A Defense of Messy Writing

Or, the value of input over output.

Recently,

Nick Potkalitsky

shared a fascinating post about “the art of imperfection” in writing. He and I must be on the same wavelength because I had already planned to write this piece on the value of messiness in writing, especially in school settings. Patkalitsky’s post draws on linguistics research by Tony Berber Sardinha, which demonstrates that AI often differs markedly from human writing in register, which in lay terms is a descriptor of the different language styles used in different circumstances (e.g., in conversation versus chemistry publications). I think Sardinha’s research probably explains why so many human readers think they can identify AI-generated text on sight: often, it just doesn’t “sound right,” especially for the situation. That may well be a matter of register.

But I digress. Potkalitsky’s point is that human writing contains “gaps, hesitations, and beautiful asymmetries” that AI does not have. Such “imperfections,” as he calls them, represent the “unique perspective” of a human individual (

Alberto Romero

, to be precise) whose “rich tapestry of experiences, emotions, and relationships” inform a style of writing “that resonate[s] with others on a profound level.” Potkalitsky concludes:

As educators, writers, and researchers, our task is to champion this mode of thinking, to celebrate its imperfections and asymmetries, and to ensure that it continues to thrive in a world increasingly shaped by artificial intelligence.

I agree with Potkalitsky: in the era of generative AI, messy writing matters more than ever before. (It also gives me plausible deniability if you find my writing unclear!) However, my reasons for saying so differ somewhat from his. To get to those reasons, I need to ask a rather pointed question in response to Potkalitsky’s argument: who cares that imperfect writing represents the unique perspective of an individual writer forging a meaningful connection with a reader? Don’t get me wrong: I certainly do. But something tells me this is not a widespread value across academia and industry.

Image courtesy of Adobe.

For decades now, the popular, and often institutionalized, assumption has been an instrumentalist and meritocratic one: that higher education ought to prepare students to enter the workforce. If students study hard and earn good grades, they ought to be employable. In most workplace settings, an individual voice and meaningful connection are not particularly valuable. I distinctly remember a former roommate—a successful engineer and MBA—complaining about what he characterized as “circuitous” writing in fiction and literary nonfiction. He wanted writing that got straight to the point: clear, informative, and actionable. His perspective is not necessarily unreasonable, nor are meritocratic assumptions about the value of higher education. People mostly need to get shit done, and “gaps, hesitations, and beautiful asymmetries” just get in the way.

What’s more, the link between writing and thinking that Potkalitsky rightly celebrates is a double-edged sword. The going assumption has been that clear writing equals clear thinking, and students generally get the point: if they can produce polished prose that satisfies their teachers, they’ll get a good grade (and again, get a good job or get into grad school or whatever). Of course, AI has thrown a wrench in the notion that clear writing equals clear thinking. Today, clear writing may not represent any thinking at all, which has led to the current literacy crisis.

However, this aspect of our current literacy crisis only really matters if we retain a product-oriented mindset in educational settings by prizing clear-writing-as-clear-thinking.¹ The crisis disappears, or at least loses much of its force, if we reorient our pedagogical mindset towards writing as a thinking process. This is hardly a new proposition: scholars in rhetoric, composition, and writing studies have been saying this for decades. Way back in 1972, Donald Murray wrote:

What is the process we should teach? It is the process of discovery through language. It is the process of exploration of what we know and what we feel about what we know through language. It is the process of using language to learn about our world, to evaluate what we learn about our world, to communicate what we learn about our world.

Instead of teaching finished writing, we should teach unfinished writing, and glory in its unfinishedness. (p. 4)

Now, by invoking Murray, it may seem that I’m contradicting my answer to the “who cares?” question I asked above. Few folks in academia and industry want unfinished writing. However, even if we accept (at least temporarily) this instrumentalist view of writing instruction that says we’re preparing students for the working world, I would imagine that many hiring managers and CEOs would readily agree that newly degreed graduates should be adept at discovery, exploration, learning, evaluating, and communicating, or what Murray calls “language in action” (p. 4). Granted, they will probably still want products that are clear, informative, and actionable, but the best of them will trust that their employees have already done the thinking that goes into those more succinct and polished artifacts.

A pedagogy of messy writing is one means of teaching students how to use language in action to discover, clarify, and extend their thinking. There will be time and tools aplenty for converting messy writing into polished products when the time calls. And for those who bristle at an instrumentalist view of writing instruction, well, the benefit of messy writing ought to translate across most pedagogical approaches.

How do we get students to trust us that we want messy writing? One way is to reward them for writing that wrestles with ideas without fussing over surface-level polish. If you see an idea emerging, however messy, in a student’s draft, encourage them to follow it, and ask them provocative questions designed to push their thinking. Bump their grade when they respond. Another way is to not ban AI in your classrooms. This may seem counterintuitive, but let’s be honest: most AI-generated writing is generic and vacuous. If a student submits something AI-generated, polished but mostly empty of thought, grade it accordingly and give it back to them with the same kind of provocative questions you would ask of intellectually empty human writing. Students will get the message quickly: AI-generated writing won’t serve them in this setting. (Thanks to my colleagues Miriam Marty Clark and Guy Rohrbaugh for this idea.)

With enough time and practice, students will become habituated to using messy writing as a thinking tool. Then we can begin to introduce means and methods, potentially including AI, of polishing writing to meet the needs and expectations of specific audiences and purposes and genres and situations. But in the meantime, messy writing may be a better marker of thinking than polished writing.

1. Other aspects of the AI-driven literacy crisis certainly remain, such as the question of truth and authenticity in our media ecology.