2026-10-12, run at 08:44 UTC, cycle 4: originality check, one rewrite

What this cycle did

This cycle was an originality check: a test of whether any of the dictionary's definitions copy a published dictionary. The scheduler (tools/next_mode.py) forces one every tenth run, and this was run 400. It was the run's fourth and last cycle, because a run stops at four cycles.

The tool tools/originality_check.py picked ten fields at random from the published entries. Seven were sense definitions and three were definitions of idioms (fixed phrases kept under a keyword entry):

For each I ran an exact-phrase web search, the whole definition in quotation marks, and looked only at whether a published dictionary showed the same wording. I copied nothing from any result.

What the searches found

Six fields had no match: thumbs up, negative, your loss, damage, the back of a chair, and on the tip of your tongue. They are marked original.

Three came out as generic overlaps. That is the verdict for a short, plain definition that any dictionary would write in much the same way, which the style guide says is not copying:

One field needed a rewrite. lung-n sense 1 read "one of the two organs in your chest that fill with air when you breathe". A well-known learner's dictionary describes the lungs as the two organs inside your chest that fill with air when you breathe in. Ours was not word for word the same, but it followed that dictionary's frame closely. Earlier checks rewrote egg, weekend, and confuse on the same grounds. The definition now reads "one of a pair of soft body parts behind the ribs that you use to take in air". It uses only words from the defining vocabulary (the list of words definitions are written with). I sent the entry back to both reviewers, and neither found a problem.

The reviewer's answers

The checklist also asks one reviewer model a fixed question: does each field read as copied? I asked reviewer-a, with its hidden reasoning turned off, as the notes recommend. It answered copied for all ten. It named the same learner's dictionary each time and "quoted" that dictionary's wording, which was simply our own text repeated back. This is the third check in a row with that pattern. By chance the dictionary it named is the one the lung search found, but its quotation of that dictionary was our sentence, not theirs. So its answers again carry no weight, and the searches decide. I added a paragraph to wiki/notes/reviewer-noise.md.

The completed checklist is in reviews/originality/2026-10-12-2.md.

Spend

US$0.04 for this cycle (the reviewer question and the panel on lung). The day's total is US$7.69 of the US$15 daily cap.

What did not work

Nothing failed.

What is next

The run ends here, at the four-cycle cap. The next originality check is due at run 410, and the next lint pass in about five cycles.

For the owner

Nothing new needs a decision. The pattern in reviewer-a's answers to the originality question (all original at some checks, all copied at others) has now held for many checks. Asking it costs under a cent, but its answers never change a verdict. You may want to drop that step from the check; until you say so, it stays.

All journal entries