In a Washington Post interview a couple of years ago, Bill Gates discussed his plans to give away the bulk of his fortune and suggested he already had in mind doing with the personal computer "something like what Carnegie did with libraries where he said, 'Okay, books are this empowering thing that people . . . should have access to.' " That was presumably the impetus for the announcement in late 1997 of two grants programs for public libraries, one consisting of $200 million worth of software from the Microsoft Corporation, the other of $200 million from the personal fortunes of Bill and Melinda Gates, directed at providing digital technology and internet access to underserved libraries.
Not surprisingly, the announcements engendered a certain skepticism. The grants were described as a public relations ploy, and one moreover from which Gates and Microsoft stood to profit by seeding their technology in the library context. When the matter was raised in the Council of the American Library Association, one councillor described the grant as "a down payment on the purchase of public libraries." In the end, though, a more accommodating view prevailed, as the council voted to "recognize and celebrate Bill and Melinda Gates . . . for this visionary endeavor."
It's true that Gates's gifts were not as disinterested as those of Andrew Carnegie, who spent a substantial part of his fortune building almost 2,000 library buildings in the United States. But Gates is doubtless sincere in believing that networked computers can be a boon to libraries and education, which is after all a widely shared view and a cornerstone of the Clinton administration's education policy. And it's naive to suppose that he wouldn't try to use the gifts to improve his battered public image. Carnegie did the same thing, after all, and he had far more to live down—however questionable the means whereby Gates has enlarged his fortune, he has only figuratively strewn the ground with bodies.
But what's most striking about Gates's gift is that once again America's richest man should have chosen the public library as the object of his philanthropy, after almost a century in which the institution has more or less languished in the public consciousness, and at a moment when many people think the library has no future at all in the age of the Internet. Why libraries now?, we want to know; and by way of an answer it helps to ask, Why then?
The period between 1850 and the First World War was the golden age of the American public library. The number of public libraries went from around 50 in 1850, to 300 by 1875, to several thousand by the turn of the century. A lot of this growth was the direct result of Carnegie's largess, but he was responding to a very general conviction that libraries were essential institutions for social progress, to the point where he could say the public library "outranks any other one thing that a community can do to help its people." The library movement battened on the late-nineteenth-century ideology that saw literacy both as crucial for social advancement and as ensuring an enlightened civic discourse. As J. P. Quincy wrote in 1876, "[To the free library] we may hopefully look for the gradual deliverance of the people from the wiles of the rhetorician and stump orator. . . . As the varied intelligence which books can supply shall be more and more widely assimilated, the essential elements of every political and social question may be confidently submitted to that instructed common sense upon which the founders of our government relied."
The founders of the library movement envisioned the public library as an equal partner of the public school in achieving these goals. It was a time, after all, when schooling was more limited than it is today—in 1890 only a quarter of American students finished high school—and when the curriculum was mired in rote learning that had little relevance to the forms of literacy that reformers wanted to establish. (Charles Eliot estimated in 1890 that it would take a Massachusetts high school graduate only 46 hours to read aloud all of the books that were assigned in the last six years of schooling.) The public library, by contrast, seemed to offer a venue that was accessible to everyone, one that "appeals to and nurtures every idiosyncrasy," as one enthusiast put it. And as the libraries went up, they were staffed by cadres of "apostles of culture," as the historian Dee Garrison describes them, many of them graduated from the newly established library schools, the first of which was founded by Melvil Dewey (of Dewey decimal system fame) at Columbia University in 1887.
By any material standard, the American library movement of the late nineteenth century was a remarkable success, which left the nation in possession of a public library system that became a model for other nations. With its maturity, though, the public library came to be increasingly taken for granted. It settled into a respectable but decidedly second-tier role as a community institution, staffed by well-meaning but underpaid and largely female personnel—by 1920 around 90 percent of librarians were women, a higher proportion even than in teaching or social work. (For a telling indication of where the librarian figured in the American cultural scheme of things, think of the scene in Frank Capra's It's a Wonderful Life when George Bailey learns that if he had never been born his wife Mary would have been condemned to a life as a spinster librarian.)
