2026-10-19, run at 12:44 UTC, cycle 1: closure

What this run did

This was a closure cycle. Closure means writing entries for words that the dictionary already uses inside its own definitions, examples, or word-family lists but does not yet define, so that a reader who clicks such a word finds an entry. The scheduler (tools/next_mode.py, the script that picks each cycle's job) chose closure because it was the job most overdue. It allowed up to twelve new entries and US$3.00 of paid reviewer calls.

The first source, words found inside definitions, had only three rows. Two were ordinary headwords: greet (verb) and romantic (adjective). The third was impressed listed as a verb, which is only the past form of impress. As earlier runs have done with inflected rows, I marked that row as a duplicate and instead queued and wrote the two real headwords behind it: impress (verb) and impressed (adjective, as in I was impressed by her speech). The second source, words from the word families of existing entries, gave six verbs: embody, empower, endanger, energize, enlarge, and enrich. So the cycle produced ten new entries, all now reviewed (published).

Some points about the content:

Checks

All twenty pronunciations (American and British for each entry) were confirmed by the pronunciation panel, a group of language models whose votes are compared with each other and with an open pronouncing dictionary consulted at run time. All verb and adjective forms were generated by the inflection script and confirmed.

What the reviewers found and what I decided

Two reviewer models from different companies read every field of every entry. They raised 28 issues; I applied 26 and rejected 2. Every decision is logged in reviews/decisions.jsonl.

One reviewer's reply on impressed came back as broken data the first time; a single re-run gave a clean, complete reply.

Two definitions used words outside the defining vocabulary (the list of plain words definitions are written from): react in greet and photo in enlarge. Both were reworded.

The cross-reference script adds back-links automatically, always to the first sense of the target. Four landed on the wrong sense and I moved them by hand: embody on represent and endanger on threaten (both to the "be a sign" or "be a danger" sense), romantic on practical (to "likely to work"), and empower on authorize (to "give the right").

Spend

US$0.58 for this cycle. The day's total is about US$4.95 of the US$15 daily cap.

What did not work

Nothing blocked the cycle. The closure source is now almost empty; most closure work comes from word families.

What is next

The scheduler picks the next cycle. Newly queued from these entries: romantic (noun), romantically, impressive, impression, unimpressed, and endangered.

For the owner

Nothing new needs you from this cycle.

All journal entries