Seed Notes: Xu Ben and the Civilisation-Diagnostician Turn
June 16, 2026
These are seed notes — observations collected from reading Xu Ben's recent AI articles, cross-checked against his earlier work, and placed within a broader pattern of how Chinese public intellectuals operate in the 2020s. This is a field being planted, not a harvest. Treat accordingly.
What Triggered This
In June 2026, Xu Ben (徐贲) — a retired English literature professor at Saint Mary's College of California, known in Chinese-language public discourse for his writings on civic education, totalitarianism, and cynicism — published an article titled "Is 'AI Cannot Replace Humans' a Rhetorical Consolation?" (AI无法取代人，是一种"修辞性的安慰"吗？). It was one of at least five articles published in Beijing News (新京报) between October 2025 and June 2026, all circling the same theme: what AI means for human subjectivity.
My first reaction was: what happened to him?
He used to write about specific things — political institutions, civic education, the mechanics of authoritarian control. Now the titles read like existential philosophy seminars. This note tracks where that shift came from and what it might tell us about a broader pattern — what I'm calling the "civilisation-diagnostician turn": a shift from analysing bounded institutional or historical problems to interpreting broad symptoms of modern humanity, civilisation, and subjectivity.
The Five Articles
All published in 新京报书评周刊 (Beijing News Book Review):
| Title | Date | Nature |
|---|---|---|
| 专访徐贲：人工智能时代，我们如何守护"完整的人"？ | 2025.10 | Interview promoting Does Humanity Still Have Hope? |
| AI时代，"自由选择"如何成为真正的重负？ | 2025.11 | Independent essay |
| AI如何改写人类的战争伦理？ | 2026.03 | Independent essay |
| AI时代，别把"人类特点"自动当成"人类优点" | 2026.05 | Extended commentary referencing the book |
| AI无法取代人，是一种"修辞性的安慰"吗？ | 2026.06 | Promotional for two new books |
These are not excerpts from his 2025 book Does Humanity Still Have Hope? (《人类还有希望吗》, 上海三联, 772pp). They form a media-publishing promotional matrix: interviews, extended commentaries, and independent essays timed to book launches. The June 2026 article already references his next two books — A Brief History of Loneliness (《孤独简史》) and Loneliness, Emptiness, and Boredom (《孤独、空虚与无聊》) — both due later that month.
The Formula
Across the five articles, a repeating argumentative structure emerges:
- Open with a popular claim — e.g., "AI amplifies human ability but cannot replace human purpose" (Huang Jensen)
- Announce: "It's not that simple"
- Deploy 3–5 theorists — rotate among Kant, Nietzsche, Sartre, Frankl, Zuboff, Life History Theory, Self-Determination Theory
- Arrive at: the real problem is human subjectivity
The war ethics article (Article 3) is the exception — it has real case studies (Israel's "Gospel" targeting system, kill-chain compression, algorithmic execution mechanisms) and genuine policy analysis. It is noticeably better than the others.
What's Actually Going On in the Arguments
The Strong Move
His critique of Huang Jensen's "human purpose" rhetoric is legitimate. "Human purpose" lumps together Newton's goals, Mandela's goals, a gambler's goals, and a Nazi's goals into one feel-good phrase. As a piece of concept analysis, this is effective.
The Weak Move
After dismantling "human purpose" as a unified concept, he replaces it with a new hierarchy: integrated long-term goals (slow strategy) vs. fragmented desires (fast strategy), drawing on Life History Theory. He does include a disclaimer: "承认动机整合水平的差异，与把这种差异归因于固定的、不可改变的内在本质，是根本不同的两件事" (Acknowledging differences in motivational integration is fundamentally different from attributing those differences to fixed, unchangeable inner essences). But the article's own language tells a different story. People on the fast-strategy end are described as "越来越成为快策略激活下的可预测反应节点——仍然在追求，却越来越难以真正拥有追求本身" (increasingly becoming predictable reaction nodes activated under fast-strategy patterns — still pursuing, yet less and less capable of truly possessing their pursuits). Humanity is framed as "一个内部高度分化的生物集合" (an internally highly differentiated biological aggregate) with individual differences in "道德结构" (moral structure) and "反思深度" (depth of reflection) that "可能远远超过大多数物种内部个体之间的差异" (may far exceed intra-species variation in most other species).
