Loading ...
Sorry, an error occurred while loading the content.

Re: [LAMI.US] Post on Human Translation -- Human still is key

Expand Messages
  • Caesariensis
    I think that is a direct result of their being programmed by people with a thing-orientation rather than a doing-orientation and ultimately going back to
    Message 1 of 4 , Apr 1, 2010
    • 0 Attachment
      I think that is a direct result of their being programmed by people with a
      'thing-orientation' rather than a 'doing-orientation' and ultimately going
      back to origins in a form of mathematical logic that did not emphasize
      relationships as much as traditional calculus because it wasn't really
      feasible. Analogue computers worked more on relationships but they were no
      good at moving information around. They were resonating circuits that had
      to be linked by plugs or digital logic and the voltage output of any basic
      component was an integration function of its inputs over time (at least an
      approximation that eventually went wild). Digital computers became faster
      and more able to copy them to any required degree of accuracy, so they
      died out.

      A mechanical version of an analogue computer working in reverse is the
      'differential' in car gearboxes (why it is called that - differentiation
      is the process opposite to integration) or the stylus on a vinyl replaced
      by a wheel that would rotate as the disc passes underneath, compared to
      its distance from the centre. Early mechanical analogue calculators worked
      exactly like that, a 'speedometer' running off the pick-up wheel giving
      the differentiation of a speed and distance from the centre that it was
      set at. One could probably do it now with a laser reading.

      I don't see how that could translate to any kind of data holding. However,
      supposing the designers of computers had used a language with extensive
      case-endings like Finnish or Russian or Japanese or even Latin or Sanskrit
      and been more concerned with language than mathematics? Might they have
      taken a very different approach that would specify relationship between
      items instead of relying on simple order and command, and thus been more
      amenable to natural language translation? The first computers arose out of
      decrypting German and (to a lesser extend) Japanese cyphers. Both these
      languages are highly inflected, though Japanese very imprecise about
      verbs. It is a language where verbs almost become nouns of possibility:
      instead of 'I am catching the bus' more 'my bus catching' (which Sanskrit
      would prefer too although more as 'bus caught me-by' although it is
      European-related language).

      How would a language where relations predominate over things (or verbs
      reduced to actions as 'things') operate? That is in mathematics, but the
      more obscure departments of set and more so group theory, but even there
      that deals with manipulating the mathematical operators themselves, they
      are expressed as 'objects'.

      I have heard that one translator at the least considered using Aymara as
      an intermediary language. I wanted to find something out about this Andean
      language but the site that had everything vanished some year ago and what
      little I can discover shows it to be still very 'thing' orientated to the
      extent of English, having no real verbs as much as descriptive adjectives
      - I am/was ... going/reading/eating etc, but tacked onto the main noun in
      what is treated as a word but since it came from a society without
      writing, who can say what is a 'word' and what a 'suffix'?




      ----- Original Message -----
      From: "TAJ" <allwaysgreatbear@...>
      Sent: Mon, March 29, 2010 18:26
      Subject: [LAMI.US] Post on Human Translation -- Human still is key

      Interesting to see the poll on human vs. machine translation. I
      study Ojibway language = Anishinabemowin and a recent post in a
      yahoo group for study of that had a message / description on the
      near complete failure of a machine translation of a simple sentence
      - The rose is pretty. Really odd how machines can't cope with
      languages that are verb centered rather than noun centered. Anyone
      else seen some odd prolems with machine translations? Thanks for
      your posts and for this group. Mii gwech. TAJ
    Your message has been successfully submitted and would be delivered to recipients shortly.