These days, my YouTube advertising feeds and my favorite YouTubers’ sponsorship pipelines promote text-to-speech software. Whisper AI is the most popular app I see. Yet my iPhone’s Voice Memo function coupled with Word’s transcription feature works just as well for me. Therefore, as AI continues to reconfigure and unsettle how we conceptualize writing, being able to speak ideas aloud carries tempting implications and imaginaries. Can voice-writing faithfully carry a writerly fingerprint when AI-assisted editing removes the texture and idiosyncrasies that make a writer’s voice authentic? Or does writing via the voice problematically give too much credence to spontaneous thought, erasing more deliberative thought that develops from writers wrestling with the cursor on an open Word document? I will start with promises. Here are three things I enjoy about dictating my writing via text-to-speech.
1. The blank page bows down to my voice.
In The War of Art, novelist Steven Pressfield urges creatives to defeat resistance by any means possible. Resistance represents the lingering, stomach-churning, slow-drip heaviness we feel in our chests when we are about to start a task that involves vulnerability or creativity. It is a more evocative externalization of the often internalized, productivity-oriented concept of procrastination.
In my early graduate school years, the blinking cursor was my own siren call of resistance. It is not that I had no ideas. It is that I had so many ideas clogged up in a bottleneck and could not decide who would own the honor of the first sentence. I therefore struggled to start all of my written assignments. However, with the rise of speech-to-text and AI’s ability to iron out transcription errors, I now never fear the blinking cursor. Words pour onto the page as fast as the speed of sound. Many of us speak at 100 words-per-minute on the slow end and a 150 words-per-minute for the fast talkers among us. Compare that to rapid typing speeds of 40-60 words per minute, and speech-to-text becomes a massive productivity unlock. Never did I imagine I could effortlessly write three thousand words over the course of thirty minutes.
Yet speed is not the problem. AI produces text far faster than humans can speak. I cannot compete. The speed of sound cannot compete with the speed of light and binary code. However, in an era where many people, particularly teachers and language researchers, raise alarms about what AI does to our voice, our creativity, and to the joy of writing, writing out loud produces a god-tier compromise between AI’s need for speed and a vocal human’s ability to play with language beyond a probabilistic, surface-level simulations of genre writing.
As many people smarter than me argue, the best writing is rewriting. And while some turn that advice into AI-generated first drafts that one can later edit, the fluency illusion of AI-generated first drafts can at times restrict the writer’s introduction of creative noise – noise that is necessary for innovative writing. The perceived fluency of AI and its smooth, sleek language gives the illusion of an optimal form of organization when preparing a paper or an essay, often closing or making invisible alternative, creative forms of style or structure.
But when I talk ideas out loud, I embrace the power of immanence to be taken in unexpected directions; to make novel connections on the fly; to find creative gems as the emerge from my mouth. Put otherwise, when pilots increase the power of a jet engine from idle to full power, the engines do not accelerate instantaneously akin to pushing a gas pedal. It can take 6-9 seconds – a potentially disastrous delay in crisis aviation. Therefore, if writing is thinking, then allowing AI to produce first drafts sets the terms of our thinking before our minds can even spool up. And that delay can doom the potential creativity of our drafts.
Which brings me to a second thing I love about speech-to-text — writing while taking the walk outside.
2. I prefer to do my writing while walking outside.
I’m a shaky, flaky, weeble-wobble of a human being. My girlfriend translates a Turkish idiom to say I have worms in my ass. Sitting still is not my strong suit. I have tried my damnedest to develop a sense of stillness, building daily meditation practices in my mid-20s. But nowadays I just accept my jitterbug nature. Rather than fighting the need to sit still, perhaps movement catalyzes my thinking.
One thing that I love about speech-to-text is the ability to write while I take a walk. I pen these paragraphs now from the green grass and shady oaks of Sunset Park in State College, Pennsylvania. Trees tower above me and birds chirp off to my left. I feel my feet gently bounce against the soft earth as I walk in circles around a park bench, snapping my fingers as I try to snap the next word into existence.