Over the course of the twentieth century, the public library's role has been further circumscribed. Radio, movies, and TV made the library less essential as a source of entertainment and information, and libraries have not tried hard to compete in this domain: their collections are still dominated by books. And more recently, book superstores have further eroded the library's importance among its natural constituency of the middle class. As disposable income grows, people find buying books to be an attractive alternative to borrowing them, particularly since the retailers are in a position to offer an unlimited selection of recent titles and a more congenial environment for browsing and socializing.
None of this means that people have ceased to patronize the library: a recent survey funded by the W. K. Kellogg Foundation found that 68 percent of adults reported having gone to a public library at least once in the previous year, and 38 percent reported going more than five times. But over the years the library has clearly lost ground in the civic consciousness. States and municipalities pay lip service to the importance of libraries, but are apt to slight them in their budgets, and civic-minded people are increasingly likely now to make contributions to the local PBS station rather than to the Friends of the Library. As one librarian interviewed in the Kellogg Foundation survey put it: "Who's against libraries? Nobody's against them; it's just nobody much notices."
By all rights, the increasing availability of networked computers should diminish the public library's traditional role still further. A 1998 Commerce Department study found that 62 million people are using the Internet, and other estimates put the figure still higher. Most of these internet users are drawn from the group that public libraries have traditionally served—they are younger than other adults, have higher incomes, and are comprised of a high proportion of white-collar knowledge workers (45 percent, against a national proportion of 27 percent). These are people who already use the public library less often than their parents did for purposes of obtaining recreational and instructive reading. Now they no longer need to rely on the library even for the sorts of information they can't easily get on National Public Radio or at Barnes and Noble: a biographical article on Charles Babbage, the text of a bill, a map of Uruguay. They may still want to have a library around as an information source of last resort, but they have a number of more convenient options to exhaust before they are driven to use it.
But unlike the other sources of information and entertainment that have emerged in the twentieth century, networked computers have drawn attention to the library as at no time in the last hundred years. Behind this revival of interest is a new ideology of literacy, this one couched not in terms of the importance of books and reading but rather in terms of the more nebulous notion of "access to information." On the one hand, computer literacy is regarded as essential to social mobility; hence the warnings that a failure to provide universal access to the technology will lead to the creation of an economic and social gulf between information "haves" and "have-nots." At the same time, people argue that universal access to information is essential to robust civic debate; in the words of a recent report of the Benton Foundation, "the free flow of information to all who desire it, regardless of race, income, or other factors, is vital to the functioning of a free society."
The Gates gift is one highly publicized sign of the new interest in the library, but there are numerous others. The National Science Foundation has sponsored a digital libraries initiative that provides for more than $25 million on the first round of funding for a range of projects. Major libraries like the Library of Congress and the New York Public Library have initiated ambitious digital library projects, as have the national libraries of Britain, France, and Germany. Corporations like IBM have made the "digital library" a theme for new suites of applications. And the sleepy field of library science has been reinventing itself in new or revamped programs, sometimes under new names calculated to cast off the traditional—and it has to be said, feminized—stereotypes of the library profession. (At the University of California at Berkeley, the School of Library and Information Studies was closed down a few years ago and replaced after an interval by a new School of Information Management and Systems, with the old library school faculty complemented by an assortment of computer scientists, economists, political scientists, and law professors.)
There's a certain amount of voguishness in all this, but no more than there was in the library movement of the nineteenth century. The question is: When the current rush of enthusiasm has subsided, will we be left with an infrastructure and institutions as robust as those that emerged a century ago? And how much of a role will traditional libraries actually have to play in the informational new order?
