I remember my mourning of the em dash. It was slicing, brutal and brash. I was simultaneously sunken with a deep depression and searing with rage. It sounds dramatic, but it’s how I felt. (And that’s kind of the point.)
When I found out that ChatGPT had developed a tendency to use the em dash, I implicitly committed myself to never using it again. I found myself omitting it entirely from my writing. Of course, though, there were the exceptions. Made in brief lapses of judgement or indulgence, they were a reclamation, if you will.
The day I discovered the em dash is not one I’d forget. Upon realizing that it could be used to finish my thoughts, stringing together sentence fragments which previously had no safe harbor in any dignified sentence (would’ve been entirely omitted without), I knew I’d found my best friend. The em dash quickly became my accomplice in blending prose with poetry. It was the conduit through which I could justify broken language and defy the draconian laws of grammar. I was able to create the kind of language that felt pieced together like Kintsugi. The kind that looks stunningly whole. Broken, still. Fragile, still. Golden, even.
Back to ChatGPT. I’ve gone back and forth in a crazed whiplash that goes something like this. One moment, I’ve retired from the em dash entirely. In another, I practice a stupid rebellion — I use it, despite. What I’ve realized, though, is that this tireless affliction of the mind is a practice in oblivion.
From an opaque, entirely soulless standpoint, I imagine this corruption of the em dash as being somewhat iterative. Syntactic structures and preferred punctuations will cycle through colloquial dialogue; certain terms will become more commonplace as others are pruned with the tangle of time. This is somewhat normal.
What’s not normal, though, is what drove me to this madness to begin with: the realization that there is something far more sinister at play. When I began to feel my relationship to the em dash dissolve into opacity, I couldn’t shake the feeling: we’ve been robbed of the ability to trust.
What we truly ought to be mourning is far less trivial and far more terrifying. ChatGPT will begin to chip away at the quality of our relationships to one another. If we can no longer trust the next text, email or love letter, to be written by its human “author,” what value will the human language hold? Can we trust one another to express our true thoughts and feelings? Or will we resign ourselves to interpersonal relations which are strung together by soulless words which were generated and rapidly predicted by a LLM?
The degradation of trust within one another, and far more perniciously, our institutions, is a challenge that will persist so long as we continue to passively accept this “unavoidable” path, which is “certainly” the “future.” But this is not inevitable. It is a choice and there is a cost. Didion writes, “We tell ourselves stories in order to live” and she couldn’t be closer to the truth. We are narrative creatures with fabricated, often focally distorted fictions that fold us into the story of our lives and allow us to breathe another day. These fictions allow us to not only understand ourselves, but others: the foundation of trust. As Stanford professor Jamil Zaki writes, “Trust is how hope lives between people.” I still carry with me the hope that our human stories will live on.
In “To love a dying art, despite,” I wrote a solemn reflection on the slow death of art like writing. That was more of a silent reminder; this one is a form of pleading. If we can no longer write our stories, or care for that matter, what will be left of us? As advancements in Artificial Intelligence continue to outpace our ability to think critically, it is now more important than ever to start asking: what makes us human?
One last note: Life is far more beautiful, more uncontrollable than an LLM could ever hope to predict. I hope you won’t let ChatGPT predict the next word in the story of your life —
— the em dash.