One set of difficulties derives from the fact that we are quite properly unwilling to settle for merely using digital technology as a way of reproducing what we used to do in print. If all that were at stake were the equivalent of typesetting and distribution of fixed editions (as is the case with works that were born in print and then scanned), digital technology could be an almost unmitigated boon. But, as a quick glance at the CV’s of many of the attendees at this symposium will demonstrate, scholars will not settle for mere reproduction of the print world. Rather, they want to produce and use works that have rich sets of links and specialized tools that allow them to do new and exciting things. The Rotunda Press’s electronic publication of Melville’s Typee is a fine example. The reader is given access to many editions, both in manuscript and encoded text, along with rich editorial commentary. This work could not be fully produced as traditional print.2

The desire of scholars to produce and use works that can only be produced digitally poses both technical and economic problems for academic libraries and for publishers who produce works that are of interest to academic libraries. The essential functionality of digital scholarship is encoded in particular formats and markup languages that are likely to become obsolete over time. As newer works embody newer gadgets, the expertise needed to provide continued reliable access to older works often becomes hard to find. All of this can be dealt with, but only with costly attention and costly intervention.3

Even when the technical problems involved in preserving digital works are relatively straightforward, libraries and publishers (whichever is doing the direct provision) have to deal with the fact that digital works require active effort on the part of provisioners in a way that is quite different from print. The Founding Fathers Papers and Founding Documents published by Rotunda provide good examples here. The papers are in XML, which will last a while and which will surely have a good migration path in the future, so the issue I have raised above is relatively unimportant. But no matter how stable the format, it costs money—power, hardware replacement, staffing the help desk—to make the works accessible every day. There is nothing very fancy about this—it’s actually analogous (if I may use that word) to the cost associated with keeping books on a shelf. But, unlike the cost of keeping books on a shelf, there is a continuous set of production activities and associated outlays of money that can be attributed to each digital publication. When a major research library “purchases” the electronic version of George Washington’s papers, it shells out several thousand dollars and it commits to paying an annual maintenance fee to cover operational costs and improvements.4

One could arrange things differently, and make the publisher liable for the annual maintenance (in which case the one-time purchase cost would necessarily be much higher) but the simple fact that the electronic resource requires ongoing annual (actually, daily and hourly) cash outlay in order to be usable means the library/publisher nexus faces a set of problems that did not exist in the print age. These problems derive much more from changes in the ways libraries and publishers do business than from increases in the cost of doing business.

Note that in the journal world we have made the adjustment to ongoing annual cost by folding those costs into annual subscriptions. This was facilitated, I suspect, by the fact that serials have always been paid for on a subscription basis. The radical change that has taken place in the digital environment is that electronic versions of serials—including the backfiles in most cases—are now hosted on the publishers’ servers. Somehow librarians allowed the associated loss of control over the scholarly record to happen without all that much fuss, perhaps because there was no immediate effect on cash flow (the annual subscription model was already in place) and perhaps because the reliability of JSTOR and the development of Portico, LOCKSS and CLOCKSS give some confidence that the journal-based scholarly record will continue to be available. Academic libraries rail against the cost of electronic journals, and rightly so. But it is principally cost, and not either the frequency of payment (annual subscriptions are fine, they are just too pricey) or the loss of physical control that causes most of the wailing and gnashing of teeth.

Libraries are much less comfortable about annual fees to maintain electronic books—that is, to maintain scholarly works with content that is at least reasonably stable.5 From an economic point of view, this aversion makes little sense, because ongoing maintenance is every bit as important and as costly for print books as for electronic books. When a research library acquires a print book, it acquires the responsibility to care for it and make the content available indefinitely. Preservation and access are bundled into one technology and there is no practical way to separate them. But the cost of future access to the print resource is large—for conventional patterns of use and storage it is of the same order of magnitude as the cost of the book itself.6 Space has to be maintained and eventually replaced, buildings have to be heated and cooled, books are shelved, re-shelved, bound, etc.

The lifetime cost of preserving an electronic record with equivalent content to a print book is much less. The electronic version is cheaper to access and “re-shelve,” takes up much less space, no more power (servers are expensive to run, but heating and cooling large buildings is even more expensive). But the time pattern of outlays to support the housing of printed books is episodic, the annual costs are folded into the general cost of running the library, and, perhaps most important, these costs are not new—libraries have learned to absorb them. Although librarians and provosts pay attention to the cost of housing their collections, they tend to view new print acquisitions as assets rather than liabilities, and there is rarely explicit consideration of the ongoing stream of future expenditures that come with a new print book. An electronic book, on the other hand, makes those future costs immediately visible, and thus looks relatively more expensive even though it is actually cheaper. Kaiserlian’s observation that someone has to be concerned about the implications of perpetual stewardship is just as true for print as for electronic resources. The difference is that with print “someone” is always the library, and the implications of perpetual stewardship, however financially unpleasant, are familiar and are already, at least implicitly, in the budget.7

At several points in her paper, Kaiserlian notes that it is difficult to get academic journals to review the electronic editions of scholarly works published by Rotunda Press. Sadly, her lament is not unique to Rotunda. The best explanation for this sort of behavior that I have heard is that in the print world the reviewer gets to keep a copy of the book, whereas in the digital world there is no similarly valued artifact that can be used in lieu of an appropriate bribe or honorarium. If this is the reason, perhaps we should give vouchers good for a box full of printed works from the relevant press, or perhaps gift certificates from a museum shop. If the reason is instead some notion that because things are digital they are less scholarly than if not, then some set of mechanisms should be developed to hold the journal editors up to public ridicule.

