SAMPLE
MACHINE NARCISSUS:
The Evolution of Intelligence beyond the Computer Age
Copyright (c) Paul Pangaro 1987. All Rights Reserved.
[The following text was written as an introduction to one
major theme of a proposed book. A "proposal"
followed, but the relation to this sample was not perfect. References
and illustrations (with one exception) are omitted.
This text has been OCR converted and is probably not be perfect;
apologies.]
The joy of creating ideals, new and eternal, in and of a world, old and temporal, robots have it not. For this my mother bore me. ---- Warren McCulloch
1 The Fantasy
Figure 1, "Our machine, the Eureka", was first published
in "Calculator Saturnalia", a short volume made up
of two, intertwined conceits. A novella, about the adventures
of a cybernetic research team and their robotic/intelligent partner
(named Olga), is punctuated by instructions for calculator games.
To leave it that the calculator games are mere amusements would
be a disservice. They consist of political satire and irony;
they are clever, instructive, and usually misprinted. This last
point makes it hard to follow the directions.
The novella is the setting for our little Eureka, a pocket calculator
(truly). As you expect from the drawing, it too is an extended
conceit. The original caption is given just below, and has metaphorical
value for a major theme of MACHINE NARCISSUS. (It is worthwhile
looking for each indicated point in the drawing as you read each
section a, b, c, .-- as this enhances the effect.)
The Eureka, however, is not entirely a joke. For Peek wants you to realize that present-day digital computers, from the smallest calculator-in-a-watch to the largest computer on earth, are all as quaint and clunky and bloody stupid as our little Eureka. And a good deal less charming. Pask is saying that there is no sense in trying to make such a crude calculating engine "smart" because any attempt will fail. You may as well try to dismantle it and carry in your pockets.
2 The Mirror
Considered to be the highest faculty of human beings, Intelligence
is at the center of our history and achievements. Efforts to
amplify human intelligence find their greatest expression in
the medium called the computer. Some researchers seek to make
machines intelligent like humans; others say they seek intelligence
in any form.
Mythology brings us Narcissus who fell in love with his own reflection,
not realizing he was seeing an image of himself. So too in seeking
to reproduce, intelligence in modern "artificial intelligence"
(AI), we think we see the universals of intelligence as independent
of human life. Instead we only see our own cultural view of what
it means to think. Culture and its view of intelligence are complements:
each delimits the other. Developments in AI are both reflections
of culture and influences upon it.
On the face of it, to consider our Eureka as a candidate technology
for smart machines is quite absurd --- and yet that is precisely
what researchers in the AI community have done since the 1950s.
Our culture at large has brought about this startling condition,
and examining the culture helps us to understand the limitations
of the current technology. That is where our story will begin.
3 The Goal
I do not mean to imply that the phrase "machine intelligence"
cannot be given meaning; but where you begin determines where
you end up. Some researchers now say that the directions of mainstream
AI cannot produce intelligent machines to any significant degree
when compared to human capacities. But they do believe there
is a meaningful interpretation to "machine intelligence"
and are working to make it.
Until recently, perhaps the last 10 years, this position would
not have been easily defended in scientific and academic circles.
It has been dismissed as pure bias or mere "intuition"
or smacking of humanistic rather than scientific values. Discussion
on the subject deteriorates into stand-off and reasoned arguments
become static statements of personal views.
However, work that is squarely in the discipline of science itself
now questions the universality of present-day computation, previously
thought to be powerful enough to make brains. Major fields of
scientific enquiry are descending on AI with powerful, new perspectives
on the problems of intelligence. Early results point far beyond
the already-predicted revolutions in computers and biology. Even
our conceptions of language and society will be changed, and
perhaps more profoundly than by any science since the Renaissance.
Our story in MACHINE NARCISSUS is the story of the imminent development
of intelligent machines as an evolution of technology and of
culture. Computer science has produced a technology, but society
and its concept of intelligence are the boundaries within which
technology develops. Some technology might aid the quest, but
which? And based on what goals and whose cultural biases? Success
in creating machine intelligence requires a major change in science
and society, and in science that change has begun.
