[Top] [All Lists]

Re: [ontolog-forum] Data Models v. Ontologies (again)

To: "[ontolog-forum]" <ontolog-forum@xxxxxxxxxxxxxxxx>
From: "John F. Sowa" <sowa@xxxxxxxxxxx>
Date: Sat, 31 May 2008 12:33:38 -0400
Message-id: <48417DE2.4050600@xxxxxxxxxxx>
Pat,    (01)

JS>> Doug is not stupid.  He wouldn't have used any sentences that
 >> he didn't feel confident he could translate to CycL.    (02)

PH> Using his translation rules, sure.  But that was my point.    (03)

I agree that his rules would be fairly straightforward.  But I'm
quite sure that the sentences on his test were "back translations"
from CycL to something that looks like a controlled natural language.    (04)

Even the examples chosen by professional linguists are similar kinds
of back translations from the representations they are proposing.
Those examples usually illustrate exactly one issue, unlike real
sentences that tend to combine many complex issues in unexpected ways.    (05)

I often make the point that almost any sentence taken at random
from a typical newspaper gets into research issues in linguistics.
To find a sample sentence, I decided to pick the 3rd sentence
from the 3rd paragraph on page 16 of Tuesday's _New York Times_.
When I turned to that page, I found    (06)

    "This is weird," Mr. Hughes said.    (07)

The syntax of this sentence is not difficult.  But the hard part
is to determine what 'this' refers to and what makes it weird.
A computer program might link 'this' to a passage in the previous
sentence:  "Hughes, getting lost as he delivered a sleeper sofa".    (08)

Getting lost might be annoying, but not weird.  What was weird is
the description of the town in the next three paragraphs.  After
that description, the following sentence summarizes the weirdness:    (09)

    "It's like a little Pleasantville," Hughes said, "It's kind
    of creepy."    (010)

The two occurrences of 'it' in this passage refer to the same
situation as 'this'.  There is no single noun or verb in that
article to which those indexicals can be linked.  Instead, they
refer to the entire situation under discussion.  But that still
doesn't explain why it was weird or creepy.    (011)

Furthermore, a computer that used a database of geographical
locations would be misled.  Hughes was referring to a fictional
town in a movie.  Following is the plot summary:    (012)

    A brother and sister from the 1990s are sucked into their
    television set and suddenly find themselves trapped in a
    "Leave it to Beaver" style 1950's television show, complete
    with loving parents, old fashioned values, and an overwhelming
    amount of innocence and naivete.  Not sure how to get home,
    they integrate themselves into this "backwards" society and
    slowly bring some color to this black and white world. But as
    innocence fades, the two teens begin to wonder if their 90s
    outlook is really to be preferred.    (013)

From:  http://www.imdb.com/title/tt0120789/plotsummary    (014)

When I picked up that newspaper, I had no idea what kind of
complexity was lurking on p. 16.  But the first sentence I looked
at led to a passage that no computer system available today could
correctly analyze.  I could have picked another sentence and
found a very different set of equally difficult issues.    (015)

I seriously doubt that Doug's job applicant was thinking of
anything like this.  And if he had, he would be smart enough
to convince Doug that he was qualified.    (016)

John    (017)

Message Archives: http://ontolog.cim3.net/forum/ontolog-forum/  
Subscribe/Config: http://ontolog.cim3.net/mailman/listinfo/ontolog-forum/  
Unsubscribe: mailto:ontolog-forum-leave@xxxxxxxxxxxxxxxx
Shared Files: http://ontolog.cim3.net/file/
Community Wiki: http://ontolog.cim3.net/wiki/ 
To Post: mailto:ontolog-forum@xxxxxxxxxxxxxxxx    (018)

<Prev in Thread] Current Thread [Next in Thread>