# Literature Synthesis Doctrine > Governs how a body of sources becomes an argument: synthesis over summary, the source-by-theme matrix before any prose, claim-source binding, and a hard citation guard — an unverifiable citation is deleted, never repaired by plausibility. Trigger on any literature review, annotated bibliography, related-work section, or evidence synthesis; whenever citations are generated or checked; and whenever a research gap or question is derived from existing work. - Skill: `forexgod21/literature-synthesis-doctrine` (Agent Skill, multi-file: 2 files) - Install (CLI): `npx skillmds@latest add forexgod21/literature-synthesis-doctrine` - Raw SKILL.md: https://api.skillmd.com/api/skills/forexgod21/literature-synthesis-doctrine/raw - Safety review: pending - Works with: Claude Code, Claude.ai, OpenAI Codex - Category: Research & Search - License: CC BY 4.0 - Author: Forexgod21 (https://skillmd.com/u/forexgod21) - Updated: 2026-09-21 - Page: https://skillmd.com/skills/forexgod21/literature-synthesis-doctrine --- # Literature Synthesis Doctrine **Attribution:** YourVisionYourCreation LLC, yourvisionyourcreation.com **Doctrine class:** YVYC library, research category --- ## Universal So-What The defining failure of weak research writing is the book report: one paragraph per source, marching in the order the sources were found, ending with "in conclusion, many authors have studied this." That structure proves the author read; it proves nothing was understood. Reviewers at every level, from a course rubric to a dissertation committee to a journal referee, detect it in one page. Synthesis is a different act: the sources are organized by what they collectively say, agree on, and fight about, and the review becomes an argument about the state of knowledge, ending at the exact edge where knowledge stops. That edge is the gap, and the gap is the entire reason the review exists: it is where the author's own work earns its justification. This doctrine converts source piles into that argument, and it does so with a citation discipline built for an era in which plausible-sounding fake references are cheap to produce and fatal to careers. --- ## The Laws ### 1. Synthesis, Never Sequence The unit of organization is the theme, never the source. A section heading names an idea, a tension, or a variable; underneath it, multiple sources converse. Any paragraph that exists because a source exists, rather than because an idea requires it, is book-report tissue and gets dissolved into the themes it touches. ### 2. The Matrix Before the Prose Before any prose is written, the corpus goes into a matrix: sources down one axis, themes across the other, findings in the cells. The matrix is the diagnostic instrument: - A source filling no cell is dropped from the review, however interesting it was to read. - A theme with one source is an anecdote, not a theme; it needs corroboration or demotion. - A column of uniform agreement is a summary paragraph; a column of conflict is a section, because conflict is where the review earns its existence. The prose is then written from the matrix, never from the reading notes. ### 3. Claim-Source Binding Every claim in the review traces to a specific source, and the binding is tight: the source actually says what it is cited for, at the granularity cited. Findings are attributed to the study that produced them, not to a review that mentioned them. Where a claim aggregates ("studies consistently find..."), the specific studies are named. A claim no source carries is the author's own inference and is marked as exactly that. ### 4. The Citation Guard No citation ships unverified. Verified means the source has been confirmed to exist and to contain the cited content, not that the reference is well-formatted and plausible. The failure this guards against is current and career-ending: fabricated references that look perfect and cite nothing. The protocol is absolute: - A citation that cannot be verified is **deleted**, never patched into plausibility. - A remembered-but-unconfirmed source is named as unconfirmed or omitted; it is never dressed in a confident reference format. - Verification is per-citation, not per-bibliography. One fabricated reference discredits every real one around it. ### 5. Source Tiers Sources carry different evidentiary weight, and weight-bearing claims state their tier: peer-reviewed empirical work, then peer-reviewed theoretical and review work, then books and dissertations, then reputable gray literature, then everything else. Tiering is not snobbery; it is honesty about load-bearing capacity. A practice recommendation resting entirely on gray literature says so. ### 6. Disagreement Is the Payload Where sources converge, summarize and move. Where they conflict, slow down and analyze: different populations, different measures, different definitions, different eras, or a genuine unresolved question. The analysis of *why* findings diverge is the highest-value content a review produces, and it is precisely the content a book report cannot contain. ### 7. The Gap Statement The review ends by naming what the assembled literature cannot yet answer, stated precisely enough that a study could be designed against it. The gap must be real: checked against the corpus, not manufactured by ignoring the study that already filled it. Where the review justifies a research question, the question maps onto the gap one-to-one, and a question the gap does not support is a question the review has not yet earned. ### 8. Currency and Seminality The corpus balances two clocks: the recent literature that defines the current edge, and the seminal work that defines the field's spine. A review of only classics is a history; a review of only the last three years is unrooted. Both failures are corrected in the matrix before prose begins. --- ## Failure Modes | Failure | Correction | |---|---| | **The book report.** One source per paragraph, in reading order. | Law one. Reorganize by theme; dissolve source-paragraphs into the ideas they touch. | | **Citation laundering.** Citing a study known only through someone else's summary. | Cite what was actually read, as what it is: (as cited in), or go read the original. | | **Quote quilting.** Prose stitched from quotations doing the thinking. | Quotes for irreplaceable phrasing only. The synthesis voice is the author's. | | **Confirmation harvesting.** A corpus assembled to agree with the thesis. | The matrix must contain the disconfirming work; its absence is a search failure, not a clean field. | | **The manufactured gap.** A gap declared by ignoring the study that filled it. | Gap statements are checked against the corpus and stated narrowly enough to be honest. | | **The fabricated reference.** A plausible citation that cites nothing. | The citation guard. Unverifiable means deleted. No exceptions, no formatting rescue. | | **Tier blindness.** A blog post and a meta-analysis carrying equal weight. | Tier stated wherever the claim bears load. | --- ## Boundary (Read Before Applying) | Skill | Governs | |---|---| | `scholar-practitioner` | The full academic writing standard: voice, rigor, structure of the written product. This doctrine governs the synthesis machinery underneath the review specifically. | | `scholar-practitioner-tempo` | The program-calibrated edition, including APA 7 mechanics. Citation *format* lives there; citation *integrity* lives here. | | `adversarial-verification-doctrine` | General claim verification. The citation guard is its application to references, run without panels: verify or delete. | | `commit-or-concede` | The truth floor. An unconfirmed source is a concession, never a confident citation. | --- ## Compliance Check (Run Before Any Review Ships) 1. Is any section organized source-by-source instead of by theme? 2. Does the matrix exist, and was the prose written from it? 3. Does every claim trace to a source that actually says it, at that granularity? 4. Has every citation been verified to exist and to contain the cited content? 5. Where sources conflict, is the divergence analyzed or papered over? 6. Is the gap real against the corpus, and does the research question map onto it? 7. Does the corpus hold both the current edge and the seminal spine? --- ## Adversarial Evaluator Gate **What would a hostile evaluator attack first?** The citation guard's cost. Verifying every reference is slow, and the pressure to skip it grows with bibliography size and deadline proximity, which is exactly when fabricated references slip through. The defense is the asymmetry of the stakes: a thin, fully verified bibliography survives review; a rich one with a single fabricated entry destroys the author's credibility retroactively and entirely. Slow is the cost of the only acceptable failure rate, which is zero. The second attack: synthesis as license for overreach, the author's argument gliding past what the sources actually establish, with the matrix as decoration. Claim-source binding is the counter: the argument may go exactly as far as the cells support, the author's own inferences are labeled as inferences, and the gap statement marks where support ends. Synthesis organizes the evidence; it does not manufacture any. --- *YourVisionYourCreation LLC, yourvisionyourcreation.com* *Licensed under CC BY 4.0*