Stephen Ullstrom

I was recently talking with a new indexer who was struggling with how much detail to pick up and include in the index. The issues were, how granular the index should be and how much of the author’s argument should be revealed in the index. This discussion got me thinking about the common advice of, “Don’t rewrite the book.”

At face value, this is good advice. It speaks to the index’s role in directing readers to information, and an index that is too large will take up more space (what a publisher probably does not want) and more importantly, will likely be more difficult for the reader to use. An effective index should provide just enough information.

In practice, however, just enough can be difficult to gauge. What constitutes rewriting the text can also provoke clashing opinions. So I want to consider these questions and see if some rules of thumb can be derived. 

On the sparse, no-rewriting end of the spectrum, simply directing readers to information without explication is usually sound advice. There may be a damning political scandal behind the entry for SNC-Lavalin, but the reader can discover that on their own. If all entries are like this, however, I think it can be difficult for readers to understand how terms are connected to each other. Such an index lacks context. I believe that readers should be able to receive a rough sense of the book’s argument and contents from the index, so there is space for some light description.

The main place for this are the larger entries, by way of subheadings. Breaking down long strings of locators is helpful for the reader, and by definition will require telling the reader something about the contents of the book. I think it can also be helpful to signal to the reader which are the more important topics by means of a few subheadings or cross-references. This does not have to be extensive, and I see it as part of guiding the reader through the contents of the book. 

Clarity should also be a guiding principle when wording main headings and subheadings. The significance of each term does not need to be spelled out, but will the reader at least understand what the term means? The temptation, to avoid rewriting, is to be brief, even to the point of being cryptic. While brevity is ideal, I think it is fine to add a few more words if that is what is needed for clarity. 

The last consideration is the audience and the purpose of the book. For books that are intended for reference, where the index will be the first point of entry, an extremely detailed index, to the point of almost rewriting, is probably ideal. Whereas books that the reader is likely to actually read first are probably fine with a lighter index. 

As with any advice, “Don’t rewrite the book” needs to be contextualized to the specific project. What constitutes too much detail will vary. And regardless of how much detail is enough, the index entries need to be clear. Lack of clarity will derail any index, detailed or not. 

Last week I wrote about internal consistency in an index, and how internal consistency can help make for a smoother user experience. But external consistency is also important, for the same reasons. The index should align with the text. 

Most obvious, of course, is alignment of content. The index should strive to neither misrepresent nor misinterpret the text, or in more extreme cases, neither ignore indexable content nor make up content. But that is not my main focus today. I am more concerned about formatting, term selection, and structure.

Spelling

One area for consistency is spelling. It is simple, yet for me, at least, easy to overlook, given how much I take certain spelling conventions for granted. As a Canadian, I usually run into this issue with American spelling in indexes for books published by American publishers and for Canadian books intended for an American audience. If I do not have a style sheet for the book and am not sure which is used, I will run a quick search for certain words which are commonly spelled differently. Spelling can be subtle, but I think alignment is still important to give the book a unified feel. At the very least, this saves the proofreader some work, if they are alert to this issue.

Locator Ranges

Locator ranges are another area for consistency. Some publishers have a preferred way to abbreviate (or not), while for other publishers it seems to vary from book to book. If you are not sure, check the bibliography or keep an eye out for ranges as you index.

Term Selection

Try to follow the text’s lead for names and terms. This can be tricky if the text itself is not consistent, but try to determine what is used most often and what the audience will be most familiar with. Cross references and double posts can also be used to add alternative entry points. We cannot assume that the reader has already read the book, but mirroring the terminology can make the index a bit easier to use for someone who has.

Structure

Following the book’s structure is not always good advice as the book may not have a clear structure or it may not be easy to mirror. So use your judgment and choose a structure that works best for the index. If the book’s structure is amenable, however, then be consistent in importing that structure. This will help the reader who has already read the book, and might help orient the reader who hasn’t. 

When considering external consistency, we are recognizing that the index is part of a larger whole. While the index has its own conventions, it also needs to take cues from the text that the index is pointing towards. The reader should be able to recognize that the index and the text belong together.

What are your thoughts on external consistency? Are there any areas that I missed?

Cutting an index to fit a tight space limitation is never fun. At least, I do not enjoy it.

By cutting, I mean shortening the index beyond what I think the index should be. Editing may lengthen or shorten the index, depending on what needs to be done to make the index its best self, while cutting is simply about removing to fit an external standard. 

Yet, I think cutting an index can still be done well. The final index may not be ideal, but it can still be elegant and appear as a coherent whole. To do this, I have developed a form of triage to guide how I cut. 

The word triage comes from the medical world, and refers to making treatment decisions in situations of insufficient time or resources. In a way, triage is heartbreaking because it means that not everyone will receive treatment or treatment in a timely manner, yet it is effective because it forces the medical personnel to start with the patients for whom they can make the biggest impact.

For an index, my goal is to deliver the best index I can by using the most judicious cuts. Which cuts shorten the most while impacting the quality of the index the least? To do this, here are five strategies and considerations I keep in mind.

1) Identify What Needs to be Kept

This step seems counterintuitive because the goal of cutting, after all, is to remove. Yet I still think it is important that the core entries remain largely intact. I want the reader to still get a broad overview of the book from the index. This means identifying the metatopic entries and supermain headings, which cover the main points of the book. Once I have identified these, I may combine or shorten a few subheadings here and there, but mostly I cut around them.

2) Remove Subheadings

I do not like long strings of undifferentiated locators or mixing subheadings with unruly locators, but I will make exceptions to cut an index as subheadings do take up space. My usual bar for when to add subheadings is six or more locators, but if I need to cut I will raise the bar to say ten locators. I will also remove subheadings for less important entries, as a tradeoff for keeping subheadings for my core entries.

3) Eliminate Multiple Entry Points

One of the hallmarks of a good index is that a reader can find the same information in two or more ways. When I cut, I see multiple entry points as a luxury. My goal is to still have at least one entry point for each piece of information, but I will look for ways to cut out repetition. The main way I do this is to remove double-posts. If information is presented as both a main heading and a subheading, I may choose one or the other. I may also use more cross-references to compensate, if it seems like the cross-reference will take up less space.

4) Raise the Bar for Passing Mentions

Another strategy I use is to raise the bar for what I consider to be passing mentions. I actually start thinking about this before I edit the index. If I create an entry that I think I might be willing to cut, I label it so that it is easy to later find. You may ask, why not just cut if the entry is borderline? I would say that there are degrees of passing mentions, and what may be irrelevant to one person is valuable to another. So I tend to err on being inclusive, while recognizing that I may need to change my standard in order to cut.

5) Be Consistent

Regardless of which strategies I employ, I try to be consistent across the index. If I am cutting an entry from a certain type or level of entry, for example, I try to do the same for all entries of that type or level. I do not want the cuts to appear random or for a reader to say, Why is this entry included but not that? I know that a cut index is full of holes, but at least by being consistent I hope to mask the damage and make it appear that the index was still deliberately and thoughtfully made.

Ultimately, cutting an index is about tradeoffs. What gets to stay and what needs to go? These are not always easy decisions, especially if the cuts occur after careful editing. Still, with some planning, I think a decent index can still emerge. The index may not be great, but it can at least still be okay and still serve the reader as best it can.