Article
September 12, 2019
0 comment
by Stephen

Index Subheadings Q&A

Subheadings are, I would argue, essential for a good index. Not everyone would agree, apparently, as evidenced by the indexes one finds with long strings of locators with perhaps the odd subheading tacked on at the end. But as a way to deliver additional information to the reader, and to break up and add meaning to those long strings of locators, subheadings are the primary tool available. Yet subheadings can also be poorly written or included, which can sabotage the whole purpose of using them in the first place. While I realize that different indexers vary on how they use subheadings, here are some of the common considerations I face about subheadings, and how I resolve them.

How many locators must an entry have before subheadings are needed?

My rule of thumb is about six locators before I add subheadings, but I also consider the context before making my final decision. For trade books I might allow up to seven or eight undifferentiated locators, under the assumption that trade books can often have lighter, less detailed indexes. I also consider how important the information behind the locators seem to be, as well as how important the entry is in the context of the index. It also depends on what type of information is being conveyed. A distinction between almond milk, soy milk, and goat milk, as subheadings under milk, for example, seems important to make even if there are only four locators.  Alternatively, there could be nine locators on the same topic with very little to distinguish them, which would make subheadings seem forced. Ideally, my upper limit for undifferentiated locators is ten. 

Going back to those nine similar locators—what if there are fifteen and breaking them down isn’t practical or possible?

I once faced this problem in a book of local history in which a few people, still alive, had been quoted extensively, with like fifteen or twenty locators each. On the one hand, I felt that it was important to index these people, as they were alive and members of the community. On the other hand, they were rarely the focus of the text, instead offering little comments here and there. For that reason, and also for space, subheadings did not seem appropriate, as it would have resulted in long lists of one subheading per locator. My solution was to use a fake subheading, appended to the main heading, along the lines of,

Smith, John, reminisces of 

The locators were still undifferentiated (we do not know what John is remembering), but I think this worked better because the reader was still provided with some additional information about what type of information this was. I think the worst scenario for a long string of undifferentiated locators is to provide the reader with no context, so if all of the locators are similar in some way, it is worth adding a little extra to the entry. 

The space constraint is really tight. Should I just delete all the subheadings?

No. Some subheadings will need to be deleted, but try to do this in a controlled, mindful manner. I like to call it performing triage on the index, which I have written about previously. Basically, my goal is to keep the most important entries as intact as possible, while making most of my cuts on the entries that are less important. Cutting this way will mean more entries with undifferentiated locators, but done right, a good number of subheadings should still be preserved and the reader should still be able to get a good sense of the book from the index.

Is there a magic formula for writing subheadings?

One of the problems with subheadings is that they take up a lot of real estate. So one temptation is to make subheadings as concise as possible. I value concision too, but I have come to value clarity more, even if it takes an extra word or two. Part of this is also making sure that the relationship between the heading and subheading is clear. If the reader cannot understand the entry, then the entry may as well be removed as it is not fulfilling its purpose.

I also try to write subheadings in plain language, alongside the author’s terminology. I don’t want to assume that the user of the index has already read the book or is familiar with the jargon, though this does vary from book to book, depending on the intended audience. For me, using plainer or simpler language is part of making the index accessible.

Do you mix subheadings with undifferentiated locators?

Ah, probably the most contentious question of them all. Personally, I do not like mixing subheadings with undifferentiated, or unruly, locators. Ask a handful of indexers what those orphaned locators signify and you will get a handful of different answers. If indexers can’t agree, I suspect readers will probably be confused too, so I’d rather have no subheadings than have a mix.

I do have two exceptions that I allow. The first is to start the entry with a broad range, such as for a whole chapter, which is then broken down in the subheadings. This tells the reader (I hope) that there is a significant chunk of text on the subject and if they want to start on the first page and read through, they can. The second scenario, which I may do if space is tight, is to leave locators for illustrations undifferentiated, though they would be indicated by the locator being in italics or boldface, for example, so they are still, in a way, differentiated. 

But what about introductory material in the text? Does that really need to be broken down into subheadings?

Introductory material can be tricky to index, as it is usually brief and covers a lot of ground. This could be a paragraph per chapter in the introduction of the book, along with a couple of pages at the beginning of each chapter, as well as possibly a couple of pages at the end of each chapter, to sum up what has been discussed. I have heard some indexers say they leave these undifferentiated, with the assumption that readers will recognize this as introductory material. I don’t want to make that assumption, so I instead gather these into a subheading labeled something like, “about,” “approach to,” “introduction,” introduction and conclusion,” or “overview.” The exact wording varies according to what seems most appropriate for that particular text. I then force sort this subheading to the top of the entry, so readers can start with the introductory material, if they want. 

