[Top] [All Lists]

Re: [ontolog-forum] LInked Data meme revisited

To: ontolog-forum@xxxxxxxxxxxxxxxx
From: John F Sowa <sowa@xxxxxxxxxxx>
Date: Wed, 11 Dec 2013 13:34:39 -0500
Message-id: <52A8B03F.7020408@xxxxxxxxxxx>
Dear Matthew, Kingsley, David, Hans, and Phil,    (01)

Since I have been talking about both natural languages and controlled
natural languages, I confess that I may have caused some people to
forget the huge differences between the two.    (02)

Summary:    (03)

  1. A CNL is a *formal* language with a precisely defined translation
     to a version of logic.  Its semantics are identical to that logic.    (04)

  2. The points that Kingsley and Matthew made are true of CNLs because
     CNLs treat words and phrases as the equivalent of URIs.    (05)

  3. But those points are *not* true of any NL.  Sometimes, a document
     on a technical subject (such as math or science) may include some
     sentences that are as precise as a CNL.  But they also include NL
     commentary -- and there is no clear boundary between NL and CNL.    (06)

  4. To use Wittgenstein's terminology, the meaning of a word in any NL
     is the totality of its uses in every possible language game in
     which that word appears.    (07)

  5. But the meaning of a word or phrase in a CNL is deliberately
     restricted to a single language game for a very narrow purpose.    (08)

> Words typically have multiple meanings, indeed just one meaning  is
> rather unusual. So you need a way to pick out which meaning you  mean
> this time  when you are using it, especially when it is out of context.    (09)

I agree, but with qualifications.    (010)

The major qualification is that there is not and can never be
a complete, precise dictionary of all possible senses for any NL.    (011)

As Sue A. and Adam K. said, "I don't believe in word senses."
The senses listed in any dictionary are rough groupings that are
helpful reminders for humans.  But the meanings of words shift
with nearly every use in every text (written or spoken).  See
http://www.kilgarriff.co.uk/Publications/1997-K-CHum-believe.pdf    (012)

> The connection between words, phrases, and HTTP URIs (which are all
> different *kinds* of identifiers) is often lost or overlooked when
> attempting to understand the principles outlined in Tim Berners-Lee's
> original Linked Data meme.    (013)

I'm glad that you emphasized *kinds*.  But I want to emphasize that
the difference is a fundamental semantic gap that no URI can bridge.    (014)

Although I sympathize with Tim B-L's goals, URIs can never convert
an informal NL to a formal CNL.    (015)

> each & every layer of the SDLC—software development life cycle, which
> is more aptly P(roduct)DLC, I(dea)DLC—has it's own unique impact on language    (016)

That is true of *every* field.  Even in mathematics, the Idea DLC is
just as vague, informal, and confused as any Idea DLC in Ontolog Forum.    (017)

If anyone doubts that claim, just look at the terms 'irrational number',
'imaginary number', and 'complex number'.  Those names reflect the long,
heated debates in the history of the subject.    (018)

> So we have dynamism in the evolution of language and word senses, we have
> broad variability in contexts in which the words are used, and we have
> dynamism in the relationships of the dialog participants to each other.    (019)

I agree.  But I'd add that dynamic shifts occur even within a single
peer-reviewed publication in science.    (020)

In fact, there is an ongoing debate in ontology:  When you change one
axiom of a theory, the relationships among all the terms shift -- and
therefore, the meanings of some or all the terms shift.  Which of the
many terms have shifted in meaning?  Which one(s) require new URIs?    (021)

If you don't change the URIs, you can't claim that URIs are precise.
But if you change them for every change to every theory, it becomes
impossible for anyone (human or computer) to choose which URI to
assign to any occurrence of a term.    (022)

> it is still possible to create formal representations of exchanges
>of meaning...  Humans can internalize the meaning of those
> representations  with sufficient accuracy to make judgments about
> their relevance and  usefulness.    (023)

I agree that formalization can be helpful.  But even without it,
people can "muddle through" to useful applications -- as the rapid
growth in technology demonstrates.    (024)

> it is still possible to create formal representations of exchanges
> of meaning...It is highly desirable to do so -- proactively...    (025)

Anybody who has spent any amount of time teaching any subject is
painfully aware that it's impossible to anticipate the open-ended
variety of possible interpretations of anything.  Most of them may
be wrong, but some of them may be highly creative innovations.    (026)

> As they say, a picture speaks a thousand words; thus, I created
> illustrations of what I am referring to with regards to...    (027)

I certainly believe in the importance of illustrations.  I use
many diagrams in my slides and articles.  But by themselves,
pictures are even more ambiguous than words.  In response to
the saying "A picture is worth a thousand words" I always reply    (028)

    A single word can often clarify a thousand pictures.    (029)

John    (030)

Message Archives: http://ontolog.cim3.net/forum/ontolog-forum/  
Config Subscr: 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 join: http://ontolog.cim3.net/cgi-bin/wiki.pl?WikiHomePage#nid1J    (031)

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