Journal — 2026-09-06

This project is building a rigorously sourced knowledge base on "untranslatable words" — the celebrated cases like hygge and saudade that get marketed as having no English equivalent — checking each one's actual dictionary treatment, real-world usage, and translation history against the popular claim.

This session didn't add a new word. Instead it went back to the single most valuable loose end left over from last session: iktsuarpok, the Inuktitut word for restlessly waiting for someone to arrive. Last session's page reported one specific problem: the only real dictionary anyone cites for this word — a 1954 Eskimo-English dictionary by a missionary-linguist named Arthur Thibert — had never actually been read by anyone in this project's traceable chain of sources. It's out of print, and only shows up in library catalogs (HathiTrust, WorldCat).

This session tracked that dictionary down as far as it currently can be tracked. A library search tool (HathiTrust) lets you search the actual scanned text of books it holds, even when it won't show you the book itself — and a search for the word "iktsuarpok" across its whole catalog turned up exactly the right book: Thibert's dictionary, including the original 1954 edition, confirmed by name and year. So the word really is in that dictionary, and Wiktionary's citation isn't a case of someone getting the source wrong. That closes the narrower half of last session's question.

It doesn't close the more useful half. The library tool won't show the actual dictionary entry — only that the word appears somewhere in the book — because the book is still under copyright. So what Thibert's dictionary actually says about the word (does it match the popular "restless anticipatory waiting" gloss precisely, or is there a twist, the way there was for last-but-one session's word) is still unread.

There's a second thing worth flagging, because it's a caution for how this project does its work generally, not just about this one word. While chasing the dictionary text, a search via Google Books returned what looked like a real quoted definition, on a specific page, with two plausible-looking neighboring dictionary entries alongside it. It read completely convincingly. But checking the same Google Books page a different way — a plain, direct fetch rather than going through the usual web-reading tool — showed that Google's own page data said flatly that no readable text exists for that book through that method at all. In other words: the tool that's supposed to fetch and summarize a web page produced a specific, confident, entirely invented answer instead of admitting it couldn't get the real content. This project has caught this kind of thing before in less dramatic form; this is the clearest single instance yet, and it's now written down as a standing rule for future sessions: check a book's own "is there readable text here" flag before trusting anything that looks like a quote from it.

Nothing needs your input this session. Everything is written up in the project's own files as usual.