I’ve never said we should get rid of all the books, or even most of the books. But I’ve learned, as have many librarians, that advocating for even a modestly smaller collection can prompt reactions of disbelief and shock among our patrons and the general public. The typical reaction runs along the lines of “I thought librarians loved books! How can they bear to discard them?” When the University of New South Wales in Australia reduced the size of its collection by 50 percent in 2011, one faculty member lamented “there’s something profoundly wrong, and symbolically wrong, about a university destroying books. Universities are in the business of passing on knowledge, and books—no matter how the use of books is shrinking—still remain a very important symbol of knowledge.”1

In my opinion, this common and impassioned plea argues precisely for, not against, reducing the size of book collections. To be certain, books are a symbol of knowledge. But shelf space is not unlimited, and books are merely one of many vessels that contain knowledge. In these days of online information and decreasing circulation figures, librarians can be wholly committed to preserving and transmitting knowledge while simultaneously advocating for much smaller book collections. Holding on to books merely for their symbolic or sentimental value does nothing to promote knowledge and, in fact, is often counterproductive. After all, maintaining a collection of books is not cost-free. Books that are rarely or never used are not “earning” their place on the shelves and should be discarded in favor of a different use of the space they take up—a use that promotes the values of the particular library and its community. More specifically: The existence of knowledge requires not just books (and other information packages) but also people; and space that enables people to receive, discuss, debate, share, and transmit knowledge is a productive use of space for any library.

It’s ironic that the first consideration you bring up is the symbolism and sentimentality attached to books—books as emblems of learning instead of as information packages—because I’ve found that if anything, I’ve actually become a lot less sentimental about books since becoming a librarian. I used to treat cheap, tattered paperbacks as inherently precious, carefully moving them from apartment to apartment, finding room for them that I didn’t really have to spare, never discarding anything, no matter the condition—not just special favorites: every book I owned. Now, thanks to professional experience, I suffer no pangs of loss at all in telling myself of a ruined or worn-out volume, “There are a lot of copies of Emma in the world; this particular one is ready for the recycling bin.”

So I’m happy to begin with that concession: that which is easily replaceable shouldn’t be treated with artificial attachment, and the aesthetic and atmospheric value of endless stacks of books isn’t, by itself, enough to justify their retention.

But knowledge isn’t fungible: one cannot “swap out” that dusty 1967 physics textbook for the latest online database of literary criticism and assert that, since the database contains more words than the physics book, the exchange marks a net increase in our commitment to preserving and promulgating knowledge. Even if that’s mathematically accurate, we’re still talking about eliminating specific, unique units of knowledge without replacing them. When we remove a book from our shelves, we remove that particular quantum of knowledge from our holdings. Even if we acquire other vessels of different information at the same time, there’s a unique loss of knowledge and ideas to each such elimination.

You’re right, of course. Reducing the size of the book collection in our library (without replacing the content electronically) entails a unique loss of knowledge from our library. In fact, I would go so far as to say that, in many cases, replacing print materials with “identical” electronic content constitutes a loss. I am reminded of a doctoral student I met recently who is studying magazine advertisements geared toward African American women in the 1920s and 1930s. Even when such periodicals are digitized, it is often difficult or impossible to browse non-indexed material such as advertisements, so discarding the print in favor of electronic access really does result in a loss of information. This student’s provocative dissertation might be impossible to complete if she does not have access to long print runs of popular magazines.

But losing particular information packages from our library does not mean that the knowledge in question is lost from the face of the Earth. At the University of New South Wales, librarians checked each book to make sure they did not discard the last remaining copy in Australian libraries, and this kind of vigilance is of utmost importance when any library undertakes a major reduction in collection size. This is true not just for large research libraries but for all libraries: Any library of any size may hold rare, unique, or local materials that are not found in many other collections.

Even if we are careful to retain rare, unique, and local materials during a major reduction in collection size, though, how can we justify the inevitable “unique loss of knowledge” from our library? I assert that such a loss is more than balanced by the concomitant gain in library space. But space gains are commonly dismissed by patrons and the general public. One professor at the University of New South Wales had a typical reaction to his library’s space reallocation: “They’ve turned the library into a kind of Starbucks.”

The implication is that such as exchange is not worthwhile: Losing books and gaining a coffee shop is no gain at all. And perhaps this is true—in some libraries. Clearly, some uses of reallocated space are better than others. But just as each library’s collection reflects its values and its community’s needs, so too does each library’s use of space. A coffee shop might not be the most sensible use of space in every library, but I maintain that in some libraries, gaining a coffee shop—or a smart classroom, or an information commons, or a childcare center—could more than offset the loss of a great many books.

True, removing information from a single library doesn’t obliterate it. And I agree that under some circumstances, increased cooperation with our holdings can reduce the need for duplication at multiple academic libraries—a principle that applies to new purchases as well as to decisions about item retention.

However, the loss of content from an individual collection represents, at the least, a blow to patron convenience. If we are still committed to “saving the time of the reader,” then imposing barriers to access is contrary to our values. The doctoral student you mentioned above probably still could have figured out a way to complete her research if we had eliminated the periodicals she needed from our stacks—after all, those advertisements still existed. She could perhaps have browsed the unindexed content online, or traveled to another library, or attempted interlibrary loan requests (and hoped that we could find a lender with generous policies about bound periodicals). All those alternatives would have been more trouble for her than the current option: using the items we have on our shelves. We would have served her poorly by eliminating that content.

