[Date Prev] [Date Next] [Thread Prev] [Thread Next] Indexes: Main | Date | Thread | Author

Re: [ba-ohs-talk] Re: Semantic web meta data


Johannes Ernst wrote:    (01)

> At 14:23 +0100 2002/08/05, Murray Altheim wrote:
> 
>> What does anyone mean
>> when they say a word? How does anyone else interpret that word?
> 
> Years ago, a fairly young Indian co-worker (went to college in US, but 
> had very traditional parents back in India) confided in me that he had a 
> major problem. His parents were a arranging a marriage for him in India, 
> and he was horrified at the prospect of getting married to someone he 
> had never seen and most likely wouldn't like. He asked me a variation of 
> "shouldn't I be allowed to find a wife for myself, someone whom I love?"
[...]    (02)

> Global ontology? Maybe ... but it is much much harder than it is 
> generally appreciated. Being an immigrant to the US, and not a native 
> speaker, I'm close to deciding for myself that any form of language 
> translation by anyone (and I'm not even talking about computers) is 
> impossible (not "hard" -- impossible) because the underlying concepts 
> don't really map. And if we can't do that very well, how can we 
> construct a global ontology? I don't see how ... maybe 50 years from 
> now, with the process being in step 7 after 20 years of trying, "and 
> then, a miracle occurs" ;-)    (03)

Yes. In my last post I mentioned Geoffrey Nunberg, a computational
linguistics faculty member at Stanford University. He was saying
last year that even within one language (one so theoretically well-
known as English), these issues are extremely difficult to map, and
that context, usage, *who* is saying *what* to *who* in what
*context* and what *voice inflection* and with what *body gestures*,
etc. makes an accurate understanding by a computer extremely unlikely.
(Here I am trying to inflect my words using asterisks -- is that some
sort of standard?) The idea that a computer might one day understand
irony, or catch the tone of someone's voice, a particular look in the
eye as something is said, extremely unlikely. You as a reader of this
message must re-create in your head *how* I would speak this message
in order to understand what I'm saying, where I'm being sarcastic,
etc. Even where I choose to put paragraph breaks has significance.    (04)

Now, on a much more limited scale computers can certainly perform
reasonably well.    (05)

But as you say, once you add the ingredient of internationalization,
well, the whole thing falls apart. How an Indian speaker speaks
Hindi, or English, how someone interprets another language (of which
they are not a native speaker), how body gestures, head nods, etc.
are necessary ingredients in understanding communication, this all
adds up in showing how truly complex our ways of communicating
really are. Certainly key word matching can be improved upon, but
the idea that "a search program correctly locates a person based on
an assortment of partially remembered knowledge: her last name is
"Cook," she works for a company on your client list, and she has a
son attending your alma mater. [...] More generally, semantics
will enable complicated processes and transactions to be carried
out automatically." [1] The fact that you never locate a *person*
on the web (only info about them, which might not even be current)
seems to have escaped the authors. Gad, what a bunch of hooey. Try
locating someone [sic] on the Web when those references are in
three different languages.    (06)

Like in the movie "Brazil", I can imagine a lot of erroneous arrests
and other such nightmares should people start believing the ability
of computer reasoning. Wasn't that movie "Wargames" about this same
topic?    (07)

Murray    (08)

[1] "The Semantic Web: A new form of Web content that is meaningful
      to computers will unleash a revolution of new possibilities",
   Tim Berners-Lee, James Hendler, Ora Lassila, Scientific American,
   May 2001, pp. 34.
......................................................................
Murray Altheim                  <http://kmi.open.ac.uk/people/murray/>
Knowledge Media Institute
The Open University, Milton Keynes, Bucks, MK7 6AA, UK    (09)

      If it wants to be a global power and a player in the
      Atlantic alliance, Europe has to get back into the
      business of making war. -- Newsweek Magazine, June 3, 2002    (010)