our discipline and a rapidly changing information culture.

Geoff Rockwell has introduced the Shandean theme and I follow him gladly in this as in other things. The paper proceeds elliptically. Its first two sections comprise a digression about the immediate future of book publishing. The next section offers a thought experiment based on these projections, followed by some notes toward a broader conception of humanities computing as a discipline poised between academy and marketplace.

First, however, a commercial.

1. After Amazon

Friends, are you tired of paying inflated prices for books because some TV-executive-turned-editor bet the farm on the wrong celebrity bio? Confused about which of five multinational media combines owns your favorite imprint this week? Wondering why you need to pay for 300 pages when you only intend to read 30? Fed up with Fed Ex charges for shipping your books to Boise or Brisbane? Frustrated because the titles you most want to read or share with others are no longer in print?

At the moment, dear reader, you are out of luck, but moments pass quickly these days. We are living through the first great explosion of electronic commerce, and while the short-term economics of the outburst may be obscene ("irrational exuberance" is a poor euphemism), its cultural consequences seem more promising, or at least less likely to disappear in the inevitable crash. This holds particularly true for information-intensive industries such as publishing. has spun off into what it hopes are richer markets, but before it turned to toys, cosmetics, and power tools, Amazon assaulted a major pillar of the publishing industry: the physical point of sale. The resulting shock was felt around the business world and generally (mis)taken for the big bang of e-commerce. Most of us are now learning to live with the fallout. Many academics enjoy Amazon's innovations on a regular basis, even if we may not fully grasp their broader implications; but such innocence cannot last forever. The information market continues to evolve. As far as information businesses are concerned, Amazon was only the first stage of a more radical process.

Pillars once uprooted have a way of crashing domino-style into other megalithic assumptions, in this case the belief that a book must be a tangible object of a certain size and mass. In a fundamental sense, today's books begin as bits. Most print works produced since the late 1970s have existed initially in electronic form, as typesetters' data tapes if not word processor files. What we call publishing today is in large measure the business of converting these bits into print on paper, transferring that material structure to consumers, and making money in the process. There is great concern lately about the fate of this enterprise. Much polemical ink has been shed by those who insist that books not be reduced to bits, an argument that often devolves into claims about sovereign authorship, reading in the bathtub, and old jam stains (see most recently Gass) but generally misses the crucial point.

Print is an undeniably useful, felicitous, and irreplaceable medium. That oldtime literacy has much to recommend it, especially in an age of digital simulation. Programming, we might remind ourselves, still depends on alphanumeric syntax and sophisticated patterns of structure and reference. Hypertextuality is not an enemy of traditional forms but their natural ally: witness Amazon. No one is about to launch a paperless society. Yet on the other hand, after half a millennium of refinement the printing press remains a cumbersome and inconvenient appliance ill suited to personal use. Not so the laser printer. Now that bookselling has been separated from the point of sale, it seems logical to extend our critical scrutiny to the stage of material production. Why do we still mass-produce books when we can print on demand?

Books that begin as bits need not remain in that state, but there may be good reason to keep them in digital form longer than we do at present. Even after Amazon, most book purchases still involve shipment of a prefabricated item from a warehouse via postal or courier service. Given existing technology, however, the information encoded in that bulky object could be transmitted in seconds either directly to the consumer or to a vendor-manufacturer who would (in every sense) make up the volume. Printing might be undertaken in a number of settings and styles, from output on a personal printer (as we may do now for software manuals and the like), to inexpensive printing and binding at a local copy shop (as in course packets), to high-quality artisanal production for books of special importance, with perhaps even a place for conventional press runs of titles with genuinely wide appeal. Mass-market chain bookstores would probably vanish from this new media ecology, but they could be replaced by small businesses combining the fine old art of bookselling (serving as informed guide to the literary world) with the estimable craft of printing and binding.

2. The coming age of print

Given these changes, the outlook for existing publishers seems challenging at best -- a fact only partly appreciated in the business. At least one major firm (Random House) seems to be betting against print-on-demand, if its decision to build the world's largest book warehouse in western Maryland is any indication (O'Brien). Conversely, Barnes and Noble has taken a 49% stake in, the most significant print-to-order publishing venture currently in operation (New York Times). The latter may be the smarter move, offering some hope that publishing will adapt rather than perish through backwardness. Print-on-demand need not spell doom for the industry. Gatekeeping and quality assurance have more to do with words in themselves than words on paper. Even a radically reorganized market would still need the services of acquisition, development, and production editors, not to mention permissions specialists, publicists, and managers.

