ONT Re: Inquiry Driven Learning Environments (IDLE's)
¤~~~~~~~~~¤~~~~~~~~~¤~~~~~~~~~¤~~~~~~~~~¤~~~~~~~~~¤
| Document History
|
| Subject: Extensions Of Mind
| Subhead: Essays And Reports On Intelligent Systems
| Contact: Jon Awbrey <jawbrey@oakland.edu>
| Version: Draft 3.04
| Created: 10 Sep 1993
| Revised: 11 Feb 2002
| Advisor: C.C. Wagner
| Setting: Oakland University, Rochester, Michigan, USA
| Excerpt: Division 4 (An OAR For Odysseus: Offline Analytic Resource)
| Excerpt: Subdivision 4.3 (Multi-Sage Deconstruction: Post*Modern Elephant)
| Excerpt: Section 4.3.2 (Prospective Needs Analysis: Talking About Tables)
| Excerpt: Subsection 4.3.2.3 (Mosaic Paradigm: Twin Tables for Every Law)
| Excerpt: Subsection 4.3.2.4 (Post Script: Alternating Currents of Inquiry)
4.3.2.3 Mosaic Paradigm: Twin Tables for Every Law (cont.)
Only a few of the relations that we need to talk about, a very small fraction
of the implicit and virtual tables that are referenced by the language over T,
will ever need to be rendered fully in the form of actual and explicit tables.
A suitable parser for the table specification language will tentatively assume,
in effect, that every term refers to a table, but that most of the properties
of these tables can be handled, in passing, by the axioms that constrain their
virtual existence -- like so many angels dancing on the head of a pin. It is
probably obvious that many of these implicit tables will entail, overlap, and
subsume others. Except in the process of responding to detailed queries and
in generating concise surveys, it is usually not vital to draw the separate
regions of the universe away from their immersion in background potentials.
In particular, the all-inclusive table X = (()) will serve little purpose
in being actualized, since all of its cards, empirically and normally
speaking, can already be found on the other tables.
A high degree of virtuosity in our relational tables is exactly the
kind of feature that we need if we wish to achieve a database system
that is truly declarative and deductive in character. The reason for
this is associated with the slogan that "Analysis Precedes Inference".
This means that the information we get out of a knowledge base through
the operation of deductive inference is dependent on the prior measure
of information that we previously put into the knowledge base, chiefly
through the operations of analysis as applied to the raw materials of
observation. We often fail to be aware of this fact, thinking that
inference can generate its own sustenance. This appears to happen
because, by the time we get around to reflecting on the operation
of our own reasoning processes, we already have very large, but
mostly taken for granted, personal knowledge bases. Inference
seems to be the active ingredient and the generative principle
because we can at least catch a drift of its current working,
while the spade work of analysis that prepared the ground for
inference is lost to awareness in the mists of distant memory.
The upshot is, that if we want to reconstruct the entire process
of knowledge development in our artificially intelligent systems,
a significant share of the work will have to be done in equipping
our systems with solid subsurface shorings and in building up the
adequate infrastructures that are afforded by analytic resources.
There are two strands convolved in the temporal extension of a database
that need to be drawn out and highlighted here. Shown below, in several
syntactic variants, are two basic axioms of classical propositional logic:
o-----------------------o o-----------------------o
| Non-Contradiction | | Excluded Middle |
o-----------------------o o-----------------------o
| | | |
| A and (A) = Ø | | (A) or A = X |
| | | |
| A & (A) = 0 | | (A) v A = 1 |
| | | |
| A · (A) = ( ) | | (A (A)) = (()) |
| | | |
o-----------------------o o-----------------------o
Interpreted for the twin table scheme, these become semantic
integrity principles with the following operational meanings.
1. "T & (T) = ()" means that the intersection of table T with table (T) is
normally empty. If a table E is determined to be empty at the completion
of a parallel table scheme, that is, "when all the chips are in", then E
should also be empty at each normal stage of development.
2. "T v (T) = (())" means that the union of table T with table (T) is
normally universal. If a table F is determined to be full or universal
at the completion of a scheme, when the empirical extension or enumeration
of examples has become exhaustive, then F should also be all-inclusive or
universal at each normal stage of development.
In sum, as they say in quantum mechanics, the relation of eternal laws
to temporal dynamics is loose enough that some infractions of the axioms
may well occur, but only the virtual and transient violation is allowed to
happen, and the tolerance is reciprocal to the moment and urgency of what's
in the offing, that is, what's at stake in the fundamental exchange relations.
4.3.2.4 Post Script: Alternating Currents of Inquiry Driven Learning
The process just described bears a remarkable resemblence to a particular
three-phase cyclic model of inquiry that was promulgated by the pragmatic
philosophers Charles Sanders Peirce, John Dewey, and William James. From
this perspective the very same recognizable stages of dynamic architecture
can be seen informing every moment of deliberate and intelligent thinking --
from a solitary individual's puzzlement in everyday thinking to organized
scientific research as a cultural activity and the work of generations.
The beginning of inquiry, as it happens in our spontaneous experience,
in a state of nature, under field conditions -- "inquiry on the hoof" --
does not run with the kind of idyllic drive that gives us rein over it
without a struggle. It is a caricature -- though a venerable one, no
doubt descended from pictures anciently painted on cave walls, for the
twilight anticipation of the next day's hunt and the dim instruction of
tomorrow's hunters -- but still a cartoon to picture inquiry as though
there grazed outside our doorsteps any such creature as a stable body
of detailed and domesticated knowledge, and as though we had a choice --
to make a beeline straight to the target or else to play the child's
game of "pin the tail on the donkey". But the life of inquiry is
more than just a diversionary way of winning the pre-fixed prizes
of an already known knowledge.
In the wild the occasion of uncertainty that instigates the struggle
of inquiry is a predicament primarily to be avoided. The bigger the
surprise, the greater the chance that it's already too late to begin
a careful investigation of the problem, and the bigger the risk that
the creature in question is presently in the middle of its very last
learning experience. It is only to anticipate these unbidden moments
of sudden doubt that we have secondarily learned to create artificial
inquiry situations for the sake of achieving instructional objectives.
These serve as moderated risk preparations for meeting with the wilder
types of inquiry. They help us learn what few general principles might
guide us in approaching the boundaries of our knowledge and the unknown
realities that impinge on us from beyond it. They work to reduce excess
sensitivity to the otherwise overwhelming experience of being in a state
where rules derived from familiar experience do not seem to apply at all.
Jon Awbrey
¤~~~~~~~~~¤~~~~~~~~~¤~~~~~~~~~¤~~~~~~~~~¤~~~~~~~~~¤