How the Adler Four-Step Analysis Method Powers Stage 0 Book Comprehension in Cangjie-Skill

The Adler four-step analysis method in Stage 0 provides a structured framework for whole-book comprehension that deconstructs source material through Structural, Interpretive, Critical, and Applicability phases before any AI skill extraction begins.

Stage 0 of the Cangjie-Skill pipeline implements Mortimer Adler's classic "analytical reading" methodology with a critical enhancement: a fourth Applicability step added specifically for skill-generation workflows. This approach ensures that downstream skills emerge from rigorous, critical understanding rather than superficial summarization. According to the repository's methodology documentation, the process creates BOOK_OVERVIEW.md as global context for all subsequent extraction stages.

---

The Four Steps of Adler Analysis in Cangjie-Skill

The methodology is fully documented in methodology/01-stage0-adler.md. Each step produces specific deliverables that feed directly into the skill-creation pipeline.

Step 1 — Structural Analysis

This phase identifies the book's skeleton: its architecture and central argument.

You must produce:

• Type of work (methodology, biography, field guide, etc.)
• One-sentence core thesis that captures the author's main claim
• 3-7 primary arguments with their relationships: parallel, progressive, contrasting, or refuting
• Central problem the author aims to solve

This structural map prevents contributors from extracting isolated tactics without understanding the author's overall framework.

Step 2 — Interpretive Analysis

Interpretive work requires re-expressing the author's concepts in your own words, not quoting or paraphrasing mechanically.

Required outputs include:

• Key terms with author-provided definitions — not dictionary meanings, but the specialized usage within the text
• 5-15 core propositions restated in your voice
• Evidence chain — the reasoning or data that links propositions together

This step surfaces whether you truly comprehend the argument or merely recognize its surface features.

Step 3 — Critical Analysis

Adler's original critical reading exposes why you might be wrong to agree with the author.

Document:

• Historical or contextual limitations — what the author could not know
• Authorial biases stemming from background, industry, or culture
• Unsubstantiated assumptions — claims presented without adequate support
• Strongest counter-argument a well-informed skeptic could raise

The Critical section directly seeds the Boundary (B) field of each downstream skill, as implemented in SKILL.md. This prevents skills from being applied blindly in contexts where the original author's assumptions fail.

Step 4 — Applicability Analysis

This skill-specific addition bridges scholarly critique and practical execution.

Determine:

• Skill-suitable items: frameworks, checklists, decision-procedures, operational principles
• Non-suitable items: pure anecdotes, raw stories, emotional content without procedural structure
• Skill count estimate and priority ranking for what to extract first

The original Adler method concludes with Step 3. Cangjie-Skill adds Applicability to ensure only material with clear execution potential proceeds to Stage 1 extraction.

---

Generating BOOK_OVERVIEW.md from Template Analysis

The repository provides templates/BOOK_OVERVIEW.md.template to standardize Stage 0 output. This document becomes the authoritative context referenced by all subsequent extractors.

Practical Template Population

# Example: Using Jinja-like substitution to create BOOK_OVERVIEW.md

from pathlib import Path

template_path = Path("templates/BOOK_OVERVIEW.md.template")
output_path   = Path("books/example-book/BOOK_OVERVIEW.md")

data = {
    "BOOK_TITLE": "Deep Work",
    "AUTHOR": "Cal Newport",
    "YEAR": "2016",
    "METHOD": "实操手册",
    "SOURCE_FILE": "DeepWork.pdf",
    "DATE": "2026-08-14",
    # ---- Structural ----

    "TYPE": "实操手册",
    "ONE_SENTENCE_THESIS": "深度工作是提升知识工作者价值的关键。",
    "ARGUMENT_1": "深度工作提升产出质量",
    "ARGUMENT_2": "深度工作需要排除干扰",
    "ARGUMENT_3": "深度工作必须被刻意练习",
    "RELATION": "递进",
    "CORE_ISSUE": "如何在信息碎片化的时代实现高质量输出",
    # ---- Interpretive ----

    "TERM_1": "深度工作",
    "DEF_1": "在无干扰的状态下专注完成 cognitively demanding tasks",
    "PROPOSITION_1": "专注的时间越长，产出质量越高",
    "ARG_CHAIN": "作者引用实验室研究以及职业案例证明专注的收益",
    # ---- Critical ----

