51 terms in 8 families. Every definition is read from the code that assigns the word, so a page can never mean one thing and the glossary another; the note beside each is how to read it when you meet it.
The grade every claim on a research page carries: how it is known, not how confident it sounds. Assigned by the same rules on every scan; the rollup numbers are computed from the graded rows.
Corroborated against an outside source, and that source is linked.
Came from material the scan actually read, with a named source.
Some support, not enough to call it settled.
Sources disagree, or the record contradicts it.
Asserted in the material, with nothing behind it.
You supplied it. The scan took it as given.
From the model's prior knowledge, not from anything fetched.
Present in the material with no provenance attached.
Something the analysis could not establish, reported as missing.
What the record of everything you have run adds that no single scan can: whether two separately run scans arrived at the same claim, and whether they agree.
One scan established it. Most of the corpus looks like this, and that is fine — it just means nothing else you have run has spoken to it yet.
Two or more of your scans, run separately, arrived at the same claim with the same figures. Neither scan could know that; the corpus is the only place it shows up.
The same claim, a different number, and enough time between them that the figure moved rather than the record disagreeing with itself. This is a trend, not a problem — the newest reading is the one to use, and the older ones are the shape of the change.
Two or more of your scans state the same claim with different numbers over the same period, so one of them is wrong. This is the single most valuable thing this page produces — worth settling before you act on either figure.
How a cause-and-effect link is KNOWN, kept apart from how much it matters. A documented link can be weak; a strong link can be inferred.
A source in the record states this causal link — a finding, a filing, an official account.
A named participant, analyst or institution asserts the link on the record.
The scan's own causal reading from the sequence or structure, labelled as its own.
Credible sources genuinely disagree about whether this link holds.
The weight of the decision an event forces on a party: two axes, five classes, three rules. A classification is displayed only where the scan earns it.
Small stakes, easy to understand. The connection is straightforward with limited wider impact.
Meaningful stakes, but the situation is well understood. The implications are evident.
Meaningful stakes with real uncertainty. The situation could unfold in multiple directions and the implications are hard to pin down.
Very high stakes. The scale of what is involved makes this one of the most important connections in the analysis, regardless of how simple or complex it is.
Unprecedented. There is no clear historical comparison — this changes the rules, not just the situation.
Organizational weight: the dollars, the authority and the stakeholders a connection touches. Not headcount. One of the two FLOW axes.
The difficulty of the correct decision and action, across three co-equal dimensions: interpretive ambiguity, analytical depth and execution difficulty. The other FLOW axis.
FLOW C cannot exist at large scale: When organizational weight grows to large scale, FLOW C reclassifies to FLOW D immediately. The prescription to establish meaning before acting does not survive scale escalation. Scale alone drives FLOW D: Large organizational weight means FLOW D regardless of complexity. A simple, well-understood situation that requires enterprise-scale response is FLOW D, not FLOW A. When debating D vs S — use D: FLOW A and B differ only in scale (both low complexity). FLOW S requires extraordinary conditions that fall entirely outside normal organizational playbooks — max 1-2 per analysis.
How directly an event reaches a party, on the intersection cards an Event Scan or an Entity Scan renders.
The event acts on the party itself.
The event acts on something the party directly depends on or faces: a counterparty, a supplier, a regulator, a rival.
The party is materially exposed through a channel the record supports.
The event reaches the party's field, not its situation.
The party is in the same world and not exposed.
What a body of studies is allowed to be said to show, and whether the question was studied as asked.
At least two INDEPENDENT named studies — different teams or datasets — point the same direction. The only grade that earns "the literature shows".
One named study only. A real measurement, but one measurement — never "the literature".
A preprint, pilot, working paper or small-n result that has not been replicated or peer-validated. Marked as what it is.
Named studies genuinely disagree. The split is the finding; the sides are named, never averaged.
A published critique, failed replication, or retraction lead targets the supporting record. A finding, not noise.
The claim circulates — in coverage, marketing or folklore — without a study behind it that this review could find.
The literature studied this question as asked.
The literature studied related questions — proxies, narrower populations, different timescales. Every finding names what was actually studied.
No study addresses the question. Saying so IS the review.
What a named method has demonstrated in practice, with the promotional claim kept apart from the measured one.
A NAMED independent study, trial, evaluation or dataset measured the outcome. The strongest grade there is.
A NAMED record documents the result in a specific real adoption — not a controlled study, but a checkable record.
Named practitioners report it from their own use. Real testimony, uncontrolled — experience, not evidence.
The claim originates with the method's own originator, vendor, certifier or advocates, and the record does not independently support it.
The record actively disagrees — a failed replication, a contrary evaluation, or named critics with evidence. A finding, not noise.
No basis found either way. The claim circulates without a record behind it.
The handful of nouns the pages use for the things the product makes and keeps.
One run of one kind of research on one subject: the unit of work, priced in credits, that produces a research page.
The material a scan actually read during its run and the sources it cites.
The list of what a research page cites, rendered only from URLs the output genuinely carries.
The standing list of signals a scan named to watch, checked on a schedule the reader sets.
Thirty named frameworks run over a subject with the same grounding a scan gets, priced flat.
The unit every scan is priced in; a scan names its cost before it runs.
The whole rubric, as the product enforces it. How the grades work →
Or browse the research catalog →