If you’re writing scientific or mathematical content, it’s not enough to get your work checked by an editor who only knows spelling and grammar. You need one who also has a mathematical or technical background, otherwise they will misunderstand your writing and make mistaken “corrections” which might make it into the final version.

I don’t just mean that they might not know technical terms like “eigenvector” or “chirality”. A non-technical editor can recognise that those are unfamiliar terms and then either skip over them or look them up to check they’re spelled correctly. I mean that you might write sentences which make sense to a technical reader but which a non-technical reader is likely to misunderstand or mistake for being incorrect. Here are five examples.

1. The author writes “IQ is normally distributed with mean 100.” The non-mathematical editor doesn’t recognise that “normally distributed” refers to the specific statistical distribution called the *normal distribution* (also known as the Gaussian distribution), and thinks “normally” is being used as an adverb meaning “usually”.

The editor mistakenly “corrects” the text to say “IQ is usually distributed with mean 100”, or “Under normal circumstances, IQ is distributed with mean 100”.

2. The author includes some samples of C code, such as

`int a=0;
if (b==a) return;`

The editor doesn’t realise that C uses a single equals sign for assignment but a double equals sign for comparison. They think the author is just being inconsistent with their punctuation. They mistakenly “correct” the second line of code to

`if (b=a) return;`

which does something entirely different (sets `b` equal to `a`, and then returns if the value of `b` is now nonzero).

A similar problem will arise if the editor doesn’t know that C uses single quotes around characters and double quotes around strings.

3. The author is a psychologist, writing about “the patient’s affect” and “affective learning”. The editor is literate enough to know that, in general usage, “affect” is a verb and “effect” is a noun, and “effective” is a common adjective. But they are not knowledgeable enough to recognise the technical terms “affect” and “affective” used in psychology, where “affect” is a noun meaning roughly “emotional expression”, and “affective” is the adjective relating to it. So they mistakenly “correct” the text to talk about “the patient’s effect” and “effective learning”, which mean entirely different things from what was intended.

4. The author gives an example of a rule which they describe as an “intensional definition”.

An extensional definition is one which simply lists the items which qualify, whereas an intensional definition is one which gives the underlying rule. “The factors of 12 are 1, 2, 3, 4, 6, and 12” would be an extensional definition, but “The factors of 12 are whole numbers which can be multiplied by another whole number to make 12” would be an intensional definition.

The editor doesn’t recognise the word “intensional” (and neither does their spellchecker, if they’re using Microsoft Word), and thinks it’s a misspelling of “intentional”. Changing it to “intentional” will either change the meaning, or just produce nonsense that makes the author look ignorant.

5. The author is a logician or mathematician, and writes that “P is true iff Q is true”. The editor doesn’t recognise the word “iff”, meaning “if and only if” (and, again, neither does MS Word), and thinks it’s a typo for “if”.

“If and only if” is logically distinct from “if”. “P iff Q” is symmetrical: it means that P implies Q and Q implies P, which means that P and Q are equivalent. “P if Q”, on the other hand, is unidirectional: it means that Q implies P, but P doesn’t necessarily imply Q.

Changing one step of a proof from “iff” to “if” is likely to invalidate the whole proof.