The ABAA was delighted to have Professor Adrian Johns of the University of Chicago lecture at the 2011 California International Antiquarian Book Fair. His talk was entitled The Promise and Peril of Universal Libraries, and Professor Johns is kind enough to let us republish the text on our blog and website. Enjoy! ******************************* The Promise and Peril of Universal Libraries Adrian Johns University of Chicago Since it was originally announced some five years ago, the Google book project has provoked a rich variety of responses. The intensity of the debate is understandable, because the implications of the project are indeed as broad as they seem to be. This really is the rare enterprise that has implications for everyone who works in academia and, in all likelihood, everyone who does not. The future of knowledge itself is at stake. So too is its past. Many of the debates that Google’s venture has ignited have taken place on the relatively familiar, if sometimes very technical, terrain of law and economics. The problems evoked have been those of copyright infringement and monopoly. The debate has also been infused with a strong dose of publishing industry jeremiad, and more recently experts in the bibliographical world have voiced disquieting revelations regarding Google’s devotion to outmoded classification criteria and surprisingly poor metadata standards. When addressing what is trumpeted as the project’s real purpose, however – to transform how knowledge itself is gained, stored, circulated, and put to use – protagonists have rested content (as is so often the case with debates about intellectual property, libraries, and media) with crude and unhistorical generalizations. My aim is to counteract that. I want to suggest that the advent of (something like) a universal library today should approached in terms of the purposes and uses of the universal library as well as its composition. If there’s a revolution here, its nature lies more in the realms of purpose and practice than in that of collection itself. So one should ask the Aristotelian question: what is a universal library’s “final cause”? What is it for? In particular, why does the Google project exist? One engineer at Mountain View suggested an answer to that question, in an encounter with an Internet aficionado named George Dyson. “We are not scanning all those books to be read by people,” he said. “We are scanning them to be read by an AI.” He was referring to what the proposed Settlement calls “non-display uses” or “non-consumptive research.” On that account, what is really transformative about this universal library scheme is that it supercedes reading itself. As it happens, I do not think that the Google Books project really is directed primarily at “machine learning.” If it were, the metadata standards would have been better, and Google would have been less insouciant about the problems with such data that have been widely acknowledged. Still, the fact that a statement like this could have been made – the fact that it is even conceivable – is very telling. And it may well prove the case that the data-mining uses of the archive will be more economically consequential than the reading uses. Yet compared to display uses – that is to say, everything that involves actual reading by actual humans – non-display uses have received scant public attention. That is not entirely our fault. The concept of a ‘non-display use’ is historically very unfamiliar. It could scarcely have been imagined before networked digital systems. Uses carried out on books treated en masse as data, without any eye seeing any page of any book, are radical departures from prior practices. There is no legal tradition to shape and constrain these uses, and little by way of ethical reflection to draw on either. When it comes to human readers, we have 300 years of copyright law to argue from and about. When it comes to machine readers, we have… nothing. Yet Google’s own conditions are strict. They include a rule that no product could emerge from such a project that would compete with Google’s own ventures. Furthermore, there seems to be no right (except a short-term one) for an author or publisher to opt out of non-display uses. A redefinition of ‘publishing’ and ‘authorship’ themselves – to say nothing of ‘reading’ – is therefore at hand. Publishing and authorship will will incorporate in their very meanings the certainty of a work’s being included in these systems. It will become part of the roles of publishers and authors to anticipate this fact, and make decisions accordingly – just as they have had to anticipate human readers’ responses for centuries. For me, this raises historical questions – questions both of history and for history. How does this prospect relate to the plans for universal libraries over time? How did they promise or threaten to affect the practices of knowledge? How in fact did they affect such practices? The answers, I want to suggest, throw real light on the problems we now face. * I want to approach them by asking about place and practice. First, Where is the universal library? The place of reading matters. It affects how we read, and to what effect. The implications of launching a universal library right now have to be appraised partly in terms of the changes under way at