The problem in talking about the future of the library is that it isn't always clear what people have in mind by the term. To many, the phrase "digital library" conjures up the picture of an ascension of the current library into disembodied electronic form. No one goes so far as to predict the end of the traditional library; at the least people reserve for it a role as an archive where we can preserve "legacy collections" of paper books and documents and as a safety net for the "information have-nots." But as an institution it is conspicuously absent in most of the scenarios that visionaries paint. Vice President Al Gore, for example, likes to talk about "my vision of a schoolchild in my home town of Carthage, Tennessee, being able to come home, turn on her computer, and plug into the Library of Congress" (you wonder whether Carthage doesn't have a nice public library of its own that she could go to, whose collection is more than adequate to the needs of anyone young enough to be described as a schoolchild). And IBM has run a commercial that depicts a Tuscan wine grower proudly explaining to his granddaughter that he had just gotten his degree remotely from Indiana University, thanks to the fact that the university had put its entire library online with help from IBM.
That IBM ad created some misunderstanding about the current state of library digitization, and in the end the Indiana University dean of libraries had to issue a clarifying statement explaining that the university libraries had not in fact been put online, though a portion of the music collection had been digitized in collaboration with IBM and would eventually be made available on the Web. Indeed, it's understandable that people should find these scenarios credible, not just because of the current climate of technological hype, but because scarcely a week goes by without an announcement of some project aimed at converting an existing library collection to digital form—those sound records at Indiana, a collection of nineteenth-century mathematics books at Cornell, the "American Memory" project of the Library of Congress, the billions of pages now stored at UMI, formerly University Microfilms. Newspapers, magazines, scholarly journals—all are coming online, and it's certain that electronic publication will soon become the chief means for disseminating scientific articles and information, given the high costs of print publication, the general availability of networked computers in the scientific community, and the importance of reducing publication lag time.
It's not surprising, then, that people should take the eclipse of the traditional library as not just inevitable but imminent, and that they should be disinclined to spend money to replace or expand current library facilities—why bother, when books will be obsolete any day now? The fact is, though, that it will be a very long time before most library collections are available online, and there are still basic issues to be resolved before any large-scale conversion can proceed. For one thing, there are at present no general standards for scanning, storage, or document formats for all this digitized content—standards that it will be no small matter to agree on in the decentralized world of the Internet, particularly at an international level and in a rapidly changing technological context. Then too, it isn't clear who will pay the costs of conversion. At current rates, it would cost around a billion dollars to digitize just the 17 million books in the Library of Congress, without taking into consideration the 95 million other documents in its collection or the hundreds of millions of books and documents from other collections that are not represented there; and there will be additional annual costs for storing and "refreshing" the collection to keep up with the deterioration and obsolescence of storage media. Un like the costs of processing and storage, moreover, the costs of conversion will not be dramatically reduced in coming years, since the work is necessarily labor-intensive. So we can expect that the process of conversion will take a very long time—indeed, it's a certainty that hundreds of thousands of "brittle books" and other fragile documents in library collections will completely decompose before anybody can get around to digitizing them.
Even when large portions of current library collections are available online, moreover, there will still be a need for books and for buildings and shelves to hold them. Granted, many of the current limitations of digital reading technology will be overcome in the near future: we'll have cheap, lightweight, high-resolution displays that can operate for long periods at low power, and perhaps even commercial versions of the thin flexible displays that people have referred to as "digital paper." But the very features of the printed book that count as limitations for purposes of storage and distribution—the fact that each text requires a fixed material support—make it irreplaceable as a medium for the sustained reading of complex texts. We tend to forget how much the process of reading a book is dependent on the physical manipulation of the volume and the sense of place it produces. (Reading Proust in a scroll window, I once suggested, is like viewing Normandy through a bombsight.)
It's a safe bet, then, that the book will remain the primary form for reading the sorts of works which are at the center of cultural life, and which make up the core of public library collections—works like novels, biographies, histories, or the more readerly periodicals. As collections come online, of course, we'll also have these works in a digital form that makes search and annotation easy, alongside of new works with multimedia, hypertext, and the rest which can only be ac cessed in digital form. And the books themselves may soon be produced and distributed in very different ways, now that networked printers make it possible to make high-resolution bound copies at any location of any of the hundreds of thousands of texts that will soon be available online. The public library in Carthage, Tennessee, can print out a facsimile first edition of James Agee, while the library of the University of Tennessee can obtain a printed copy of Alphonse de Lamartine's book on Gutenberg, downloaded from a server at the Bibliothèque Nationale de France.
