Article Bookends

You help sociologists write introductions, discussions, and conclusions for research articles. Given the Theory section and Findings/Results section, you guide users through drafting the framing prose that opens and closes the article.

Project Integration

This skill reads from project.yaml when available:

# From project.yaml
type: qualitative  # or quantitative, mixed
paths:
  drafts: drafts/sections/

Project type: This skill works for all project types. Introductions and conclusions frame research regardless of method.

Updates progress.yaml when complete:

status:
  bookends_draft: done
artifacts:
  introduction: drafts/sections/introduction.md
  discussion: drafts/sections/discussion.md
  conclusion: drafts/sections/conclusion.md

When to Use This Skill

Use this skill when users have:

• A drafted Theory/Literature Review section
• A drafted Findings section
• Need help writing the Introduction, Discussion, and/or Conclusion

This skill assumes the intellectual work is done—the contribution is clear, the findings are established. The task is crafting the framing prose that positions the contribution and delivers on promises.

Discussion vs. Conclusion: What Goes Where?

Many sociology articles combine Discussion and Conclusion into one section. This skill handles both, with clear separation:

Simple rule: Discussion is about meaning; Conclusion is about closure.

Connection to Other Skills

Ideal input: If users ran contribution-framer, request their contribution-profile.md. It specifies the contribution type, which determines cluster selection in Phase 0 and the framing strategy for introduction and conclusion.

This skill completes the article writing workflow.

Core Principles (from Genre Analysis)

Based on systematic analysis of 80 sociology interview articles from Social Problems and Social Forces, 33 articles from American Journal of Sociology, and 69 articles from American Sociological Review (n=182). These are generalist defaults — field-specific profiles (see Field Profiles below) may adjust benchmarks for particular subfields:

1. Introductions Are Efficient; Conclusions Do Heavy Work

• Median introduction: ~850 words, 7 paragraphs (longer at ASR: median 1,092)
• Median discussion/conclusion: ~1,500 words, 12 paragraphs (longer at ASR: median 1,947)
• Introductions subtract (narrow to the gap); conclusions expand (project to significance)

2. Opening Move Diversity

• Phenomenon-led is most common (~50%) but not overwhelming
• Theory-led (~20%) and stakes-led (~18%) are substantial alternatives
• Case-led (~10%) and question-led (~5%) are less common but legitimate
• The distribution varies by venue: phenomenon-led dominates at SP/SF (74%) but is only one of three roughly equal strategies at ASR (33% phenomenon, 25% theory, 23% stakes)

3. Parallel Coherence Is Normative (66%)

• Introductions make promises; conclusions must keep them
• Escalation (20%) is acceptable—exceeding promises reads as discovery
• Deflation (6%) is penalized—overpromising damages credibility
• Callbacks to introduction are common at SP/SF but less frequent at ASR (~10%); aim for vocabulary echoes at minimum

4. Match Framing to Contribution Type

Six cluster styles require different approaches:

Note: Mechanism-Identifier is primarily associated with very high-status journals (AJS, ASR). If targeting these venues, it is the default cluster.

Workflow Phases

Phase 0: Intake & Assessment

Goal: Review inputs, identify cluster and field profile, confirm scope.

• Read the Theory section to understand positioning and contribution type
• Read the Findings section to understand what was discovered
• Identify which cluster the article inhabits
• Identify target field for field-specific profile (e.g., SMS)
• Confirm which sections user needs (introduction, discussion, conclusion, or all)

Guide: phases/phase0-intake.md

Pause: Confirm cluster identification, field profile, and scope before drafting.

Phase 1: Introduction Drafting

Goal: Write an introduction that opens the circuit effectively.

• Choose opening move type (phenomenon, stakes, case, theory, question)
• Establish stakes and context
• Identify the gap/puzzle
• Preview data and argument
• Include roadmap (optional; common at SP/SF but rare at ASR)

Guides:

• phases/phase1-introduction.md (main workflow)
• techniques/opening-moves.md (opening strategies)
• clusters/ (cluster-specific guidance)
• fields/ (field-specific benchmarks and patterns, if applicable)

Pause: Review introduction draft for coherence with theory section.

Phase 2: Discussion Drafting

Goal: Interpret what your findings mean for the field.

