Agent Skills
› brycewang-stanford/Awesome-Journal-Skills
› soctheory-argument-development
soctheory-argument-development
GitHub用于压力测试社会学理论手稿的推理，将命题转化为经得起反驳和反例检验的有效论证。通过审查前提、推论、担保及竞争理论，确保逻辑严密性，而非定义概念或划定边界。
Trigger Scenarios
命题存在但连接推理薄弱
审稿人质疑前提无法推出结论
未处理明显的竞争性解释
论证未经过反例或困难案例测试
Install
npx skills add brycewang-stanford/Awesome-Journal-Skills --skill soctheory-argument-development -g -y
SKILL.md
Frontmatter
{
"name": "soctheory-argument-development",
"description": "Use when stress-testing the reasoning of a Sociological Theory (ST) manuscript — turning stated propositions into a valid, warranted argument that survives rival theories and counter-cases. Develops and audits the argument; it does NOT define the concepts (soctheory-theory-construction) or bound the theory's domain (soctheory-boundary-conditions)."
}
Argument Development: Warrants, Rivals, Validity (soctheory-argument-development)
When to trigger
- Concepts and propositions exist, but the reasoning that connects them is thin
- A reviewer would say "the propositions don't follow from the premises"
- You have not engaged the obvious rival explanation
- The argument has not been tested against a counter-case or a hard case
At ST, logical soundness plays the role statistics play at empirical journals. The argument is the evidence. This skill makes it valid before reviewers find the gaps.
The argument audit
Walk the manuscript's spine — premises → mechanism → propositions → conclusion — and check each link. Reed's "Justifying Sociological Knowledge" (ST 2008, 26(2):101–129) frames the central demand: every theoretical claim carries an implicit warrant, and you owe the reader the warrant, not just the claim.
- Premise check. Are the starting premises stated, defensible, and consistent with the
assumptions from
soctheory-theory-construction? An unstated premise is a hidden bet. - Inference check. Does each proposition actually follow from the premises + mechanism? Name the inference type — deductive derivation, abduction (the "best explanation" move central to ST theorizing; cf. Timmermans & Tavory, ST 2012, 30(3):167–186), or analogy — and check it is doing legitimate work.
- Warrant check. For each major claim, state the warrant: why is this a good reason to believe the claim? If the warrant is "it seems plausible," it is not yet an argument.
- Rival-theory check. Name the strongest rival explanation and show, on theoretical grounds, why yours is better (more parsimonious, more general, dissolves an anomaly the rival cannot). Defeating a strawman convinces no one.
- Counter-case / hard-case check. Construct the case where your theory would seem to fail.
Either the argument handles it, or you have found a boundary condition (hand to
soctheory-boundary-conditions).
Toulmin per proposition
A compact discipline for each load-bearing proposition:
- Claim — the proposition.
- Grounds — the concepts/premises it rests on.
- Warrant — why the grounds support the claim (the mechanism does most of this work).
- Backing — the tradition or prior reasoning that licenses the warrant.
- Rebuttal — the condition under which the claim would not hold (a candidate boundary condition).
A proposition that cannot fill warrant and rebuttal is an assertion, not a theoretical claim.
Handling rivals without data
Because ST does not test, you adjudicate rivals theoretically:
| Rival situation | Theoretical move |
|---|---|
| Rival explains the same cases | Show your account is more parsimonious or unifies more |
| Rival rests on a shakier premise | Expose and challenge that premise |
| Rival and yours seem observationally equivalent | Specify a conceptual difference that matters, even if untested |
| Rival is a special case of yours | Subsume it and show the added range |
Checklist
- The premises are stated and consistent with the theory's assumptions
- Each proposition demonstrably follows from premises + mechanism (inference type named)
- Each load-bearing claim states its warrant (not "it seems plausible")
- The strongest rival theory is named and answered on theoretical grounds
- At least one counter-case / hard case is constructed and handled (or yields a boundary condition)
- No empirical "test" stands in for an argument; data, if any, only illustrates
Anti-patterns
- A chain of plausible-sounding claims with no stated warrants
- Engaging a strawman rival instead of the strongest one
- "Future research could test this" used to dodge a current logical gap
- Abduction invoked as a license to assert rather than to reason toward the best explanation
- Inconsistency: a later proposition silently violating an earlier premise
- Smuggling an empirical result in as if it settled a theoretical dispute
Output format
【Spine】premises → mechanism → propositions → conclusion: valid? [Y/N + gaps]
【Warrants stated】per load-bearing claim: yes / missing [...]
【Strongest rival】[named] → why yours wins (theoretically)
【Counter-case】[case] → handled / becomes a boundary condition
【Inference types】deductive / abductive / analogical — each legitimate?
【Next step】soctheory-boundary-conditions
Version History
- 1839142 Current 2026-07-05 14:27
