10. The human contribution

The project’s rules require describing the owner’s role objectively, and the knowledge base records it in enough detail that the facts are not in doubt.

Tom Gally — a lexicographer, translator, and professor based in Japan, as the companion report describes him — conceived the project and wrote its mission statement, which fixed the subject (the untranslatable-words phenomenon, popular and academic), the two phases (build a knowledge base, then examine individual claims against evidence), and the standards (accuracy above coverage; under-claim when in doubt; cite what you read; record disagreements; write up negatives). He supplied the framework document that describes how an unattended project keeps its memory, its honesty, and its owner informed — a persistent wiki, a baton file, an entry protocol, a daily report, an owner’s inbox, and an integrity check — together with a starter kit of templates and a link-checking tool (Appendix A). The repository’s rules of operation, in place before the first session, fixed the schedule on which sessions ran, the rule that every session must end with its work merged, the prohibition on reading the sibling project’s files, and the requirement that no session end having done nothing; the project’s own rules, which its files mark as owner-set, added the rest. On the project’s first day he authorized a spending ceiling of two and a half US dollars per day for paid services; the project never spent any of it.

During the run the project wrote a plain-language journal entry for him after every session; the record does not show when or whether he read them, only that he never intervened. The owner’s inbox, which any session would have treated as outranking its own queue, stayed empty from the first session to the last: he issued no mid-course instruction, corrected no finding, and chose no word. On 8 September 2026, after sixty-one sessions, he turned off the schedule and asked for three things: that any research still in progress be wound up, that this essay be written, and that the entire knowledge base be published beside it in readable form. His instructions, as the close-out session recorded them in its report to him, specified the essay’s length, tone, and readership, the earlier report it should resemble, an appendix on method, and an independent check of its claims and links before publication.

The record shows no case study designed, source read, page written, or result altered by him. His contribution was to decide that the project should exist, what it should value, where its lines lay — accuracy above coverage, cite what you read, under-claim when in doubt, record disagreements rather than resolve them — and when it should stop. Nothing in the record suggests that division of labor was the wrong one, and the record, in the form of the mission statement and the framework, is published here alongside everything it produced.