BookTok's Canon: When Evaluation Becomes Testimony
Brandon Baker's five-star review sits on the cover of *Shy Girl*, a "buzzy BookTok sensation" that Hachette pulled from publication in March after evidence emerged it was AI-generated. In his blurb, Baker praised the novel as "brutal, uncomfortable, and visceral." But when the scandal broke, he posted a mea culpa video explaining he loved the book precisely because it was *none of these things*: "When I read it, I was kind of, like, very fatigued on indie extreme horror... I just can't handle the most brutal things put to paper." The contradiction is total—he raved about a book for qualities it didn't possess and he couldn't tolerate—yet neither Baker nor his followers registered this as disqualifying. The comments commiserated not about the book's merit but about their own embarrassment: "i read it too last year and also thought the writing was off but assumed that was the author's writing style 😭 it's so unfortunate because i really did enjoy the story and i'm lowk embarrassed i gave it 5 stars." That a book where "the writing was off" merited five stars tells you everything about BookTok's evaluative logic. Judgment has collapsed entirely into testimony about emotional experience. The question is no longer "What does this book do?" but "What does this book do for me?"
This isn't merely popularization. It represents a structural transformation in how literary merit gets determined, and the scale is staggering. According to industry reports, roughly 59 million U.S. print books sold in 2024—one in thirteen—"could be tied" to BookTok. When the overall market shrank one percent in the first quarter of 2023, BookTok-viral authors saw sales jump 43 percent. According to publishing industry data, Colleen Hoover sold 14.3 million copies in 2022, reportedly a 661 percent increase over her 2022 sales figures. According to industry observers, she reportedly outsold the Bible in 2022. In 2022, 70% of books sold were backlist titles, a 68% increase from the year before, meaning BookTok resurrected old books faster than publishers could produce new ones. The industry has responded predictably: publishers now pay upstream for the *promise* of virality before marketing even begins. We are watching canon formation happen in real time, except the mechanisms producing this canon have nothing to do with critical judgment and everything to do with algorithmic amplification of affective testimony.
The phrase "evaluation is testimony" captures BookTok's critical architecture. On the platform, a reader's emotional response to a book isn't evidence in service of judgment—it *is* the judgment. Videos with the highest engagement involve BookTokers clutching novels verklempt or reading passages aloud while crying. One reviewer of Sarah J. Maas's *A Court of Mist and Fury*—a series whose hashtag has over two million posts—begins: "I was sitting right here, five minutes ago, and I finished the book, and I'm feeling every emotion possible… I'm feeling the depression and ecstasy of the [faerie kingdom] Hybern." These reviews claim, implicitly: this book is good because of how it made me feel, and it will make you feel that way because we're the same kind of reader. The reader's appraisal caves into their emotional response regardless of the book's ambition, craft, or complexity. If a book is emotionally pleasurable, that it's badly written "devalues it no more than spit on a fire."
The best literature rewards opacity, surprise, the friction of encountering minds unlike your own.
This isn't entirely new—Samuel Johnson appealed to "common readers" who read for pleasure over pedantic critics centuries ago. What's new is the systematic replacement of any evaluative distance between reader and text. BookTok's critical language presupposes a reader equally knowable across anyone who can read at all, a predictable range of emotional responses rather than any particular mind. Yet the best literature rewards opacity, surprise, the friction of encountering minds unlike your own. When relatability becomes fiction's primary metric, the algorithm shifts fundamentally. The author becomes a service provider; the protagonist must reflect the reader's emotional needs; the book succeeds insofar as it delivers therapeutic comfort rather than cognitive challenge.
The economics confirm what the critical language reveals. Publishers pay for predicted virality because TikTok's algorithm has proven more reliable than traditional gatekeepers at moving units. Many large BookTok accounts receive two thousand dollars per video, and twice that with publisher usage rights. The platform's affordances—the For You Page, Duet functionality, hashtag aggregation—allow content to circulate regardless of who created it, which feels democratic until you recognize that circulation depends entirely on engagement metrics optimized for emotional intensity, not textual quality. Videos showing readers crying get amplified; videos analyzing narrative structure don't. The algorithm doesn't care whether a book is well-written; it cares whether the video about the book generates watch time, shares, saves, comments. BookTok's "widespread cultural impact off platform," as Barnes & Noble described it when launching their 2022 BookTok Challenge, translates to a system where affective response drives sales with unprecedented efficiency.