    "LIMITATION_1": "忽视了创意工作对分散注意的需求",
    "BLINDSPOT_1": "作者是学术背景，可能低估了灵感的偶发性",
    "UNPROVEN_ASSUMPTION_1": "所有人都能通过训练达到深度状态",
    "STRONG_OPPOSITION": "在高度协作的团队中，深度工作可能降低整体效率",
    # ---- Applicability ----

    "SKILL_CANDIDATE_1": "时间块规划框架",
    "SKILL_CANDIDATE_2": "噪声过滤清单",
    "ESTIMATED_SKILL_COUNT": "≈5",
    "PRIORITY_1": "时间块规划框架",
}

# Simple substitution (real implementation may use a templating engine)

content = template_path.read_text()
for key, value in data.items():
    content = content.replace(f"{{{{{key}}}}}", str(value))

output_path.parent.mkdir(parents=True, exist_ok_ok=True)
output_path.write_text(content)
print(f"Generated {output_path}")

The template enforces complete analysis before extraction begins. Missing fields indicate incomplete comprehension.

---

Feeding Critical Analysis into Skill Boundaries

The Critical section's output directly populates skill metadata. This creates guardrails for AI system behavior.

# Example snippet from a generated skill file (SKILL.md)

boundary: |
  - 作者的时代局限：本书写于信息技术高速发展阶段，某些工具已过时
  - 作者的立场盲点：专注于个人生产力，未充分考虑团队协作的复杂性
  - 未被证明的假设：所有知识工作者都能通过时间块实现深度工作
  - 反对意见：在创意密集型任务中，分散注意可能更有利

These boundaries prevent skill misapplication. A productivity skill derived from Deep Work, for instance, carries explicit warnings about collaborative contexts and creative work — limitations discovered during Stage 0 critical analysis.

---

Key Files and Reference Implementation

File	Purpose	Location
methodology/01-stage0-adler.md	Complete Adler four-step methodology documentation	methodology/01-stage0-adler.md
templates/BOOK_OVERVIEW.md.template	Standardized template for Stage 0 deliverables	templates/BOOK_OVERVIEW.md.template
SKILL.md	Meta-skill showing how Critical analysis feeds Boundary fields	SKILL.md
README.en.md	Pipeline overview with Stage 0 positioning	README.en.md

The pipeline reference in README.en.md lists "Whole-Book Comprehension (Adler Analysis)" as the foundational first step, establishing dependency order for all downstream processing.

---

Summary

• Adler four-step analysis in Cangjie-Skill Stage 0 consists of Structural, Interpretive, Critical, and Applicability phases — the fourth being a custom extension for skill workflows.
• Each step produces mandatory deliverables captured in BOOK_OVERVIEW.md, which serves as global context for extraction stages.
• The Critical section directly feeds Boundary (B) fields in skill definitions, preventing uncritical application of author methods.
• Applicability analysis filters source material for actionable, procedural content suitable for AI skill generation.

---

Frequently Asked Questions

What makes Cangjie-Skill's Adler analysis different from standard analytical reading?

Standard Adler analysis stops after three steps: Structural, Interpretive, and Critical. Cangjie-Skill adds a fourth Applicability step specifically to bridge the gap between scholarly critique and practical skill generation. This ensures that only material with clear execution potential — frameworks, checklists, and decision procedures — proceeds to Stage 1 extraction, while anecdotes and purely emotional content are filtered out at the comprehension phase.

Why does the Critical step feed directly into skill Boundaries?

The Critical step exposes contextual limitations, authorial biases, and strongest counter-arguments that would otherwise remain invisible in extracted skills. By piping these insights into each skill's Boundary (B) field, the system creates automatic guardrails against misapplication. A user deploying a skill receives explicit warnings about when the original author's assumptions may fail, preventing blind application in unsuitable contexts.

How complete must BOOK_OVERVIEW.md be before proceeding to Stage 1?

The template in templates/BOOK_OVERVIEW.md.template enforces completeness through required fields for all four steps. Empty or placeholder values indicate inadequate source comprehension and block progression. The document functions as a contract: contributors must demonstrably understand structure, accurately re-express arguments, identify genuine weaknesses, and map actionable content before any skill extraction begins.

{
  "mcpServers": {
    "instagit": {
      "command": "npx",
      "args": ["-y", "instagit@latest"]
    }
  }
}