actual, physical, non-universal libraries. The project happens to coincide with a time of critical pressures on those. At Syracuse University, the dean of libraries recently announced that physical libraries were “kaput.” Syracuse had moved to store volumes not just off-site, but 250 miles away. And that’s not the worst case. The worst case seems to be that of Cambridge (UK), where old medical volumes are now being put in “dark storage.” This means that they will be held not only offsite, but actually inaccessible. Short of invoking nuclear war, it is difficult to think of any way in which dark storage can be a good thing. If the universal library is to lead to a place of reading that is no place or any place, what does that imply? The associations are ambiguous. On the one hand, knowledge created in no particular place may be regarded as objective because it could have arisen in any place. But on the other, practices of reading are never placeless in reality. Universality means reading in lots of actual places. This being so, it may be argued that there are real advantages to having it occur in a site designed for the practice. That’s in part what a physical library is. So what is the relation between such specific places and universality? To answer that, you have to understand the history of universality itself. (a) Ancient universality The history of universal library projects is not continuous from the ancient world, yet many (perhaps all) of the later incarnations of the dream have referred in some way to the Library of Alexandria as its fons et origo. Founded in about 300 BC, the Library reportedly amassed some 500,000 scrolls. It depended on trans-national trade to obtain them: the pharaoh’s officers reportedly collected works by taking scrolls from merchant ships and copying them (the captains got the copies back, not the originals). But as significant as the sheer accumulation of material was the accumulation of site-specific reading practices devoted to that material. Textual scholarship, authentication techniques, and bibliographical classification practices are all said to have originated here. Various other ancient libraries existed, some of them reputedly very large, like this one at Ephesus. But Alexandria’s is the dominant example, and no other has had anything like its iconic status in subsequent history. (b) Renaissance universality The Renaissance saw the revival of the ideal of a universal library. You now did not need to go somewhere unique in order to experience the illusion of literary universality. There was more than one universal library, and some belonged to private citizens. John Dee had one at Mortlake, outside London, and Gabriel Naudé gave rules for setting one up in his Instructions concerning erecting of a library, written originally for the President of the Parlement of Paris. But it was soon clear that no one building could house a universal library. So attention began to shift to the idea of a virtual place. This meant creating a kind of distributed library – through print, and then through periodicity. Conrad Gesner’s Bibliotheca Universalis was meant to be a virtual library of this kind – a volume surveying the entirety of the domain of print. Of course, the pages of one book could no more hold universality than the shelves of a library could. So Gesner had to project successor volumes. By the late seventeenth century, this had been formalized into the notion of the Bibliothèque Universelle (note the French), now produced in parts – essentially a hybrid of periodical and encyclopedia, devoted to noticing and summarizing the most notable of recent publications across the Continent. In that form Jean LeClerc’s journal became one of the major forums of the Enlightenment. The reading-room of this universal library was the boundless space known as the Republic of Letters. Access was supposedly unlimited; but it had elaborate codes of civility, hierarchy, organization, and exchange. Some of the bibliographical systems invented for “libraries” like Gesner’s were in fact derived from these codes – especially the commonplacing schemes of scholarly readers. That is, readerly techniques led to classification sciences, which were in turn put to new uses by readers. (c) Enlightenment universality Probably the most famous image of a universal library is the 1785 design for a royal library created by visionary architect Etienne-Louis Boullée. 100m long and 30m high, it would express the king of France’s ambitions as universal monarch. (Apparently there was a strict classical dress code for readers.) Although Boullée’s was never built, there were real libraries in the Enlightenment that embraced an ideal of universality. Most notable was the university library at Göttingen. The librarian there designed an acquisitions policy – the first for any such institution; it defined universality in terms of coverage of specific research fields. The library printed journals (and catalogues), extending the reach of this policy across the Republic of Letters. Many of the practices of modern research libraries descend ultimately from Göttingen. At the same time, in France the Marquis de Condorcet was developing this notion of universality to define the virtual space of print itself. Condorcet proposed that print should create a universal library of Enlightenment