What are some of the common questions or issues you face? Do you have different solutions to these problems? Feel free to reply and let me know.

Article
July 17, 2019
0 comment
by Stephen

Translating the Text for the Index

Words matter in the index.

This may be stating the obvious, but I think it is worth repeating. The terms and phrasing used need to be readily understood by the user of the index. 

The text, of course, is the first place to look for terminology that can be transposed into the index. But as indexers, we also need to be careful about what we use from the text and how. Not all books contain a wealth of terminology that is suitable. 

I often see two main scenarios when alternative terms should be used.

The first is when the author’s terminology should and can be used, and it is also mostly jargon. If we can assume that the audience of the book will also understand that jargon, then there is no problem. But sometimes the audience will include non-specialists as well, and for that audience I think it can be helpful to include other terms, probably as cross-references, for these other readers. I also try to write the subheading using plain language, so that the concept or reference is understandable without having first read the book.

The second scenario is to put in the jargon that the author either does not use or buries. I like to call this translating the text. It is when the author, for the sake of being creative or engaging the reader, uses descriptive language that is enjoyable to read and conveys the concept, but the actual term or phrase that captures the aboutness is either missing or not obvious. This can happen throughout the text, from the title to chapter titles to headings, as well as how sections and paragraphs are written. To give a simple example, I once indexed a book titled Igniting the Internet, which to me sounds like it could be about viral social media posts or celebrities online. Instead, as the subtitle tells us, it was about Youth and Activism in Postauthoritarian South Korea. There are times that indexing requires us to read further or read between the lines to understand what the text is about.

There is the concern of taking this translation too far and of introducing too much new terminology. We do not want an index that is divorced from the text. This is where cross-references play a role, to link and point towards preferred terms. We also need to make sure that any terminology we introduce will make sense to the audience, rather than introducing terms that make sense to us. For one book I googled sales terminology for a sales book that employed a highly descriptive writing style. It was not terminology I would ordinarily use, but the terminology fit the book.

So when indexing, pay attention to the language used. Match the book if possible, and do not be afraid to do some translating if need be. The reader will only be able to find information if they can first understand the index.

Article
July 10, 2019
0 comment
by Stephen

Editing the Index on the Go Redux

Last week I tried to puzzle my way through the concept of editing the index while the entries are being created, as opposed to saving most of the editing for the end of the indexing process. As I wrote, some indexers say that they save a lot of time by working like this, which I accept in principle while finding that it can be difficult to implement. 

Coincidentally (I promise I hadn’t recently checked the conference line-up), the first speaker at the Indexing Society of Canada’s conference in Ottawa last week was Anne Fifer, with a presentation titled “Edit Your Index Without Pain.” I did find Anne’s talk helpful for clearing up some of my questions and misconceptions, so I am going to revisit the topic with some of Anne’s insights. I thought you might find it helpful too.

First, to set expectations, editing is still required at the end of the indexing process. It is not possible to eliminate editing entirely, but the amount can be reduced. Anne said that often she just needs a day for the final edit. 

Anne’s approach seems to be two-fold. The first part is to use the full capabilities of her indexing software to streamline indexing. Specifically, extensively using labels to highlight entries that need different sorts of editing or extra attention, so that these are easy to find later on. Anne also discussed custom dictionaries for accurate spell check, which I can see to be helpful if one is working on a lot of medical indexes like Anne does, but I don’t really see the benefit for books that lack such specialized terminology. I don’t see myself making custom dictionaries for the books I typically index.

The second part to Anne’s approach is to lay the necessary groundwork for the index at the start. Part of this is making sure that all necessary resources are in easy reach, which is a good reminder that the details of our work environment can make a difference. The largest component is to establish the index structure from the start, which Fred Leise discussed in more detail in his talk (Fred also has an article in The Indexer 34(4)and has presented a webinar with ASI on his approach to structure, if you want to learn more). The benefit of this is that once established, the structure does not need much editing later on, if at all. It also provides the framework to slot entries within as one works through the text. Thinking through the structure upfront was my biggest takeaway from the conference, and something I will be adding to my indexing practice going forward.

