Worked sample · printable PDF
HUM-04 · Mechanical reasoning, historically
Connect a historical calculating/logic machine to a modern LLM claim; separate continuity from marketing.
Specialty centers · Machines & Meaning
Twelve printable humanities-and-AI briefs you can assign this term
Free humanities briefs about what a language model reveals — and does not reveal — about language. Students connect a historical calculating machine to a modern marketing claim without Whig history, and they measure something falsifiable about voice under AI assist. Every brief has its own PDF, and all twelve are in the bank PDF. No account. No student data leaves your school.
Non-claims about mind and meaning
Every major artifact ends with an explicit non-claim list about mind, meaning, consciousness, or authorship as essence. Measure something falsifiable when you study voice or style — or admit you are doing interpretive work without a metric. Corpus ethics: only texts you have rights to use.
How to run one
- Students choose the brief. You approve the research question, corpus rights, and measure (or interpretive frame) in writing first.
- Two scheduled checkpoints. Score with the six questions below.
- Mentor is advisory — school-supervised, never one-to-one unsupervised.
Featured packets
Featured · printable PDF
HUM-01 · Authorship after autocomplete
Design a study of voice change with AI assist; measure something falsifiable about style or revision.
Catalog (10 briefs on the page)
- HUM-02 · Corpus you can defend
Build a small, documented corpus; justify inclusion/exclusion like a digital humanist.
- HUM-03 · What tokenization erases
Show a meaning-bearing distinction that breaks under BPE/WordPiece — and why it matters for a text you care about.
- HUM-05 · Translation and loss
Compare human and machine translation on a short literary or historical passage; annotate what is lost.
- HUM-06 · Citation theater
Audit an LLM's citations on a humanities query; design a verification protocol students must follow.
- HUM-07 · Close reading vs. distant reading
Run a simple distant-reading pass, then return to close reading — show what each method cannot see.
- HUM-08 · Language ideology in prompts
Analyze how prompt wording encodes norms about "proper" language; propose ethical prompt practice for classwork.
- HUM-09 · Multilingual center brief
Study model performance across languages relevant to your program; refuse English-only conclusions.
- HUM-10 · IB / AP Research scope lock
Write the research question and method section that fits the official rubric without scope inflation.
- HUM-11 · Public humanities deliverable
Adapt the essay into a public-facing explainer (talk, exhibit text, or podcast script) that stays honest.
- HUM-12 · The claim you will not make
End with an explicit non-claim list — what your computational result does not prove about mind, meaning, or authorship.
Six-dimension rubric
- R1 — Is one text set, one question, and one claim type (empirical vs interpretive) named?
- R2 — Could another student rebuild the corpus step and method from the appendix?
- R3 — What does close reading show that the computational pass cannot?
- R4 — Are tokenization artifacts, sampling bias, and prompt effects recorded?
- R5 — Does the essay keep computational limits visible to a humanities reader?
- R6 — What does the student explicitly refuse to claim about mind, meaning, or authorship?
Feedback
Scope unclear? Standards language off? Tell us here — you do not have to leave an email.