by abandoning authorship as its principle of classification. Instead, the tree of knowledge would become the card catalogue of a limitless virtual library defined, as at Gottingen, by subjects. The law of publishing itself should be transformed accordingly: Condorcet would eliminate authorial copyright and replace it with a system of periodicals, each defined by its specialty. This was to be the practical realization of “Enlightenment” – the dispersion of “light” from central sources, uncorrupted by individual pride. It made the Republic of Letters into the counterpart for Adam Smith’s Republic of Trade. Condorcet’s scheme was not to be put into effect, of course – until the French Revolution, when it proved a disaster. Romanticism then repudiated the idea wholeheartedly, elevating individual genius in its place. But of that more in a moment. (d) Modern universality Since around 1850, communications and information media have proliferated. Schemes of universality have proliferated with them. E.g., in the 1890s French bibliophiles published an essay on “the end of the book,” illustrated by the visionary artist Albert Robida (these pictures come from Willa Silverman’s The New Bibliopolis, 2008). Robida imagined a world with universal “clichéotheques” storing “storyographes” on Edison’s cylinders, the sounds being transmitted through “telephonoscopes” into vending machines and homes. The new culture would come with its own origin myth (Edison, like Fust, threatened with burning at the stake), and anti-pirate measures. The best known actual proposal of this kind was Vannevar Bush’s Memex project. But much more interesting was Paul Otlet’s Mundaneum, because it really was operational in Brussels in the 1920s. Otlet (a pioneer of scientific bibliography) envisaged extending bibliographers’ classifications into books themselves, cataloguing their facts, languages, arguments, images, and references. He then created a physical space - an office – filled with networked screens. Staff used these screens to access pages from a central store, juxtaposing print with films, music, and other media. Otlet called the whole thing variously “the radiated library” or the “Universal Book” – and also a web (réseau), connected by “links.” Information held within it would be accessed through an “electronic telescope” the eyepiece of which stood on one’s desk. His Mundaneum was meant to become a universal “edifice containing all books and all information.” Citizens could send in questions, which staff would answer by telegraph. It began brightly, associated with utopian hopes for the League of Nations. But by the mid-’30s it had lost state support. What was left was abandoned until it was rediscovered in a forgotten room at a local university in the late 1960s – after which almost all of it was destroyed. There were several schemes along similar lines, all premised on the use of new technologies (typically microfilm) to collate information from various media into a single scopic device. The effects – where there were any effects – were not all good. On the one hand, there was the STC microfilming, which revolutionized the study of the early modern era and is the basis for today’s EEBO. But on the other, the study of the modern era was set back drastically, as massive piles of newspaper stock were destroyed on the premise that microfilm was sufficient. Moreover, modernity has highlighted the fact that a library can be universal in scope but not in access. Who does have access, and under what terms? Who polices it? How does it extend beyond national (and other territorial) boundaries? The open-access movement in science reflects such questions. The other side of this issue of publics is of course publishing itself. What is its role in the age of universal libraries? A good case can be made that publishing is the business of managing credibility. It is the enterprise of ascertaining and affirming the worth of particular authorial works, thereby singling them out for advantageous production, exposure, and distribution, and of sustaining their authority and authenticity from then on. In previous centuries, publishers employed (or, earlier still, were) printers, and handled the actual manufacture of books; but this is an ancillary function. What is essential is the practice of judgment, at a corporate, collective level. The question is what role this practice has in a world where archiving is normatively universal, but searching often pesky; and how to provide for it economically in a world where openness is normatively universal too, but access in fact limited and partial. So universality has a long history, involving destruction as well as preservation. Aspirations of placelessness have repeatedly come up against the need for actual places, with their distinct publics and practices. We need a historical typology of reading practices across these places – not least because reading – and the knowledge it facilitates – is being changed by the intercalation of virtuality now, so that in a digital universal library we shall have to work to coordinate our reading practices with the habits of original readers. For example, in the sciences, printed journals are now an anachronism. Scientists do not employ them in their research, as they did before about 1990. Instead they use digital