R U Cooked or R U Cooking?: Revisiting Platonic Binaries in a Neoliberal Age
by Loren James Smith| Xchanges 21.1, Fall 2026
Ross vs. Plato: Death Match
Ross argues that Plato’s belief in a single, absolute truth and his reliance on binary thinking deeply shapes how writing is taught. For those unfamiliar, in Platonic works like Phaedrus, Gorgias, The Republic and elsewhere, he essentially argues that there is an unchanging truth to reality which knowledge is based upon. In Plato’s view, this reality can only be accessed via the mind, because the physical world around us is ever changing.
This influence of truth / untruth, Ross argues, encourages the writing teachers of today to treat writing as something that can be judged as “right” or “wrong,” pushing students toward a single correct way of writing rather than allowing for multiple, context-dependent meanings. Ross also loosely ties this binary of correctness to the way that Standard English is privileged over other English dialects, accents and languages. Ross writes:
“However, recognition that Plato’s influence is one of the primary factors that encourage teachers to push students toward a “correct answer” can provide insight about the deeply imbedded tendency for writing to be judged as “right” or “wrong.” Subscribing to Plato’s fixed, transcendent truth enables one to make simple judgments of writing as ‘right’ or ‘wrong’ because all writing is for the singular purpose of nearing Platonic truth. Critics who are influenced by Platonic models may contend that they are not judging writing by labeling it “right” or “wrong” but, rather, are teaching students to communicate in standard modern English” (np).
The binary of right / wrong writing and its relationship to standard English is certainly still problematic. Hamilton’s 2025 article in Xchanges issue 19.2, “Empowering Voices: A Graduate Student Instructor’s Introduction to Linguistic Justice” gives us a modern history of how Standard Academic English (SAE) marginalized writing students from diverse linguistic backgrounds, and how the field responded to it over time. Hamilton, like Ross, discusses writing studies resistance to SAE by discussing process pedagogy and social-epistemic rhetoric and pedagogy. However, Hamilton fills in some important gaps by also charting important developments such as code meshing, translingualism, and multimodal literacy.
Although Hamilton doesn’t discuss Plato, the binary of good / bad persists in classrooms and the world in many ways. I’ve mentioned Feng’s Xchanges article, “Bad Sex vs No Sex” already, but Feng has an example that illustrates this point. There, Feng is writing about Utah’s abstinence program as it appears in three sexual educational texts. Feng finds that these abstinence-based education texts present heteronormativity as morally “good” while marginalizing other groups. They write:
“The main binary of “future versus sex” erases identities that exist outside its margins, but because the emphasis is on the morally respectable nature of heteronormative timelines, the identities that exist within the category of “sex” are also marginalized. This binary model denigrates teen parents, queer people, adolescents, rape victims, pornography users and anyone engaging in premarital sex” (np).
It will likely come as no surprise to those studying the homogenization of language that the dominance of Standard American English (SAE) (also referred to as Dominant American English (DAE) or White Mainstream English (WME) reflects the kind of binary rhetoric Ross and Feng describe. A contemporary example of exclusion can be seen in the linguistic bias large language models (LLMs) and AI systems show. Since it’s widespread adoption in 2022, there have been copious authors who have covered how ChatGPT and other models prefer / enforce / empower SAE, WME, or DME. Thompson & Pokhrel in “GenAI: The Impetus for Linguistic Justice Once and For All” for example discovered that, even when expressly told not to do so, ChatGPT reflects western, male-dominated, white, middle-class understandings and expectations about language ideologies. These findings corroborate work from Lester, Hsu, and others.
Though the following authors aren’t focused on linguistic bias, Fang, et. al reported that seven different LLMs exhibit notable discrimination against females and individuals of the Black race. Shieh et al. across 500,000 generated outputs reported that LLMs are more likely to omit characters with minoritized race, gender, and/or sexual orientation identities or relegate them to subordinated roles as opposed to dominant ones. They also discovered that stereotyping across language model–generated outputs disproportionately affected minorities. Finally, Locke et al. illustrated that gender bias was prominent when generating images of professionals in CanvaAI, OpenArt, and Midjourney. Bias in examples such as this is deeply problematic, particularly when, according to a new Gallup poll, more than half of college students are using AI weekly (Gallup).
In Plato’s allegory of the cave, prisoners are chained down so that they can only face a wall, where they watch shadows cast by objects behind them. Over time, they come to believe that those shadows are reality itself. Plato uses this allegory to claim that truth (the objects casting shadows behind the people) is something that cannot be directly accessed via our senses.
There’s a lot of parallels that could be made about the shadows, the prisoners and truth in Plato’s allegory. Here’s one: if the output of LLMs are the shadows, students are the ones casting these shadows and writing instructors are the prisoners forced to watch them, does that make Plato’s “forms” or “truth” the White Mainstream English LLMs are trained on?
Hell no.
Ross writes that “Plato’s fixed, universal, transcendent truth is a pedagogy that values grammatical correctness, vocabulary, and the presence of analysis over the development of thought and the quality of content.” Fixed, grammatical correctness AI excels at. It can also perform, to a limited a degree, literary and textual analysis as Mozaffari et al. indicate. What are we left with when grammar and analysis have been usurped by a machine? AI does not think, feel, nor understand human beings. It cannot do these things because it’s not sentient. What Ross tells us to focus on, though she doesn’t frame it this way, is human; “Multiple-truth epistemologies” where discovery, critical thought, creation of knowledge, and distinct opinions are fostered in an environment rife with experimentation and equality. The path Ross charts for these outcomes are paths that are well trod: writing classrooms with theory, practice, and action which take the form of both in-class, and out of class reading, writing, discussion and assessment practices like, though not named, labor-based grading.
I have a lot of thoughts about writing pedagogy I’d like to add to Ross here, but first I’d like to show you one example of how deeply embedded binary rhetoric is in western capitalism. Though I’ve given you a taste of it above, in the following section, I will discuss the testing industry and why a platonic view of rhetoric is a lucrative philosophy for the neoliberal world inside and beyond academia.
