Some Thoughts on Indexing

N. Sivin

Draft 1


I indexed my first book, but after it came out and I compared my index with those in Science and Civilisation in China, I had to admit that a professional could do better than I had done. For several books after that, including those I edited, I was lucky enough to engage Muriel Moyle, who did SCC’s indexes. She was probably the best indexer in England, as well as Dorothy Needham’s sister. She was a very old-fashioned lady, probably born about 1900. Among her many virtues, she would not accept more than ₤1 an hour for her work. At that stage, one made indexes on index cards, alphabetized them by hand, and then typed them into a manuscript.

When I finished Traditional Medicine in Contemporary China, in 1984, I tried indexing again. By that time I was used to writing on a computer, so cards were no longer necessary. I made that index partly because the book was so technical I thought an indexer would have trouble understanding it, and partly because I wanted to include all the technical terms, with definitions. It took me more than three solid working days, plus time for checking. By that time I had studied a number of indexes, and had a clear idea of how to proceed. The two 1995 volumes probably took me about a day each. The indexes I prepared were ordinary Word files that looked exactly like the final product--in fact both volumes were photographed directly from the Word files I prepared.

With the new astronomy book, Springer asked me to use the indexing facility built into Word. It involved inserting index entries in invisible fields right in the text. Where the name Shen Kua appeared, I would put “Shen Kua” into a hidden index field. If I wanted a sub-entry (e.g., “education”), that went into the field too. That was more trouble, so it took three or four solid days, with a lot of final fixing up. Once all the entries are in the file, it takes less than a minute to automatically generate the index.

It is much easier to do this if your whole book manuscript is in one file. If you have a reasonably recent computer, with at least 1GB of memory, that is an extremely convenient way to edit it. Just copy all of the original chapters into one file. It may be 2 to 10MB long, depending on the size of the text and the number of tables, graphics, etc. Word will have no difficulty handling it. On the other hand, the “Master Document” feature of Word is buggy and unreliable.

The point is that Springer, in addition to publishing the book on paper, also publishes it online in its Springer Link facility, to which many libraries subscribe. The electronic versions are divided into chapters; having the index entries built in makes it no more work to generate an index for each chapter. A look at my Granting the Seasons will give you an idea how the index turned out.

It is natural to think of indexing software that can simplify the job for you. Such software exists, but it is a waste of money, and will not save you time. On the other hand, you may want to look at Word’s built-in indexing facility. It is one of the many remarkable things that Word can do that most people do not learn about because they don’t get a good manual and read it (the best manual I know of is Special Edition: Using Microsoft Word <version no.>).

The Work Involved in Various Choices

The advantage of doing an index yourself is that you know the book, and what you want in the index, better than anyone else. The disadvantage is that you have to learn a process more demanding than, say, making a bibliography, and it takes time you may prefer to spend on research or with your family. Good professional indexers can make a basic index with little or no input from the author. If you want to control the process, you can do it more efficiently than if you do the dog work as well.

What do I mean by control? Take a good look at the index to The Way and the Word. You will see several characteristics of an index designed to be as useful as possible to scholarly readers:

1. When there are more than a couple of entries under one item, they have sub-entries to help readers find what they are looking for (look at “rhetoric”). Nothing is more discouraging than an entry with nothing but a dozen page numbers after it.

2. Every name of an ancient person includes his dates and, if you wish, brief information on what he is known for.

3. Every entry for a Chinese term gives characters and a translation; those for Greek terms include translations. This makes the index useful as a glossary.

4. There are adequate cross-references so that someone trying to look up the wrong term will find the right one, and people who look up an ancient book by its translation are directed to the title in the original language.

No indexer without help will be able to work out all such entries, when they should appear and what should be in them. But you can mark up a printed copy of your MS, putting in the margins what should go into such entries.

If you study that index, and look up some of the entries, you will get some idea what is involved.

Making a choice

Unless you have plenty of free time, I don’t suggest you make the index for your first book. Publishers usually have a list of professional indexers they recommend. When you contact one or two, ask them for a list of, say, ten books they have indexed, and look at a couple of them. See if they are the kind of index you want. These days the professional rate for a 300-page book is about a thousand dollars or so, depending on what detail you want. It is normal for the publisher to charge the cost of indexing against your royalties, so you don’t have to pay the money up front.

As an experiment, read the long chapter on indexing in The Chicago Manual of Style, which most authors have. Then spend one full working day indexing as many pages as that gives you time for. Start with page one and type the entries one by one into a file, in the order that they come up in the book. That will give you an idea of how much work is involved, and help you decide whether it is practical to do the job yourself. If not, you can also send this fraction of an index to the indexer to give him or her an idea of what you want. If you want a scholarly index like the one in The Way and the Word, you will also need to read through a printed copy of the manuscript and write cross-references and other data (say, authors’ identifications and dates that do not appear in the text) in the margins.

By the time you have carried out this experiment and your book has come out, you should have a good idea of what you want to do with your next book.

Finally, whether you make your index or have it made, I suggest that you specify to your editor at the press that the running header on every page of the index should specify what letters are included (e.g., FI–GR). That makes it immensely easier for readers to find what they are looking for in the index. I am still annoyed at Yale University Press for blankly ignoring my directions to do that. If a publisher tries to limit the number of pages in your index, as some try to do to save a couple of dollars, simply tell them you insist on a useful index and ignore them.

Final Steps

1. Keep in mind that changes which result in repagination have to be finished before you begin making the index. That means checking everything thoroughly in advance. For one thing, if your file includes page cross-references, or a table of contents, it is a good idea to update them just before you begin work; the simplest way is to select the whole MS (ctrl-A) and repaginate with F9.

2. Your index, once you generate it, will need to be carefully proofread. For instance, characters in the index may not always be readable until you reformat it using your regular Chinese font.

3. When everything in the index is satisfactory, so you can be sure that there will be no further change in page divisions, you can modify the running heads of the pages to show what letters are indexed on each page. This is not difficult, but you do have to learn how to do it. Read up in a good Word manual about getting rid of the “same as previous” link in heads; you will need to unlink all of those in the index, and get rid of the “separate first page” choice in the index formatting dialog.

To conclude, you will probably find it is worth while to index at least one of your books, to learn how the process works. But your first book is probably not the best one to begin with.

Copyright reserved by Nathan Sivin. Please email
nsivin! at sas! dot upenn! dot edu! (but use normal form and omit the exclamation points)
with comments and corrections.

Back to Nathan Sivin's home page Back to the Department of History and Sociology of Science home page

Use your browser's "back" button to return to your last location.

Last Modified 2008.6.12