The disclaimer says "this is not elitism." The vocabulary says: some people are predictable reaction nodes, others are deepening their cognition with AI at unprecedented speed. The hierarchy is not argued. It is built into the choice of metaphor.
AI as Trigger, Not Object
These articles are nominally about AI. But AI is not treated as an independent technical or philosophical object — there is almost no engagement with questions like whether AI has or could develop subjective experience, what machine learning architectures actually do, or how specific AI systems change specific practices. Instead, AI functions as a trigger for a humanistic self-understanding crisis: it is the thing that forces the question "what makes humans special?" back onto the table. This is not necessarily a flaw — it is a choice of genre. But it means these articles belong to the tradition of humanistic defence literature, not AI philosophy. Recognising this reframes the criticism: the question is not "why doesn't he engage with AI?" but "what does it mean that AI is being used as the occasion for reasserting humanistic authority?"
The Unevenness Pattern
Xu Ben's quality has always been uneven — this is not new. His 2015 essay collection Decadence and Silence: Perspectives on Cynical Culture (《颓废与沉默》, 东方出版社) drew the same criticism on Douban: the theoretical first section is strong, the second and third parts degenerate into repetitive op-eds. Multiple reviewers noted the book reads like two different authors. I read parts of it myself a few years ago and felt the same fracture — the first section reads like someone who has genuinely lived inside the problem of cynicism and come back with a map; the later sections, especially the pieces on American university culture, read like a retired professor sighing into his coffee about kids these days. The shift is not subtle. You feel it in your hands as you turn the pages: the weight of the argument drops out, and what's left is opinion wearing the costume of analysis.
The more precise way to describe this is not "he's bad at some topics" but that he deploys different evidence regimes depending on the subject matter, and the quality tracks accordingly. When the topic has concrete cases, technical mechanisms, or policy questions to anchor the argument — as in the war ethics article, or the cynicism theory chapters — the writing achieves high argumentative density. When the topic is inherently abstract ("what is a complete human?", "what is subjectivity?"), he defaults to cycling through canonical thinkers (Kant → Nietzsche → Sartre → Frankl) without the kind of grounding that makes the concrete pieces work. The result is not incompetence but a consistent asymmetry: his evidence regime for applied analysis is rigorous; his evidence regime for philosophical synthesis is rhetorical.
The Publication Trajectory
This is perhaps more revealing than any single article:
| Period | Work | Self-positioning |
|---|---|---|
| ~2000 | Academic papers — e.g. commentary on Arendt's civic thought | Researcher |
| 2015 | 《颓废与沉默》 — cynicism essay collection | Public commentator |
| ~2023 | 《西方人文经典讲演录》(4 vols) — "from Ancient Greece to modernity" | Educator / "I'll fill the gap in Chinese students' humanities training" |
| 2025.05 | 《人类还有希望吗》(772pp) — AI-era humanistic enlightenment | Civilisation diagnostician |
| 2026.06 | 《孤独简史》+ 《孤独、空虚与无聊》— two books in one month | ... |
The trajectory: researcher → commentator → educator-for-a-nation → civilisation diagnostician. The four-volume humanities lecture series is the hinge point. That is where the self-positioning shifts from "I study things" to "I teach the nation." Once that identity is established, the move toward diagnosing civilisational crises becomes natural.
Three books on AI and the human condition published within thirteen months (May 2025 – June 2026), totalling well over a thousand pages, plus a stream of media articles. A caveat: academic publishing in China typically involves a one-to-two-year lag between manuscript delivery and publication. These books were likely written over a longer period than the publication dates suggest — the 2025 book may have been drafted as early as 2023, and the two 2026 titles probably overlapped with the lecture series period. The clustering is partly a product of publishing schedules, not necessarily of writing speed. Still, the sheer volume of output across books and media articles in this period is notable.