4 The Paradigm
In a culture which has successfully reduced the world to its
atomic parts and their causal interactions, it is inconceivable
that the same reductionist approach would fail when applied to
intelligence. But the evidence against reductionism in machine
intelligence is growing, and the pressure to change is coming
not from one field but from no less than four: physics, mathematics,
biology and cybernetics. Each are providing new evidence, experimental
as well as theoretical, as to why and how a revolution in AI
is required. And AI, within itself, is evolving and responding
with new models and new technologies.
AI developed from the context of the present scientific milieu.
Although there are other perspectives not in the mainstream (as
our story will tell), AI is mainstream science. The context of
any enquiry has implicit influence on the nature of the search.
Modern parlance might call this context "the paradigm",
but modern parlance would also miss the essence of that term
--- that the implicit framework is not known --- for a paradigm
is the underlying structure that influences but is not known
explicitly. The irony is that, once known, the paradigm exists
no longer. But while it exists, it forces researchers' hands
and only with revolution can real change occur.
Ironically too there have been countless side-effects (read:
advantages) in the search for AI in the field of AI. Given the
very hard problem of making machines "smart" in that
way, computer scientists have worked hard to create problem solving
tools. Powerful techniques and computer languages, software development
workstations and even expert systems are entirely within this
category. These technical advances cloud whether the goal is
being approached. As raw power increases it is easy to say (1)
advances have been made and (2) further advances require only
better technology.
But smart machines, built on the paradigm of the Eureka? First
1 expose the paradigm in our culture's notions of intelligence,
and its effect on efforts to embed intelligence into computers.
Once exposed it is disposed, and we can more clearly discern
a new context of interpretation for machine intelligence in modern
research.
5 The Computable
Alan Turing created the "mathematics" behind digital
computers --- and the quotation marks emphasize that the term
has a specialized context of interpretation. After all, one person's
mathematics is another's alchemy. Ironically Turing didn't intend
his earliest work quite the way it was taken on by computerists.
But under the circumstances they couldn't help it.
Turing made a mathematical proof. I don't say "discovered"
a proof because that would imply it was a universal waiting to
be found. And if anyone at all could have found it, it would
not have been such a good gift.
The value of the gift was to reduce all "computation"
to one type, or class, of computation. No matter what you were
trying to do (balance a checkbook or compute the orbit of the
sun) the calculation could be described in an abstract form.
In Turing's sense we have two types of fodder for calculation:
that which is generally recognized as the "data" of
conventional problems (the amounts in dollars and cents, positions
of the heavenly bodies); and a representation of the desired
calculation (the goal, in the form of a procedure). This is good
news in the short term, but bad news in the long term.
What would be the advantage of this abstract description, other
than an aesthetic consistency? Here is the genius of Turing:
once the descriptions were made consistent, each description
could itself stand alone without any reference to an individual
knower, or a particular calculating engine. There is no reference
to physical embodiment, yet the procedure is completely defined.
The concept of an abstract machine is invoked to simulate the
calculation of any specific machine. A specific machine for the
specific problem need never be built. Only the data of the problem,
and the data that describes the appropriate machine to "solve"
it, is needed. Here is the culmination of years of reductionism
in science: the complete, stand-alone, disembodied calculation.
Turing took the opportunity to move from thought experiment to
technology: he constructed his abstract machine in physical form.
Turing's computer was fed the data for the calculation (about
our checking account, for example) as well as the disembodied
description of a machine that, if it existed in physical, restricted,
hard-wired form, would compute what we wished (i.e., the balance
of our checkbook). Today we have a name for this disembodied
description: "software." Although that physical hard-wired
computer which only balances checkbooks does not exist, its results
do. A machine is created by its computation being performed,
even though the machine itself is not created, except virtually.
Doesn't it have a Frankenstein monster appeal, even at this primitive
stage of its capability? (Would the Eureka be scary, rolling
down the hall at you and in the dark of night?)
6 The Implications
The first implication is that you can make money by going
public with a company selling expert systems based on the Eureka.
No, sorry, we need to develop the argument ~a bit further for
that one.
