The Em Dash Problem: Jumping to False Conclusions, Suspect Correlations, Specious Inferences, and Avoidable Fallacies
Why faulty assumptions and limited evidence can lead you astray both in problem solving and life!
I was just thinking about writing this newsletter — when out the blue — an article popped up in my news feed stating dramatically…ready…wait for it…that using the em dash (the punctuation mark “—” ) is an indicator, or may be suggestive, of AI generated content. Oh no!
That's a really strange coincidence: Did Google somehow read my mind or listen to my conversations — who knows, probably? Also, darn!! — that really sucks — because I really like punctuation. In fact, I tend to over punctuate!!!! Honestly, I'm a really big fan of the em dash in particular. You probably know it from your freshman high school English class: It is not only a punctuation mark, but also a writing tool that can be very stylistically useful when deployed properly, much like the semicolon; however, it seems like the em dash has been flagged as suspicious — what will those fancy, highfalutin (alive and dead) writers use instead?! Oh, and by the way, we may possibly need to reevaluate the works of Emily Dickinson, Stephen King, Kurt Vonnegut, Philip Roth, James Joyce, in fact any one who's ever used them — hmm, maybe they were secretly using AI before it was ever invented — I’m being sarcastic, of course (or am I???)
But irregardless (hahaha, that's funny, I just used a nonsensical word, or to be more specific, a nonstandard word), it’s a real conundrum — use em dashes and be flagged as, oh my gosh, guilty of using AI generated content or be true to your English teacher's admonition that em dashes can aid in making writing — wait for it — I'm going to do it again — more dramatic! Ah, we know that some writers (again, like unapologetic me) love fancy punctuation, but other authors, and much more successful ones at that, see it as completely superfluous. The wonderfully gifted but sparsely punctuating author Cormac McCarthy famously avoided quotation marks commas and semicolons even capitals like the plague because he saw them as unnecessary and interrupting to the main flow of writing even though sentences can seem to run together without them and one can get confused and lost in the, oh whatever. I can't even follow what I just wrote.
All silliness aside, this sketchy accusation regarding the use of the em dash as indicative of possible AI generated content first spread through the Internet faster than the tragic Palisades and Eaton fires last January here in California. Warning, personal opinion to follow: The failure to prevent, contain, manage, or stop those deadly and destructive fires before they raged and destroyed the lives and property of so many people is an utter disgrace, showing that proactive problem solving abilities were sorely lacking in certain key local municipal management positions, among many other reasons, and I know that may others agree!! But I digress substantially for my main point. Sorry.
Regardless of the rank stupidity and ill-informed actions which can sprout from dangerous conclusions derived from specious premises, certain assumptions or premises can, without much consideration of their validity, take on the ring of truth, until more rational minds begin to examine those claims. The em dash “problem” is inconsequential in the “Grand Scheme” of things, yet it provides genuine insight into the sort of distorted thinking that pervades the internet, life in general, and sometimes problem solving in particular.
At this point, you may be asking, "Why is this em dash thing a problem?" Good question! My response is this: It’s a problem (meaning an obstacle to overcome) because it presents an opportunity to examine and reflect on certain common patterns of "jumping to conclusions" which are based on faulty or incomplete evidence that has not been fully evaluated. This kind of "fact distortion" prevents movement towards accuracy and truthfulness in common discourse. Falsely labeling a writer for using a common punctuation mark is wrong. Hedging by saying something might be true based off a correlation, without other substantial evidence, is also wrong.Initially, like a clueless Sherlock Holmes run amuck, writing detectives, AI writing detectors, publishers, teachers, etc. declared the use of em dashes a “clue” to fake writing, or more importantly, true authorship. This assumption even took on a type of authoritative, judgemental tone. And, sometimes they were correct, but for the wrong reasons. This is what happens when a single piece of data becomes an overly important indicator without proper context. Furthermore, this kind of jumping to conclusions or assumptive thinking often hinders problem solving efforts.
My initial reaction to this while em dash “gotcha” thing was this: Anyone who believed this should read any one of the hundreds of books on rhetoric and composition which encourage the proper and judicious use of the em dash to enhance writing. Now, to be fair, the overuse of em dashes may point to AI generated content — but guess why: Because the artificial intelligence systems themselves have been trained on writing by human beings! Accordingly, they (AI systems) have copied a punctuation device used frequently by writers in the past. The em dash, unfortunately, has fallen out of favor recently with the “less is more” writing crowd that eschews everything but the bare minimum in communication at the expense of any personal sense of writing style.
That kind of simplistic writing, what I call “business minimalist”, is short, direct, to the point, standardized, and completely and utterly lacking in any interest, aside from the information it purports to telegraphically communicate. Ugh! Thankfully, most of these minimalists (although there are some notable exceptions) didn’t exist heavily among literary types, or we would have some pretty boring legacy writing. And don't think that just because Ernest Hemingway wrote minimally that he was not capable of breathtaking, complex, lengthy prose, and cumulative sentences.
Yes, there are clues to AI generated writing. Statistical patterns exist to determine probabilities of such content. Since one of the main drivers of artificial intelligence systems is pattern matching the statistical likelihood of the next word, the analysis of writing suspected of being derived from an artificial intelligence chatbot will yield some idea, but most likely, not definitive proof.
Watch out for of applying a part to the whole, or, more specifically, the overgeneralization fallacy of making broad claims based off of limited evidence. It must be acknowledged that many people before me were writing about the em dash problem, including many worried writers who challenged the em dash conclusion. Thankfully, there has been a very recent trend correcting this and acknowledging the “—" for precisely what it is: a great little punctuation mark!
Some rather simple observations to remember:
- Correlation is not cause and effect. 
- Inferences that lack solid evidence lead to specious conclusions. 
- Fallacies of thinking get amplified through repeated exposure. 
Don't let your assumptions about what may be causing your problem(s) go unexamined for flaws, viewpoint distortion, mistaken correlations, overgeneralizations, undergenralizations (yes, it is a word), and false conclusions based on faulty premises or incomplete evidence that uses correlation as cause and effect .
In conclusion, though this em dash problem might seem like a trivial complaint about rooting out (or not) AI generated content, it is a marker of a much larger issue impacting effective thinking.
My general advice:
Stay focused on solving those problems which are causing you the most trouble. Don't give up! There are solutions: You have to be relentless in your search for those answers.
Happy Problem Solving!
Evan

