Discipline Guides

🌐 English · 中文 · 🏠 Home

Discipline Guides

How the Humanities Writing Companion adapts to your field — and how to make it adapt better.

This skill grew out of one specific humanities dissertation (in art history / philosophy / digital cultural studies). It is not a one-size-fits-all polisher. A history chapter and a philosophy chapter fail in different ways, so the skill needs to know which kind of writing it is reading before it gives feedback. This page explains the mechanism that lets it do that, walks through each humanities field it targets, and shows you how to test it in fields it doesn't yet know well.

---

1. The core idea: discipline is a routing variable, not metadata

Most writing tools treat "your discipline" as a label on a form. This skill treats it as a switch that changes what counts as a mistake.

When you say you write history, an anachronism (projecting "the individual" or "the nation" onto a 14th-century actor) becomes a foundation-level error — something to fix before any sentence-level polish. When you say you write analytic philosophy, a hidden inferential step or a sloppy use of "necessarily" becomes the thing to catch first. Same paragraph, different field, different critique.

The rule the skill follows internally: "Discipline is not metadata — it is a routing variable. Read it every time you give critique." Discipline-specific concerns sit at Layer 1 (Foundation) of the four-layer critique, above structure, evidence, and sentence-level issues.

Three layers of declaration: L1 / L2 / L3

When you onboard, the skill asks for your discipline in three layers and records them in a small config file (_writing-config/discipline.md, or 学科档案.md in Chinese projects):

Layer	What it is	Required?	Example
L1	Your main discipline	One required	Literature · History · Philosophy · Linguistics · Art studies · Religious studies
L2	A specific subfield, inheriting L1	Optional, recommended	古代文学 (classical Chinese lit) · 经济史 (economic history) · 分析哲学 (analytic philosophy) · 艺术史 (art history)
L3	Cross-disciplinary field(s), often more than one	Optional	Intellectual history · Classics · History of science · Cultural studies · Media studies · Digital humanities · Gender studies · Postcolonial studies · Environmental humanities

The three layers stack additively, they don't replace each other:

• L1 only → load that discipline's methodology dimensions.
• L1 + L2 → load L1's dimensions, then add the subfield's extra constraints (e.g., 古代文学 adds philological / manuscript-tradition concerns on top of Literature).
• L1 + L3 → this is where it gets interesting. An L3 field inherits the dimensions of all of its parent L1s, plus a field-specific overlay. Intellectual history (思想史) = History + Philosophy + an intellectual-history overlay. You get all three sets of concerns at once.

What "overlay" means, concretely

Think of the overlay as a thin extra checklist layered on top of the inherited L1 concerns. When you declare intellectual history, the skill loads History's concerns (anachronism, source handling, historiographical positioning) and Philosophy's concerns (conceptual rigor, exegesis vs. intervention) and the intellectual-history overlay (which method — Begriffsgeschichte or Cambridge School? have you paid the "transport cost" when a concept migrates across periods?).

In the devil's-advocate mode, each layer's critique is even tagged by its source so you can tell where an objection comes from:

[Mode D · methodology · L1 History]      Source bias — whose interest is served by remembering it this way?
[Mode D · methodology · L1 Philosophy]   When you say "necessarily" here — logical, metaphysical, or moral?
[Mode D · methodology · L3 Intellectual history]   Which method? Begriffsgeschichte or Cambridge School?

What happens when nothing fits

If your field matches no L1 / L2 / L3 entry, the skill runs a fallback protocol: it asks for your object of study (text? past event? concept? image? practice?) and your primary method (close reading? archival? argument analysis? ethnography?), infers the closest L1 plus relevant overlays, and tells you it's a best-fit approximation — then invites you to refine the declaration as the project develops. It will not silently give you generic critique when you have a real discipline.

Two more routing behaviors worth knowing:

• Straddle naming. When a passage does history on the surface but philosophy at its foundation, the skill says so out loud — "let's critique both layers separately" — instead of flattening it.
• Migration logging. Theses sometimes change frame mid-project. If you re-declare your discipline, the skill updates the config and logs the change in your revision log.

---

2. Per-discipline guidance

Below: for each field, what it's typically used for, the specific concerns the skill watches for, and any special handling. The six L1 disciplines come first, then the cross-disciplinary L3 cases.

