Burden of Proof
How to Read Proofs Like a Pro
In last Thursday’s Shelf Life about the editorial order of operations I talked about proofreading a bit and today I’m going to talk about it some more. This is called being on a tear. Today’s Shelf Life is going to be a deep dive into proofreading—how to do it. At least, as much of a deep dive as I can fit into approximately 2,000 words or fewer. Given that I have a limited amount of space and time in which to impart a skill I’ve been working on for a long time, I’ll try to keep it to the most important bits and not get into the weediest weeds about ligatures and kerning and stuff.
A proof, in the publishing sense, is a preview version of a publication as it will go to the printer. It is laid out in print-ready pages with running heads and/or running feet, folios (a fancy way to say page numbers or pagination), margins, fonts, and everything else. Proofs are made from manuscript; the manuscript is your word-processed document, like your .docx, before it has been typeset. The proof usually comes in .pdf format and has been typeset, or composed, into near-final pages.
Proofreading is the act of reading that proof to look for any remaining errors leftover from all the rounds of editing before the proof is released to the printer. Sometimes the printer will send back bluelines (or “blues”), which is an untrimmed sample of the printed pages that is made on the actual press, so you can read them one more time.
It is a truth of publishing universally acknowledged that the later you make a change the more expensive it is to make that change. When you make editorial changes to your manuscript, that’s free. No one charges you money to open a Word doc on your computer and change a word here or there. Or, no one should be. At proofs, changes are more expensive because anything you notice must be passed along to the compositor (or typesetter) to introduce that change to the proof. The typesetter charges for this, usually by the affected line of text (that is, three corrections in the same line of text would result in a single correction charge) or by the change (that is, three corrections will result in three correction charges).
If you want to make a change to a text when it’s already on press, that’s even more expensive. The very latest I’ve personally had to make a change to a book before publication was after printing but before binding. All copies of the affected signature (that is, the block of 32 pages in which the error occurred) had to be thrown out, reprinted, and manually re-collated with the rest of the signatures before binding. That was an expensive typo, indeed.
This is why you want to catch the last few critical things at proofs. You want to have made most of your changes at manuscript—when it was free—and any lingering changes you want to take care of before the file is released to printer, at which time it becomes very expensive to make the change.
I’ve never seen a proof that required not one single correction. Even if the manuscript was squeaky clean, errors can be introduced during typesetting. And anyway, no manuscript is that clean—no matter how many times you and everyone else has gone over it.
Therefore, when you proofread, you look for two types of things that require correction:
Errors that existed in manuscript and were missed during editing, and thus appear in the proof.
Errors that were not in the manuscript and introduced during composition, and thus appear in the proof.
When I read a proof, I don’t go through it just once. I may only word-for-word read it one time, but I’ll go through the pages multiple times for different things. Here’s how I do it.
First, I review any documents that came with the proof. If the publisher has provided the copyeditor’s style sheet I’ll read that. If I have the publisher’s style manual on file, I’ll give that a once over just to make sure I’ve got it fresh in my mind—when I proofread on a freelance basis, I had multiple publisher clients at a time so I had to ensure I didn’t mix up their style manuals. If I received a letter from the publisher with any notes about the text, I’ll read that too.
Next, I read the front matter carefully, word for word—all the Roman numeral pages at the front. During this close read, I’m looking for specific things:
Title matches on the half title page, full title page, and copyright page.
Author’s name matches in all instances, including middle initial if present.
Publisher’s name and imprint is correct in all instances. Their logo and location information is formatted correctly on the full title page.
First folio (page number) appears on the first page of the table of contents, and is correct. That is, if the table of contents begins on the seventh page, that page should be numbered vii and not i, even though it is the first page with a folio.
Copyright boilerplate looks true and correct and has no typos, wrong information, or overlooked placeholder text. For example, the copyright year is correct.
Next, I remove the table of contents from the text. If I’m working in Adobe Acrobat I’ll extract the pages to a second, separate PDF or if I’m working on a hard copy I’ll make a second copy of the pages that contain the contents. I put the contents side-by-side with the rest of the text and then check each item on the table of contents against the corresponding item in the text to ensure:
Chapter title, subtitle, headings match on the table of contents and on the page on which they appear.
Page number reference on the table of contents is correct.
Once I’ve verified the table of contents is correct, I check the running headers (or footers) throughout to make sure:
They are formatted correctly for verso and recto pages, with the folio on the outside.
Verso and recto running heads are correct throughout; for instance, if they are author name on the verso and book title on the recto, I check spelling of both throughout and make sure author’s name is always on verso and title is always on recto, and that they are never switched.
Pagination is correct all the way through; no skipped page numbers or duplicated page numbers.
Blind folios—that is, pages that do not have a folio, like blanks between chapters and full-page images, are appropriately counted so that the next folio is correct. For instance, if Chapter One ends on page 47, and page 48 is blank with no folio, Chapter Two should begin on 49, not 48.
