The Trace Paradigm runs one method across every investigation. This is the standing key to the notation, the source grades, and the reading conventions used in each piece. Every investigation links here and carries only its own case-specific terms and its evidence matrix, so the apparatus is written once and maintained in one place.
The method in one paragraph
The investigations reconstruct hidden structure from involuntary traces. This is the clue paradigm of Carlo Ginzburg and the abduction of Charles Peirce: reasoning from the marks a system leaves to the cause that produced them. The work runs under an apophatic limit that marks where reconstruction must stop, rather than overreaching into conspiracy. Claims are graded, the strongest competing explanation is stated, and the reasoning is visible in the body rather than hidden behind a verdict.
How the markers work
A marker leads the statement it governs. It applies to the text that follows it, up to the next marker. It is written as a label, sometimes with grades or sources, then an arrow, then the governed claim. For example, FACT, L2, Reuters → means the statement that follows is a fact at grade L2 sourced to Reuters.
FACT → verified by L1 to L2 sources.
ASSESSMENT → a reasoned conclusion drawn from the traces, weaker than a fact.
INFERENCE → in Module B, a logical deduction from position or context, where the claim follows by necessity although direct documentation is absent, stronger than an assessment and distinct from a hypothesis.
HYPOTHESIS, N/100 → consistent with the traces but not proven, carrying a confidence score out of 100.
ANOMALY → a finding not explained by current models.
STRUCTURALLY UNKNOWABLE → cannot be established, by the design of the object or by the available evidence.
APOPHATIC LIMIT → the boundary beyond which reconstruction yields to honest acknowledgment of its status.
BEHAVIORAL PATTERN → a recurring, observable behavior of the system.
DISCIPLINE → a methodological guard against over-reach.
SPECULATIVE → an unfalsifiable reading, capped at a confidence of 60.
REPORTED → claimed by sources but not established, status open.
Source hierarchy
L1 → court documents, financial records, treaties, declassified intelligence, official institutional records such as OFAC, the EU Council, the IMO, national courts, and the US Senate.
L2 → major reporting and research: Reuters, AP, the Guardian, the Wall Street Journal, Lloyd’s List, the KSE Institute, CREA, peer-reviewed work.
L3 → expert commentary and processed open-source intelligence, such as Starboard, Windward, and Navy Lookout, with the source’s positionality accounted for.
L4 → unverified claims, or single-source social-media open-source intelligence without independent corroboration.
Machine-mediation grade
M0 → primary trace, a direct artifact with verified custody: court record, financial filing, original document or photograph with provenance.
M1 → processed trace, human-mediated handling of primary traces: investigative journalism with named sources, peer-reviewed research, provenance-disclosed data.
M2 → synthetic narrative trace, AI-produced or heavily AI-assisted content presenting as fact, provenance often unclear, cannot carry a claim above SPECULATIVE without independent M0 or M1 corroboration.
M3 → weaponized synthetic trace, AI content built to deceive and attributed to false sources, excluded unless the deception is itself the object of study.
Confidence scores
Each hypothesis carries a score out of 100. 85 and above is well supported by converging traces. 65 to 84 is probable, with alternatives addressed. 50 to 64 is balanced between competing models. Below 50 is weak support, retained for tracking. Speculative claims are capped at 60.
The evidence matrix
Each investigation ends with a matrix of its key claims. The columns are: the claim, the source, the L-grade and M-grade, the confidence, and the strongest alternative. The matrix is where a claim stated plainly in the body is shown with its grade and its strongest competing explanation, so the reader can weigh it rather than take it on trust.
Case terms
Terms specific to a single investigation are defined inside that investigation, not here. In the Russian cases, for instance, these include the clan-rent state, the clan-rent state interface, and the sanitary regime. The standing apparatus on this page does not change from piece to piece; the case terms do.
The Trace Paradigm. AlexL Framework v2.5.1. thetraceparadigm.substack.com
Originally published:
https://thetraceparadigm.substack.com/p/how-to-read-the-trace-paradigm