journals, e-print vehicles, databases, etc. This has consequences, not all of them good (the horizon of conceivability has contracted, and the haste of consensus-formation accelerated). We should not shoehorn current scientific ‘texts’ into a model of reading (that of printed pages) that their target community does not in fact practice. By misrepresenting reading practices it would obscure those consequences. But by the same token, so would relying on digital images of seventeenth-century tracts. The question for history is therefore: what do we do about this? How do we ‘translate’ between the reading practices of our universality (which is in fact embedded in our technology, society, and culture) and the reading practices of their universality (embedded in their technology, culture, and society)? These broad questions have to do with how knowledge is arrived at, what it is, how it is circulated, to whom, and what those who encounter it can and will do with it. To put it a bit hyperbolically: they have to do with the fate of civilization. * Now, I want to bring this to earth by showing one way in which the history of universal libraries has shaped our predicament. Think of the terms that are at stake in our own time: principles of authorship and creativity; the practices and institutions of research; the travails of the publishing industry; the virtues of mass access. All of these have their historical trajectories. Those trajectories came together at an earlier moment too – the moment when Romanticism encountered the advent of a mass culture industry and the origins of the research university. This was the early nineteenth century, roughly the generation from 1800 to 1830. Like our own, this period saw a media revolution coincide with radical pressures on the academy. Publishing was changing radically. First iron presses and then steam transformed it into an industrial enterprise. A commercial realm of mass print came into being, centering on the principle of copyright as the enabling device of speculative capitalism. The publisher – the master of this kind of lettered property – appeared as a distinct social kind. The power of the system was exemplified by Sir Walter Scott’s novels, which sold in unprecedented numbers. And at the same time the cozy world of the academy was being upended by the post-Jacobin regime in France and the Humboldtian reforms in Germany. It turns out that in those same years a bitter struggle raged over proposals to create and sustain a universal library, housed in the “public” institution of the university, and devoted to collecting the output of industrial publishing. The debate was not just about collection and selection, but also about use – that is, it was partly about the practice that became research, and the institutions that it would define. The bid for universal libraries emerged from the widely accepted claim that printing had made Enlightenment itself possible. It was reckoned essential to collect the output of the press in order to preserve existing knowledge and facilitate future progress. Enlightened societies ought to have libraries aspiring for universal coverage, in some sense of that term. And they must be public, too – again, in some sense of that term. Anyway, they must not be the preserve of castes devoted to priestcraft and mystery of state. In England the approach to creating a universal library was eminently pragmatic. It rested on the ancient universities. Their libraries claimed a right to get a copy of every book published in London. This claim descended ultimately from a private agreement of 1610 benefiting Oxford’s then-new Bodleian Library. But more directly it derived from the 1710 Copyright Act. This statute had not only instituted the world’s first law of copyright, protecting titles registered at Stationers’ Hall in London for 14 years; it also gave nine libraries the right to demand free copies of these titles, all on the best paper. For decades this right lay neglected. But around 1800 the university libraries resurrected it. They suddenly claimed an exhaustive right to collect all newly published works. The new demand reflected a growing awareness at Oxford and Cambridge that they needed to change, and that research might become part of their raison d’etre. There were just two problems: they had never made this sweeping claim before (although they said their right was 200 years old); and they did not want to pay for what they collected. So the possibility of a universal library – and hence of perfecting Enlightenment – was intimately linked to the fate of copyright. And contemporaries accordingly saw the bid for universal libraries in terms of a tension between those two cardinal principles of England’s post-Glorious Revolution sense of nationhood: public knowledge and property. Briefly, the conventional story about copyright is that its status was determined for Britain (and subsequently for the United States) in 1774. At that point a landmark decision in the House of Lords known as Donaldson vs. Becket resolved that it was a form of artificial monopoly, with a term limited by statute. Subsequent battles are generally portrayed as mopping-up operations. Yet Romantics were not so sure. They liked to appeal to the concept of genius to explain originality, and this was a difficult concept