On the linguistic component of inaccurate editing, linguists can offer something. "Technical" failure presumably plays a part in all inaccurate editing, in some cases very obviously (e.g. 25, 38). We can move from description to prescription, and propose linguistic guidelines by which copy editors can avoid certain inaccuracies. A very small percentage of the hundreds of changes made by copy editors result in inaccuracy. There are few editing rules which are essentially bad, which in themselves necessarily cause inaccuracy. Deleting or inserting negatives are such non-rules, for the obvious reason that a statement and its negation cannot be compatible. And few rules are obligatory, e.g. updating time adverbials.