Anne also discussed editing in small increments. Part of this is doing a small edit at the end of each day for a specific list of common issues that she looks for. Part of this is also isolating and editing in batches the previously labeled entries. This means that a lot of spelling, formatting, or other common issues are already dealt with by the time Anne reaches the final edit. As I wrote last week, I find it difficult to focus on editing at the end of the day, but I like the idea of being self-aware of issues I commonly need to fix and of proactively fixing them in batches. This could also be done at the start of the day, for example, if I don’t want to do it at the end.

All told, I think Anne has some practical tips to make editing easier. I hoped you learned something too, and are able to become more efficient in your indexing.

Article
July 3, 2019
0 comment
by Stephen

Editing the Index on the Go

With our income as indexers often proportional to the amount of work we can complete, efficiency is key. I want to consider today a practice that I know some indexers use, which is editing as the index entries are created. In theory—and I have heard some indexers say they manage this in practice—by the time you reach the end of the book, very little editing remains to be done. 

I have to admit that while I try to do this, I still tend to split indexing into two distinct phases. There are aspects of the editing-in-the-go approach that I struggle with. So this is more of a self-reflection than a how-to. Maybe you will get some ideas for what works for you. 

There are definitely a couple of advantages to editing as you go, which I try to keep in mind as encouragement. The main one for me is that the content is dealt with while it remains fresh in my mind, which means less going back and double-checking later in the indexing process. This should save time during the final edit, as most of the issues in the index will have already been dealt with. 

Still, I find there can be some barriers to putting this approach into practice. I often find that creating the entries and editing the index requires me to think in different ways, and it can be hard to switch back and forth between entry mode and editing mode. After several hours of entering entries, having to go back and edit can also seem too taxing for my tired mind. And sometimes, especially for more complicated texts, I find I don’t really understand the text until I’ve had a second look at a later date. In this instance, editing is very much part of my process for understanding the text. I also suspect that editing on the go requires a clear sense of what the final index should look like and what constitutes a good final draft.

When I find editing-as-you-go most helpful is for books that are organized into discrete chapters or sections. If I know that armadillos are primarily discussed on pages 34-63, then a quick edit of that one entry, before moving on to the next section, does save time later on. Unfortunately, not all books are this well organized, and a quick edit is still needed at the end to make sure the entry still holds together. Also, something I should do more of, and to avoid the excuse of tiredness, is to do a quick edit after every chapter, to clean up any obvious errors or redundancies. 

Another form of editing which can be done early in the indexing process is to establish consistent wording, particularly for elements like glosses. I may not always know what I want to use right away, but once I do, it can be worth taking a moment to go back and fix, and then be consistent going forward. This can also be done to consolidate subheadings, if I decide partway through to change the wording or how I am gathering locators.  

Otherwise, I have to admit that if I am not certain about a particular entry, I tend to just keep going. I will put in my best guess or use some sort of placeholder wording, and I may write a separate note for myself about the issue, and then I will come back later during the editing phase. There are elements of indexing on the go that I use and can better incorporate, but I do like the editing phase. I find that I view the index differently while editing, which is valuable in its own right. 

How do you index? Do you edit as you create the index entries? Why do you do it, and what works best? Let me know, if you have a moment. I am curious to know what works for you.

Article
June 27, 2019
0 comment
by Stephen

Why Correctly Identifying the Metatopic Matters

I have been reluctant to write about metatopics. It is already extensively discussed by Do Mi Stauber, Janet Perlman, and Margie Towery in their books, as well as, I assume, elsewhere. Appropriately handling the metatopic is also one of the trickier aspects of indexing to master. I am not sure what I can add to the discussion that is new. I also suspect that reading about the metatopic only goes so far. The best learning comes from taking theory and applying it to real books, though perhaps that is simply how I learn best.

Still, I was recently talking with a new indexer who was struggling with the metatopic, and a misconception came to light that might be worth highlighting. Specifically, this new indexer was feeling overwhelmed by the multiple metatopics which she had identified in the book. The misconception was that the metatopics identified were actually main topics, or what Margie Towery might call supermain headings. These are clearly important topics that need to be in the index, but they do not encompass the whole book, which is what differentiates main topics from metatopics.

I like what Janet Perlman says about the metatopic in her book, Indexing Tactics & Tidbits (Information Today, Inc., 2016):

“…The metatopic is intertwined with aboutness. When the indexer can make a cogent statement of what a book is about, whether it be a word, a phrase, or a sentence, he has identified the metatopic of the book.”

There is usually only one metatopic in a book, sometimes two if a book has dual subjects. It may be a complicated metatopic, requiring a sentence or three to explain, but it should still be a single topic or concept that encompasses the whole book.