to square with an artificial, limited copyright. It remained possible to argue that literary property, far from supporting genius, was incompatible with it. What sparked this possibility into a fierce controversy was the fact that copyright in practice became wedded to the bid for universal libraries. As a result, by 1812 an alliance of respected authors and publishers could be heard denouncing copyright itself as “fatal to literary property.” The London book trade had interpreted the 1710 law in the spirit of contemporary political economy. The booksellers treated it as a quid pro quo: the libraries’ free copies were a fee charged for the law’s anti-piracy protection. So they reckoned that they remained at liberty to decline protection and not pay the fee. By not registering a given book, they could take the risk that it might be pirated, and avoid donating the nine copies. And this became standard practice. As a result, the books offered to the libraries were those most likely to be pirated – plays, sermons, and the like. Scholarly works – law books, for example – increasingly stood outside the system. Worse still, in the case of multi-volume works a bookseller might register just one volume, guessing that that would be sufficient to deter pirates. Depositing that one volume would then almost compel the libraries to pay full price to get the rest of the set. So deposit and copyright had long been linked – but in a way that in practice risked turning the libraries into universal repositories, not of learning, but of hackwork. Piracy looked like becoming the de facto definition of Enlightenment. In practice, the proffered works were so unpromising that the libraries frequently did not even bother to retain them – a selectivity that all sides defended as virtuous as well as necessary. Resentment within the universities burst forth when the court of King’s Bench ruled that a publisher could collect penalties against a piracy even if a work were not registered. At a stroke, this 1798 decision removed all incentive for registering books, and hence for their being liable to collection by the libraries. And the motivation to avoid depositing increased when the Act of Union raised the number of benefiting libraries to 11. Unsurprisingly, the number of titles deposited fell sharply: in the whole of 1803 Cambridge received just 22 – less than 1/20 of what it thought it should get. If universal libraries were really an essential tool of Enlightenment, then what was this but a serious cultural crisis? Cambridge led the charge to solve the problem. It began demanding a copy of every single book published, explicitly hoping to create what it called a “universal library.” And it also demanded a copy not just of new works, but of reprints – a real departure, and one destined to prove consequential. In 1812 the university won a test case on the principle. Suddenly the publishing industry faced an enforceable universal collection practice. It was this predicament that it denounced as “fatal for literary property.” The universal library would apparently rest on universal piracy: property would be seized without compensation to build it. It amounted, or so the publishers said, to a prohibitive tax on knowledge itself. The publishers were not alone. They found allies among some of the foremost authors of the time. This was primarily because the prospect of a universal collection practice twinned to copyright ran counter to a prized Romantic conviction. As Wordsworth most famously declared, truly worthwhile knowledge (that produced by geniuses) was ipso facto unpopular. Real knowledge (or art) therefore stood apart from the new world of mass, copyright publishing – the world with which the universal deposit was aligned. Genius produced very small print-runs (50-100, say). It was therefore disproportionately punished by a practice that demanded the same donation regardless of run. That argument revealed deep-level assumptions about creativity, commerce, and cultural value. The Romantics maintained that the link between copyright and deposit was more than merely contingent. They thought that copyright had brought into existence a culture of print that necessarily neglected originality and genius in favor of popular appeal. “There are… few works of great genius adapted to the general reader,” they maintained: “whatever is deep, whatever is abstruse, whatever appeals to the highest qualities of the mind, or the most difficult subjects of intellectual acquirement” was necessarily “fitted to interest a very limited number of readers.” Valuable copyrights therefore generally attached to derivative works: they estimated that 95% of bestselling authors were “mere repeaters.” In other words, copyright, intended to bolster creative originality, had generated a realm of print that negated it. By contrast, the readership for advanced works of mathematics or natural science was often 200 or less – a level well below any copyright calculus. This practice of universal collection therefore militated against it. With the current law in effect, they demanded, “Can any Author or Publisher be insane enough to embark in an expensive publication?” Clearly the answer was no. So “the man of genius, or of science, or learning, dies in obscurity; and his talents or acquirements are buried with him