influences.sgit.ai / register / Karl Popper & falsifiability

Karl Popper & falsifiability

traced person + idea full

An influence on Dinis Cruz — entry 10 of 25 in his register.

The shortest entry with the longest reach. Popper may be the influence that explains the shape of all the others' entries — including the trace table you are reading them in.

Block 1 The anchor

The Logic of Scientific Discovery — Karl Popper — 1959

No canonical URL. Not every anchor has one address — a book, a practice or a personal era does not — and this site links what is linkable rather than inventing a canonical home for something that has none. The wider library below carries what can be linked.

First published in German as Logik der Forschung (1934). The anchor is the demarcation criterion rather than the whole philosophy of science.

Block 2 In his own words

No first-person statement about this influence exists in the corpus yet. What follows is the site's reading of the evidence, not Dinis Cruz's words — and it stays labelled as such until his briefing document arrives.

The reason this entry is short and load-bearing at once: it is not a topic in the estate, it is the format of the estate's outputs. A Wardley map is a claim about where a component sits, refutable by pointing at it. A risk in the risks register is a claim with a named acceptor and a named mitigation, refutable by checking whether the mitigation exists. A trace row on this site is a claim that a pattern appears at a path at a version, refutable by opening the file.

Those three formats were not designed together and they have the same shape. That is either convergence or a single inherited habit, and Popper is the most economical explanation.

The uncomfortable corollary is the one this site has to live with: an influence claim is barely falsifiable. You cannot open a file and check whether a talk caused a design decision. The trace table is the best available substitute — it makes the consequence checkable even when the causation is not — and the disclosure page states the limit rather than letting the format imply more rigour than it has.

Block 3 The principle

A claim earns its status by being refutable. If nothing could show it to be wrong, it is not saying anything.

Block 4 The trace table

Pattern from the anchorWhere the estate implements itVersionStatus
A claim must be refutable to be worth makingWardley mapping as practised here — a map is a claim about position and movement, refutable by pointing at a misplaced component—implemented
State the conditions under which you would be wrongThe risks register's falsifiable-statement discipline — an accepted risk with a named acceptor and a named mitigation—implemented
The demarcation applied to inspiration itselfThis site's trace tables. An influence entry is written so that it can be checked against the repositories and found wrong—implemented
Publishing the refutations, not only the claimsPartial. The gaps are published as build specs and the absent rows are kept, but no claim on this network has yet been retracted in public—partial
A test that could falsify an influence claim itselfNowhere, and probably nowhere possible. See the resonance above—absent

5 of 5 rows carry no version. The format's own rule is that a trace table without versions is an opinion — these rows name a file the corpus scan reached but not the release it was read at. The generate-or-date rule.

This table on its own, with its sources → · as JSON · as markdown

Three of these rows point at sibling sites rather than at code, because that is honestly where this influence lands: it shaped the format of the artefacts, not a feature. The fourth row points at this page, which is either the strongest evidence in the table or the weakest, depending on how generously you read it.

Block 5 The gaps, as build specs

Patterns from the anchor the estate does not implement yet, written precisely enough that an agent could pick one up as a work item. This is what makes an influence page forward-looking: it is provenance and backlog in the same document.

G1 · build spec

Nothing on this network has ever been publicly retracted

Every site here publishes its gaps and its open questions, which is the cheap half of the discipline. The expensive half is withdrawing a claim that turned out to be wrong, in public, with the date. A build spec: a standing /shipped/ row type for this was wrong, here is what replaced it, used the first time something qualifies. A register that only ever adds is not practising the principle it cites.

G2 · build spec

The counter-influence entry does not exist

The most Popperian thing this site could publish is an influence that was rejected — something widely assumed to have shaped the work and did not, or was outgrown. It needs Dinis Cruz rather than a build, and it is Q4 in the briefing queue. Until one exists, the register can only confirm.

Block 6 The checklist

What to ask of new work in this influence's light. The checklist is the influence made operational — the part an agent can run without having consumed the anchor work.

Block 7 The wider library

The rest of the work, linked and never rehosted, each item with one line on what it adds. The full union of every entry's Block 7 is at /library/.

The corpus evidence

What the mining run found, by path. These are the files that put this entry in its tier — the claim on this page is checkable against them, which is the whole point of the format.

Path in the corpusWhat it carries
(corpus-wide, ~7 files)falsifiability as an explicit standard, most visibly in the maps-as-claims framing and the risks register's discipline

Paths are as recorded by the mining run behind the commissioning pack (v0.33.62, 25 August 2026). This repository holds the website, not the corpus, so they are cited rather than resolved — R2 in the comms queue.

Where this sits

No nesting and no sibling site: this entry stands on its own. The influence map draws every relation the register records.