Yet if our myths about the end of the Cold War are any guide, economies based on centralized production diverge significantly from those predicated on consumer demand. These differences could make a difference not just for publishing but for the literary culture of the next century. I cannot discuss these implications fully within such a limited digression, but a few key points need mentioning.

First, there may be bad news for the publishers of Books in Print unless they too are prepared to shift bitwise. A print-on-demand publishing model would likely expand the number of available titles by a very large factor, especially if selection continues to attract customers, as it does for Amazon and its competitors, and as the profit axis shifts from mass to niche markets. Though some titles would still become unavailable (the term out of print seems about as useful here as horseless carriage), this outcome would probably be accidental rather than programmatic -- and much more easily reversible. There is no overwhelming reason why a print-on-demand market should depend on artificial scarcity. Publishers would have an economic incentive to keep titles available and much less substantial reasons to withhold them. Authors might assert their right to republish personally any properties not kept in circulation, suggesting some interesting shifts in the relationship of writers and publishers.

This grand expansion of the catalog would have many implications. Publications might generate revenue far longer than they do now, allowing authors to publish fewer, better books (or to profit more highly from a line of mediocrities). The number of people able to sustain themselves by writing might actually increase, unthinkable as that seems. On the other hand, competition for readers' attention would certainly intensify, suggesting that conglomeratized publishers might turn at least initially to mass-market promotion and celebrity culture as checks on market diffusion. Market changes would not come without resistance and even transient reversals. In the long run, though, it is probably impossible to confine readers to a short bestseller list or a thin catalog as eligible alternatives mount into hundreds of millions. Sooner or later, today's mass-market publishing paradigm seems sure to collapse.

In light of these prospects, statements like the following seem both poignant and revealing:

"I don't really see myself doing anything dramatically different in changing the way that St. Martin's publishes and acquires books," Mr. Witte said. "It all starts with the books that you buy and I want to be as accessible to the editors as I can be." (Carvajal C17).

The source of these remarks is George Witte, recently named editor in chief at St. Martin's Press after his predecessor resigned to protest publication of an infamous political biography. Under the circumstances Mr. Witte's endorsement of the status quo seems inevitable -- this is damage control, after all -- but there is an interesting resonance in his second-person phrase, "the books that you buy." At the moment this "you" does not include you or me unless one of us happens to work in New York publishing. The present system is centrally planned and controlled. Bestseller lists may measure selected retail volume, but retail operations are heavily circumscribed by wholesale decisions. There is no regular worst-seller list based on tonnage turned to pulp. To be sure, even a demand-side market must depend on editorial decisions; but since those choices would no longer commit millions of dollars for printing, warehousing, and distribution, the influence of another "you" -- the person who pays for the right to print one copy -- should rise in importance. This shift of emphasis could eventually reconstitute book culture as well as the print market. It probably has significant implications for copyright, textual structure, and genre as well as for the profession of writing. You may be wondering what any of this has to do with computing in the humanities; but that is the point.

3. Computing? Humanities?

I offer these speculations without any warranty of predictive value. Even if my suggestions prove more visionary than prophetic, however, they should have some interest as foundations of a thought experiment. Consider this question: Does print-on-demand publishing count as humanities computing? Construe the second term any way you like -- as text-based analysis, as interdisciplinary methodology, as theory and practice of digital multimedia, or as some new, hybrid science called humanistic informatics. Would the shifts in the literary market I have outlined hold any significance for these ventures? Why or why not?

On first consideration, why not may seem obvious: because the delivery of printable texts via the Internet does not appear to use digital resources in any technologically ambitious or progressive way. On-demand publishing may seem orthogonal not just to humanities computing but to computing in general, at least if we associate that subject with something other than mundane utility. From a scholarly perspective, the data processing needed to bitstream books seems about as trivial as what goes on in car engines or microwave ovens. It is hard to imagine a research course in household computing (these days even at MIT).

