commissaire inspector dottore

academic and personal thoughts on detective-inspector novels

essays may contain spoilers

on bibliography

Bibliography is just about the most out-of-date thing an academic can do.

Not that it was ever cutting-edge. When I was in graduate school (1979-83) we thought of bibliography as a pis aller, something you did if you didn't have the chops to do literary criticism or theory. Mind you, we used bibliography constantly, and depended on the information generated by bibliographers to study our fields and write our criticism or theory.

We were spoiled because so many people were doing bibliography, of course. Research libraries hired humanities bibliographers whose life's work was the building and documenting of text collections. Academic presses published print bibliographies. Professors compiled those bibliographies; they counted for tenure; dissertations could be made of such projects. Major libraries did in-house bibliographies of their resources. I worked at the reference desk of our research library when I was a graduate student, and at one point, compiling a guide for our users, I annotated an annotation of a bibliography of bibliographies of bibliographies. This was a surreal experience, but in retrospect it was probably a useful thing to do at the time.

Bibliographical work has largely been automated in the ensuing 35 years. You can turn on your device – I guess, if you're reading this, you already have – and within a few minutes you can assemble an ample, customized list of citations on any topic you can think of, and many you can't. Yet the quality of these instantly-generated reading lists varies inversely to the ease with which machines create them.

At one absurd extreme is the general search function at my university library, which invites you to type in a term and then returns a massive list of items: most of which the library doesn't hold, some of which it can't retrieve, many of which consist entirely of citations to other items on the same list, and none of which has been thought over by a human being to see whether it merits inclusion.

This is not going to be an anti-technology screed. Neither is it going to be a "there's always a tradeoff" piece of equanimity. I think it's possible to have the digital advantages of speed and power combined with the human advantages of curatorial attention. It's just that somebody has to put in the work. Relying on algorithms for access to information means that, much of the time, you don't know what you're getting or what you're missing. But it should be possible for people to supplement algorithms with the kind of bibliographical scholarship that prevailed till recent decades.

Bibliographies are notoriously, soul-sappingly, perpetually incomplete. Even in fields where all the texts are archived and all the authors safely dead, bibliographies can never be finished. Of the making of many books there is no end, and of the making of lists of many books there is scarcely a beginning.

They are also inevitably full of errors. I've consulted a lot of books in the making of cid, but I have also drawn from many library databases, and the first rule of library databases is that an astonishing amount of material in them is incomplete or incorrect.