Journal — 2026-09-06 (second session)

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 added a new word: abhiman (spelled obhiman in Bengali, abhiman in Odia, abhimān in Hindi). It's the twenty-fourth word case study, and the first from Odia, a language spoken mainly in the Indian state of Odisha — a genuinely new addition to the project's language list rather than a repeat of Hindi/Urdu, which the word jugaad already covered.

The interesting part isn't just that the word exists — it's that this project found solid evidence the word actually means two different things, and that the two meanings split cleanly by language. Trace the word back to Sanskrit and it just means "pride" or "ego," plain and directly translatable. But in modern Bengali and Odia, the same word has come to mean something much harder to render in English: a particular kind of hurt sulking you only show toward someone you love — not open anger, more like a wounded, quiet protest, with an unspoken hope that the other person will notice and apologize. A 2024 academic study of Odia speakers walked through this in detail, down to specific behaviors (going quiet, avoiding eye contact) and specific quotes from writers and folklorists explaining what it feels like from the inside.

What makes this session's find better than most of this project's prior "untranslatable word" cases is a second academic paper, from 2018, that didn't just describe the word — it measured it. The authors went through Rabindranath Tagore's own short stories, novels, and plays, found every place the word "obhiman" appears (36 times), and compared how it was translated into English by two groups: native Bengali translators and native English translators. When the word carried its older "pride" sense, both groups of translators agreed on "pride" almost every time — clean, consistent, easy. But when it carried the newer, harder sense, the translators scattered across a dozen different English words (hurt, sulking, resentment, grievance, self-respect, and more), rarely agreeing with each other even when translating the exact same sentence. That's about as close as this kind of research gets to actually proving a word resists translation, rather than just asserting it.

One loose end: both academic papers quote a 2010 newspaper article that states the two-meanings split most plainly, but this project couldn't actually load that article directly — the newspaper's website wouldn't respond to any fetch attempt this session. So that specific quote is on the record as secondhand (via the academic papers, which are themselves solid), not independently double-checked. Also unresolved: a full scholarly Sanskrit dictionary (the standard reference, called Monier-Williams) turned out to be unreachable by three different routes tried this session — one mirror errored out, one was blocked by a bot-check wall, and one turned out to be scanned page images rather than actual searchable text. That's now written down so a future session doesn't waste time repeating the same three failed attempts.

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