History · 史学 (L1)

Typical use — Chapter review and devil's advocate on narrative-heavy drafts; checking whether a story has quietly turned into an argument (or failed to).

What the skill watches for:

Concern	The question it asks
Causal reasoning	Is temporal sequence being smuggled in as causation? "X then Y" is not "X caused Y."
Source handling	Primary or secondhand? Is attribution clear? Are the actors' own interests in being remembered a certain way accounted for?
Anachronism	Are modern categories — "intellectual," "nation," "individual," "the public" — projected onto periods where they didn't yet exist?
Narrative-vs-analysis ratio	Has the chapter slipped into telling a story and forgotten to do analysis?
Counterfactual stress-test	If X had been absent, does the causal claim still hold — or is this just a chronicle?
Historiographical positioning	Which tradition are you extending or arguing with? Annales? Cambridge School? Subaltern Studies?

Standard arc (Mode J, plan-only): historiographical positioning → narrative → analytic argument → broader significance / re-periodization.

Special handling & honest limit: the skill's "foundation critique" layer (scholarly contribution / conceptual explanatory power) has a mild philosophical tilt, and may under-rate history's distinctive "newly discovered source" kind of contribution. If you see this, it's a known soft spot — report it.

Philosophy · 哲学 (L1)

Typical use — Running the four-layer critique on a core argumentative passage; formalizing an argument; pressure-testing against the strongest objection.

What the skill watches for:

• Rigor of derivation — is every step from premise to conclusion accounted for, or are there hidden steps?
• Conceptual distinctions — are key concepts defined, and are the distinctions real rather than merely verbal?
• Cross-theoretical transplantation — when a concept is grafted from one thinker onto another framework, has its meaning shifted, and is the shift acknowledged?
• Steel-manning — does the chapter face the strongest opposing view head-on, or only knock down weaker ones?
• Abstract–concrete respiration — long stretches of pure abstraction with no example are a red flag; examples test whether an abstract claim has bite.
• Exegesis vs. intervention — are you reconstructing what Thinker X said (demands textual fidelity) or arguing a new position via X (demands philosophical defensibility)? Different evidentiary standards.
• Modal scope — when you write "necessarily," is it logical, metaphysical, nomological, or moral necessity? Conflating these is a classic foundation-level failure.

Standard arc: concept-puzzle → conceptual analysis → defense against strongest objection → consequences for downstream debate.

Special handling: the skill leans toward "dialogic advancement" (acknowledge predecessors, then critique). The analytic tradition often prefers to directly assess argument validity without the courtesies. If you're analytic and this grates, that's a real stance conflict — calibrate the devil's advocate higher and tell us.

Literature / Literary criticism · 文学 (L1)

Typical use — Checking that interpretation stays anchored to the text; assessing close-reading citation density; simulating a disciplinarily literate reviewer (Mode D).

What the skill watches for:

• Close reading vs. interpretation — does the reading sustain the interpretive claim, or does interpretation float free of the text? Every claim needs textual anchor points.
• Theoretical scaffolding — does the frame (psychoanalytic, postcolonial, deconstructive…) illuminate the text, or has the text become a pretext for the frame?
• Quotation as evidence — are quoted passages evidentiary or decorative? Does the analysis engage what's in the quote, or just orbit it?
• Author / implied author / narrator — is the chapter conflating the three (especially in narrative)?
• Genre awareness & form-meaning fit — are genre conventions respected or read against (and if against, is the contrarian reading earned)? Are formal features read as carrying meaning?
• Intertextuality — are allusions analyzed, not merely spotted?

Standard arc: theoretical frame → close reading → generalization back to the frame / re-reading the frame through the text.

Special handling & honest limit: the skill's distinction between "analytical exposition" and "aesthetic exposition" isn't fine-grained enough yet — in literary studies, aesthetic judgment is itself legitimate exposition, not a lapse from rigor. A known area to push on.

Art history · 艺术史 (L2 under Art studies)

Typical use — Keeping formal description and interpretation cleanly separable; verifying attribution and dating claims; making sure figure references do analytical work.

