Choose your words carefully

44 Indexing
Choose your words carefully
In his second article, Bill Johncocks considers how the terms you put into an
index — and the ones you leave out — influence its usefulness to the reader.
Last issue, I suggested putting yourself
in the reader’s shoes when compiling
an index. This time, I want to look at the
next obvious question: what to index?
Resisting the temptation to cut corners
Casual acquaintances often ask indexers
whether they actually have to read
the whole of each book. It does sound
time-consuming but, nevertheless, the
traditional method is indeed to read
right through a document, pausing after
each section to ask what significant
information it includes (if any), whether
readers will want to retrieve it and what
terms they might think of if they did.
Today, this foolproof approach has a rival
because, with even parish magazines now
drafted on PCs, your computer can select
the terms for you automatically, can’t it?
Certainly, a computer will find all the
occurrences of a word very efficiently.
If we are indexing a guidebook, it will
instantly find every use of ‘Rome’.
These will include many significant
mentions, but also duplication and
some completely useless references
(such as ‘as we saw in the Rome chapter’
and even ‘available everywhere except
in Rome’). It won’t find synonyms (‘the
Capital’, ‘the Imperial City’ or ‘on the
banks of the Tiber’) and the first result
might look something like this:
Rome, 112, 127, 130–3, 134–48, 150,
152–184, 186, 212, 222, 236–7, 261,
267–8, 283, 291, 303, 311, 327
This is hardly user-friendly (ask
yourself how you’d like to scan 17 page
ranges to find which bus will take you
to the Trevi Fountain) but you often see
such horrors in published indexes.
To use a fishing analogy, Internet
search engines are like using a net — if
not a hand-grenade — while a good index
is like fishing with a fly: precise and
targeted (though it should yield rather
faster and more certain results). Another
entry that needs breaking up would be:
Italy, 130–186
That’s because nobody wants to scan
57 consecutive pages to find what they
want. However, you can go too far the
other way:
Coliseum,
getting there, 183
history, 184
Communicator Summer 2006
opening times, 183
plan of, 183 …
This is simply a waste of space, giving the
reader no more useful information than:
Coliseum, 183–4
Returning to the entry for Rome, it’s
usually accepted that five page references
are the most you should inflict on your
readers under any one heading while,
turning to that for Italy, I’d suggest any
range of more than four or five pages
deserves breaking up with subentries.
So what a computer produces
unaided isn’t an index at all: in fact
it’s a concordance. These have a long
history; they originally listed word
occurrences in the Bible and the first
person to publish one was — some
might think justifiably — burned at the
stake. But concordances survive to this
day and they even seem to meet the
expectations (though not the needs) of
the Internet generation.
There’s an old story of a doctor who
left to attend a conference and asked
his secretary to index the contents of
his filing cabinet while he was away by
making entries under ‘all the words you
don’t understand’! The result might well
have been useful, but it wouldn’t have
been an index. The novice indexer’s habit
of taking this easy way out is known as
‘word spotting’ and again it’s something
a computer could potentially do.
The essential point is that indexes
don’t just list keywords: they identify
and collocate concepts.
Choosing your terms
Well, that’s a lot on what to avoid. What
should you include? Here are a few ideas:
 Products covered and model variations
 Purpose and specification
 Parts
 Controls and techniques for operation
 Actions users might wish to perform
 Substrates and materials
 Conditions affecting operation
 Troubleshooting by problem.
One sensible convention is not to index
under the main subject. This is to
avoid putting nearly everything in the
same place. If your document is about
xylophones, you don’t really want half
the index to appear under X.
It seems that readers expect us to
index headings, and why not, provided
they’re informative? A section title
like ‘tuning the receiver’ is a good
candidate for an index entry although
‘The next steps’, say, wouldn’t be.
Diagram captions too are a rich source
of potential index terms. Try to be
specific and avoid phrases beginning
with weak words. The stories of car
manuals where ‘how to change a bulb’
is indexed under ‘H’ may be apocryphal
but their lesson is plain enough.
Other conventions govern the form
of the chosen word. Usually, concrete
terms — the names of things that can
be counted — are given in the plural;
nouns and noun-phrases are preferred
to adjectives, and inversion is used
only where it usefully collocates related
concepts (thus ‘eclipses, lunar’ and
‘eclipses, solar’ but not ‘holes, black’).
Some rules are so familiar that we
sometimes forget they are rules, like
inverting personal names. Children soon
learn that, to find books by Roald Dahl
in the Library, they look under ‘D’ not ‘R’,
though again the Internet has different
conventions and company names are not
inverted (so put W H Smith at ‘W’).
Similar, common-sense rules govern
alphabetical order and the forms of
cross-references and page locators…
but let’s leave those for another time.
Don’t worry if your first index is too
long. It’s far better to over-index now
and cut later, even though space is
usually tight. If you give the document
the index it deserves, then cut it down
to fit the space available, the cuts will be
rational and the result more usable. It’s
also best to edit your index after a break
of a few days whenever possible.
This has all been worthy but fairly
basic stuff: for those of you impatient
to progress faster, next time we’ll look
at the popular technique of embedded
indexing. C
Bill Johncocks is a freelance Accredited
Indexer living on the Isle of Skye, who
specialises in embedded indexing of
technical documents.
E: [email protected]
W: www.technicalindexing.com
Society of Indexers: www.indexers.org