Broader Pattern: Not Just Xu Ben
The observation that triggered these notes is not really about one person. It's about what looks like a cohort phenomenon among Chinese public intellectuals in the 2020s. The following table is a set of preliminary observations, not a demonstrated conclusion — it maps impressions from reading these figures' public output, not a systematic comparison. It is included here as a hypothesis to be tested, not a finding.
| Figure | Primary mode | Observed AI-era tendency |
|---|---|---|
| 徐贲 Xu Ben | Civic education / cynicism critique | Civilisation diagnosis |
| 项飙 Xiang Biao | Ethnographic description ("the disappearance of the nearby") | Social-mood diagnostician — names feelings, stops before explaining causes |
| 秦晖 Qin Hui | Institutional comparative history | Cross-border overreach (Middle East commentary influenced by personal network) |
| 刘擎 Liu Qing | Public philosophy / media-friendly theory | Media-intellectual par excellence |
| 王笛 Wang Di | Micro-social history (teahouses, street vendors) | Counter-example — actively resists grand narrative |
To move this from impressions to evidence, the first step would be narrowing to a comparable corpus: e.g. 2018–2026 interviews and essays from 新京报书评周刊, 澎湃思想市场, and 上海书评. The thing to demonstrate first is whether a measurable trend exists — whether argumentative scale increases, whether civilisational vocabulary ("主体性," "完整的人," "文明危机," "意义感") becomes denser, whether conclusions shift from "this mechanism works like X" to "modern humanity / civilisation / subjectivity is in crisis." Only after that trend is established does it make sense to compare individual trajectories.
Defining "Civilisation Diagnostician"
The term needs to be operationalised if it is to do analytical work. Tentative criteria:
| Dimension | Specific analysis | Civilisation diagnosis |
|---|---|---|
| Argumentative scale | Institution, event, policy | Civilisation, humanity, the human condition |
| Evidence type | Cases, archives, fieldwork, data | Conceptual concatenation — rotating through canonical thinkers |
| Author posture | Analyst of a bounded problem | Interpreter of an era's symptoms |
| Conclusion form | "This mechanism works like X" | "Modern humans / civilisation / subjectivity is in crisis" |
A scholar can move between these modes. The hypothesis is that certain figures have shifted predominantly toward the right column in the 2020s, and that this shift correlates with identifiable pressures.
Five Hypotheses for the Turn
-
AI shock hypothesis — AI erodes the knowledge monopoly that gave intellectuals authority; they retreat from "I know things" to "I interpret meaning"
-
Media mechanism hypothesis — Book-review media and interview platforms reward civilisational-scale questions over specific institutional analysis; scholars adapt to the market
-
Age-cohort hypothesis — Scholars in their 60s–70s naturally shift from "what is this problem?" to "what does my life's work mean?"
-
AI-tutor hypothesis — AI tools (especially sycophantic ones) lower the cost of cross-disciplinary synthesis while inflating confidence in one's ability to integrate fields one hasn't mastered. (This applies reflexively to the present note: AI collaboration can sharpen a research hunch, but it can also make an under-tested synthesis feel more coherent than the evidence yet warrants.)
-
Institutional despair hypothesis — When public intellectuals judge that institutional reform is no longer plausible, speakable, or actionable, they relocate the burden of historical agency onto the individual subject. The turn toward "subjectivity," "the complete human," and "meaning" is not simply an expansion of humanistic ambition — it may be a displacement produced by the closure of political exits. What was once "how do institutions fail?" becomes "how do individuals preserve integrity?"; what was once "why can't society organise effective action?" becomes "why are modern people lonely, empty, and bored?"
This reframing has a generous reading and a dangerous one. The generous reading: when large structures cannot be moved, preserving individual capacities for judgement, shame, empathy, and precise language genuinely matters. The dangerous reading: it risks converting structural failures into problems of personal cultivation — as though people becoming more "complete" or more "subjective" could compensate for systems that remain broken.
Xu Ben's trajectory fits this pattern closely. His earlier work — civic education, totalitarianism, cynicism in public life — addressed institutions and collective political culture. His recent output — "the complete human," "subjectivity," "loneliness," "emptiness," "does humanity still have hope?" — has relocated the same underlying anxieties onto the individual. The surface looks like humanistic expansion. The substrate may be political contraction.