What the skill watches for (inherits all of Art studies · 艺术学):

• Description vs. interpretation — does the formal description (composition, material, technique, iconography) sustain the interpretive claim, and are the two structurally separable?
• Provenance and dating — are attributional claims evidence-backed, or leaning on catalog tradition?
• Materiality — has the chapter engaged the medium's material conditions (paint, marble, celluloid, digital substrate), or treated all art as abstract content?
• Contextualization — are patronage, production conditions, and intended viewing context treated as constitutive of meaning, or as background decoration?
• Reception history — is the work's later reception distinguished from its original context, and are anachronistic readings flagged?
• Medium-specific vocabulary — music, cinema, and painting have different formal vocabularies; is the chapter using the right one?

Standard arc: description → contextualization (provenance, materiality, reception) → interpretation → consequences for the history of seeing / hearing / perceiving.

Special handling: as an L2 subfield, art history inherits the full Art-studies dimension set; if your work adds e.g. conservation or material-culture concerns, declare them and they're treated as additive.

Religious studies · 宗教学 (L1)

Typical use — Source-language checking against originals; situating a reading within long interpretive traditions; keeping emic and etic claims distinct.

What the skill watches for:

• Source-language rigor — when citing Greek, Latin, Hebrew, Arabic, Sanskrit, Pali, or Classical Chinese, are translations checked against the original? Are interpretive translation choices flagged?
• Tradition awareness — within Pauline studies, Quranic exegesis, Confucian commentary, etc., is your position situated relative to existing schools?
• Insider–outsider (emic vs. etic) — is your stance (devotional, agnostic, critical, comparative) acknowledged where it shapes interpretation? Etic claims that ignore emic understanding misread; unexamined emic claims drift into apologetics.
• Cross-tradition comparison — are categories like "mysticism," "salvation," "ritual" defined within each tradition, not just from the comparison's frame?
• Practice vs. text — does the chapter over-privilege text where lived practice matters more, or vice versa?

Standard arc: text-philological work → tradition-positioning → interpretive argument → bearings on contemporary scholarship.

Classics · 古典学 (L3)

Inheritance: Literature + History + Philosophy + Religious studies + Archaeology.

Typical use — Work where the argument turns on a specific word in the original, and where the manuscript tradition is part of the evidence.

Overlay the skill adds on top of the inherited L1s:

• Textual criticism — is the manuscript tradition acknowledged? Are textual variants relevant to your interpretation discussed?
• Philological rigor — are translation choices defended? Is the original consulted where the argument hinges on a word?
• Reception history — is later (medieval, early modern, modern) reception distinguished from the ancient context, and is your own modern lens acknowledged?

Intellectual history · 思想史 (L3 — History + Philosophy)

This is the flagship cross-disciplinary case, and the clearest illustration of the overlay idea.

Inheritance: History and Philosophy — you get both full L1 dimension sets.

Typical use — Tracking how a concept moves across periods and traditions; reconstructing the actual problem-space a past thinker was addressing rather than projecting today's concerns onto them.

Overlay the skill adds on top of History + Philosophy:

• Method declaration — are you doing Begriffsgeschichte (concept history, Koselleck), Cambridge School contextualism (Skinner / Pocock), histoire des mentalités, or something else? Each carries different evidentiary standards.
• Context vs. text balance — reading the text in its context, or imposing context on the text? Both errors are common.
• Avoiding presentism — judging past thinkers by present standards, or reconstructing their problem-space on its own terms?
• Concept migration — when a concept travels (medieval → early modern; Greek → Arabic → Latin), have you paid the transport cost and tracked what changed?

Standard arc: method declaration → context reconstruction → text analysis → concept-migration narrative.

Special handling — layered, tagged critique. Because three layers apply at once, devil's-advocate objections come tagged by source ([... L1 History], [... L1 Philosophy], [... L3 Intellectual history]), so a presentism worry, a modal-scope worry, and a method-declaration worry never blur into one undifferentiated note.

Other cross-disciplinary fields the skill knows (L3)

Each inherits its parent L1s and adds a focused overlay:

