Every now and then indexers’ mailing lists get enquiries something like this:
I want to make an index to a document on my computer. Can I get a program that will list every word in the document in alphabetical order with a page number? Then I can cross out the words I don’t want, and have an index.
Why is this a mistake? Obviously a document is made up of words. If you pull out all the words and put them in alphabetical order you have an index, don’t you? Well, no — you have what’s correctly called a concordance .
But what’s wrong with that? Let’s leave aside the fact that the concordance will be as big or bigger than the document it is supposed to be indexing. Let’s leave aside the fact that the only way to find which words you want to keep is to refer back to the document over and over again for each one. What’s the fundamental error here? It is in the assumption that only words carry meaning.
But the meaning of a document is not wholly contained within the words. It is inherent in the way that words are arranged in sentences, paragraphs and even chapters. “Napoleon beat Wellington at Waterloo” has a different meaning to “Wellington beat Napoleon at Waterloo”, and this will show up in a proper index. Where a concordance would have the bald words Napoleon, Wellington and Waterloo, a correct index might have:
Napoleon, defeat at Waterloo
Wellington, victory at Waterloo
Waterloo, Wellington’s victory at
It puts the words in a context, and alerts the user to what they can expect to find when they look up the heading.
There’s another word in the phrase above — beat. Is that going to go in the concordance too? If so it’s going to find itself mingling with drum beats, police officer’s beats, the Beat generation from the 1950s, and other beats that have nothing to do with the topic of victory or conquest.
Meanwhile, other words like victory, conquest, win, lose, defeat and vanquish are sitting
elsewhere in the concordance in splendid isolation, with nothing to indicate that they’re really talking about the same thing. A correct index, on the other hand, would boringly group all these references together under the same concept, like this:
Allied victories
Toulouse
Trafalgar
Vimiero
Waterloo
Concordances violate one of the basic principles of correct indexing: index the concept, not the word. A ‘simple’ concordance is actually an index of stark incomprehensibility.