Thread Links Date Links
Thread Prev Thread Next Thread Index Date Prev Date Next Date Index

Re: SUO: Re: Lattices, Modules, Registries




Jon,

Two things must be carefully distinguished:

  1. Normal use of natural language for communication between
     human beings who are more intent on discussing some serious
     subject matter than playing language games.

  2. Linguistic analysis of language use for the purpose of
     developing a theory of language, semeiotic, or ontology
     for possible implementation in computer systems.

When using language, metonymy is a common and indeed preferred
technique for dispensing with needless pedantic verbiage for
marking obvious distinctions.

But when doing linguistic analysis, as for example, when writing
axioms for an ontology, it is absolutely essential to note and
distinguish each and every use of all figures of speech, especially
metonymy and metaphor.  Failure to note such distinctions is a
common source of fallacies in reasoning, bugs in computer programs,
and disasters in ontologies.

ja> But when I sat down to the Metonymous Computer to play, it
 > laughed, and it laughed, and it just kept laughing if I did not
 > play numbers and numerals straight.

Indeed.  I recommend Section 2, entitled "Signs of Signs" in the
following article:

    http://www.jfsowa.com/ontology/ontometa.htm

One of the examples caused by the kind of confusion you mentioned
is illustrated by the following example in that section:

    Q: What is the largest state in the US?
    A: Wyoming.

The confusion is caused by not distinguishing the size of the area,
from the size of the population, or the size (in alphabetic order)
of the name.  The computer did not laugh, the programmer cried, and
the rest of us can sit back and laugh.

To summarize,

  1. When we are in a hot discussion of whether or not to apply
     lattice theory to the question of how to organize and use the
     modules of an ontology, a discussion of the fine points of
     metonymy and metaphor is an inappropriate distraction.

  2. But when we are in the process of formalizing the definitions
     of the lattice operators and their application to the issues
     of designing the software for processing ontologies, then those
     points are not only appropriate, but essential.

  3. Most of all, however, metaphor, metonymy, and other figures of
     speech must be recognized and carefully distinguished whenever
     anyone is translating natural language specifications into
     programs, axioms, rules, or other notations used by computer
     systems.

Bottom line:  I strongly urge you to apply your analytical skills
to analyzing and exposing the errors embodied in the axioms of SUMO,
OpenCyc, WordNet, etc.  There you will find the mother lode of
confusions and fallacies that need to be exorcised.

John Sowa