This dismissal defines computing far too narrowly, though. Those who know the difficulty of designing markup languages, databases, and encoding protocols might make a strong contrary case. Even in purely instrumental applications, digital technology is never as simple as non-specialists imagine. Beyond this, the transformations I suggest are not simply mechanical. Publishing on demand implies more than just moving texts through the wires, just as retailing involves more than transfer of goods at point of sale. Demand presumes selection which in turn implies browsing. Before a title can be delivered a prospective reader needs to locate, identify, and examine it. My scenario applies not simply to traditional texts (books) but to a vast meta-text that we presently call by many names: the library, the archive, the catalog, the set of books in print. When that meta-text comprises hundreds of millions of items, the representation of options will be far from trivial. Solutions for this problem will require considerable sophistication. On some level they will still involve computing in the original sense (simple mathematics) but it seems better to separate these basic functions from computing of a different sort. Sherry Turkle suggests an important ground for distinction:

The personal computer culture began with small machines that captured a post-1960s utopian vision of transparent understanding. Today, the personal computer culture's most compelling objects give people a way to think concretely about an identity crisis. In simulation, identity can be fluid and multiple, a signifier no longer clearly points to a thing that is signified, and understanding is less likely to proceed through analysis than by navigation through virtual space. (49)

Turkle's remark signals a momentous shift in perspective, stemming from her reading of the Internet as the postmodern form of information technology. In Turkle's view navigation succeeds analysis as primary mode of signification in spatialized, social information systems. After the Internet we understand computing differently; and indeed this "we" includes some virtually present company. Though he writes in a very different vein, Willard McCarty echoes Turkle when he warns of the problematic tendency toward disambiguation implicit in purely analytical uses of literary computing ("Interdiscipline"). Analysis does not close off the space of alternative possibilities. Likewise, by introducing "ergodics" or path work as a principle of textual expression, Espen Aarseth reveals one of the ways in which navigation may constitute a form of understanding (5).

If there is any significant connection between the changes I anticipate in the literary market and what we do in humanities computing, it will grow out of insights like Turkle's, McCarty's, and Aarseth's that treat computing as a complex signifying practice. We must be prepared to see beyond "transparent understanding," and take up problems of simulaton, navigation, and other challenges of "virtual space" as primary problems for humanities computing.

The work might come naturally. A skeptical, ambiguating approach has long been characteristic of the humanities, even before the poststructuralist turn from which Turkle draws her inspiration; but one may legitimately wonder what this way of thinking has to do with market changes and business models, which after all depend on that disambiguating practice called bookkeeping. Perhaps publishing on demand qualifies as an interesting application of computing, but that interest may seem better situated in a school of business or department of social science than in any branch of the humanities. Historians, philosophers, and literary scholars might benefit considerably from a new book market, but they may be no more concerned about the details of its implementation than they are about present online bookselling. Humanists care deeply about books, of course, but the books in question come mainly from research libraries and academic presses, neither of which (ostensibly) has much connection with retail publishing. Why bother about the marketplace?

Within a certain narrowly defined, technology-averse definition of the humanities, this objection may be unassailable; but for humanities computing it seems very hard to sustain. Consider Willard McCarty's earlier attempt to define our enterprise, introduced by John Unsworth at the start of these discussions:

Humanities computing is an academic field concerned with the application of computing tools to arts and humanities data or to their use in the creation of these data. ("What Is Humanities Computing?")

The "data" of the humanities are for the most part written records -- books, articles, periodicals, broadsides, manuscripts, notebooks, theses, musical scores, pamplets, diaries, manifestoes, et cetera. By and large we find these objects in the library or archive, as relics of the past. But we cannot fix our attention solely on what has gone before. Data is past participle, that which is given, but in the humanities we tend not to accept the given without skepticism or inquiry. We interrogate sources, we argue derivations, we construct genealogies: we navigate as well as analyze. We are intimately concerned with the way discourses are framed and articulated, socially and materially.

This tendency to look beyond initial premises into the origins of discourse underlies the latter half of McCarty's definition: the use of computing tools "in the creation of these data." This phrase reminds us that humanities computing is not simply a retrospective undertaking but one that looks forward as well. We need to care about new books as well as old, and thus by extension about innovative ways of circulating forms of writing. It remains to be seen, however, what such prospective engagement might mean in practical terms and how this engagement affects our disciplinary schemes.

4. Academy and market

Because this brief sketch can only accomodate limited discussion, I apologize in advance for omissions, including one that is both categorical and deliberate. I intend to overlook an area in which I have an obvious interest and which several contributors to this seminar have singled out: experimental and creative work in electronic media, including hypermedia narrative, cybertext, virtual architecture, and other forms of digital art. Work of this sort may contribute significantly to humanities computing and I would not exclude it from our schemes, but it seems best left aside if our main concern lies with the traditional literary marketplace.

