PEs and EAs

A former colleague who’s applying for an entry-level production editing job e-mailed yesterday with this inquiry about EAs and PEs:

Would you clear something up for me? I took their proofreading test, have to return it today, and don’t get something. The test consists of a setting manuscript (word doc marked up by the copyeditor) and corresponding first page proofs (no markings). Some of the copyeditor’s alterations are reflected in the proofs, some are not. There are also errors in the proofs which do not appear in the setting ms. The directions say to mark corrections as pe’s or ea’s, but my understanding from Chicago is that ea’s and aa’s are changes made to the proofs, therefore all discrepancies I find between the setting ms and the first page proofs would be pe’s. Or are pe’s only errors in the proofs which do not appear in the setting ms?

Did you follow that? I couldn’t. But I answered it anyway with the following tract, which I’m posting here in case anyone else out there is confused on this issue. And judging from the incorrectly marked up proofs I have to decipher from time to time, there are more than a few professional proofreaders who don’t quite get this concept:

An error is marked PE if it’s something the printer should have done but didn’t. In most cases this refers to an actual mark by the copyeditor that was missed, but it also includes corrections the compositor was expected to make on his or her own and that the CE could not possibly have marked in advance, i.e., page makeup issues that only arise once the text is typeset. The CE can’t tell where lines or pages will break, so how could he or she mark them? The compositor is expected to use a dictionary to break words and to scan through the typeset pages for inauspicious breaks.

For example, if a word that’s connected to another word by an em dash is also broken at the end of a line, that’s a PE. Widows and orphans are PEs. Stacked words are PEs. None of this is stuff the CE could have anticipated, and the comp should have known to fix it.

An EA in this context denotes any new correction you, the proofreader, are suggesting. For example, if the CE overlooked a misspelled word and you caught it, that’s an EA. If the CE marked something up but created a new error in the process, which the comp dutifully set, no matter how obviously wrong it was, that’s an EA. You must assume that the comp has a dictionary but can not actually read English. This is, in fact, often the case.

So as you look at each mark you’ve made on the proofs, you have to decide whose fault it is—the CE’s or the comp’s. And as you’re representing the publisher’s side in this test, if there’s a gray area you’ll probably want to blame it on the comp. As a compositor, I sometimes cross out the “PE” and attach a cranky Post-it if I think they’re full of it, and when I proofread I’m more inclined to blame the CE for any ambiguous problems. Usually the source of the error is clear, though.

Note that you should make these marks at the very outside edge of the paper, as far from the corrections as possible. They’re not intended as a slap on the hand to the person who fucked up. Rather, the production editor has to count them and figure out whom to bill. Some proofreaders interlard them with the corrections, so it looks like

(lc)|(PE)|(del)|(EA)|

This infuriates everyone. It makes it very difficult for the comp to see what needs to be fixed, and it makes tallying the charges especially tedious for the production editor. I like to use two types of pencil–red for the corrections, graphite for the charge marks—but you probably shouldn’t do this for a test.

In general, the most important rule when reading proofs is to introduce no new errors. Well, maybe of equal importance is keeping the bill down. If it’s a PE, fire away, but be very conservative with your EAs. If you want to make an editorial change, especially if it affects more than one spot, consider posing it as a query. And try to avoid any change that will alter the flow of text over more than one page. Anything that reflows will have to be proofed again by a production editor, and they hateses that.

So there. Any questions? Alternate theories?