How does Word cope with a grammar quiz in The New York Times?
The New York Times has a grammar quiz today (free registration required for some of the NYT web site).
” When we stumble over sound-alike words, readers accuse us of turning our editing chores over to a computer program (and not a very sophisticated one). “
Which immediately had us wondering how Word 2007 (with the US language setup) will cope.
There’s a selection of nine blocks of text – a careful human reading will show the problems. Word 2007 doesn’t pick up any of the mistakes (though it does note a few concerns like “it’s” vs “its”).
It highlights an inherent limitation in any grammar checking software – the inability to detect a mis-spelling because the ‘wrong’ word is a legitimate spelling itself.
For example: “glass and steal radio studios ” instead of ‘steel’.
In this example the additional complication is words that sound the same but have different meanings – known in the trade as Homophones. Homphones were once more common in newspapers when urgent copy was ‘phoned in’ from reporters in the field. These days copy is usually sent electronically so the culprit is probably mis-typing combined with rushed use of the spell-checker (a problem for any publisher, including Office Watch).
In book publishing we’ve noticed an improvement in text quality over the last decade or so, marred only by the occasional word that’s clearly out of place. A sure indicator of an over-reliance on computer checking.
Grammar checking might get smart enough to detect words that seem ‘out of place’ based on common usage but we’re not there yet.
As NYT editors probably remind staff from time to time – there’s no substitute for careful reading by a human.