The following is a partial list of where we end up as researchers
in the present milieu, carrying on from the present scientific
traditions, adding Turing Computability to a fair amount of chutzpah:
- Problem solving is breaking large problem into smaller ones.
- Solving small problems means extracting the variables and plugging them into procedures.
- Procedures are what we use to think.
- Knowledge is what is in the procedures.
- Computer languages can represent and execute procedures and hence can have knowledge.
- Putting procedures into computers is called programming.
- Computers can be smart if they have the right programming.
Got it? It surely took a great deal of cultural and scientific
development to invent and engender that scenario, told of in
Part 1 of MACHINE NARCISSUS. Part 2 of our story will show how
each of the above points is interpreted in computer science,
including the present fashion in expert systems. Alternatives,
the newest starting points, will be the focus of Part 3.
7 The Limitations
Would that we did get the above points without any restrictions.
One problem is that Turing machines, in their abstract and theoretical
form, perform any computation of the Turing Computable type.
This has the advantage that we now only need one machine and
anything (do you detect a climax coming?) could be computed-
It takes no PhD to realize that the crux of the matter lies
in the definition of Turing Computable, a topic which must be
taken up again in some detail.
But what has this to do with human intelligence? Some would feel
that thinking and the execution of procedures as described are
quite different things. Aside from superficial likenesses, there
seems to be no reason, other than naive desire, to push the analogy.
But what if there were some hint that the brain worked along
the same lines as Turing machines? Or, at least, that brains
and computers could be compared in some way? You must forget,
during this leap, that you are assuming that the brain is the
only item of interest in a discussion of intelligence. The impact
of this error (paradigm?)is a major topic of our Story, and is
dealt with briefly a few sections below. Before that, I present
the stepping stones for the leap that brains and computers have
similarity. Much of this book, and the remainder of some century
of research (perhaps this One) will be about knowing their differences.
8 The Nervous
Warren McCulloch and Walter Pitts put neurophysiology and
mathematics together to look into the computational structures
of the nervous system. With the powerful Turing metaphor so close
by, they sought to take advantage of it. Knowing about neurons
and theories respectively, McCulloch and Pitts made a theory
about networks of simplified neurons, showing that these networks
could compute anything that Turing machines could. Their purpose
was to show a strict relationship between neural nets and Turing
machines. Putting it in the crude and possibly misleading terms
of "the layperson,"
However, the community of computer scientists, soon to be researchers
in the new field of artificial intelligence, took over from there.
With the personal confidence that Turing Computability was all
encompassing, they became Sorcerer's Apprentice. Turning the
reasoning backwards in an absolute fit of scientific reduction,
their dream is simplified to:
A computer could be a brain.
After all, the brain is a "meat machine", isn't it?
It has been a slippery slope for smart machines ever since.
9 The Culture
This feels like a kind of wish fulfillment, having all this
theory and the advancing technology too. The euphoria is understandable,
though the stock market price of expert systems firms may be
in doubt. Fundamentally, every desire within this community of
research is to bring mentation down to the level of computation.
Is it a reasonable wish? According to Turing mathematics, it
is. Now no one would easily imagine that what we have here is
a theory of mind --- this is mathematics, only. And although
logical and others consider mathematics the center of the universe
there is not much agreement on that point from physicists, cyberneticians,
or biologists.
A theory does not itself make an artificial mind.
Researchers in many fields have had the instinct (or more) that
Turing Computability is a trap. However workers in AI have been,
until recently, unstoppable. Some concession to parallel computers,
where many processors are working at once, has been made all
along, but usually in terms of efficiency, not inherent limitation.
But at last a strong response has come from physicists, extending
the notion of computation to forms which are not contained in
the Turing model. (The world has caught up with Turing himself
who, after his early work on computability, explored how living
organisms execute procedures beyond the limits of his earlier
formulation.)
Now the concept of Quantum Computability subsumes that of Turing
Computability and the physics of this is the subject of [Chapter
x]. Because physics is a respected field and plays the same "objectivity"
game as mathematics, and because these fields have currency with
the computer scientists, this work is having widespread impact.
