比较两个理论不能停在”都谈了 X”。有用的对比要揭示假设、适用范围、证据、预测的差别。下面 15 个 Prompt 把 AI 变成方法论老师，逼你穿过表面相似。
适合哪些场景
准备综合考的研究生、理论论文作者、tutorial / 研讨课负责人，以及写比较 essay 的哲学、社会科学、自然科学本科生。
什么时候不建议这样写 Prompt
两个理论的一手文献都没读时不用——只靠二手综述基础太薄。也不要在 AI 训练数据外的冷僻或非常新的理论上盲信。
Prompt 结构公式
理论对比 Prompt 一定要带这六个要素：
- 角色：AI 扮演谁——研究导师、同行评议人、考试教练、辩论对手、图书馆员。
- 上下文：水平、学科、deadline、论文数量、引用风格、课程或项目。
- 目标：一个具体交付物——12 道题、1 页文献矩阵、5 条反论、4 周复习计划。
- 限制：字数、深度、允许的来源类型、跳过什么、绝不主张什么。
- 输出格式：编号清单、表格、JSON 或分级块（E / M / H），能粘到 Notion / Anki / Word。
- 示例 / 信号：1-2 段参考或反例（“不要像维基那样讲”）。
这套 Prompt 适合用在哪
- 综合考准备
- 理论论文 / 章节起草
- Tutorial 讨论设计
- 文献综述理论段
- 本科比较 essay
15 个可直接复制的 Prompt 模板
1. 逐条假设矩阵
默认对比，揭示真正的分歧。
You are a theory professor. Compare {Theory A} and {Theory B} on their core assumptions. Output a 5-row table: assumption topic | Theory A position | Theory B position | where they overlap | where they diverge. Cite one foundational text per theory.可替换变量： Theory A, Theory B
优化建议： 重叠被夸大时追加：“Be precise: surface convergence often hides deep divergence. Note any cases where both use the same word for different concepts.”
2. 适用范围对比
For {Theory A} and {Theory B}, identify the scope conditions under which each is meant to apply. Where do they overlap, where does only one apply, where does neither claim authority? Cite the theorists or texts that explicitly bound the theory.3. 证据基础对照
What kinds of evidence does {Theory A} typically marshal? What about {Theory B}? Compare on: empirical type (experiment, fieldwork, archive, formal model), prototypical cases, and what counts as a refutation.4. 预测分歧
On the question "{question}", what does {Theory A} predict and what does {Theory B} predict? Where do the predictions diverge most sharply? What empirical observation could decide between them?5. 共同案例检验
Take 3 well-known cases in {field}: {case 1, case 2, case 3}. For each, how does {Theory A} explain it? {Theory B}? Where do their explanations differ, where do they converge?6. 谱系与影响
Sketch the intellectual lineage of {Theory A} and {Theory B}: predecessor thinkers, the problem each was reacting against, and the contemporary descendants. Note any place where Theory A descends from a critic of Theory B (or vice versa).7. A 立场批评 B
Write a 200-word critique of {Theory B} from the standpoint of {Theory A}: which assumptions of B does A reject, why, what implications follow. Use the vocabulary a serious proponent of A would use.8. 对称反向批评
Now reverse it: write a 200-word critique of {Theory A} from {Theory B}`s standpoint. Maintain the symmetry; do not stack the deck.9. 综合还是不可通约？
Can {Theory A} and {Theory B} be synthesized, or are they fundamentally incommensurable? Give the case for each answer in 100 words each. End with which case is stronger and why.10. 词表对齐
List 6 key terms used by both {Theory A} and {Theory B}. For each: definition in A, definition in B, whether the terms mean the same thing. Flag any words that look like cognates but refer to different concepts (treacherous terms).11. 实证测试设计
Design a study that would distinguish {Theory A} from {Theory B}: the question, the empirical setting, the measurable outcomes, and what result would favor each theory. Constraints: feasible for a {grad student / lab} to run.12. 可视化示意图大纲
Describe a Venn / 2x2 / cladogram diagram showing the relationship of {Theory A} and {Theory B}. List the regions / cells / branches and what should be placed in each. The reader should be able to draw it from your description.13. 常见误读
List 4 common misreadings of {Theory A} (especially as filtered through critics or textbooks) and 4 common misreadings of {Theory B}. For each: the misreading, why it persists, and the corrective text or passage to read.14. 比较 essay 段
Write a 200-word body paragraph for a comparative essay arguing "{thesis}" using {Theory A} and {Theory B}. Show me how to acknowledge the strengths of each while supporting my thesis. Cite at least 2 sources by author + year.15. 综合考备考摘要
Produce a 1-page exam-ready summary comparing {Theory A} and {Theory B}: 3 assumptions of A, 3 of B, 3 points of agreement, 3 points of conflict, 2 cases each best explains, and 1 sentence on why this comparison matters for {field}.容易踩的坑
- 只比表层主张，不比底层假设。
- 同一个词在两个理论里用，没核对含义是否一致（模板 10 帮你抓）。
- 一方写得公允、一方写得肤浅——这是切牌。
- 可以诚实地承认不可通约时硬要综合。
- 只引二手批评者，不读一手原典。
- 把”更新”等同于”更对”——老理论在重要领域常仍有效。
- 忽略 scope——很多”分歧”其实是关于不同领域。
优化技巧
- 模板 7 和模板 8 永远成对跑，保持对称。
- 尽早用模板 10——“伪共词”是大多数比较混乱的源头。
- 锚到一手文献——AI 默认综述二手。
- 综合考用模板 15 把书单里每对理论各跑一次。
- 答案就是”不可通约”时直说，比硬综合更诚实。
- 可能时把理论对比配以实证测试设计（模板 11）。
- 模板 12 的图表大纲存档当 tutorial 视觉教具。
FAQ
- AI 对理论对比有多准？: 经典对（功利主义 vs 康德伦理、凯恩斯 vs 货币主义）很准。冷僻或当代争论较弱，需对照一手文本。
- 一次对比该多长？: Essay 段 200 字；综合考答案 800-1200；论文章节 5000+ 分多节。
- 要不要选一方？: Essay 要；文献综述通常不要。综合考列各方优势后表态。
- AI 能判哪种理论对吗？: 不能。它描述立场和分歧；判断要看证据和价值观。
- 两个理论几乎不重叠怎么办？: 用模板 2（适用范围）——比较有时归结为”它们回答不同问题”。