This article from a hot, cooking scholar has the distinction of my both agreeing and completely disagreeing with it (if that’s a distinction at all). The argument the author, Rick Anderson, makes is that books with factually wrong content should nonetheless be bought by libraries because these books’ wrongness may have potential use for researchers. He writes
I’m here to argue that it’s not the purpose of a library to give people access to high-quality documents, but rather that the purpose of the library is to give people access to useful documents — or, more to the point, to connect its users with the documents they need in order to do their work.
So far, I completely agree. Research libraries have two purposes: one is to be a collection of records, the logical conclusion from which is that they should try to collect everything under the sun. The disabling factor to this natural impulse is the fact that libraries have budgets, and so even being a depository of records means that there are things libraries won’t spend money to acquire, catalog, and store. The second purpose of research libraries is to be responsive to its community, in this case, researchers. It is this dual mandate that distinguishes a research library from, say, a public library whose one overriding goal is to meet the needs to the community in which it is located.
To my mind, then, there is no arguing with the proposition that libraries should buy things that their community wants to use. However, upon examining Anderson’s argument, it is clear that he is not advocating for the purchase of books for which there are definite, somebody-at-the-front-desk-is-now-asking-for-it sort of use. Instead, from the examples that he employs and the research question he contemplates, what Anderson means is that libraries should buy books for which the librarians can imagine somebody in their community at some unspecified time needing the books for their research. For instance, he writes,
“I don’t see how a library could support research into modern political discourse without offering patrons access to books like [Ann Coulter’s] “Slander” . . . [Likewise, Michael A. Bellesiles’s]“Arming America” is also potentially useful as a cautionary example for undergraduate journalism and history students: I can imagine a professor profitably assigning students to examine the arguments made in the book, chase down some of the mis-cited sources, and explain what they find.
While I don’t dispute that librarians buy materials they think their patrons will need (what else are they going to do?), I’m of the opinion that “buy things that you can imagine being used” is not a sufficient standard for collection development. I can imagine a potential use for just about any material that exists. Instructional guides for programming languages nobody codes in anymore? A researcher working on a history of technology may need that. Another edition of a classic work that the library already has 10 other editions of? A scholar of the book may want to use it to study how the classics are advertised as being relevant for a new group of readers. Toilet paper from the Kremlin? Differential quality of manufactured goods as experienced by the political elites vis-a-vis commune worker.
The likelihood that any of these research needs actually materializes varies by the material, the imagination of the selector in question, and how accountable is the selector for her purchase decisions. What they don’t vary by is the specification that “bad” books have the potential to be useful and deserve for libraries to expend their resources acquiring them — that specification is simply too broad of a criterion to be helpful. Definitions, standards, and guidelines are powerful insofar as they provide principles for including some things and excluding others. If the set of the things they allow for is “everything”, then we haven’t gotten much farther with the standards or guidelines than we got without them.
In the same vein, to provide a workable guideline, the author needs to clearly demarcate the marginal cases, not the “clearly in” or “clearly out” examples, but the worst “in” and best “out”. In the specific case of Anderson’s proposal, the marginal cases can raise some uncomfortable questions. For instance, if Anderson is willing to countenance the inclusion of “bad books” because of their helpfulness, what is his stand on “bad books” that may harm? Should a library buy books that contain instructions on how to make explosives or other harmful weapons because some researchers may wish to study terrorism, political protests, or other legitimate topics? If yes, then how does a library balance that against concerns that it may be enabling a terrorist, violent political protest, or other illegitimate uses? Should a library buy books that endorse Andrew Wakefield’s claim that autism is linked to vaccination and gastrointestinal disease, a claim that has been thoroughly discredited in the medical profession and may well be labeled “factually wrong”? If Anderson is OK with a library buying these book because researchers may want to study “why people believe weird things“, how does he feel about the potential fallout from dropping vaccination rates and loss of herd immunity? How many “factually right” books should a library buy to overwhelm the “factually wrong” books so that library patrons cannot justly claim that they were misled?