At this same time, AI itself has been toying with new computer
architectures.
So it is all moving ahead smoothly, or so it seems. The solution,
however, is not merely an expanded definition of Computation
with a capital "C". It also requires a shift of emphasis,
away from what goes on "in the brain", to what happens
"between the organism and its environment". And suddenly
we are in the world of the New Biology.
10 The Organism
Everyone who studies biology knows that it is about living
things, and part of that study is the relationship of the "organism"
to its "environment". Our descriptions of this relationship
include the concept of the organism's model of the environment:
an inner-constructed view of the world which is the means by
which the organism operates, plans, and survives.
Plausible? Yes indeed, it is an excellent model. For that, in
turn, is all that it is: our model of what we conceive of as
happening in the organism. We do not, and can never, know what
is inside the organism and by what basis its operations are conceived
and carried out. Instead, and almost without our knowing it (one
of those paradigms creeping in?), we project the notion that
the organism builds a model of the environment.
Who are the "we" of the above paragraph, who posit
these models? We act as individuals, surely, but always in the
context of the culture. To say that organisms build models of
the world is a cultural statement rather than "the way it
is." Why do we have that particular view of organisms? Because
clearly we humans do it that way, and so met other living things...oops,
falling into the trap again. Just because we 88 humans have the
power to describe a world in language, we forget the subjective
nature of those descriptions.
Why must it be the case that our description of our experience
of building internal models of the world is actually what we
do in our nervous systems?
Descriptions are created in the context of purpose: to predict,
to build a technology, to fulfill a cultural desire.
So, why complain? Only if the descriptions generated mislead
or inhibit certain developments (considered, of course, to be
desirable).
11 The Gap
We might consider the eye an "input" device: we
say that when we see a beautiful flower, it is because the eye
receives the image of the flower as input. We might also consider
the hand as an action or "output" device (even in a
restricted context, say of movement). But just because we attribute
the status of input or output does not make it so; the words
used can cause us to think that the nervous system takes input
from the world and provides output to it.
A new body of work, in biology primarily, warns against this
as mis-stated. Consider the connection between a neuron and a
synapse in the nervous system: Can you imagine spreading apart
that connection and calling the neuron an input device, and the
synapse an output device? It seems absurd, but from the point
of view of the nervous system, any attribution of input or output
is from an observer, not from the nervous system. To say of ourselves
that "our own eye receives input" is an observer statement.
To the nervous system, what happens in the retina is a perturbation;
but it is no more "input" than what a synapse has to
say to a neuron.
What alternative view can replace the conventional view and provide
new capabilities? That provided by New Biology (and affirmed
by a variety of other fields, perhaps even literary deconstruction)
is that the nervous system is a closed system. Nothing external
to it (as the observer sees "external") can determine
what happens in the nervous system; it can only trigger changes
determined within its structure. (The implication is that the
nervous system is "deterministic" and in important
ways, it is; it is not however usefully predictable. The full
implications must be reported in [Chapter y].) If the behavior
of the nervous system is not determined by the environment, then
the nervous system cannot construct a representation of the environment
"internally." The observer may say that an interaction
between the nervous system and the environment causes a structural
change in the nervous system, and that this change constitutes
a representation; but for the operation the nervous system, this
structural change is not a representation of the environment.
To quote one maestro of this reasoning, "To talk of representation
in the operation of the nervous system has no sense."
Why bother with this reasoning? Sounds like pure pedantry on
first reading (or maybe on second, when the reasoning appears
more firm but no more helpful). To build an artificial intelligence
that *s like a nervous system, the implications are major. The
amount of energy spent in AI in the areas of knowledge representation
and expert systems is huge. Perhaps the direction to follow is
not that of AI, which is merely building on the current cultural
paradigm. Would it be better to first approach from theory, then
create the technology? Turing could create the technology because
he had a theory. That the theory has allowed the technology to
be taken too far is ironic.
The implications of the reasoning in New Biology might not be
savory to some. Reversing the example of placing a gap between
the synapse and neuron, consider that the environment is the
medium that closes the gap between eye and hand. In that sense
the environment is a component of the "mind"; just
another synaptic gap. The so-called mind-body "problem"
takes on another perspective.
