Ambrose Li

Why your Chinese-language index must have headings

(updated )

How do you look up a word in an index? Let’s say the index is 10 pages.

Easy. You flip to the index, take a glance to see roughly where your word should be, then start looking until you find it or until it’s obvious it isn’t there. You can do this because you know the alphabet and you know where each letter goes in relation to any random letter.

Now imagine you need to look up a word in a Chinese-language index (Figure 1).‍[Note 1] Not so easy anymore. The Chinese “alphabet”‍[Note 2] contains more than 10,000 “letters”‍[Note 3]. No one knows the whole alphabet, let alone where each letter goes in relation to any random letter.

man¹
See (mɐn⁶).
mɐn³
See (mɐn⁶).
See .
17
man¹ 35; mɐn³ 35; mɐŋ¹ 35–36
See .
而今
60
61
See .
Figure 1. A partial hypothetical index, sorted by computer, for a reference book I have that has no index‍[Note 4]

In short, an index in Chinese looks like alphabet soup. To look up any word you will have to scan through the entire index.


So how could anyone look anything up?

If you flipped through a properly-indexed Chinese-language book, you’d notice that the index has headings.

What the headings do is to break the index down into manageable chunks. They impose order on the alphabet soup so that you know roughly where to start looking.

In short, headings help reproduce the experience of using an English-language index.

The headings do this by grouping the headwords by stroke count, the number of “strokes” needed to handwrite the first letter of the headword. The indexer puts that word under a heading for that stroke count. The reader then can look up any word by counting strokes. This is the same system used by traditional dictionaries.‍[Note 5]

For example, say you know you want to look up the word 咗 (dzɔ²). The letter contains eight strokes, so you’d look up the word under the heading “8 strokes”. Figure 2 shows how our hypothetical index might look like in a proper index:

man¹
See (mɐn⁶).
mɐn³
See (mɐn⁶).
6 strokes
而今
60
8 strokes
See .
10 strokes
17
11 strokes
See .
man¹ 35; mɐn³ 35; mɐŋ¹ 35–36
12 strokes
61
15 strokes
See .
Figure 2. The index in Figure 1 grouped by stroke count

Now you can just count and start looking. You don’t have to look through 10 pages any more.


But what if you counted wrong?

Remember that 10-page-long real index that you need to scan? Without headings you will need to scan the entire index a second time if you can’t find a word just in case you missed it. If you still can’t find it but aren’t sure you didn’t miss it again, you will have to scan the entire index a third time. That’s 30 pages of small print to count.

But with headings it’s not so daunting any more, because the counting is usually just off by one.

Say you counted nine.‍[Note 6] You wouldn’t find the word under “9 strokes”. So you’d look in the section before and the section after. Say you tried “10 strokes” first, then “8 strokes”; the word is right there when you looked under the third heading.

Unless your counting is really off, you don’t need to look under more than three headings.

Compare that to the 30 pages that you’d have to go through without headings. Having to scan 30 pages to look up any word would be torture.


That’s why you can’t take those headings out in a Chinese-language book, even if your style guide says indexes shouldn’t have headings. You don’t want to torture your reader.

Notes

  1. You can also check out this site’s index for a possibly more realistic example.
  2. It’s not technically an alphabet, but let’s pretend it is.
  3. Again, ideograms technically aren’t letters, but let’s pretend they are.
  4. This hypothetical index is based off 論粵方言詞本字考釋 [lœn⁶ jyt⁶fɔŋ¹jin⁴ tsi⁴ bun²dzi⁶ hau²sik¹, On the study of historically correct orthography of words in the Cantonese dialect], published by Chung Hwa Book (Hong Kong) in 1998, ISBN 962-231-947-5.
  5. Dictionaries designed for speakers of a specific Chinese language might organize its entries by pronunciation (this makes the dictionary useless for people that speak a different Chinese language), Some dictionaries might organize its entries by some kind of code, for example Changjei code (a system used by professional typists) or the 4-corner code.
  6. Which is perfectly possible, because contains a stroke that looks like an upside-down L, which looks like two lines but counts as one.

Tags

  • #Chinese languages
  • #headings
  • #indexing
  • #internationalization
  • #perceptible information (universal design)
  • #stroke count
  • #tolerance for error (universal design)