in the grave!” With this we approach the core of the argument against the universal library of the Romantic Age. It was really about the prevailing culture of print itself, and how that culture impinged on knowledge – knowledge in the making and the reading. Copyright was allegedly representative of a new age in which “political economy” (that is, the economics of Malthus and Ricardo) was the queen of the sciences. Collecting everything would make the great libraries into testaments to this new age. Properly, the campaigners thought, collecting should be an active practice, involving selectivity, discrimination, and responsibility. It should be an act of polite civility, too, carrying meaning for those works and authors selected. If this principle were abandoned, then a universal library in practice would become disastrous. Who would sacrifice themselves searching for “success in the more difficult branches of science,” they asked, “if the reverence and celebrity which in enlightened ages have attended Authorship are destroyed, by giving equal preservation and the same place of distinction to whatever the Press vomits forth?” It would exert a leveling effect that would destroy true authorship, leaving only hackwork. Ultimately, universal deposit would become an evil even for the very libraries themselves. They would find themselves “overgorged” with frivolous books. And since it was human nature to be depressed by excess, the crammed libraries would provoke in readers – here they meant university professors and students – only a debilitating “depression.” It might indeed be useful to have one repository of all books published, they conceded, but it should not be public. Perhaps it could be confined “in special custody” at the British Museum. So for Romantics the universal library threatened to extinguish the most valuable creativity of all. The demand to collect knowledge was destroying the practice of producing it. Thomas Longman himself testified that he had cancelled a prestigious edition of Humboldt because of the deposit. This conviction became the basis of a momentous alliance between science, scholarship, and poetics, on the one hand, and metropolitan commerce on the other. In particular, the libraries’ demand for reprints posed a direct threat to the most fashionable form of genteel scholarship of the day: antiquarianism. Today antiquarianism has a reputation as dry; but around 1800 it was anything but. Burkean Tories and Jacobin Levellers alike found in antiquarianism a route to recovering their own versions of a hallowed past, recorded in artifacts: inscriptions, stonework, manuscripts, fragments of stained glass. Antiquarianism depended on the issuing of reproductions of such objects. Literary antiquarianism in particular, of the kind favored by Scott and the Roxburghe Club, virtually was an enterprise of reprints. So opposition to deposit arose from this influential community, which allied itself with the publishers. So it was that London’s Romantic intelligentsia and its leading publishers rose up “en masse” against the prospect of a universal library. The campaigners included novelists like Isaac Disraeli, natural philosophers like Thomas Young, and physicians like William Lawrence and John Abernethy (violent enemies in everything but this). Faced with near-unanimity, for a moment Westminster looked set to abolish the deposit. But then fate stepped in in the form of the Prince Regent, who dissolved Parliament before it could do that. When it met again, the campaign’s leader – an eccentric ultra-Tory named Sir Samuel Egerton Brydges – had lost his seat. Brydges was a poet and novelist, as well as an aesthetic bibliographer and the proprietor of a private press devoted to reprinting works of literature that he felt had been obliterated by the commercial forces of the copyright industry. For him the campaign had been a heartfelt issue, and he had devoted himself body and soul to it. With him gone, the campaign had no lynchpin. It collapsed with stunning suddenness. The clash between universal libraries, publishing, and genius thus ended with the libraries on top. But it left a more mixed and resonant legacy. At the peak of their campaign, the publishers had proposed a compromise. They would accept a universal deposit, they said, but in return demanded a blanket doubling of the copyright term to 28 years. In 1814 Parliament passed a new copyright law enshrining this – it was the first statutory departure since 1710. It also gave the universities everything they wanted. Or rather, it gave them more than they wanted: thanks to a last-minute, one-word amendment, it extended their explicit right not just to new works, but to those much-contested reprints. What emerged, then, was not just the principle of copyright libraries – the closest things to universal public repositories until now – but also the first breach in the century-long principle that copyright protection was for a fixed term. It was still for a term, but that term could no longer be assumed to be fixed. And the consequences would be especially clear for anyone wanting to reproduce old works. * The process of lengthening copyright terms that ensued is one with which we are painfully familiar. In the United States, the first Copyright Act, in 1790, introduced the same term as the