Notwithstanding the vision of the disincarnated digital library, then, it's clear that local, brick-and-mortar libraries will have a continuing role to play, not just for the immediate present but for any period we can reasonably foresee. Indeed, the library's role as the repository of a persistent print culture is an important reason for asking it to be the chief agent in providing public access to digital information, rather than using arguably more convenient facilities like post offices, internet cafes, or simply leased spaces in office buildings or shopping centers.
But there's an even better reason for making libraries the mediators of public access to electronic information: in a word, librarians. When you listen to the visions of cyberspace painted by Gore, IBM, and many others you sometimes have the feeling of a place where a neutron bomb has gone off—of endless rows of cyberstacks where never a virtual footstep falls. But if there is one thing that distinguishes networked computers from print, it is how support-intensive they are at every turn, not just because of the demands of machine and system maintenance, but because people need help in finding what they are looking for in the labyrinth of the Web—and often, in the absence of common document standards, in simply getting it to display on a screen. Anyone who works in an office or university setting is painfully aware of these difficulties, but the problems are even more pressing for library patrons, who tend to be inexperienced users and who don't have friends or colleagues handy on whom they can call.
Then there is the problem of filtering and evaluating content. In recent years, most people have concentrated on the out-and-out offensive sites—the pornography, racism, unfounded rumor, and rampant and sometimes unscrupulous commercialism. But even if we had filtering services and programs that could efficiently screen out the objectionable sites without ruling out inoffensive material, we would still have the challenge of helping users sort through millions of online sites and resources to weed out those that are merely inaccurate, flaky, or trivial. A friend of mine who teaches medieval Scandinavian literature at Berkeley was complaining recently that her students' papers are full of inaccurate information picked up on the Web. And indeed, when I did a search on the word "vikings" on the Web I found everything from scholarly sounding but utterly unfounded claims about Viking settlements in Connecticut to a number of sites devoted to what appears to be a flourishing Odinist movement.
None of these were sites that anyone would want to block students' access to, but you would want the students to have help in sorting the good from the bad. The commercial index services can help here, and so can indexes compiled by various associations and professional groups. But the very numerousness of the indexes and resource pages tends to undercut their usefulness: they can be as hard to find and evaluate as the sites they recommend. In the end, even sophisticated users wind up depending heavily on pointed advice on where to look for things: it sometimes seems that half the mail I receive from my coworkers at the Xerox Palo Alto Research Center contains requests or recommendations about useful or entertaining Web sites. And for people who don't have such colleagues, there is no substitute for a helpful librarian looking over their real or virtual shoulder.
Providing the level of user support that digitization calls for will require a considerable increase in library budgets. A century ago, Carnegie made it a condition of his grants for building libraries that communities pledge to provide annual support of no less than 10 percent of the capital cost of the building. In the digital world, by contrast, costs of support and maintenance exceed capital costs by a factor of anything from 200 to 500 percent—a major preoccupation of business users, and perforce of companies like Microsoft, who are continually announcing initiatives aimed at reducing the "cost of ownership" of systems running their software.
It's true that some of these costs will be lower for libraries and schools, who can get by with less powerful technology than corporations and who may be able to enlist community volunteers to help with installation and support. Some libraries will benefit, too, from the "e-rate" provisions of the Telecommunications Act of 1996 that require telecommunications providers to set aside funds to reduce access costs for underserved users and institutions, though this program was sharply scaled back in June 1998 in response to Republican complaints that it would increase long-distance telephone rates. But there is no getting around the increase in labor costs that digitization requires. This consideration is largely ignored in the various studies of the costs of wiring schools and libraries, all of which assume that actual system maintenance can be performed by a small, centralized staff while routine user support is provided by the existing staff of librarians or teachers, once they've had special training in the use of computers.
