
I gave a video presentation on copyediting last week. What started out as a live demo became a demonstration of manuscript preparation and a discussion of the tools and processes copyeditors use to do what we do.
Exciting stuff, I know. What can I say? Word nerd, here.
The questions ate into the demonstration time, but I enjoyed them. Shining a light on the approaches editors use and the tools writers can borrow to polish novels is time well spent.
After answering a question about whether one uses a space before and after an em dash—no, no spaces around an em dash in US usage, and yes to spaces around an en dash in UK usage—someone asked what I thought about the em dash being a sign of heavy ChatGPT use. (My apologies to the attendee for my awkward paraphrasing.)
What? I’d never heard such a thing.
He then linked to this article: "Gen Z can easily spot ChatGPT writing! Here’s how”.
The article asserted that ChatGPT writes articles full of em dashes and emojis. An abundance of them has increasingly become a sure marker of non-human writing. Don’t believe me? Google it.
Leave it to some large language model (LLM) to take a legitimate and elegant punctuation mark and soil it amid a vomit of emojis. Darn, AI.
It took me over a year of study in a certificate program to learn how to use this sexy, versatile punctuation mark. And in no time, the ChatMonster turned it into this quarter’s Fake News.
According to the article, Gen Z is calling it the “ChatGPT Hyphen.”
Please, no.
The em dash is a legitimate punctuation mark. This Merriam-Webster article describes its utility in showing a shift in the intent of a sentence, going deeper as a parenthetical or signalling an abrupt cutoff.
So, what happened? When did it become ChatGPT’s supposed minion?
I turned to LinkedIn and saw that I was late to the party. My fellow editors had already risen to the challenge.
In his LinkedIn article, Grant Shumaker points out that LLMs are trained on a large volume of writing, much of which is formal in nature, such as academic, newspapers, and magazines. Since LLMs don’t think but rather reproduce patterns, it’s no surprise that we see them overindulging.
In his LinkedIn post, Rodrigo Fuentes agrees that what we’re seeing is pattern-mimicking. He further points out that LLMs often shift between parentheses, em dashes, and commas and suggests that ChatGPT thinks the preponderance of em dashes is a happy medium between the other two.
I'm curious if em dashes visually stand out because of their size. Named for their “m” letter length, the em dash takes up more space. Its horizontal length demands more lateral eye scanning than the diminutive comma or the vertical parenthesis.
In addition to em dashes, Shumaker suggests that eagle-eyed readers look for lists of three as a sign of ChatGPT. Of course, we’re all taught the value of three in essay writing. Are we really going to stop that practice?
For that matter, what about the serial comma? Some suggest “too many commas” is a sign of AI. Perhaps it’s a sign of long sentences with introductory clauses, restrictive clauses, appositives, coordinating conjunctions, or a list of more than two items. Come on. Maybe it’s a sign of a copyeditor.
I want to stand in honor of the writer and their story, in honor of their words. Writers suffer enough judgment, some of it self-inflicted. Let’s not add unfounded accusations about their writing tools, or lack thereof. That’s a path toward paralyzing creativity.
No one’s writing articles accusing celebrities of using ghostwriters.
Can we just let writers write, free from fear of being deemed a robot? If the content helps you, then consume it and move on. If the novel entertains you, read it and share your joy.
There is little utility in shaming English majors, copyeditors, and conscientious writers away from using proper punctuation to tell a good story or write an informative article.
ChatGPT is going to do what ChatGPT’s going to do. There is no pretending any of us has the power to change the trajectory of generative AI. Not while our bread and butter programs are top-loaded with AI tools so in our faces that we can barely navigate the scroll bar.
The last thing we should do is look at each other with suspicion. Seriously. Suspicion of what? Using Microsoft Word out of the box? Performing a Google search? Using a grammar checker?
Hurling accusations of AI use at writers shifts blame to the user, away from the developers. I reject victim shaming. How did using software become morally or criminally repugnant? I only wish developers gave us more of a choice.
The current focus on detecting ChatGPT tells and singling out em dashes and other “clues” seems designed to scare people into watching more monetized videos and signing up for subscription services to prove humanity. ::cough, cough::
I’m in favor of creativity and letting imagination soar. If a story sings, it sings. We are far past being able to tell how it got there. I’m just glad that it’s on the page, and AI or not, if em dashes paved the way, then them’s the rules—style rules.
Yikes! I use the em dash heavily. So does Joyce Carol Oates! So, NO!