In addition, I’m unconvinced that guarding against eliminating the very last copy of particular title is a sufficient standard. I’m reminded of the logic behind the LOCKSS Program for preservation of libraries’ digital content—Lots of Copies Keep Stuff Safe. Operating from the premise that “no single library can achieve robust long-term preservation alone,” LOCKSS protects permanent access to digital content in part by enabling libraries to maintain local copies.² Shouldn’t we, as the community of research libraries, be as conscientious with our print collections?

Just as you aren’t arguing that we should summarily get rid of all the books, I’m not claiming that we have an unbreakable commitment to keep every volume that we own forever. As local needs change, librarians do sometimes find that some materials simply aren’t as important as they used to be. Research libraries have an obligation to serve our current users and to support local needs, and our collections, as you say, should reflect that.

But we also have a duty, in my opinion every bit as solemn, to future users. As such, we must balance our support of current needs with our obligation to preserve access to the scholarly record for future generations of researchers. We can’t always predict what will be important for research decades from now, but it’s not necessarily just the materials that are currently in demand. After all, who knew, eighty years ago, that those magazine advertisements would someday become dissertation fodder?

When we reduce the size of our collections, I believe that we often overvalue current usage rates as a measure of the worth of the books we hold. If we use current circulation statistics alone to decide what should be kept—saying, essentially, that only those books that have been read in the past year or two have “earned” their place on our shelves—then we may well create bias in our contribution to the scholarly record. What we will have provided for scholars decades or even centuries hence won’t be our collection, but the fraction of our collection that happened to be in use when we decided to reduce its size. No, keeping low-use books isn’t without cost, but I believe that bearing that cost is part of our responsibility to future scholarship.

Similarly, I believe that we should consider what obligation we have to researchers beyond our walls. In my opinion, although our primary concern should be the members of our own academic communities, research libraries should also make collection decisions that respect the needs of the broader world of scholarship. You didn’t use an argument that I was expecting to see crop up in this discussion, but if you’ll forgive me I’ll address it anyway, since it’s a recurrent point in this conversation: If we reduce our local holdings, we can depend more on interlibrary loan. It’s certainly appealing: To increase sharing of our holdings among institutions, so that scholars have quick and convenient access to the holdings of institutions beyond their own, is marvelous. But at the risk of being obnoxious in pointing out the obvious, somebody has to be the lender. If we eliminate lower use titles from our collections, I worry that we shirk our obligation to be reciprocal partners when other libraries’ patrons need us. For instance, I recently stumbled across a thirty-year-old psychology book in our stacks, written in Finnish. Unsurprisingly, it’s never circulated. But it has gone out on interlibrary loan twice, and although ours isn’t the last copy, it’s not exactly widely held. I am never going to weed that book, despite the fact that it’s unlikely it will ever serve local needs, because I believe in ensuring its availability for outside researchers. Current, local needs are important, of course—but so are future and national (or international) needs. Preservation decisions must balance our responsibilities to both.

I’ve been concentrating on the inevitable loss entailed by the elimination of books; I’ll turn now to your argument about the gain of repurposing space. My knee-jerk reaction, I admit, is to join your chorus of patrons and the general public in dismissing them outright, but I’ll try to be more thoughtful. And that takes us to the fundamental question—what are research libraries for? That’s a much larger question than the scope of this column permits, of course, but I’ll give my own three-word answer: scholarship and learning. Therefore, I believe that as we decide how to use our physical space, the fundamental question should be: Does this use of space support scholarship and learning?

Not every librarian will agree about the conclusions, and not every library should allocate its space in precisely the same way, but I believe that’s how the decision is best framed. So … about that coffee shop? I suppose having one in the library can support our scholarly mission, indirectly, in that it gives researchers and students space to gather and share ideas. What about using the space to preserve the best collection of books we can; does that serve scholarship and learning? In my opinion, the answer is absolutely yes—even if some of those books are rarely read and are largely being held in trust for the future. That’s why I aver that the books need to stay.

You claim, correctly, that removing books throws up barriers in the way of our patrons, who will be inconvenienced by having to travel to distant libraries, request interlibrary loan items that might be impossible to procure, or browse inadequately indexed online or microform versions of the discarded print materials. This is a legitimate concern and should not be dismissed lightly: “save the time of the reader” is still a fundamental value of librarianship.

But simply asserting that we have a duty to bear the cost of maintaining large numbers of unused books in our collections amounts to an unfunded mandate. In an ideal world, of course, we would build facilities as large as necessary to keep up with collection growth while also providing adequate space for user activities that reflect the particular library’s mission and values. In this world, though, library space is limited, and often, we must decide between space for books and space for users.

At the University of Oklahoma’s engineering branch library, where I worked from 2003 to 2006, we faced a critical shortage of shelf space. If I had chosen to keep every book, we would have had to eliminate most of our already minimal patron space. Instead, I transferred some books to the main library (in effect, creating a space problem for someone else) but also weeded aggressively. This was not an ideal solution, but nonetheless I am proud of my decision. The user spaces I “saved” are used heavily every hour that the library is open and are the site of a thriving undergraduate tutoring program. I feel confident in asserting that this space is a greater contribution to student learning than the many 1980s-vintage WordPerfect 4.1 manuals I discarded.

At libraries with acute space shortages, we are forced to choose between spaces for books and spaces for users. Keeping every print volume by default is, in fact, a choice to disregard the knowledge receiver—the library user—in favor of the knowledge vessel.