This produces a remarkably narrow canon despite BookTok's self-conception as democratic liberation from elite taste. The platform's most-recommended "classics" are almost exclusively short books by white male authors—Oscar Wilde's *The Importance of Being Earnest*, John Steinbeck's *The Pearl*, Daniel Keyes's *Flowers for Algernon*—selected seemingly because users recognize the titles from high school syllabi and can signal seriousness without reading anything long or linguistically difficult. One BookTok user described *Flowers for Algernon* as "super old but brilliant," the kind of book "that no one actually reads unless assigned, except, perhaps, a precocious middle schooler." These recommendations sit alongside Colleen Hoover and Kristin Hannah, half a star separating Leo Tolstoy's *The Death of Ivan Ilyich* from Emily Henry's *Happy Place*. Both are "easy reading"—one because of length, one because of content—and easy reading is the chief aesthetic criterion. Recent online discourse about BookTok readers who skip long paragraphs or avoid books written in the third person suggests the logical endpoint of a system that privileges readability and relatability over textual complexity.
The genre that dominates BookTok—romance and its offspring, "romantasy"—reveals how algorithmic amplification and emotional identification reinforce each other. Romance historically developed as a genre where readers built counterpublic communities to resist cultural marginalization, creating spaces for emotional labor and fan-driven meaning-making. BookTok converts these resistant practices into behavioral data. The emotional authenticity that made romance communities resilient now makes them algorithmically valuable. When readers post videos crying over *A Court of Mist and Fury*, they're performing emotional labor that the platform extracts as engagement metrics, which publishers use to determine what to acquire and promote. The community's greatest strength—its capacity for collective feeling—becomes its exploitable vulnerability.
This creates what researchers call "algorithmic subjugation." While BookTok grants romance readers unprecedented cultural legitimacy and mainstream visibility—bookstores dedicate entire tables to "As Seen on TikTok"—this legitimacy operates within systems fundamentally indifferent to community welfare. Researchers note that pending TikTok bans at government levels reportedly threaten the platform's use in multiple countries, revealing how precarious algorithmically-mediated community actually is. More insidiously, the platform fragments solidarity even as it creates visibility. TikTok's algorithm curates each user's For You Page individually, so "BookTok" isn't a unified community but millions of personalized feeds that happen to overlap. Users experience this as intimate connection—a vast internet book club—while the platform harvests their affective responses as data.
Publishers and booksellers have entered this space not as transparent commercial actors but as participants claiming authentic connection with readers. They study BookTok to understand taste, then produce books designed for virality. They pay influencers while maintaining the fiction of organic recommendation. They've moved so far upstream that manuscripts get acquired based on predicted TikTok performance before anyone knows whether the books are good. The *Shy Girl* scandal is diagnostic: Hachette promoted it as a BookTok sensation and featured a BookTok influencer's blurb on the cover before pulling it after evidence emerged it was AI-generated. The system worked perfectly until it didn't, and even then, the community's response wasn't outrage about quality but embarrassment about having publicly loved something with "off" writing.
What gets lost in this transformation isn't just critical rigor but the range of literary experience available to mass readership. When evaluation collapses into emotional response, books that don't deliver immediate affective payoff disappear from circulation. Opacity becomes a liability. Cognitive friction—the experience of struggling with difficult prose, sitting with ambiguity, encountering perspectives that don't flatter your identity—gets systematically filtered out. The algorithm rewards books that make you feel seen, not books that make you see differently. This produces what looks like democratization but functions as enclosure: more people reading, but reading an ever-narrower range of texts that reproduce rather than challenge their existing emotional frameworks.
The institutional consequences are already visible. When Colleen Hoover outsells the Bible, when 70 percent of sales are backlist titles resurrected by viral videos, when publishers commission books based on predicted TikTok performance, we're not witnessing taste liberated from elite gatekeepers. We're witnessing its capture by algorithmic systems that encode very specific assumptions about which textual qualities deserve amplification—assumptions that privilege therapeutic comfort over aesthetic challenge, emotional identification over intellectual surprise, relatability over the irreducible strangeness of other minds. BookTok believes itself democratic while systematically narrowing what counts as literature worth reading. The gatekeepers haven't disappeared; they've been replaced by infrastructure that mistakes engagement metrics for literary value and calls the result freedom.