These are not mutually exclusive. They probably interact. But the fifth hypothesis may operate at a deeper level than the others: it addresses not just why the topic changed but why certain topics became unsayable, which would make the diagnostician turn partly a symptom of discursive constraint rather than intellectual choice alone. It also cannot be verified through authorial confession. It would need to be tested indirectly: by tracking whether institutional vocabulary, reform-oriented causal explanations, and collective-action imaginaries decline over time in a defined corpus, while subjectivity, moral psychology, and individual self-preservation become more prominent. It should also be allowed to fail. If the same authors continue to discuss institutional design, collective action, and reform pathways in concrete and sustained ways during the same period, then "institutional despair" becomes at most a partial explanation, not the hidden master key.
The Core Question
The most productive reframing may be: don't ask what subjectivity is (that's Xu Ben's question); ask why so many people suddenly started talking about subjectivity in the 2020s (that's the intellectual-history question).
Why did Chinese-language public intellectuals in the 2020s shift from specific institutional, historical, and social analysis toward diagnostic discourse centred on "subjectivity," "the complete human," "civilisational crisis," and "meaning"?
This can be decomposed into three layers:
- Text layer — How do keywords, titles, conceptual combinations, and argumentative structures change across a defined corpus over time?
- Media layer — How do book-review platforms, interview formats, and new-book promotion cycles reward "big questions" over bounded analysis?
- Knowledge-sociology layer — When AI erodes the knowledge monopoly that underwrote intellectual authority, how do public intellectuals pivot toward "meaning-interpretation rights" as a new basis for relevance?
This shift — from engaging with the content of a concept to asking why the concept has become urgent at this particular historical moment — is where the research potential lives. It is also where the comparison with earlier waves becomes possible: the "civilisation decline" anxiety of 1890s European intellectuals, the "alienation" discourse of 1950s existentialism, the "end of history" moment of the 1990s. Each wave produced its own version of "what makes humans human?" — and each wave's answer tells us more about the anxieties of its producers than about the question itself.
Xu Ben is an entry point into this question. He is not the whole answer.
Materials for Future Work
- 爱思想 (aisixiang.com) — Archives extensive collection of Xu Ben's essays and op-eds, chronologically sortable. Potential corpus for diachronic topic analysis.
- 新京报书评周刊 — Primary media partner for Xu Ben's recent work; also publishes interviews with 刘擎, 许纪霖, and others in the same cohort.
- 澎湃思想市场 — Another major platform for public intellectual discourse.
- Douban reviews of Xu Ben's books — reader responses that track quality perception over time.
- Xu Ben's Caixin blog — Contains older material including table of contents from the Oxford-published Cynicism and Jokes (《犬儒与玩笑》, 牛津大学出版社 2018), which cannot be published in mainland China.
If this ever becomes a real project:
The most feasible entry point would be a diachronic text analysis of scholar interviews across 2–3 platforms (新京报, 澎湃, 上海书评) from 2018–2026, tracking shifts in keyword frequency and argumentative scale. AI tools can handle the coding and extraction; the interpretive framework is the human part.
Possible quantitative features to track:
- Conceptual abstraction ratio — references to canonical thinkers and abstract categories per 1,000 characters, compared with concrete cases, institutions, technologies, or policy mechanisms. The ratio may be a proxy for whether an article engages with AI as a technical object or uses it as a humanistic backdrop.
- Pronoun and stance markers — frequency of "我们" (we), "人类" (humanity), "应当" (ought to), "警惕" (beware) — quantifying the shift from analytical to pedagogical/prophetic voice.
But that is future work. For now, these are seeds.
Assembled 2026-06-16 through human-AI collaboration and revised against the original articles. Several AI systems helped generate, check, and sharpen the analytical frame, but the final judgements, remaining uncertainties, and responsibility for publication are mine. This note is therefore also implicated in the phenomenon it describes: AI-assisted synthesis can clarify a research hunch, but it can also make provisional patterns feel more stable than the evidence yet warrants.