Theory, on the other hand, needs more immediate discussion. I have already mentioned Turkle, both of whose major studies (The Second Self as well as Life on the Screen) seem excellent models for our theoretical agenda and are both quite notable for their awareness of the Internet as a social and commercial phenomenon. Similarly important work has been done recently by Janet Murray, N. Katherine Hayles, Donna Haraway, and Roseanne Stone. All these figures have had or should have enlightening things to say about literature, publishing, and emerging information spaces. The fact that all are women is both unsurprising and indicative. Turkle's emphases on bricolage, emergent forms, negotiations of identity, and social applications of technology suggest that humanities computing could be a strong alternative to more rigidly defined, instrumental approaches that alienate many women (and men) from other forms of computing.

It also seems noteworthy that only one of the theorists mentioned above (Hayles) is primarily a literary critic. As Espen Aarseth suggests, the attempt to derive theories of new media directly from literary poststructuralism appears to have reached its limit ("Humanistic Informatics"). Though I think our field should address itself to the immediate future of the book, we need theoretical approaches that no longer struggle primarily against the hegemony of writing, and especially of print -- a standard by which some of my own earlier work may be found wanting. Jay Bolter and Richard Grusin's concept of "remediation," focusing on the interaction of coexisting technologies, seems potentially more valuable in this regard. However, any kind of theoretical abstraction risks losing touch with material and social phenomena, as perhaps in this case:

What remains strong in our culture today is the conviction that technology itself progresses through reform: that technology reforms itself. In our terms, new technologies of representation proceed by reforming or remediating earlier ones, while earlier technologies are struggling to maintain their legitimacy by remediating newer ones. (Bolter and Grusin, 61)

Simply to equate technological interaction with "reform" skates perilously close to meliorism and glosses over the frictions that arise between institutions in the process of change. The visionary strain may serve better in thought experiments than in media theory. The transformation from mass-market publishing to publishing on demand might indeed improve upon the present system, but one wonders if champions of the status quo like Mr. Witte of St. Martin's will share this view, especially if near-term profits are compromised. "Reform" is an inherently contentious idea, not an abstract good. Change seems likely to create as many wounds as it remedies or heals.

My quibbling aside, remediation theory may shed much light on the coming developments in the literary marketplace. I think however we will also want accounts that emphasize imperfection, resistance, and even failure. In this respect McCarty's notion of "modelling" seems a better frame of reference:

The point of modelling is not to establish the truth directly, since models are never true; it is to achieve failure so as to raise and point the question of how we know what we know.... ("Interdiscipline")

Strictly speaking, of course, McCarty's statement is not theoretical, at least not in Bolter and Grusin's sense. By "modelling" he means not abstract representations of cultural forms but the implementation in analytical software of assumptions about specific acts of writing. I am taking his remark out of context for a particular reason: to illustrate how much better we may be served by theories that are strongly grounded in application. It follows, as Geoff Rockwell argues, that humanities computing must be as much a productive as a reflective discipline (Rockwell). We must not simply write about changes in publishing and literary culture, but must make material contributions to emerging practices.

In this respect the more pragmatic side of our textual computing ventures may represent at least the equal, if not the senior partner, of theory. Management of very large corpora, whether in commercial publishing or academic archives, will demand considerable development of current markup, editing, and presentation tools, most immediately SGML and XML. The Text Encoding Initiative and centers like HITC at the University of Bergen, STG at Brown, and IATH at the University of Virginia are obvious leaders in this regard, especially as they become more engaged in problems of direct relevance to the publishing industry.

The more pragmatic side of the hypertext community (not the literary wing which Aarseth so distrusts) also offers some useful examples. Jim Rosenberg's description of electronic document spaces in terms of transactions rather than static structures suggests important new avenues for the design of browsing systems (Rosenberg). Catherine Marshall's work on annotation opens up new ways of understanding the ways readers develop and represent links among texts (Marshall). Mark Bernstein's experiments with automated systems for building and analyzing link structures -- work sadly overlooked by humanists -- might have significant applications for very large catalog systems such as those implied by demand-centered publishing (Bernstein et al.).

