Fractal Semantic Graphs: Meaning Through Connectivity
The Universe. An atlas of the ideas this estate is built from, each one with its anchor. Not the argument: the universe the argument stands on.
What this is, in one paragraph. The estate proved a method on one document: pull out what it says, anchor every extracted claim to a verbatim quote at known bytes, connect the claims with named verbs, and the document becomes something you can ask questions of. This volume runs that method over the whole corpus, and prints the result as a book. Every entry carries its quote and its source. Every connection says whether the corpus stated it or this volume's author drew it. And the two chapters that matter most are the ones about where the corpus contradicts itself and where it says nothing at all.
The chapters
| Chapter | What it does | |
|---|---|---|
| 00 | Fractal Semantic Graphs: Meaning Through Connectivity | The locked title, what this volume is, what a reader will know at the end, the five editorial decisions recorded so they can be argued with, and the disclosure of how it was written. |
| 01 | How to read this atlas | After this chapter you will be able to open any entry in the book, read it without a key, and check it against the source in three hops. |
| 02 | The meaning core | After this chapter you will be able to state, in one sentence and without hedging, what this whole estate claims, and to say precisely what would have to be true for the claim to be wrong. |
| 03 | The grammar of edges | After this chapter you will be able to look at a graph somebody else drew and say, with reasons, which of its edges are wrong. You will also know why the estate treats a vague edge as a correctness bug rather t |
| 04 | Anchors, scopes and bridges | After this chapter you will be able to connect two systems that disagree about what their words mean, without asking either of them to change, and to say exactly how much of the connection actually holds. |
| 05 | Confidence, evidence and reality | After this chapter you will be able to say how much weight a claim will bear, and to show your working: not a score, but a path somebody else can walk and disagree with. |
| 06 | Provenance, correction and time | After this chapter you will be able to answer a question no document set can answer: how much of what I believe rests on claims that have since been corrected? |
| 07 | The fractal principle | After this chapter you will be able to test whether a system that calls itself fractal actually is, using a check that takes about a minute and does not require you to trust anybody. |
| 08 | The machinery | After this chapter you will know what it actually costs to hold a corpus this way: which parts are built, what each one buys, and which of them you would need first. |
| 09 | The estate, the vaults and the network | After this chapter you will be able to tell, for any claim in this book, whether it is running or only argued, and to check which without asking anyone. |
| 10 | What the atlas found | After this chapter you will know which of this estate's ideas are load-bearing and which are decoration, where it disagrees with itself, and what it does not say. None of it is asserted. The rankings are comput |
| 11 | Colophon | What this volume is made of, what was cut, what it got wrong, and what a reader who wants to argue with it should argue with first. |
The eight regions
Grouped by neighbourhood, not by the alphabet. A concept sits in one region and links freely into all the others.
| Region | Concepts | What lives there |
|---|---|---|
| The meaning core | 10 | Where the thesis lives: a node is an address, the edges are the meaning, and the same value differently connected is a different thing. |
| The grammar of edges | 12 | The rules that make a graph queryable: verbs with distinct inverses, no generic edge, paths that read as sentences, and the query rendered instead of the graph. |
| Anchors, scopes and bridges | 13 | How meaning crosses a boundary without anybody agreeing: reference points with no authority, scopes that override, and bridges instead of merges. |
| Confidence, evidence and reality | 13 | What a claim will bear, computed: the two ladders, the blast radius, the named absence, and the twin where the graph touches something real. |
| Provenance, correction and time | 8 | Who said it, what rested on it, and what happens when it turns out to be wrong. |
| The fractal principle | 14 | The title's first two words, as a testable claim: composition with local override, projections, altitude, and the map that a graph becomes when it gains position. |
| The machinery | 13 | What the estate actually built to hold all of this: the extraction, the core graph, the identity ledger, the instrument and the deterministic transformer. |
| The estate, the vaults and the network | 7 | Where the ideas are running, where they are only argued, and the honesty positions that separate the two. |
The peaks, computed
Not declared. This is how many connections reach or leave each concept in this atlas, and the caveat travels with it: degree measures this volume's reading of the corpus, and most of its edges are authored rather than anchored. The full ranking, the thin ones, and what both mean →
| Concept | Degree | What it says |
|---|---|---|
| meaning through connectivity | 9 | What a thing is emerges from the edges traceable from it, not from anything stored inside it. |
| a named absence beats a hidden one | 8 | An absence that is stated can be counted, queried, assigned and closed; an absence that is hidden silently supports whatever rests on it. The lack of |
| classification is a query, not a judgment | 7 | A node type is a required pattern of typed, directed paths that a node either matches or does not; the content does not decide the type, the paths do. |
| never render the whole graph | 7 | The default view is never the graph; it is the result of a query, because a diagram of everything is rarely useful. |
| the chain of custody | 7 | Claim, graph node, file, commit, official source with the hash of the retrieved bytes: every link walkable by a reader holding nothing privileged. |
| declared meaning | 6 | Meaning asserted in advance by a schema, which instances either conform to or fail. |
| documents are projections of graphs | 6 | The graph is the truth and the document is a view of it, rendered in the context of use. A skill file, a compliance standard and a consolidated legal |
| every edge is a verb | 6 | An edge is not a line but a claim, and a claim needs a verb, stated in both directions. |
How to check any of it
Everything is in this folder and nothing in it needs privileged access.
data/universe.json— every concept, edge, alias, distinction, contradiction and gap, with the source of each anchor and the SHA-256 of every carried source at the time this book was built.build/validate.py— the gate. It re-reads all 165 quotes out of the sources they name, checks both ends of every edge, refuses any edge stored under an inverse verb, and fails on an orphan.build/gen_chapters.py --check— fails if a chapter has drifted from the data.- the markdown — the source of truth for the text. These pages render it in your browser and the PDF is printed from it, so neither can disagree with it.
figures/figures.json— the 8 figures, each naming the page it photographs and the estate version it was taken at.
Written by an AI, and disclosed. This volume was written by Claude, an AI system made by Anthropic, as one session of a three-book commission. The session read the corpus, chose the concepts, wrote the prose, authored the graph and built the gate that checks it. The selection is a judgement and it is the author's; the anchors are not a judgement and are checked mechanically; the authored edges are the softest material here and are marked as such. The full disclosure and the editorial decisions →