To do this check of the running heads, I view the proof in spreads. In Adobe Acrobat, I do this by navigating to the “View” menu and then the “Page Display” submenu and then toggling on both “Two-Page View” and “Show Cover Page in Two-Page View.” This shows me the text pages two-up, so that the verso pages are on the left and the recto pages on the right—just as it would look if I had the book open in front of me.
While checking running heads/feet and folios, I also scan each spread for correct formatting. The gutter margin (the margin on the inside, toward the spine of the book rather than the free edge) should be slightly wider than the edge margin—this is because some of the gutter margin falls into the bind (in other words, it’s eaten by the spine when the reader opens the book). While checking spread formatting, I want to ensure:
Text block on each page of the spread is positioned appropriately so that the larger side margin is to the center. You’d be surprised how often the compositor neglects to turn on verso/recto alternation and all the pages have the larger margin on the left.
Text block is the same height across the spread. That is, the top of the text block and the bottom of the text block are aligned. If the text blocks are not the same height, composition should shift lines around from other pages during proof corrections so they are the same number of lines on verso and recto facing pages.
Note that each spread will not be the same number of lines tall; for instance, if the number of lines of text per page per the template is 36, a spread might be 36, 35, or 37 lines tall. They have to match across the spread but not from spread to spread, usually. Unless the editor is super finicky.
Next, I read the rest of the front matter (foreword, preface, introduction, and so on). I also read the first chapter. During this initial read of the front matter and first chapter, I make notes of what I see that might or might not need correction. For instance, if I notice a series without a serial comma, I won’t mark to add it right away unless the style sheet specifies always use the serial comma. Instead, I’ll make a note and keep reading to see if it was omitted accidentally or if it’s omitted consistently throughout.
I also use this preview read to acclimate myself to the author’s voice and determine how heavy I think the corrections will be. Are there spelling and grammatical errors on every page? If so, I might write back to the publisher at that point to ask how heavy they want the corrections. They may ask me to stet all but the egregious mistakes at this late stage.
Stet is proofreader shorthand for “let it stand as it is set,” or, “don’t make any change; leave the text as is.”
Once I have a good handle on how much I will have to mark up, I go back to page i (the half title page) and start reading, end to end. Every single word. This is here I mark errors of content like misspelled words and obvious grammatical errors like subject-verb-agreement mistakes; and errors of style like missing serial commas, number style (that is, what are the rules for spelling out numbers versus using numerals), and so on.
During this read I am also looking for errors introduced during composition that I did not find earlier during my high-level review of front matter, contents, running heads/feet, and folios.
I ensure font face and size are consistent and correct throughout. For instance, if the book is set in 10 point Minion Condensed and there’s a paragraph in 10 point regular Minion (not Condensed) for no reason, it’s my job to notice that and point it out. The spacing between lines of text and between paragraphs should be consistent throughout. I also ensure paragraph indentation is consistent.
If any characters dropped out during composition or got mixed up with another, similar character (character conversion error), I will note those. I may not be able to identify the correct character from context; I will mark to replace with the correct character if I can tell what it should be or leave a margin note for the editor if I can’t. Conversion errors I see often include 3 replacing ≥, m replacing μ, a replacing α, e replacing é, and the dreaded empty rectangle character that means “software couldn’t figure out what this was supposed to be.” Also sometimes you will see a character with its diacritical mark next to it as a separate character, like this:
O^ or n~
Instead of like this:
Ô or ñ
So be on the lookout for those, too.
When the last line of a paragraph spills off the bottom of the page and onto the next page, that’s a widow (a lonely partial line of text by itself at the top of a page). When the last word, or partial word, of a paragraph spills over onto its own line, that’s an orphan. Widows and orphans should be marked for correction. Although not strictly a widow or an orphan, a recto page that ends with a hyphenated word is a no-no—a word should not be broken outside of a spread. Inside a spread (from the end of the verso page to the beginning of the recto page) is allowed.
During composition, the compositor will have used their software’s auto-hyphenation function to insert breaking hyphens at the ends of lines as needed to ensure word spacing looks good throughout and the text block is justified (a nice, square block on all sides). As proofreader, I check to ensure the auto-hyphenation is correct and that words are broken correctly—between double letters, or between syllables if there are no double letters in the word. I also look out for hyphen stacks, which is when more than three lines in a row end in a hyphen. If that happens, comp should adjust the text to break up the stack.
Final word on proofreading: When marking proofs that have been composed by a professional typesetter—someone who is charging you for the corrections—mark any error introduced during composition as “PE,” or “printer’s error.” This tells the compositor they should not charge you to fix this error, because they caused it.
If you have questions that you'd like to see answered in Shelf Life, ideas for topics that you'd like to explore, or feedback on the newsletter, please feel free to contact me. I would love to hear from you.
For more information about who I am, what I do, and, most important, what my dog looks like, please visit my website.
After you have read a few posts, if you find that you're enjoying Shelf Life, please recommend it to your word-oriented friends.