# Brief — Part I (Introduction), Part X (Frontiers), Part XI (Discussion & Conclusion) + abstract
Files: manuscript/sections/part01_introduction.tex (12–18 pp), part10_frontiers.tex (15–25 pp), part11_discussion.tex (10–15 pp),
sections/abstract.tex (≤ 300 words, no \cite). Follow WRITER_COMMON.md. Read first: CLAUDE.md, RESEARCH_PLAN.md, protocol/PROTOCOL.md,
reports/PHASE1_CORPUS.md, reports/PHASE4_STRUCTURAL_CHANGE_TOTAL.md, reports/*.md, and the written parts (II–IX, specialist
syntheses V–VII if present) — the introduction states the research problem, the contributions (corpus, ontology, knowledge graph,
structural-change evidence, emerging-topic framework, release), the scope and the structure of the paper; the frontiers part
(X) consolidates the open problems collected in literature_synthesis/*/open_problems.csv and Part IX's unresolved questions into
testable research directions and competing hypotheses, with explicit statements of what the corpus cannot decide; the discussion
(XI) states contributions, limitations (coverage, classifier noise, right truncation, LLM extraction, no full text, no Semantic
Scholar), generalisability, implications, and the human-review status (which checkpoints H-0…H-5 remain unsigned — read
reports/HUMAN_REVIEW_CHECKPOINTS.md). Numbers only via macros. No causal language about ChatGPT. The abstract summarises corpus
size, methods and the main verified findings with macros (the title page uses it).