Before I embraced text-to-speech, the only way that I could write was to sit down in a chair, oftentimes in front of a screen, and grind my wrists into dust while my back slowly heaved cries of quiet desperation. Nowadays, when I write a first draft, I don’t need to sit still. I can open my body and let in the power of movement as extended cognition. I can compose in nature. Or put differently, I knew it was in my nature to compose. What I did not realize was my ability compose in nature.
When I walk, I furthermore bring more ammunition into my gunfight against resistance. Walking dissolves anxiety. And so, as I take step after step, I spit out sentence after sentence. And before I know it, I have a wall of text composed of my own ideas in my own voice — a heap of iron ore ready to be smelted into clean prose. And it only took 15-30 minutes.
3. It makes me feel alive.
My explorations into philosophical Daoism, or daojia, has pushed me to think through my own writing process in the form of a compound, sheng wen (生文). If you search this term in contemporary Chinese discourse, a likely hit would be an abbreviation for shengchengwenben (生成文本) — Tencent’s AI text generation program. However, when I think through the classical Chinese of sheng (生) and wen (文), I feel like I am holding an irreconcilable yet generative contradiction.
On one hand, I read sheng (生) as a character signaling life beginning, unfolding, becoming — a becoming that is not singular, but co-enacted among ten thousand things. Some philosophers refer to this as fang sheng (方生). 生, put otherwise, appreciates imminence and constant recomposition of life — a recomposition that cannot be disciplined, planned, or known in advance. To know with certainty the unfolding of life in advance is not the problem. Articulating that certainty is. For as the Daodejing opens, “The Dao that can be told is not the eternal Dao. The name that can be named is not the eternal name.” Life therefore unfolds simultaneously, continuously, and without a priori knowledge of its direction or future movements.
We can contrast 生 with wen (文). Many associate wen (文) with the literate, the scholarly, the learned, the teachers. 文 is a common character for words associated with writing, education, culture, and patterns. 生文 therefore reflects Daoism’s skepticism towards language. As one line in the Zhuangzi says, “Dao is obscured by partial understandings.” Yet Gia-Fu Feng translated the characters xiaocheng (小成) as “partial understandings,” which I only partially agree with. From a Western perspective, we can see understanding as something held, some embodied knowledge or truth. But xiaocheng (小成) does not gesture to someting cognitive per se, but rather an act of closure, achievement, or accomplishment. In the case of 文, 小成 demonstrates how writers fix and translate their situated, embodied way of knowing into writing via names, propositions, language. And the problem with names and language is that they always open some meaning-making possibilities and foreclose upon others. This is the tragedy of writing.
As scholars, many of us build energy, enthusiasm, and sense of flow by immersing ourselves in 生. And organizing that flow of experience into data, which we analyze and write up in a research report, translates 生 into 文. Yet when we write something into 文, we cannot help but ultimately transform and limit our ability to understand. This is the paradox of 生文. We advocate for transparent 文 to accurately or systematically represent 生, while much of what makes 生 energizing to begin with defies representation altogether. We try our best, through precision and creativity of language, to articulate our systematic engagement with the unfolding of 生, while ultimately foreclosing on other potential meaning-making assemblages. When we choose theories, terms, names, or analytical techniques disciplined through the lineage of 文, we also struggle to articulate otherwise that “this analysis is not the world, but based only on a very selective interpretive frame of the world disciplined through language and language’s fractal disciplinary unfolding.
When I use text-to-speech to write outside, I not only defeat the blinking cursor by putting words onto the page. I not only get the privilege of writing in the green embrace of the outdoors. I can also produce sentences, ideas, and turns of phrase on the edge of imminence. When I do not plan, when I let go of the affective filter that shapes and extrudes my words into disciplined shapes, then new metaphors, new assemblages of language, new avenues of thinking become available. As I walk and as I invest my breath into these words, I can sense my body becoming alongside my mind. When I sit in front of a computer, I feel my mind disassociating, my mind severing connection with my body, my body crying silent sobs. When I walk and write, our wen 文 joins the flow of 方生 to produce not only ideas, but aesthetics that are uniquely situated – uniquely ours, and uniquely beyond the purview of AI.
Leave a Reply