Communications on a
National Project
Ken said:
"The semantics we express are not artificial semantics or n-tuples, or mental gymnastics or peusdo-semantics of the kind Noam Chomsky accused linguists of espousing. We did not invent these semantics, rather, we found them lurking within all the ideas that have already found expression. I call these "perceptual" or "sensory" semantics."
<Comments from Paul Prueitt>
Many frameworks for observed
semantic primitives has been formulated, Zachman, Sowa and Ballard are just
three from a class of perhaps 50.
In many cases, there is an
underlying assumption that language itself encodes the primitives
perfectly. Ballard, Sowa and
Zackman reject, with qualifications, this notion, as do I.
We conjecture that the structure of
human language has arbitrary elements to it that makes it a poor mechanism for
precise knowledge indexing using a computer only. John Sowa points out that poor language use should not be
held as a reflection about the potential value of natural language, rather as a
reflection of its poor use in specific circumstances.
So, for example a numerical system
to keep a correspondence between new words and words developed to reflect a
numerological significance was developed for the Arabic language [*] as mnemonic, or
memory, aid and as a means to shape poetry.
The ReadWare technology makes use
of a framework composed of an oppositional scale framework, which we represent
(not in the way ReadWare does, but in a way that is consistent) in Figure 1,
and an elementary categories framework based on the power set of the set of
types:
{
identification, manifestation, ordering }
Figure 1: The
abstraction of an oppositional scale and the 2*2 framework
In the oppositional scale representation we (BCNGroup) collapse similar instances of the various oppositional scales into a single category. So dissipative, out, up, positive, appreciation etc are treated as a category called “doupa”. Like wise we define the category eidnc. The fist problem is immense, and this is that the oppositional scales have paradox within the possible and useful semantics. So up can mean down and down mean up, ect. This specific work follows the work of Dimtri Pospelov on oppositional scales. This work was not published by Pospelov but was communicated to us personally (*). Other assessment tools use oppositional scales also.
In the ReadWare framework three categories of types, { identification, manifestation, ordering }, are used to form a power set having 8 elementary categories of type. These 8 elementary categories of type are then crossed with the four elements of the general oppositional scale (ReadWare calls this bi-polar). A mapping is then made to letters appearing in natural language.
Clearly, the assumption that words and letter patterns have semantic significance will produce results that are interesting but also fail to provide the type of information theory that Dick Ballard believes he has identified with his 18 semantic primitives:
The fundamental enumerations of the
3*2*3 framework matrix is as follows:
independent (I), relative (R), mediating
(M)
physical (P), abstract (A)
occurrent (O), continuant (C), universal
(U)
The 18 cells are then derived (by
Ballard) as
{ process (IPO),
script (IAO), object (IPC), schema (IAC), measure (IPU), definition (IAU),
participation
(RPO), history (RAO), juncture (RPC), description (RAC), interaction
(RPU),relativity (RAU),
situation (MPO),
purpose (MAO), structure (MPC), reason (MAC), law (MPU), formalism (MAU) }
The derivation is straight forward, for example
“mediating / abstract / universal” is rendered as “law”.
A foundational problem that we, as
a community face, is in how to use frameworks that are useful in a specific
context and how to not use these same frameworks in context that are not
appropriate.
Beyond this issue of context, which
Klausner helps us understand, is the issue of non-stationarity and
novelty.