Too new to translate

« previous post | next post »

Thanks to reader JR, we are able to bring you the harmonic convergence of two recent LL memes, namely singing-dancing figurines and the vagaries of machine translation. The English advertising copy for the Ozaki iMini Pet ("Dock + Radio + Alarm + Speakers + Dancing Pet") is so badly written that it was probably created by a human pretending to be a translator, though it's possible that the human pretended to write a machine-translation program instead:

iMini is built in the rhythm decoding chip MJ1191 of the programming embedded system, and to integrate the HIPS skeleton; No matter you play any kind of music, MJ1191 always make your pet in dancing for you at once.


The folks who marketed this product line in Catalonia decided to go with a different strategy — they reproduced the English-language copy, and boasted that the iMini is

tan nou!! que no ens ha donat temps a traduir la pàgina, vine a veure'l!!!

Anybody who forks over $99 for one of these is probably not convinced by the ad copy, anyhow…



8 Comments

  1. Victor Mair said,

    January 8, 2010 @ 6:27 pm

    According to Joaquim Roé, the Catalan means "So new!! We didn't even have time to translate the page, come and see it!!!"

    [(myl) Sorry — I took it for granted that LL readers were all able to read Catalan!]

  2. Richard said,

    January 8, 2010 @ 7:07 pm

    [(myl) Sorry — I took it for granted that LL readers were all about to read Catalan!]

    I hate to go off-topic, but … I understand the highlighted text to mean 'all going to read', but I don't have this idiom meaning 'all able to read', which I assume is what you mean; is it a simple typo or really a construction that can have this meaning? I can kind of see how it might be the latter, and if so, it looks as though it could pose some interesting questions for students of grammaticalisation.

  3. Gary said,

    January 8, 2010 @ 7:43 pm

    I think Mark probably was speaking Catalan into the high sensitivity mis of electrets condenser microphone, and to integrate the voice activation from decoding chip MJ1191 to translate into English.

    [(myl) Actually, Mark was experimenting with writing a comment in a cell phone's browser, under time pressure, with the usual overactive predictive spelling correction in force… What came out as 'about' was supposed to be 'able'…]

  4. John said,

    January 8, 2010 @ 8:04 pm

    Well there is "I'm not about to read Catalan" meaning something like "I'm not willing to put the necessary effort into it", so it could be related, but that's the only construction I'm aware of where 'about' can take that meaning. And on reflection even that's not very well phrased, "I'm not about to learn Catalan" would make more sense.

  5. David Eddyshaw said,

    January 9, 2010 @ 8:04 am

    *Of course* all LL can read Catalan!
    Who can't?

  6. John Cowan said,

    January 10, 2010 @ 5:16 pm

    "You are linguist, no? Listen, and try to understand." –Roman Jakobson, explaining why he's going to give a lecture on Bulgarian folk poetry in Bulgarian to an American audience who mostly speak only English

  7. Nanani said,

    January 11, 2010 @ 10:15 pm

    >so badly written that it was probably created by a human pretending to be a translator

    I'm saving that burn for future use.

    I'm going to guess it was actually MT, and presume that no human ever checked the result before it left Japan.

    [(myl) If it was MT, I don't think it could have been any vaguely modern statistical MT program, since they make heavy use of a model of probable word sequences in the target language, and things "always make your pet in dancing for you at once" are not likely to emerge from such a process. But that sort of English is exactly what you get from a speaker of another language, with almost no knowledge of English, fitting together fragments from a bilingual dictionary. Or maybe from a poor-quality old-fashioned rule-based transfer system.]

  8. Nanani said,

    January 14, 2010 @ 12:17 am

    Clearly not a modern program, agreed. It "smells" like MTto me mostly because the passage doesn't fit Japanese grammar, ruling out direct-substitution-with-a-dictionary. The kinds of errors on display here aren't the same ones native Japanese make.

    New hypothesis: They had a human translator write it, but the translator doesn't usually work in English and is a native speaker of some other language.

RSS feed for comments on this post