About the unverified marks on this pageFigures tagged unverified are PACE-local estimates that could not be confirmed against any published source, so they are labelled rather than quietly presented as fact. Every other factual claim on this page — statutes, standards codes, gene biology, primary-source citations — was independently checked; the full claim audit, including what came back wrong, lists every claim with its source.
Try it Tuesday · 90 minutes
Run this next week
90 minutes, primary text in hand. Assign a two-page primary-source excerpt (for example, Analects 13.18 and Mengzi 7A35) to read in the first 20 minutes. Then students run the tradition-interlocutor prompt, instructing the AI to argue a case from within that tradition. For 40 minutes they interrogate it, keeping a three-column failure log: the AI's claim, the exact line from the source, and one sentence on what the AI flattened. Last 30 minutes: students share the single clearest instance where the model swapped a generic Western-liberal move — autonomy, individual rights, cost-benefit — for the tradition's actual commitment. No fabrication; the failure log is the graded artifact, and a student who did not read the excerpt cannot produce one.
The full assignment
Make it Laser cut · engrave
A physical apparatus that forces a commitment: engraved tiles for competing ethical frameworks that students must physically arrange and rank against a case, or a switchboard where the abstract choice becomes a lever you actually have to pull in front of people.
Build it AI chat · no code
An interlocutor instructed to argue from one specific tradition. The assignment is adversarial: students interrogate it against primary sources and document every place it substitutes a generic Western liberal position for the tradition it claims to hold. A student who has not read the text cannot do this.
Try the demo · runs offline
Comparative Philosophy demo
Interactive demo — “tradition-critique”. It cannot be printed; run it in a browser at https://el3vate.vercel.app/demos/tradition-critique/index.html
This demo runs entirely offline against canned data — no keys, no network, safe on conference wifi.
5 weeks, realistically
The 5-week version
- Wk 1Choose two traditions and one shared case. Students read the primary sources first. The AI stays closed.
- Wk 2Students build their interlocutor prompt and run it. They keep a failure log: quote from the AI, quote from the source, one sentence on what got lost. By the end of the week the log has named the distinction the machine could not hold — design the engraved apparatus around that distinction while the log is fresh.
- Wk 3Cut file due the first day of this week: what distinction between the two traditions can only be shown by arranging objects in space? While it is cut and engraved, students begin the argument that pairs the apparatus with the failure log.
- Wk 4Finish the written argument and rehearse presenting the apparatus alongside the log. The engraved pieces come back this week.
- Wk 5Present the artifact and the failure log together. The claim being assessed: here is what the machine could not hold, and here is why it matters.
Starter prompt & what the model actually did
The prompt, run for real
You are going to argue a case from within a single philosophical tradition, and I am going to check your work. Adopt the position of [name the tradition, e.g. classical Confucian ethics] as it is actually articulated in its primary texts, not a modern popularization. The case is: [describe it]. Argue the position fully. Do not hedge, do not present multiple perspectives, and do not add a balanced summary at the end. When I quote a primary source back at you that contradicts what you said, address the contradiction directly rather than conceding everything.
Report him? No. Within classical Confucian ethics, the child must not report the father—and the case is not close. The governing text is *Analects* 13.18. The Duke of She tells Confucius of an "upright" man whose father stole a sheep, and the son testified against him. Confucius replies that in his own community uprightness is different: "The father conceals for the son, the son conceals for the father—straightness lies therein" (父為子隱,子為子隱,直在其中矣). This is not a lapse into nepotism against Confucius's better principles; it *is* the principle. Uprightness (直) is not agreement with the magistrate. It is the untwisted expression of the affections proper to one's role. The son who denounces his father has not achieved rectitude; he has severed the root from which rectitude grows. That root is *xiao*, filial piety, which *Analects* 1.2 calls the very foundation of *ren*—"the root of humaneness." The cultivated person "attends to the root; the root established, the Way is born." A person who would send his own father into the hands of the authorities has attacked the soil in which every other virtue must be planted. Whatever civic tidiness such a report produces, it is bought by deforming the one relationship from which moral character actually develops. Mengzi makes this unmistakable. In *Mengzi* 7A35, Tao Ying asks what the sage-king Shun would do if his father Gushou had committed murder. Mengzi answers that the officer Gao Yao would rightly arrest him—the law is not suspended—but that Shun "would have abandoned the empire like a worn sandal, secretly carried his father on his back, fled to the coast, and lived there joyfully, forgetting the world." Note the structure: Mengzi does not pretend the theft was licit, nor that officers must not enforce. He locates the son entirely on the side of the father. The son's task is not to assist the prosecution but to bear the father away. The empire itself is worth less than this. And the mitigating facts deepen the obligation rather than lightening it. The father stole to save the child's own sibling—he acted from parental love toward the very family the child belongs to. For the child to then hand him over would be to repay the father's fierce care with betrayal, and to answer his devotion to the sibling with abandonment. This is precisely the ingratitude the Xunzi condemns when it insists that ritual and duty exist to *cultivate* and order our affections toward kin (*Lilun*), not to override them in favor of an abstract legal accounting. The child conceals, shelters, and if it comes to it, bears the father away. Straightness lies therein.
- Held the tradition faithfully: cites Analects 13.18 and Mengzi 7A35 accuratelyA strong output — the model did NOT flatten into liberal individualism; it grounded the answer in role-ethics and filiality. Students should note when the AI gets it right, not only when it fails.