No good purpose is served when academics use arbitrary screening devices to determine what is worth reviewing and what is not.8 In the case at hand, the works carry the imprimatur of a distinguished press, and the associated peer review, so the refusal of journals to review digital products makes no sense at all. Of course, it may be the case that a traditional review in a print journal cannot do justice to the electronic work, but if that is the case, the print journals could create quite simple websites with the requisite functionality, and explain to their readers how to use them. (The M-Publishing Group at the University of Michigan Library would be glad to help.)

The documents that underlie the vast majority of Rotunda Press’s publications (both historical and literary) are squarely in the public domain, and there is a powerful argument for a freely available, indexed, searchable version of those documents, especially in light of the fact that in many cases federal money was used to help to support the works in the first place. Recently, the Office of Science and Technology Policy asked for comment on proposals that works supported by federal money should be made openly accessible. The emphasis of the OSTP query and related efforts has been on the scientific journal literature, but exactly the same public policy case can be made for the humanistic book literature.9 To its credit, Rotunda is working with federal agencies to create such a resource for the founding documents, but in the meantime, with some exceptions, these out-of-copyright works are behind a pay wall, and a pretty high one at that.

The logic of open access for scholarly work extends well beyond works in the public domain and works funded by government, and is an instance of a standard economic argument regarding the production and pricing of public goods.10 The best statement of this argument as applied to the products of scholarship was made by the builder of the Rotunda himself, Thomas Jefferson. It is often quoted, and yet still bears repeating:

If nature has made any one thing less susceptible than all others of exclusive property, it is the action of the thinking power called an idea, which an individual may exclusively possess as long as he keeps it to himself; but the moment it is divulged, it forces itself into the possession of every one, and the receiver cannot dispossess himself of it. Its peculiar character, too, is that no one possesses the less, because every other possesses the whole of it. He who receives an idea from me, receives instruction himself without lessening mine; as he who lights his taper at mine, receives light without darkening me.11

Which is to say, the ideal price for scholarly work is zero. Of course, there remains the minor detail of paying to create and produce the work. This is not the place to take on the economics of scholarly publishing, but I note that the money now spent by academic institutions and their supporters is sufficient to provide for the production and distribution of the scholarship currently published. It would thus be technically and economically possible to organize a system in which the costs of publication and distribution were prepaid, presumably from library budgets, and the works were then distributed free of charge. The net effect on library budgets would be either zero or somewhat positive. But I digress.12

The problem faced by Rotunda Press in providing perpetual stewardship is exemplary of a problem faced by all digital scholarship. Analyzed carefully, the differences between perpetual stewardship of print works and electronic ones are mostly (but not entirely) illusory. Stewardship requires the provision of ongoing resources, and it is expensive. Because libraries have much more experience with this kind of stewardship—with taking responsibility for providing permanent access to the published scholarly record and associated source material, I am persuaded (and would like to persuade readers of these comments) that academic libraries are best suited to take on the task of preservation of electronic publications—be they relatively simple or complicated—produced by academic publishers. Rotunda Press, which has done an excellent job, is far too small to exhaust scale economies associated with digital preservation. For this reason and for reasons of branding and visibility, they have limited their range of content to a few areas.

The separation of stewardship from direct provision of access adds an unnecessary and complicating layer to the ecosystem of scholarly publishing. As an alternative, the Press could be a quasi-independent element of an academic library, responsible for its own editorial functions, but relying on the library for provision of the perpetual stewardship that electronic publication requires, and using the library “brand” to advertise its ability to provide such stewardship. Such an arrangement would be efficient in the simplest sense: neither the library nor the publisher would be required to learn how to do things that are not already part of its natural compass of activity.

Let publishers make things public and let librarians preserve them and assure permanent access. I’m a great believer in market transactions and division of labor. But taking the preservation function from the library and giving it to the publisher, who then must preserve and make accessible works that are licensed back to libraries, is not efficient. Rather, it is an unexamined natural consequence of extending a print model to digital technologies, and it would behoove us to consider rearranging our institutions to the job more reliably, less expensively, and less painfully, and with explicit recognition (which the libraries have, and the publishers often do not) of the broad public and scholarly value derived from digital scholarship. The result should be more and better books.