Tricks for reading and correcting proofs

Some parts of a writing project are exhilarating; some parts (at least for me) are grueling; and some are stubbornly perplexing.  One part is important but very, very tedious, and I’m deep in that part now:* checking proofs. Fortunately, there are some tricks to make dealing with proofs easier.

In case you haven’t yet had the pleasure: the “proof” is the all-but-final version of your piece of writing, typeset exactly as it will appear in the journal (or as a published book, or whatever). “Checking” proof means what it sounds like: going through the proof in search of any errors or other problems introduced during the typesetting process – or the (hopefully rare!) errors that have snuck through revision and copy-editing undetected.**

Checking proof is mind-numbingly boring, and it’s hard to do effectively. That’s because you know your work too well: by the time you’re in proof stage, you’ve written, rewritten, read and re-read it dozens of times. It’s therefore very easy for your eye to jump right over any errors, simply because you know what you meant to say. Checking proof is an extreme case of a skill all writers need: the ability to see their work not as themselves, but as their reader. (In The Scientist’s Guide to Writing, I call this “reader simulation”, and I discuss ways you can achieve it.)

So how can you check proof effectively? How can you read page after page of your own writing so carefully that you’ll notice every tiny flaw? You need tricks – ways to enforce that reader simulation. Here are a few; and if you have a trick that isn’t on my list, please leave it in the Replies.

• Read the text out loud. Read it word by word, exactly as it’s printed, and don’t try to fancy it up with your best theatrical moves – you’re looking for errors, not staging Hamlet. The point of reading out loud is to force yourself to hear exactly what’s on the page.
• Read backwards – word by word, but starting at the end of the document and proceeding right to left, bottom to top, until you reach the beginning. This forces you to ignore the content and meaning of what you wrote (the proof stage is far too late to change that), and to focus on individual words.
• Use a pointer, like a pencil. This is my own approach. It’s less hard-core than reading out loud or backwards, but it’s easier – if you’re able to be strict about it. Run the pencil along each line of text, and read at its tip. Run it along the whole line, from the very first character to the last, and slowly – much slower than you’d normally read anyone else’s text, much less your own. The point is to check every word, when the pencil tip tells you to. [A piece of cardboard with a one-line-of-text-sized slot cut in it can work similarly.]
• Pay special attention to elements that are prone to typesetting errors. Once upon a time, a manuscript going into typesetting would be completely re-keyed, and errors could be anywhere. That’s no longer true; most of your text will be copied directly from the copyedited digital files. That means that regular text will rarely get messed up. But the same isn’t true for tables, figures, equations, special characters (like Greek or Hebrew letters), and the like.
• Check document design as well as text and figures. By “design” I mean the arrangement of items on the page. Check placement of figures, association of figures with their legends (is the right legend with the right figure?), indenting of lists, that sort of thing. At the finest level, check hyphenation of technical or foreign words that might not be in a standard hyphenation dictionary.
• Check and recheck. Unless you’re extremely practiced, you’ll find it hard to check for small errors (like misspelled words) at the same time you check for larger ones (omitted blocks of text). One read-through for the big picture, and another character-by-character, is wise.
• Find a unicorn. Believe it or not, there are a few people out there who enjoy checking proof. I have a retired colleague who loves checking proof, and who’s at this very moment checking a book-length set for me for the third time. (The Scientist’s Guide to Writing, Charles Darwin’s Barnacle and David Bowie’s Spider, and now the Scientist’s Guide second edition.) Each time I’ve given her a bottle of Calvados (her preferred tipple) in thanks, and having a second set of eyes on my work makes me feel like the luckiest writer in the world. If you can find a unicorn like mine, befriend them!

Once you’ve found something that needs correcting, you’ll need to indicate that to the typesetters. There are standard symbols for this communication, but they’re designed for checking proofs on paper and are thus less important than they used to be. It’s now routine to mark on a PDF proof using the sticky-note comment tool. A clear and unambiguous direction to the typesetter will usually work just fine, whether it uses the standard symbols or not.

Will an error occasionally slip through?  Yes – perhaps you remember the collective egg on the face of the authors who wondered, in print, “should we cite the crappy Gabor paper here”. You don’t want that to happen to you; so take proofs seriously, and use all the tricks you can find.

© Stephen Heard  August 18, 2021

 After writing this post, I discovered that I’d written it before – two years ago. It was sort of fun to discover that most of this new version is the same, but not all of it.

Image: Page 241 of 338 of my interminable set of proofs. Own work, photo CC BY 4.0

---

*^In particular, I’m about 2/3 of the way through the proofs for the second edition of The Scientist’s Guide to Writing. Proofs for a journal paper are unpleasant. Proofs for an entire book are a whole new level of unpleasant. This set runs 338 pages, every character of which needs to be checked. Three guesses why I’m writing a blog post instead of finishing.

**^These should be rare. The proof stage isn’t meant as one last chance to revise, and most publishers reserve the right to charge you for “author’s alterations” – that is, for requests to change material at the proof stage for any reason other than a typesetting error. There is, however, one uncommon but important wrinkle. A few journals do quite extensive copyediting, but don’t show authors those copyedits until the proof stage. This is both bizarre and inappropriate, but that doesn’t seem to stop it from happening. (I’m looking at you, Conservation Biology.) When it does, checking proofs also involves looking for errors introduced during copyediting, not just since copyediting; and these can be surprisingly numerous and surprisingly major. Ugh.