Correctly identifying the metatopic is important because this impacts how the metatopic is handled in the index and how the whole index is structured. This new indexer was having trouble because she was trying to follow the advice to disperse information away from the metatopic entry. This is generally good advice, and the reasoning is, since everything in the book, theoretically, falls under the metatopic, all main topics should stand on their own as main entries. In this approach, the metatopic should be a signpost to the main entries via cross-references, and should only gather general information that are not able to stand alone. It is usually not a large entry. This approach works fine when there is a single metatopic, but not if you are trying to disperse information for multiple topics. As this indexer learned, if you are trying to disperse all of the information from all of the main topics, the index quickly loses all sense of structure and some information simply does not stand alone very well. 

The flip side to dispersal is gathering, and while gathering the whole book under the metatopic entry is not practical, gathering under individual main topics is. This is why we need to distinguish between the metatopic and supermain headings, because these different types of information and entries are handled differently. As I have discussed previously, entries should contain or direct readers to everything they need to know about that topic. Overly large entries can be split up into smaller chunks that are still meaningful, but we should see dispersal and gathering as complementary techniques for structuring the index. They each have their role, and the starting point is correctly identifying the different components of the book and how those components translate into the index.

The metatopic plays key role in the index, but only if it is first correctly identified.

Article
June 5, 2019
1 comment
by Stephen

Who to Serve: the Client or the Reader?

Who is the index for?

This may seem like a trick question. You are probably thinking, “The reader, obviously.” So let me rephrase the question.

As the indexer, who are you trying to please?

I hope you are still trying to please the reader, but when writing the index, the first person or people to see and approve the index are often not the book’s readers. 

What I am getting at is the distinction between the client and the reader, and how the needs and wants of each may not be fully aligned. 

As freelancers, we primarily work with the client, which can be either the author, editor, or both. If you are the author writing your own index, this may not be an issue if you are self-publishing, but if you are working with a publisher, you may be working with an editor or proofreader on the index. 

Ideally, our vision for the index aligns with the client and the reader. But it can happen that the client wants something different. What happens when that point of difference violates indexing best practices? What do we do when we think that the reader will no longer be well served? Where does our ethical obligation lie?

There is a line of thought in business that the client is always right. So long as we get paid for our work, and the client is happy enough to come back for our services again, who really cares what the end product look likes? The client is paramount.

What this line of reasoning misses, though, is that the index, in this case, is not actually for the client. Sure, the client is paying us, but the people who will actually use the index are the readers. 

Now, I am not suggesting open conflict with our clients. Money is persuasive, and I like being paid as much as anyone else. At the same time, keeping the ultimate audience in mind, I think we should be willing to push back a bit when clients request changes that are detrimental. Perhaps the client simply does not understand and is willing to learn. Perhaps the client does have a point, and we can offer an alternative solution that is a better fix for the problem. Perhaps there is some other way to compromise. 

Gently pushing back can also be a reminder, to ourselves as much as to the client, that we are professionals. Being a professional does not always mean that we are right, but it does often mean we have a broader understanding of the situation. We should not be afraid to make our voice known, even if ultimately we accede to the client’s point of view. This also involves knowing what is worth speaking up for, and when to stay silent.  

Ultimately, I want both the client and the reader to be happy with the index. This requires taking both seriously, and not being afraid to explain my position when needed.

Article
May 29, 2019
0 comment
by Stephen

On Double-Barrelled Entries

If the index structure can be likened to a skeleton, and the information that the index seeks to convey the lifeblood, then perhaps all the small details that shape and bind an index together can be considered the muscle and sinew. It is this attention to detail that can elevate an index. One specific type of detail, which I have been considering and using in some of my own recent indexes, are double-barrelled terms. 

By double-barrelled terms, I mean terms that bring together similar and related concepts into a single heading or subheading. The different components are usually connected by an and or a forward slash.

At the main heading level, I think this works best for concepts that are clearly related in some way. These could also be variations of a word, and if the terms would sort side-by-side anyway. Think architects and architecture, or canoes and canoeing. These terms are not quite synonymous, so if going the double-barrelled route, make sure that there is overlap in how these terms are discussed in the text. I find it also helps if each term would be a relatively small entry on its own. Otherwise, if the terms are used in distinct ways or would form substantial entries by themselves, then two separate entries are probably the better option.

Double-barrelled terms can also work as subheadings. The first situation would be the same as discussed above for main headings. The second situation (which does not work as well for main headings) is to link together related events. For example, in a family history I recently indexed I had the following subheadings: cancer and death; wedding and married life.

