The following message was sent by Tim Stewart to members of the Dictionary Society of North America:
I am pleased to announce a video interview with Igor Mel’čuk. The interview was done by Ian Mackenzie. Ian has published papers with Professor Mel’čuk, and he is also a professional film maker (see rimba.com). The film is copyrighted by Ian and can only be used for non-commercial purposes. It is a little over two hours long, and it is unedited. Ian and I both agree that the interview works well as is. The first hour is on lexicology, and the second is about Igor’s personal history. It describes his experiences growing up in Stalin’s Russia, and working as a translator years later for Andrei Sakharov.
I haven’t finished watching it, but the material on lexicography is so interesting I thought I’d go ahead and post it. Some random notes I took as I watched: Mel’čuk started by working on machine translation. He noticed the word “heavy” was one of the most difficult to translate into Russian — how should you distinguish between senses? He realized it depended on the modified noun — losses, rain, etc. Similarly, “take” could be rendered according to its object (walk, shower, etc.). He called these “lexical functions”; this got him started working on dictionaries. Is Russian коса ‘braid,’ ‘peninsula in river,’ or ‘scythe’? Look for the lexical context (hair, field, etc.). The key to a language is in its dictionary, not grammar; “being a linguist, you have to be a lexicographer.” He worked on an explanatory combinatory dictionary; primary rule: never use vicious circles (select = choose, choose = select), which can be OK for people, but not machines. He has never consulted a Russ.-Russ. dictionary, just (e.g.) French-Russ. and Russ-French. In an Eng.-Eng. dictionary, nobody looks up the hardest words to define, like “I.” (When he talks about the OED he says “Cambridge” when he means Oxford, but hey, the dude is 90.) When you define words, you need a predicate: authority of X among people Y in domain Z. Anna Wierzbicka invented semantic primes and insists definitions must be written using only them, but this is impossible. In English you say “Congratulations!” but in Russian you have to say “Поздравляю тебя!” (“you cannot say ‘Поздравления,’ it sounds stupid.”). There’s no word for ‘privacy’ in Russian — the concept doesn’t exist. He shows a definition (“lexical network”) of “kiss” that a computer can read easily but a human needs special training for. (At 45:24 you can see the screen showing the entry for авторитет ‘authority, prestige.’) He discusses problems associated with the Russian word рука ‘hand, arm’: you can’t define a single one, you have to define the pair (and then define the single object as one of the pair). The same goes for anything that occurs in pairs (shoes, skis, etc.). Idioms have separate entries in his dictionary; he says English lexicography is the best in the world in this regard. Thanks for sharing it with me, Parry!
Barbara Partee had a post at Language Log: No word for “privacy” in Russian? that agrees, citing her Russian husband.
The robo-transcript starts by “here is professor eager, igorithmetic…”
Anna Wierzbicka invented semantic primes and insists definitions must be written using only them, but this is impossible.
Not impossible, just very very difficult.
In English you say “Congratulations!” but in Russian you have to say “Поздравляю тебя!” (“you cannot say ‘Поздравления,’ it sounds stupid.”)
In any case, idiomatic syntax is outside her scope: semantic primes are an artificial language for expressing the meanings of words and fixed phrases and have nothing to do with (un)idiomatic syntax. (We say “Happy birthday!” but we sing “Happy birthday to you!”)
The Russian Wikipedia article has much more about his family and life, e.g., how the father avoided Stalinist purges (going to Tuva) and this:
От голодной смерти в Свердловске спасли родственники из Одессы, дважды в неделю сдававшие кровь за паёк
What is the difference between a semantic prime and a seme?
In any case, idiomatic syntax is outside her scope
The two sentences have nothing to do with each other; as I said, I was just jotting down things he said as I listened. The Wierzbicka material and the “Congratulations!” part may have been five minutes apart.
e.g., how the father avoided Stalinist purges (going to Tuva)
Tuva being nominally independent from the USSR at the time; it wasn’t annexed until 1944.