original UK law, of 14 years renewable for another 14. But in the mid-1820s Noah Webster (who had been a prime instigator of that first law) toured Britain and heard of the campaign that had just ended. He returned to Washington and lobbied for an extension, explicitly hoping to replicate Britain’s 1814 statute. The result was the 1831 Copyright Act, which doubled the US term to 28 years. Extension has had a continuous history here ever since. So what originated with a Romantic rebellion against universal libraries in the early 1800s has now given rise to a central element of the new rebellion against them in our own day. The problem of “orphan” works that has dogged the Google project is a distant legacy of that earlier clash. In that sense, far from being a unique product of the Internet age, the Google conflict is a consequence of the technologies of the Internet being applied to desires and concerns that are older. Its problems are through-and-through products of the history of universal library purposes. In particular, the problem of old knowledge (reprints) in a universal library is as critical now as it was then. But there’s also a difference. In 1810, an age when economics was not yet the all-pervading authority that it now is, the scope of the question was seen to be deeper, broader, and more cultural. Disraeli, Longman, and their allies believed that it was central to the fate of knowledge itself. Universality, all sides believed, would affect readers. It was supposed to – that was what it was for. It would either generate new knowledge via research, or lead to enervation in the face of a mass of mediocrity, created by copyright publishing and merely stored up in the universal library. The result, antagonists thought, would be a deadening of critical instincts and an apathetic reliance on the homogeneity of mass culture. Genius was bound to be a victim of universality, on this account. A modern age characterized by political economy, copyright, and the new collecting would kill creative authorship. It was a reactionary conception, to be sure, but it was not only that. It commanded a good deal of support from scientists too, and from professionals of the book. And they may have had a point. How will readers today make new knowledge out of the undifferentiated reservoir of digitized mediocrity? Will they even want to make new knowledge out of it? And what long-term trends might we unintentionally spark in a bid to pacify our own warring factions? These are among the questions that emerge when we identify the purposes of universal libraries as foci of inquiry. I don’t think we can answer those questions now. But a focus on uses may help us answer a related one. Such a focus suggests an approach – albeit a speculative and perhaps hyperbolic one – to the problem of the future of the book itself. Perhaps an answer is to be sought, not in the architecture of the codex, let alone in the sentimentalism of “look and feel” that were until recently so widely invoked, nor even in the anxieties about monopoly or metadata swirling around Google. It may lie, instead, in the pairing of practice and object – the dyad formed by the conjunction of reading with books – a pairing that of course is always in some place. The importance of this dyad was stressed in an important work that appeared in English just as the debate over universal libraries was gearing up: Immanuel Kant’s What is Enlightenment? Kant’s paradoxical assertion was that a private place of reading and writing were essential for the public realm’s authority. It initiated a long tradition of arguments along similar lines. For this tradition, in order for public reason to persist, participants need to be able to retreat into a private space – to disconnect. But this possibility is lost when an important purpose – and perhaps the purpose – of the universal library is to subordinate reading to “non-display uses.” That is, if the purpose of mass digitization is to convert not only works, but work (viewings, uses, readings) into data. These data are then merged, sorted, combined, analysed, mined. To paraphrase that engineer, what AIs will read is not just books, but us. The process is not necessarily sinister. Often it is useful and convenient. Overall, I myself would be inclined to say that it is a good thing. But a certain caution is in order. In the first six months of 2010 Google received over 4,000 requests from US government agencies of one kind or another for data on “users”, and another 1,300 from the UK. Many of these requests are said to relate to criminal cases. But it is not hard to imagine the practice expanding. In the UK, to cite what I suspect is a relevant comparison, raw data on phone-tap authorizations have been made public since the first term of the Blair government – and it turns out that the practice is very widespread indeed. In just the one year of 2008, for example, various state agencies, down to the level of local authorities, made more than 500,000 requests for phone surveillance. What happens when claims of security or morality tempt authorities (or their subcontractors) to ask for data about ‘reading’ – data that a universal library devoted to this practice and this purpose must necessarily have? Where does the authority of public reason then abide? Does it still exist at all?