This last example suggests perhaps the most exciting application of humanities computing -- in its oldest sense, by the way -- to the next age of print. With many millions of titles potentially available, readers and publishers will need new ways of describing and associating texts. The most significant challenge to browsing and navigation lies in the disappearance of the bookstore or library shelf. As Amazon and its competitors demonstrate, the first logical replacement for the shelf is some system of information retrieval -- e.g., Amazon's various devices for suggesting other titles the shopper might add to her cart. Presently these suggestions are based on sales data -- people who bought book X also bought book Y -- a fairly crude instrument for eliciting textual relationships. What if readers had access not simply to a retail database query but to full-text searching algorithms -- the sorts of sophisticated tools familiar to computing humanists? These systems might more usefully answer questions like: Who elses writes this way? Where did these ideas come from? and How can I find out more? The possibilities for such applications remind us that navigation and analysis are not mutually exclusive -- and most importantly, that McCarty's notion of empirical modelling may have direct relevance to the marketplace as well as the academy.

5. Conclusion

None of these speculations are meant to elide the real, enduring, and beneficial differences between scholarship and commerce -- I do not argue for I also need to acknowledge that much of what we believe as academics differs diametrically from business doctrine. At the risk of proving that this paper is simply a footnote to McCarty's (though worse things could happen), I turn to him yet again:

By nature humanities computing... challenges issues of ownership, which is to say, reveals that many [source materials] are held in common and there is much to be gained from sharing them. If its real potential is understood, humanities computing can be quite threatening to the status quo. ("Interdiscipline")

As the status quo evolves away from mass-market publishing -- and as distinctions between the commercial marketplace and university research libraries begin to blur, at least in terms of sheer volume -- we may find ourselves pressed to defend our ways of thinking and our differences from the commercial culture.

This task demands a very clear sense of the moment. We should I think maintain a balance between retrospect and prospect, especially at a time when the past in the form of book culture may be in for considerable renovation and re-formation, if not remediation and reform. As my defense of theory should indicate, we also need to maintain our capacity for detachment and generalization, faculties which work best when tempered by practical engagement. The possibilities for such engagement -- numerous already and likely to expand considerably in the next decades -- may represent the greatest innate strength of our undertakings. Whatever we call ourselves (and it seems wise to keep up several aliases), humanities computing/multimedia/hypermedia/humanistic informatics clearly comprises the core of a discipline. If we hew to our agendas, especially our commitment to practicing what we teach, we may prove in the long run to be the heart and future of the humanities.


Thanks to Mark Bernstein of Eastgate Systems for correcting my understanding of bestseller lists.

Works Cited

Aarseth, Espen. Cybertext: Perspectives on Ergodic Literature. Baltimore: Johns Hopkins UP, 1997.

Aarseth, Espen. "From Humanities Computing to Humanistic Informatics: Creating a Field of Our Own." Seminar: Is Humanities Computing an Academic Discipline? University of Virginia, October 22, 1999.

Bernstein, Mark, Jay David Bolter, Michael Joyce, and Elli Mylonas. "Architectures for Volatile Hypertext." Hypertext '91 Proceedings. San Antonio: Association for Computing Machinery, 1991. 243-60.

Bolter, Jay and Richard Grusin. Remediation: Understanding New Media. Cambridge: MIT Press, 1999.

Carvajal, Doreen. "An Insider is Named to Head St. Martin's." New York Times 22 November, 1999. C17.

Gass, William. "In Defense of the Book: On the Enduring Pleasures of Paper, Type, Page, and Ink." Harper's Magazine. 299 (November, 1999): 45-51.

Marshall, Catherine. "Toward an Ecology of Hypertext Annotation." Hypertext '98 Proceedings. Pittsburgh: Society for Computing Machinery, 1998. 30-39.

McCarty, Willard. "What is Humanities Computing? Toward a Definition of the Field"

McCarty, Willard. "Humanities Computing as Interdiscipline." Seminar: Is Humanities Computing an Academic Discipline? University of Virginia, November 5, 1999.

O'Brien, Dennis. "Rewriting the Chapter on Big." Baltimore Sun 31 October 1999: A1.

"Barnes and Noble Gets Stake in Web Publisher." New York Times 3 November 1999: C4.

Rockwell, Geoffrey. "Is Humanities Computing an Academic Discipline?" Seminar: Is Humanities Computing an Academic Discipline? University of Virginia, November 12, 1999.

Rosenberg, Jim. "The Structure of Hypertext Activity". Hypertext '96 Proceedings. Washington: Association for Computing Machinery, 1996. 22-30.

Turkle, Sherry. Life on the Screen: Identity in the Age of the Internet. New York: Simon and Schuster, 1995.