I think this works because it shows a clear chronological progression. This also works when the two discussions happen on the same page or across a small span of pages. Separate subheadings would have simply duplicated the information. As with main headings, it also helps if there are only a few locators to gather together. If there are a lot of locators—say an entire chapter or section each for the wedding and then married life, then separate subheadings make more sense.

If you need to shorten the index, double-barrelled terms can also be another tool you can use for merging entries and cutting down on duplication. It will probably not make a big difference, but can still be useful if just a few lines need to be cut or if used with other methods for shortening. 

Otherwise, I think elegance is the main reason for using this technique. Double-barrelled headings and subheadings can bring terms together in a way that is less didactic, and can provide a more nuanced understanding of the book than through standalone entries. Double-barrelled terms are not always applicable, but are useful to keep in mind for the right book.

What do you think of double-barrelled entries? Do you use them in your own work?

Article
May 22, 2019
0 comment
by Stephen

On Rewriting the Book in the Index

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. 

Article
May 15, 2019
0 comment
by Stephen

External Consistency in Indexes

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?

Article
May 8, 2019
0 comment
by Stephen

Internal Consistency in Indexes

Consistency is key in the quest to provide a smooth user experience. How the index works should be largely invisible to the reader. What we do not want is the reader to find errors or inconsistencies which cause them to question whether their interpretation of the index is correct or whether they are finding everything that should be available. 

In this post, I am going to write about internal consistency, which is making sure that the index is internally consistent with itself. Imposing consistency begins at the start of the indexing process, with selecting which formatting conventions to use and making initial decisions about what kinds of entries to capture. It then continues to be applied during the editing phase, when decisions are reaffirmed and double checked. 

Wording

Consistent wording is a subtle way to add consistency. This is part of gently shaping the reader’s expectations around preferred terms and how to recognize similar information.

If two or more synonymous terms are available, pick one as a preferred term and use it consistently, with a cross-reference if necessary. For a recent book I indexed on the annual encampments of the American Canoe Association, I choose to consistently use the word “meet” for each individual encampment, as in Jessup’s Neck meet and Grindstone Island meets. I could have chosen the terms “camp” or “encampment” instead, which were also used in the text, but I thought meet gave a better sense of community (an important theme in the book), it was short, and by using the same term I hoped to indicate to readers that these dozen or more entries throughout the index were all the same kind of thing. If I had also used the term camp, as in Jessup’s Neck camp, I was concerned that some readers might have questioned how a camp was different from a meet. 

Consistent wording can also be used in subheadings to indicate similar content. Using the same subheading across several entries, as such “accommodation” in a guidebook, indicates to the reader that the same kind of information is provided for all of the locations discussed. Within a single entry, similar wording can gather subheadings so that the reader can easily see all of the related information. For example,

Los Angeles: population growth; population projection

Glosses

A variation on consistent wording is being consistent with glosses. An index can have glosses for different types of entries. Within each type, though, try to provide the same information in the same order. Otherwise, I think inconsistency is visually disruptive and can cause the reader to spend too much time rereading entries to make sure it says what they think it says. For example, the following is not a good idea. Pick a style and stick to it.

Henry VIII (king of England)

James II of England

Louis VI (French king)

Louis XIV (king of France)

Victoria, Queen

Locators

Another basic place to impose consistency is with locators. Are ranges consistently identical, if used at all? Are typographical elements, such as bold or italics, used consistently to indicate figures, tables, or other illustrations?

Treatment of Topics

The last area I will discuss is the treatment of main headings and topics. Consistency here can happen in a couple of different ways. The first is to make sure that all examples of a certain type of information is picked up. To give a simple example, if I decide to index dog breeds, then I should make sure to pick up all breeds mentioned, so that the reader is not wondering, “I see the entry for Dalmatians, but what about Labradors? I am certain I saw Labradors mentioned in the book.”

Another way to be consistent is to treat similar topics similarly. If I decide that some supermain headings should have subheadings to indicate their importance and to break down a longer range, then ideally all supermain headings should have subheadings, even if some ranges are shorter. This is not always possible, if a topic is important but only discussed on 2-3 pages. But I think that consistent treatment is important, when possible, to help the reader identify similar topics and relative importance. 

These are the main areas for internal consistency that I think about when indexing. This can appear to be nit-picky, but I do think that it is the small details that can elevate an index and make for an easier user experience, without the user necessarily knowing why it is a nicer experience. When you index, what areas of internal consistency do you pay attention to? Feel free to reply and let me know. I am curious to know what other areas we should pay attention to.