Nonfiction books generally come with an index somewhere in the back. You’ve seen them. In It’s All Relative, it comes after “Acknowledgments”, “Appendix”, and “Notes”, but before “About the Author”. It begins alphabetically like this:

Abrams, J.J., 38

Adam, 222

Topic, page number. Topic, page number. A standard, trustworthy format. Occasionally it becomes florid in its prose and detail:

ancestors:

  AJ’s skepticism about, 72

  do they really want their stories told, unvarnished? 27–30

  earthy descriptions of, 22

But the point of an index is to allow you to look up a section of the book by topic. Suppose that a month after you finish a book you think, Hey, there was that funny bit about a horse and a bagged lunch. I want to read that again. Without an index you’d have to skim through the entire book to find the bit you remembered. With an index, you only have to look up horse in the index, and it’ll direct you to all the pages on which a horse was referenced. Indices are alternate pathways into a book, side windows that open the book up to nonlinear exploration. Even in the age of Internet searching, they’re very helpful.

Lately I’ve been asking a question I’ve never thought to ask before: Who writes indices?

My first thought is that it’s the author. She knows her book best. For her final manuscript, she’ll comb through her whole work (which she’s already done several times in her editing process) and she’ll list all the topics and salient mentions of things that she imagines someone would want to look up later. Then she spends a few days on epic doses of caffeine, matching each topic with a page number, and arranging all of that data in the Standard Index Format. After this point she’ll have to be very careful editing the actual text of her manuscript, lest the page numbers shift and she’ll have to go through this whole sordid process all over again.

My second thought is, why does it need to be the author? Anybody in the publishing chain should be able to read the manuscript, take notes, and arrange her data into an index. It may not be the same index that the author would have come up with, had she done her own index; different topics may be included in the list or overlooked. And not all authors, having poured countless nights of toil and sweat into a book, additionally have the patience to do the clerical work of arranging an index for it.

Ancestors: AJ’s skepticism about, 72. Here A.J. (styled tersely as AJ—every punctuation mark in an index takes space!) is rendered in the third person. It’s not My skepticism about. Either A.J. is writing the index impersonally, or someone else is writing the index for him.

Maybe index-writing is a task delegated to people low on the publishing totem pole, junior editors or aspiring writers. Maybe they’re screened for a high score on “Attention to Detail” on the Standard Publishing Industry Diagnostic Assessment, and then they’re set to work, indexifying the manuscripts laid out on their desk on the fourteenth floor of a downtown New York building. Maybe they work with other indexers, comparing lists of notes and keywords and topics. Maybe their combined efforts are delivered to the author for proofreading before the entire thing goes to print.

Some works, by their nature, are more difficult to index. Fiction, generally, is written as a narrative and meant to be consumed from front to back in a linear fashion. I don’t think I’ve noticed any indices in works of fiction. After the main text, the book usually ends, although there might be an About the Author paragraph, or a More Books by the Same Author page, or possibly a teaser chapter for her next book. If I want to find, for example, that spot in the story where Jill Colborn trades awkward banter with Tiffany the Earth Specialist, I’ll just have to read through the story again, skimming for clues about where I am in the plot, because the author isn’t going to provide an index entry for that.

I’m not sure if The Dictionary of Obscure Sorrows counts as a work of fiction. Although the words it defines are concoctions dreamed up by author John Koenig, it does style itself as a reference work. And it does have an index, which consists of all of its words listed in alphabetical order with the page numbers on which they are defined. But if I wanted to look up that disquieting pang of realisation that your importance to somebody else may not be equal to her importance to you, I could not do that with its index. I’d have to conduct a linear search, definition by definition, to find the word I’m looking for (falesia).

David Belaire’s A Guide to the Big Band Era, a reference book I bought in The Last Bookstore in Los Angeles in March 2016, is more curiously structured. Its main section lists the hit songs on the radio for each season from 1935 to 1946 with commentary, so it is, in its own way, an index for popular culture. But it also has its own index, Index of Songs and Instrumentals, which lists the records in alphabetical order with their artists and years of popularity. A meta-index, perhaps. A second-order index for popular culture.

Theoretically, you can have arbitrarily many layers of indices, each index pointing to the next larger index until finally the last one points directly into the text. This appeals to me in a Douglas Hofstadter kind of way.