Methodology
Why every entry has a confidence label, and what each tier really means.
Four confidence tiers
Attested
The etymology is supported by at least two peer-reviewed academic sources and a primary historical record (Meiji-era dictionaries, newspaper archives, government documents, etc.). The first attested year is known within a 5-year window.
Confident
Supported by at least two specialty dictionaries(e.g. Nihon Kokugo Daijiten, Kojien, OED). No direct primary source, but established consensus among lexicographers.
Plausible
Supported by at least one specialty dictionaryplus credible secondary sources (academic web resources, scholarly blogs by named etymologists). May have competing hypotheses we note.
Uncertain
Only secondary sources available, or sources disagree. We include these because marking them is more honest than omitting them. Reader discretion required.
Why mnemonics weren't enough
Memory aids are useful for recall, but they often invent plausible- sounding origins that have nothing to do with the actual history of a word. A learner who memorizes a mnemonic for arubaito rooted in “a robot working” may remember the meaning, but they won’t know that the word came into Japanese from German Arbeit in early-twentieth-century student slang.
Corrections
Every entry has a Report Correction link. We respond within 24 hours. Source-cited corrections are accepted; preference claims are recorded but not auto-applied.
Limitations
We are a small, source-driven publication, not the final word. For words with deep contested histories (e.g. early kanji forms), we cite competing schools rather than picking a winner. Where the truth is genuinely lost, we say so.