Thursday, September 02, 2004
Manhattan Project to Integrate Human-centric Information Production
Stratified complexity and the
origin of mental/social events, (Prueitt 2002)
Conjecture on Stratification
A discussion between information scientists
Paul,
Basically,
I'm sympathetic to Ben's comments.
As
he notes in bead 50 in this
series – people have proposed various schemes of semantic primitives.
During
the era of "classic" AI research in the 70s Roger Schank, for
example, was well-known as an advocate of constructing meaning from a small set
of semantic primitives.
And,
as I recall, he had no qualms about inventing new primitives as the situation
demanded.
But
I'm inclined to be skeptical of such schemes.
They may provide workable solutions to practical problems in specific
domains, but it is not at all obvious to me that searching for a finite set of
atomic meanings is a worthwhile endeavor in the context of understanding
language and the mind in general.
Bill Benzon
Bill,
There is a misunderstanding about what it is that I am proposing, one that Ben’s first note [48] has focused on and reinforced. In his second note [50], one sees the path forward:
I think this is a reasonable approach, but that the atoms will never
be as unique or clearly-defined as in the chemistry case: I think you can
isolate many different "conceptual periodic tables" underlying human
concepts as expressed in language. [50]
So we have to walk out of this box, and realize that the conjecture is about ontology, not our understanding of things.
Ben and I have been here before, so perhaps he will help clarify what the situation is.
I am not proposing specific methods for discovering periodic tables.
The conjecture is about the possibility that knowledge science will be founded on an ontological commitment that these substructures exists in great stability, like the periodic table of physical atoms, and that the great diversity and seemingly endless possibility seen in the structure of things has a strong analogy to what we see from physical chemistry.
So this leads us to the issue of how many sets of substructural stabilities are there. For example, in one person the formative structures are very stable after puberty, so that if a second language is learned after puberty the second language will be spoken with certain specific perturbations of the sounds produced. So do we have “two” periodic tables in the production of speech in one person? The answer maybe “yes”. Does a social group create a unique periodic table that controls some aspects of the generation of communication? (see Pribram and Bradley’s paper.)
Ben’s comment about openness to new atoms is important. But the Conjecture would suggest that any specific substructure, once formed, [1] is not open, for reasons that can be explained in objective terms. Where the openness is seen is in how the combinatorics are expressed. Stratification theory makes this clear, and distinguishes Readware letter semantics form what Ann Wierzbicka talks about (see [54]).
The Readware substructural ontology is not perfect, but there is a principled argument being made that such substructural ontologies are in place because of the invariances in the need for communication, the shape of the mouth, the nature or brain phenomenon such as electromagnetic coherence, etc.
The deep definition of ontology is in play here, as an origin of causes - not something that can be "seen". There is a parallel that Peter Kugler makes, privately, regarding ontology expressing as form with function in real time. In the same way, the gene expresses as form with function in an environment. The parallel between formative ontology and the science of genetics has some merit.
One can place the Conjecture on Stratification up against the Hard Artificial Intelligence Conjecture, and ask the government program managers to choice between them.
It seems far better to try to ground computer science in a study of real physical phenomenon such as exhibited in all systems that are in fact balanced in “stable” phenomenon far for thermodynamical equilibrium. This is the nature of the Process Compartment Hypothesis.
[1] I want deeply to develop a set of mathematical theorems relating Stu Kaufman’s work on autocatalytic sets and catalytic “ignition” to the formation of substructural invariances. I just have no time, as the efforts that I make to fund the National Project leaves very little time for this work. My hope is that I will life long enough to do this work in an environment where continuous mediation and contemplation is possible.