Note from Paul, October 17, 1999

Dear colleagues,


Although theoretical issues are often discussed in the Foundations of Knowledge Management in the 21st Century, we are most concerned about the proper scientific and logic grounding of computational Knowledge Management (KM) technology.


This grounding is made in the framework of a tri-level architecture. We will say at the outset that the tri-level architecture is a convenient myth, which we use to create the basis for a new information science. The tri-level architecture assumes that the measurement problem, and the consequences of measurement - e.g. representation, is sufficiently managed so that some new economic value from knowledge management using the tri-level architecture can be demonstrated.


The measurement problem and a proper solution to the problem of representation are easy.


What we know about measurement and representation, in the context of the human perceptional system, is informed by Pribram (1971; 1991) and his colleagues. We look at neuropsychology through the eyes of his work, because Pribram is the only neuroscientist who is able to claim a theory of the whole brain, except perhaps Freeman or Edelman or Grossberg.


From our interpretation of Pribram's work we feel able to ground the tri-level architecture in experimental neuroscience and in what Pribram, and many of his colleagues, refers to as quantum neuropsychology. Pribram's work suggests to us that "perceptional measurement" results in a cross level transformation of energy that is subject to some class of patterns that depend on co-occurrence of energy distributions.


During a course taught by Pribram at Georgetown University (Spring, 1999), he described a series of three sets of Fourier - Inverse Fourier transforms of the optic flow.


The three transforms work like this:


The process starts with an inverse Fourier at the retina lens. Here the scattered light of the environment is focused into a retinal process (not a 2-D or 2 and one half-dimensional image as Marr and others claim) that builds an energy manifold (a forward Fourier transform). The retinal process is a metabolic process that has complex protein conformational reaction circuits. Note that quantum mechanical processes are involved in the absorption of protons by redopsin molecules, in the retina, and that this single class of events must be the gateway events in the metabolic circuits that produce a single distributed manifold and coherent awareness.


The resulting energy manifold is sampled by the axonal dendrites of the Lateral Geniculate Nucleus (LGN) half way between the eyes and the visual cortex. The neurons of the LGN provides a sequence of non-linear processing in rout to a re-spreading of energy into the layers of the cortex. This spread is the second Fourier transform and becomes linear as the information, including timing information, is encoded in the frequency spectrum. The reason why it is a Fourier is due to the underlying physics of lens and the species (Darwinian) need for effortless data fusion. In the linear spectral domain data fusion is merely concurrence of energy fields. The process control mechanisms merely need to push the two energy fields together. Nature finds an exceedingly simply solution to the difficult problem of data fusion.


The third linear transform produces object consistencies that we perceive as objects in the world. This third transform occurs over a period of time and involves movement in space-time. The author interprets the neuropsychology to mean that the third transform acquires code from metabolic processes occurring in the limbic system and in the association cortex and spreads this code across many brain regions.


Code selection involves the induction of a set of basis elements for the Fourier, or Fourier like, spectrum (as channels or dimensions in phase space) and a phase value from the underlying metabolic circuits. The forward transform distributes the results of visual processing and integrates context and pragmatics into the field. The field then must be sampled by one more inverse transforms from the spectral domain to produce a specific recognition of object invariance.

It is important to note that the linear transforms may actually be the non-communitive composition of many non-linear transforms. Each of these transforms are "cross scale", and involve the measurement of "beable" (term taken from David Bell's work on non-locality in quantum mechanics) like phenomena, from substructure.


The viewpoint is not a mainstream viewpoint. Most mathematicians are trained that linear transforms are merely approximations to the more important non-linear transforms. Our analysis suggests that linear is desirable because of the efficiencies of linear processing in the spectral domain. Thus the conjecture, if true, would be counter to the milieu that results from a uniformity in the training of mathematicians.


A case can be made that human induction delays the other wise regular propagation of radiant energy (see Eccles, 1993). Perhaps it is also true, that in highly intentional systems, the duration of emergence is extended to allow self image to interact more strongly during the formation of a symmetric barrier to action. Eccles claims that this symmetry is globally linked via some holistic mechanism to produce a min-body interface and to thus establish the means for spiritual awareness by biological systems. He argues that the human synaptic structures have unique, to human, characteristics that intensify this possible spiritual awareness. The mechanisms also produce cognitive ability far in excess of non-human animal through a multiple level transformation of energy distribution. In this way, plans and goals are incorporated into the substance of the resulting mental compartment. This delay should be empirically observe as irregularities in processing of regular metabolic activities occurring in the neurons and associated gilia. So the case being made can be tested in a rigorous fashion.


Perhaps it is important to note that the framing of the issue of dynamics in terms of linear and non-linear might be less relevant, than the main stream supposes, to the ultimate cause of understanding what we do not now understand about perceptual processes in complex systems. The cross scale process is neither linear nor nonlinear because the definition of the linear or the non-linear systems requires that a set of observables have been nominated before the formal mathematical problem is set up.