Library of Babel

STEPHANE MALLARMÉ FAMOUSLY wrote that everything in the world exists to end up in a book. And today, it would seem, every book exists to end up on the World Wide Web.

Coverage of Google’s agreement, announced this month, with a host of major research libraries (including Harvard’s, where I work) to digitize their collections and make them available on the Internet, might lead one to believe that we have entered a final phase in the long-sought emergence of a universal library – a compendium of knowledge at once comprehensive, densely cross-referenced, and instantly accessible.

And yet, as happened with the advent of the Web itself little more than a decade ago, we’ve lost sight, amid a flurry of splendid mystifications, of the impossibility of this goal, as well as the pitfalls that lie along the way. Op-ed pages in recent days have blossomed with talk of Flaubert and Alexandria, Borges and Babel. Such metaphysical speculation notwithstanding, Google’s project will begin modestly with a few tens of thousands of volumes belonging to elite institutions. Alternately smitten with and horrified by Google’s ambitions, we tend to overlook the challenges that libraries, books – and the word itself – face on their way to a digital future.

Librarians caught their fi rst glimpse of this future in microphotography, which after World

War II promised to shrink burgeoning collections to manageable size. But by and large the technology never fulfilled its promise. Books after all are physical things; when exposed to the rigors of mass-production photography or scanning, they turn fragile and unwieldy. These are mechanical problems, however, many of which have been solved by decades of library experimentation with the high-volume imaging of books. Though microfilming has often been destructive – as literary abattoirs “disbound” books with an automated guillotine – robotic scanners of the kind Google will likely use can turn the pages of books with all the

care of St. Jerome. The physical care of the collections that will serve as the source of Google’s digital library can be assured. Less certain is the matter of selection: What books will Google take first?

At Harvard, the first selection will consist of 40,000 volumes chosen randomly from the stock of the Harvard Depository – a facility outside Cambridge where the overflow of the university’s libraries is stored in archivally correct conditions. These off-site books tend to be the least-used titles in the collection, sent to the depository because they circulate so rarely. The possibility that these books will make up the first chapters in Google’s initiative is enticing, and would do much to calm fears that the new digital library will look too much like the shelves of Barnes & Noble.

. . .

More signifi cant than any of these concerns, however, is that the products of microfilming and digital imaging are notoriously unlike books, lacking the user-friendliness that has made the codex – the format that emerged to replace the scroll in the Roman Near East about two millennia ago – the most important information technology (perhaps the most important tool of any kind) of the last three thousand years.

Yet I’m not concerned about the codex. For despite the cries of many critics, traditional books and their modes of reading will not be supplanted by the digital library.

The book in its traditional form is a memory machine of surprisingly compact and enduring power. It carries in its bindings, its covers and the materials out of which it is made traces of its origin and travels, both as an artifact and as a repository of images and ideas. As a physical object it has what the 20th-century philosopher Walter Benjamin called an “aura,” consisting of the host of ritual and metaphysical associations it calls to mind. When some of us recoil from the idea of the digital library, what we mostly fear is the loss of this aura – a loss that began long ago, of course, perhaps in Gutenberg’s own time, or perhaps in the mid-19th century, when the book ceased to be a handicraft and became a mass-produced commodity.

But the book as we know it carries within itself something more concrete: its own archeology. Dependent on ever-changing technology, e-books are relatively ephemeral; and although this need not be so, they tend to obscure their own origins and inner workings. Seeking to tame the ghosts of the past, the digital future may end up erasing its own history.

The very power and promise of Google’s project may carry this peril further. Accessed through Google’s justly famous searching technology, the texts of all those books would become continuous, omni-conversant, cross-referenceable down to line, word, and letter. The Google library ultimately is not 40,000 or 4 million discrete books, but one book: a seamless singularity of text, ramified and interconnected.

The book itself would thereby seem to disappear; and perhaps the individual books in the library – an as-yet-unknown number of which never will be scanned because they are too fragile or are composed in formats that defy digitization – will begin to seem the accidental artifacts of a bygone techne, like the flickering frames of old silent movies.

Technology will transform the library again and again, as it has throughout history. Books in the libraries of Alexandria were as different from today’s paperback as paperbacks differ from today’s Web pages. The library has absorbed the shock of such transformations through the work of readers willing to think about the ramifications of change, willing to adapt and persevere, willing first of all to remember.

As we enter another cycle in the history of text and technology, it’s this – the perception,

creativity, and memory of all of us as readers and thinkers – that will determine the outcome. Only through such agency will we ensure that there is truly room for “everything in the world” in the library to come.

Matthew Battles works at Harvard’s Houghton Library as coordinating editor of the Harvard Library Bulletin and is the author of “Widener: Biography of a Library” (Harvard).