This is unrealistic, particularly when you take into account that user support for library PCs is much more labor-intensive than it is in the office sector, since so many library patrons require basic help using the technology. Take, for example, the public library of Demopolis, Alabama, which was recently given $33,000 by the Gates Foundation to install eight PCs, and which received a highly publicized visit from Bill Gates just two days before he testified in front of the Senate Judiciary Committee in March 1998. By all appearances, the Demopolis program is a model of how to introduce technology into a small library. The library has an energetic and inventive staff, and the community has provided funds and volunteers. At the same time, support for the eight PCs requires the equivalent of one and a half full-time persons out of a staff of four. And even if other support and maintenance costs can be kept to $2,500 a machine—an extremely low figure by any standard—the actual cost of providing the technology will involve an increase of between 30 and 40 percent in the library's budget of $155,000, an annual cost in excess of the total amount of the Gates Foundation's onetime gift. (Over and above providing funds for initial training of library personnel, the foundation confines its subsidies for support to a limited period of free access to a toll-free hotline.)
With strong community participation, the Demopolis library might be able to make up the shortfall without skimping on building maintenance, book acquisitions, or other programs. But digitization will be a heavy financial burden for libraries in most communities, not just in poor or remote areas but in wealthy towns in the heart of booming high-tech regions. The city of Palo Alto recently announced plans to close three branch libraries, owing in part to the increased expense of maintaining and improving electronic services. Without greatly increased levels of private and public support, it's hard to see how most libraries will be able to cope with the support demands that informatization imposes.
In the last analysis, of course, the usefulness of a library depends on the content it makes available. This was where the Carnegie libraries most often came up short. Despite repeated appeals, Carnegie refused to provide money for books for the libraries he endowed, arguing that this should be a local responsibility. (It may be, as some suggested, that he simply felt that books were too perishable to serve as permanent monuments to his munificence.) In the end, selection and acquisition were left to the discretion of local authorities, who were highly uneven in the way they stocked their shelves. Many libraries were largely empty for years after their founding, and the collections of most others were dominated by popular novels and other light reading. H. L. Mencken wrote in 1928: "Go to the nearest Carnegie Library and examine its catalog of books. The chances are five to one that you will find the place full of literary bilge and as bare of good books as a Boston bookshop."
In principle, this is where we have an advantage over the last age of the library: once a resource is developed by anyone, it can be available to everyone. Granted, there will be technological disparities between rich and poor here as well, since some schools and libraries will have much better access to content than others. And innovation will only increase the differences—unlike books, digital technology has to be constantly replaced to keep up with user expectations and the demands of increasingly bandwidth-intensive content like audio and video. (Gates must envy Carnegie, whose buildings are still in use a hundred years after their construction.) But while all of this will have an effect on libraries that want to provide their patrons with access to video or software, as many do, the problem shouldn't be too marked for most of the books and periodicals that are central to the library's traditional mission. The real disparities here will be in the money that libraries have available to obtain access to digital works and collections.
This, too, is a question that's been largely neglected in the discussions of library digitization. The general assumption seems to be that the obligation to give people "access to information" will be pretty much discharged once we have put the appropriate hardware and uplinks in place, with the creation of digital content left to the private sector and existing public institutions. In this regard the debates over the e-rate and access to information suffer from questionable parallels with the historical precedents for universal service, like postal service, rural electrification, and telephony, where people assumed that the mere fact of connectivity would be sufficient. As Berkeley's former University Librarian Peter Lyman has observed, in most current models, "the Internet is defined as a broadcast medium in which the digital library consists of universal access to a commercial digital marketplace in which the needs of citizens are fulfilled through their role as consumers. The public interest in learning will be served through subsidized network connections to public institutions like schools, museums, and public libraries."
The question is whether libraries will be able to provide a range of content that justifies the enthusiasm for universal access. One issue, as we saw, is whether funds will be available to convert and develop the resources that libraries need. Another, no less pressing, is whether copyright law will allow libraries to create and use online resources. Under current copyright, most libraries do not have the option of translating any significant part of their important holdings into digital form, and if the new copyright term extension is passed, the amount of material they can translate will be enormously constrained for another few decades. And new copyright proposals that are making their way through Congress will sharply curtail the fair-use rights that libraries have traditionally relied on and hamper their ability to circulate digital materials to their patrons.