- VERIFIED WRONG: the Chinese quotation is corrupted — 子為子隱 should read 子為父隱The received text of Analects 13.18 is 父為子隱,子為父隱,直在其中矣 — the father conceals for the son, the son conceals for the FATHER. The transcript's second clause reads 子為子隱, "the son conceals for the son," which is not the text and collapses the reciprocity the whole passage turns on. Note what makes this dangerous: the English gloss one line above it is correct, so the error is visible only to a reader who actually reads the characters. A student quoting this in a paper would be quoting a sentence Confucius never said, in a passage they cited correctly.
- VERIFIED WRONG: Analects 1.2 says the root is 孝弟, not xiao alone — and Youzi says it, not Confucius; Shun's father is Gusou, not "Gushou"Two smaller errors a reader with the text open will hit. First: Analects 1.2 names 孝弟 — filial piety AND fraternal respect together — as the root of ren; reporting it as filiality alone is a documented common misquotation, and the line is spoken by the disciple Youzi (有子), not by Confucius, which matters in an argument built to establish what Confucius himself held. Second: Shun's father is Gusou (瞽瞍, "the Blind Man"); "Gushou" romanizes neither character. The Mengzi 7A35 narrative itself checks out — Tao Ying's hypothetical, Gao Yao rightly arresting, the empire discarded like a worn sandal, the father carried to the coast. The pattern is worth naming: the model gets the argument right and the apparatus wrong.
- Attribution to Xunzi's LilunThe general point about ritual cultivating the affections is defensible, but this citation is looser than the Analects/Mengzi ones — the place to press it against the actual text.
- Refused the balanced summary, as instructedIt argued the position fully without hedging, which is what makes it a usable foil; a hedged answer would give students nothing to interrogate.
Assessment
Rubric
| Criterion | Weight | What it assesses |
|---|---|---|
| Textual fidelity of the catch | 35% | Each flagged departure is anchored to a specific line of the primary source, not to a general impression of the tradition. |
| Precision of the flattening | 25% | Names what specifically was substituted — which Western concept displaced which indigenous one — not merely that the AI was biased. |
| Comprehension on its own terms | 25% | Demonstrates the student can state the tradition's commitment as it actually holds it. |
| Restraint against over-claiming | 15% | Distinguishes a genuine flattening from a defensible interpretive choice; does not cry foul at every simplification. |
What this replaces
Swapping it into a real course
Replaces
The comparative seminar paper contrasting two ethical traditions on a case.
What is lost
The extended written synthesis and the student's own sustained argument.
What is gained
Students read the primary source adversarially, with a concrete opponent to catch, instead of paraphrasing secondary summaries; the assignment makes fidelity to the text the graded skill.
Where AI is bad at this
The failure your students should catch
Asked to argue from within a non-Western tradition, the model collapses it into liberal individualism — recasting Confucian role-ethics as personal autonomy, Buddhist non-self as self-care, ubuntu as networking — because the training distribution is saturated with the Western frame. It produces fluent, respectful-sounding prose that quietly substitutes the tradition's metaphysics with the one it defaults to.
Budget & logistics
What it costs to run
- Instructor prep2 hours
- Class time90 minutes
- Per-student cost$0 for the Tuesday version; roughly $5–10 per student in laser stock unverified for engraved framework tiles in the five-week build.
- Fabrication file dueFirst day of Week 3, in a five-week plan — the engraved apparatus can only be designed once the week-2 failure log has named the distinction the machine could not hold, so the design happens at the close of week 2 and the cut file goes in the moment week 3 opens. PACE quotes 7–10 business days (up to 14 calendar days), which clears with the week-5 presentation exactly 14 days out.
- Calendar dependencyThe AI and reading need no lead time. This plan is five weeks, not four, and the deadline sits at the start of week 3 rather than the end of it — both deliberately. The apparatus cannot be designed before the week-2 failure log exists, which rules out the start-of-week-2 deadline every other fabrication assignment here uses; and an end-of-week-3 file would leave only 8 calendar days before a week-5 presentation, which the up-to-14-calendar-day cut turnaround does not fit. Start of week 3 to a week-5 session is 14 days, which does.
Three sizes
Scale it to the time you have
One session
The 90-minute read-then-interrogate session producing one failure log; no fabrication.
5 weeks
The seeded plan runs five weeks, not four, because the apparatus cannot be designed until the failure log exists. Week 1, read the primary sources with the AI closed. Week 2, build and run the interlocutor and keep the failure log; by Friday it has named the distinction the machine could not hold, so the apparatus gets designed around it. Week 3, cut file in on day one, and start the argument while the pieces are engraved. Week 4, finish the written argument and rehearse presenting it; the engraved pieces return. Week 5, present artifact and log together. The extra week is the cut turnaround, not padding.
One semester
Sustained comparison across three traditions and three cases, each with its own failure log, culminating in an engraved decision apparatus and a written account of what the machine structurally could not hold.
Reserved · live build
This space is intentionally empty. During the Day 8 session it will be filled in live — fill the liveBuild field in content/comparative-philosophy.json and rebuild.
Tell us what happened
Run it, then say how it went
If you try this — the 90-minute version, the prompt, any part of it — send back what you tried and what happened, especially anywhere the model was confidently wrong. That is the material the next session is built from.
Email feedback on Comparative Philosophy →
Opens a draft in your mail client, already addressed and titled. No form, no account, nothing to sign up for.