Proof Reading Tip

Owe carp! You're spiel chequer wont yelp ewe hare. The human brain is amazing. Computers may trump us on repetitive tasks, complex computation, and search or retrieval of stored data, but we have the advantage when it comes to intuitive error correction and filtering redundant information. This is why we mostly still have, for example, human CCTV operators rather than delegating the work to computers. (Although there’s still on-going research in artificial intelligence to achieve this.)

Our innate ability to correct errors is useful in daily life, from reading damaged or worn road signs whilst driving to understanding obscure SMS txt msgs. The drawback comes when we need to switch this mechanism off. Unlike computers, we don’t have any toggle buttons to change our settings. Experienced proof-readers are trained to overcome this ability, but it’s still tiring and time-consuming for large documents.

When I first published ‘Quack, Quack, Quack. Give My Hat Back!’ I gave a copy to my niece. My brother-in-law read it to her every evening for two weeks (apparently she liked it!) and then he decided to point to each word whilst reading it to help her to learn to read. It was only then that he noticed an error. Various people had read through draft versions of the book before I had published it but hadn’t noticed it. This error was rather annoying as I’d just had a hundred copies printed, but the surprising thing is that even when I tell people the page number where the error occurs, a lot of people can’t immediately spot it. Sometimes they will read out the page and automatically read what the text should be rather than what is actually written down. I have corrected the error and any copies on sale should now be the revised first edition.

It’s one thing to point to every word in a short children’s story, but it’s another thing entirely to have to do this for a 136,000 word novel, which is the approximate word count for ‘The Private Enemy’. Again, multiple people read the draft version and although they picked up plenty of mistakes, there were still some that were missed, and again I have had to bring out a revised first edition. (Luckily I hadn’t printed out so many this time. If you have a copy, look for ‘Revised first edition’ at the bottom of the copyright page to see if you have the corrected version.)

How can a writer combat this rather frustrating issue? Spellcheckers can help where words are misspelt but not when the misspelling forms a different word that happens to be in the dictionary. Grammar checkers can irritate fiction writers who have dialogue that should emulate the grammar of their characters. (Unless, of course, all the characters in the novel are extremely well-educated, articulate people with pedantic speech.) However, even grammar checkers can’t help if the typographic error is grammatically correct.

For example, a long time ago I wrote a story where, in one scene, a character snubs another character by refusing to shake hands. The snubbed character was left with his hand extended. At least that’s what I thought I’d written, and every time I re-read that scene I saw the word ‘hand’ because I knew that was the word that was supposed to be there. A friend of mine read the story and asked what I had meant by ‘head extended’. How can he extend his head? Both ‘head’ and ‘hand’ are nouns, so the sentence was grammatically correct, and while it doesn’t make sense for a human to have a head extended, it may be valid for an alien or artificial life form (or a giraffe).

This accidental act of replacing a word with another word that has a similar shape is the type of error that causes me the most problems because it isn’t flagged by the spellchecker, and whenever I re-read the passage I always see the word as the intended word rather than the actual word. My English teacher was highly amused when I wrote ‘they exploded the house’ (it should’ve been ‘explored’), and I often trip up with release/realise. Sometimes I simply miss out part of the word but the truncated form is also a valid word (for example, ‘image’ instead of ‘imagine’).

The problem here is forcing my brain to read what’s actually written rather than what it believes is written on the page. How can I emulate the slow point-at-every-word technique on a document that’s over 100,000 words long without getting thoroughly bored and giving up? I’ve discovered that the solution to this is to get the computer to read the book instead because the computer will read exactly what’s written, not what it thinks should be there, and my hearing (though not brilliant) is better at detecting errors. This is partly due to the artificial tone used by the speech synthesizer that helps to concentrate on the actual words, but it can also have a significantly different pronunciation of visually similar words.

So now I convert my novel to HTML*, so that each chapter is in a separate HTML file, and then I open each file in turn in Konqueror with the Jovie text to speech plugin installed, and use the Tools > Speak Text menu item. Then I just listen without looking at the words. It takes a while to get used to the artificial voice, its pronunciation of less common words can be a little off, and it can’t cope with homographs (such as ‘read’ or ‘live’ or ‘wound’), but the errors stand out. Unfortunately it can’t help with homophones (such as ‘new’ and ‘knew’) and it shouldn’t be used as a substitute for a professional reading, but it’s a useful tool to use alongside other tools, such as a spellchecker.

Also, from a literary point of view, it’s no bad thing to hear how well your writing stands up to a stilted monotone reading.


* Since I write my novels using LaTeX, I can convert to HTML using TeX4HT. If you don’t use TeX, some word processors have an export to HTML function.