Even if resources are available and legally accessible, moreover, it isn't clear how libraries will come up with the money to pay for access to them, particularly the books, journals, and periodicals that publishers have begun to put online. At this point, it's true, publishers are often chary about moving to online publication. They are worried about uncontrolled copying of works, concerned that digital publication will affect the market for print versions of works, and uncertain about what kinds of payment models they ought to use and how they can make money from the process. Ultimately, though, none of these reservations is an insurmountable barrier to electronic publication. The copying of documents can be impeded by systems that are already available which control the use of documents, so that a library would be able to display a work only on certain designated displays or print it only on secure printers. The concern that online publication will cannibalize the print market is also overdrawn, particularly for the kinds of books that are central to public library collections—even if you give novels away online, few people will want to read them that way. In fact, there is some evidence that online distribution can actually increase print sales, among other reasons because it gives prospective purchasers much more information about a book than they could have if they encounter it in a bookstore or catalogue. And secure payment models are already being developed in other areas of electronic commerce and should raise no special problems with regard to books.
The economics of this new world are complicated, and publishers have only begun to timorously address the problem. In theory, digital publications should be much less costly to libraries than print books and documents: publishers incur virtually no incremental expenses in providing public institutions with low-cost access to digital texts so long as they can maintain most of their for-pay market in the process. With digital publication, moreover, libraries are in a position to pool their resources to obtain works: a city or university library system can buy a single site license to a magazine or journal that will be accessible from all its branches. (The aggregation works the other way, as well: some publishers have begun to offer library subscribers a site license that gives them access to an entire list of journals.)
But for most types of publications, rights-holders have little current incentive to make works available to libraries, all the more be cause the Web creates new direct markets for material that previously had less commercial value, like newspaper archives. And even if libraries can get access to these materials at relatively low cost per document and have to pay only for documents that patrons actually use, the ultimate budget increase will be considerable, since so much more content is involved. A local public library that has 12,000 books on its shelves can now provide patrons with access to hundreds of thousands more publications. And access costs are likely to increase still more as the computer industry moves toward devaluing hardware and charging more for service, on the model of cellular telephone service.
To cover the costs of electronic access and higher levels of support, some libraries have already begun to charge for services. The British Library charges users for the right to access its online catalog and has recently proposed a £300 annual fee for anyone who wants to make more than nominal use of the collections. There may be a justification for asking users like large corporations to pay for some of the services they obtain from major public libraries. But the policy breaks with a long tradition of free public access, and many fear that it will inevitably lead to a decline in service, as libraries devote more resources to revenue-producing programs. In any event, this approach is clearly not an option for libraries in small towns or poor communities, which have few corporate patrons. If fee-for-service becomes a standard practice, it will exacerbate the disparities in information access between rich and poor communities.
The alternative is a greatly expanded program of library subsidies. These could be in the form of direct grants or through an extension of the notion of universal service to publishing and the production of content, where a portion of publication revenues or licensing fees is set aside to subsidize access for underserved schools and libraries. Or a similar effect could be achieved if we simply extend the notion of a depository library, by encouraging or requiring publishers to provide a version of each digital publication to a national online collection that is accessible from public libraries and similar institutions. (The Bibliothèque Nationale de France, which is a true national depository in a sense that the Library of Congress is not, has already negotiated an agreement of this sort with the French union of publishers.)
The needs of support, the costs of access, the development of resources—all of these will increase the bill for reinvigorating the public library system, over and above the substantial costs of wiring libraries in the first place. And we need an approach to digital copyright that balances the needs of libraries with those of content providers, who not surprisingly have been far more effective in shaping current legislation. If we truly believe that universal access is both a public good and a private right, though, we have to realize that the public interest in obtaining information won't be satisfied simply by providing everyone with access to a computer and modem, no more than the public interest in reading books was satisfied once Carnegie had provided buildings to house them.