Correcting Without Relitigating

D · Plate 14 in the Corrections series · Other Communications

The strongest pull in writing a correction is towards explanation. The error had causes, some of them mitigating, and the writer knows them. Putting them in the correction feels like completeness. It is almost always a mistake, and the reason is structural rather than moral.

A correction has one job: to replace a false thing in the reader's mind with a true one. Every sentence that is not doing that job competes with the sentences that are. Explanation is the most competitive kind, because it is interesting, and a reader who becomes absorbed in how the error occurred may finish the text without registering what the corrected fact is.

The shape of the failure

The recognisable version runs: the claim was based on a preliminary figure supplied under a previous methodology, which has since been revised, and while the underlying direction remains accurate, the specific number quoted should have been. By the time the sentence reaches its correction, the reader has been offered three reasons not to treat the error as an error. The correction is in there, grammatically, and it will not survive being read.

The same content is not forbidden; it is misplaced. Anything explaining the cause belongs after the corrected fact is complete, in a separate paragraph, with a sentence of its own. Readers who want the explanation will find it. Readers who only need the fact will already have it.

The difference between cause and defence

There is a real distinction between explaining a cause and mounting a defence, and it is detectable in a draft. An explanation of cause describes a mechanism and could be written by someone with no stake in the outcome. A defence assigns weight: it tells the reader how seriously to take the error, which is not the writer's decision to make.

Phrases that assign weight are easy to find once looked for. Minor, technical, isolated, does not affect the overall conclusion. Some of these may be true, and where the scope of an error genuinely matters, it can be stated as a fact about scope rather than as a characterisation. The affected figures appear in one table is a scope statement. The error was minor is a verdict.

When the original argument was also wrong

Occasionally an error is not local: the incorrect fact was load-bearing, and the conclusion built on it does not stand. This is the case where the temptation to relitigate is strongest, because the writer has to correct a fact while a whole argument wobbles behind it.

The correct handling is to do both, separately and in order. Correct the fact in the correction, completely and without qualification. Then state, in its own sentence, that the conclusion drawn from it is being reconsidered, and say when that reconsideration will be published. Attempting to rescue the conclusion inside the correction produces a text that argues while claiming to be a repair, and it usually costs more than the original error.

The general rule underneath all of this is that a correction is not a place to be persuasive. It is a place to be accurate and to stop. A correction that reads slightly more severely than the writer thinks is fair is close to right; one that reads as fair to the writer is usually already tilted, because the writer is the only reader who knows how easily it could have been worse.