L3 field	Inheritance	Overlay highlights
Cultural studies · 文化研究	Literature + History + Sociology	Power-as-wand check (name the mechanism, don't wave "power"); positionality that's substantive not boilerplate; defended generalization range.
History of science · 科学史	History + Science + Philosophy	Internalist vs. externalist (or inseparable?); Whig-history vigilance; technical accuracy — get the science wrong and credibility evaporates.
Media studies · 媒介研究	Literature + Cultural studies + Philosophy of technology	Medium as substantive variable vs. mere channel; production vs. reception claims; tech-social determinism vs. co-construction.
Digital humanities · 数字人文	Any L1 + Computation	Data reproducibility; tool transparency; algorithmic bias acknowledged as shaping findings, not neutral.
Gender studies · 性别研究	Literature + History + Cultural studies	How gender is conceptualized (identity / performance / structure); historicizing gender categories; intersectionality as intersecting not additive.
Postcolonial studies · 后殖民研究	Literature + History + Cultural studies	Whose voice is centered vs. theorized-about; translation as politics; the import cost of European frameworks read onto non-European material.
Environmental humanities · 环境人文	Literature + History + Science	Anthropocene / Capitalocene framing as a position not a neutral given; more-than-human agency; justified scale-jumps.

Humanities-adjacent fields are welcome too. Communication studies (the media-ecology tradition — Innis, McLuhan, Postman, Carey) and educational research (history / philosophy of education, critical pedagogy) are supported in their humanities-style sub-traditions where prose is the argument. Their empirical / quantitative sub-traditions (effects studies, learning-outcomes research) are explicitly out of scope — the four-layer critique still works on the prose, but the discipline dimensions won't fit.

Always-on: case-study analysis

Whenever you use a case study as a method — in any field — an extra set of dimensions loads on top of your discipline's: case-selection representativeness (why these cases?), two-way movement between case and theory (does the case modify the theory, or just illustrate it?), internal tensions among cases, analytical value of narrated detail (cut the encyclopedic background), and calibrated generalization (typicality? exception-illuminating-rule? deviant case?).

---

3. Test it in your field — and tell us what breaks

The skill knows the fields above with varying confidence. Its comfort zone is media/cultural studies, philosophy, and art history; its known soft spots are flagged honestly throughout. The single most valuable thing you can do is use it for real in your discipline and report back — especially when it gets something wrong.

The feedback that helps most (in order):

1. 🥇 Misfit reports. A mode that's unusable for your kind of writing; advice that violates your discipline's conventions; an "AI cliché" that's actually legitimate usage in your field; a citation format incompatible with your target journal. More valuable than "I found it useful."
2. 🥈 Discipline-specialization proposals. A new critical dimension, a discipline-specific AI-trace pattern, a new working mode, or a script tool your field needs.
3. 🥉 Concrete success stories — but say what specific problem it solved, not just "it's great."

A worked example of a useful misfit report: "I'm in empirical sociology. Mode G (blind reading / promise-delivery) assumes the intro makes argumentative promises the body must deliver — but my methods section makes procedural commitments, not argumentative ones, so Mode G's checklist misfires. In my field the right check would be 'does the analysis follow the pre-registered plan?'" That tells the maintainers exactly which module, exactly why, and what the right behavior would be.

Why the maintainers want this even when it's negative: the skill's design is deliberately not neutral — it has positions (thought over format, engineering rigor, voice preservation, anti-sycophancy, reflexive self-awareness) that may simply not fit some traditions. Cross-domain testing isn't meant to sand it into neutrality. It's meant to (1) map the skill's real boundaries, (2) make it better inside them, and (3) help communities beyond those boundaries fork their own versions. "This discipline, please detour around" is itself a valuable, responsible finding.

What you can expect back: within ~48 hours a maintainer reads your report and responds — not necessarily "I'll change it," but always "I've seen it, here's what I think." Misfit reports get triaged into absorb-into-main vs. discipline-specialization extension, with reasoning. Genuine design-philosophy disagreements get an open "agree / disagree + why" rather than being ducked.

Where to go

• 📋 Full per-discipline test scenarios + the report template → docs/cross-domain-testing.md. It has suggested "entry points for testing" in History, Sociology/Anthropology, Literary studies, Philosophy, Law, Economics/Policy, and Communication/Media — plus known potential misfits for each, and adaptation notes for different kinds of writing (degree theses, journal submissions, conference talks, coursework, public writing).
• 🤝 How to contribute → the Contributing guide. Note the bilingual coupling rule (changes to SKILL.md must also land in SKILL.zh.md) and the bar for a new working mode.
• 💬 Post a report → open a topic under Discussions · Show and tell.

---

The roster of fields above is a starting point, not a fence. If your discipline isn't listed, run the fallback protocol, see how the best-fit approximation holds up, and send a misfit report — that's exactly how the next L1/L3 entry gets written.