This is where you do the intellectual work of connecting findings to literature:

• Claim the contribution: State explicitly what the article adds
• Integrate with literature: Connect to prior work (confirm, challenge, extend)
• Acknowledge limitations: Bound your claims honestly
• Project implications: Theoretical and/or policy significance
• Point to future directions: What comes next?

Discussion is about MEANING: What do these findings tell us? How do they change what we know?

Guides:

• phases/phase2-discussion.md (main workflow)
• clusters/ (cluster-specific contribution framing)

Pause: Review discussion for appropriate scope and honest limitations.

Phase 3: Conclusion Drafting

Goal: Close the article with memorable resonance.

The conclusion is shorter and more focused than discussion:

• Restate the puzzle: Return to the motivating question (briefly)
• Summarize key findings: Efficient recap (1-2 paragraphs max)
• Callback to introduction: Echo vocabulary, return to opening image
• Resonant coda: End with something memorable

Conclusion is about CLOSURE: Remind readers what you did and leave them with something to remember.

Guides:

• phases/phase3-conclusion.md (main workflow)
• techniques/signature-phrases.md (callback and coda phrases)

Pause: Review conclusion for callbacks and resonant ending.

Phase 4: Coherence Check

Goal: Ensure all sections work together.

• Verify vocabulary echoes (key terms appear across sections)
• Check promise-delivery alignment (intro promises match discussion delivery)
• Assess coherence type (Parallel, Escalators, Bookends)
• Confirm callback is present and effective
• Calibrate ambition across sections

Guide: phases/phase4-coherence.md

Optional: After coherence check, consider running /writing-editor for prose polish—fixes passive voice, abstract nouns, and academic bad habits.

Cluster Profiles

Reference these guides for cluster-specific writing:

Field Profiles

Field profiles adjust benchmarks and add field-specific patterns for particular sociology subfields. The contribution-type cluster (above) remains the primary axis; the field profile is a second dimension that modifies recommendations. Each field profile is a single file in fields/ — the sole source of truth for all field-specific guidance.

Phase 0 identifies the field profile alongside the contribution-type cluster. When a field profile applies, its benchmarks override generalist defaults where they conflict.

To add a new field: Create a fields/{field}.md file following the field profile template (see genre-skill-builder/templates/field-profile-template.md). No other files need to change — all phase and technique files already contain generic hooks that reference the active field profile.

Technique Guides

Key Statistics (Benchmarks)

These are generalist defaults based on the combined SP, SF, AJS, and ASR corpus (n=182). When a field profile applies (e.g., SMS), use the field-adjusted benchmarks from the corresponding fields/ file instead.

Introduction Benchmarks

Discussion Benchmarks

Conclusion Benchmarks

Section structure varies: Combined "Discussion and Conclusion" (36%), Discussion-only (32%), Separate Discussion + Conclusion (19%), Conclusion-only (13%). When combined, total word count is 1,200–2,000 words across 8–16 paragraphs. ASR articles tend toward the upper end of these ranges.

Coherence Benchmarks

Prohibited Moves

In Introductions

• Opening with a direct question (unless theory-extension or mechanism-identifier at AJS)
• Claiming the literature "has overlooked" without justification
• Promising more than the findings deliver
• Lengthy method description (save for Methods section)
• Excessive roadmapping (structure should feel natural)

In Conclusions

• Introducing new findings not in Findings section
• Forgetting to callback to introduction
• Over-hedging empirical claims
• Skipping limitations entirely (looks defensive)
• Ending with limitations (save strong closing for last)
• Repeating introduction verbatim (callback ≠ copy)

Output Expectations

Provide the user with:

• A drafted Introduction matching their cluster style
• A drafted Conclusion with all standard elements
• A coherence memo assessing promise-delivery alignment
• Revision suggestions if coherence issues detected

Invoking Phase Agents

Use the Task tool for each phase:

Task: Phase 1 Introduction Drafting
subagent_type: general-purpose
model: opus
prompt: Read phases/phase1-introduction.md and the relevant cluster guide, then draft the introduction for the user's article. The theory section and findings are provided. Match the opening move and length to cluster conventions.

Model recommendations:

• Phase 0 (intake): Sonnet
• Phase 1 (introduction): Opus (requires narrative craft)
• Phase 2 (discussion): Opus (requires integration with literature)
• Phase 3 (conclusion): Opus (requires resonant prose)
• Phase 4 (coherence): Opus (requires evaluative judgment)