Closed nervous systems imply a forced solipsistic existence for
us all; but this is not really so. Instead of remaining monads,
we share with each other the status of observers. We agree that
we see the flower, and even agree that it is beautiful. We enter
each other's synaptic gaps and thus, through language, intertwine
our nervous systems in a dance of agreement.
12 The Convolution
To imagine a change coming about in a field of science is
not difficult. The constant evolution in all fields of endeavor,
and the acceleration of that change in this century, is reported
daily. in the media (albeit by hyperbole). Within science the
stories of revolution have made news, careers and reputations.
What is particularly difficult about our story is that it represents
revolution not just in computer science or artificial intelligence,
but in physics, biology and mathematics as well.
To have a major change within a defined field is part of the
excitement of intellectual endeavors in science; individuals
become scientists in hopes of bringing about such revolutions.
For major fields of science to also converge upon each other,
not for the sake of taking over another's domain but because
there is no choice in the matter, is very, very rare.
The details of how each f~ield fire my pattern is the story of
the book itself. The focus of each field, and the revolution
that each represents, is surprising. I just presented the some
surprises ~in biology, and the next sections do the same for
physics and mathematics.
13 The Uncertain
At its extremes, the field of physics appears more concerned
with description than anything else. Implementation and technology
seem irrelevant when discussing the composition of distant galaxies
or the existence of atoms or molecules or quanta or quarks, whatever
happens to ~be fashionable in the current century. To expect
that physics has anything to say about the nature of computation
is to expect that the molecules of the Eureka ~better not come
apart or we've got troubles in our pillars. Not very helpful,
really.
The physics of things would also help insure that the computation
itself were robust. We hope that it will be consistent and give
the same result for the same data every time, for example. We
do want determinism somewhere. Physically, the Eureka achieves
this because of its mechanical, and hence physical, prowess;
if it holds together (why do you doubt this?) it calculates consistently.
Electronic computers appear to do it by physics too; after all,
the electrons are connected to the ions, the ions' connected
to the... A familiar tune. But knowing the physics at this level
is not a useful description of the calculation. That would require
consideration of what the electrons, et al, represent in the
calculation. This requires a symbolic representation, and quickly
we are out of the physics and into the computation science of
the problem. (Here we have one reason why the physicists want
some way to move back into the spotlight.)
Physics in the 20th century has been dominated by quantum mechanics,
and we still hear stories of how you might actually appear on
the other side of a wall, just due to the physics of things,
though no one has ever seen this. To say "physics"
in the modern age must invoke some sense of the uncertain, indeterminate,
probable, and even the unlikely. But, how might such indeterminacy
be at the heart of an argument about intelligent machines?
Quantum Computability requires a look at 20th century physics'
current metaphor: quantum mechanics. Oddly enough, quantum computers
might compute some things faster than conventional, deterministic
Turing machines, and, of course, they compute some things which
Turing machines cannot. Quantum computers are capable of simulating
any arbitrary, finite physical system --- including the universe
itself. Hence, in the conclusion of one main proponent of this
work) all this does is merely to "make computer science
a branch of physics
14 The Illogical
Mathematics too is part of the story and not unexpectedly.
The realm of logic and representation has often been its triumph,
though the struggle between mathematicians and physicists is
sometimes forceful,
The logic provided by modern mathematics is not the trivial sort
of Mr Spock of Star Trek; it admits to the subjective and multiple-world
sort that humans sense every day. Computers though derive from
a tradition of representation and calculation that is impoverished
by comparison. Their technology is the very newest, but not their
techniques. In this sense computation science is as new and as
primitive as can be.
One contribution of George Boole's was to apply logic to do arithmetic,
two subjects that are not the same. When binary numbers arose
as an easy technique in which to interpret Boolean logic, computer
technology jumped ahead because here was a single framework in
which to represent and compute forms that were logical ("if
a and b, then always c") and arithmetic ("z = x + y").
Single framework means simpler technology, and computers could
therefore be built from simpler elements.
So the only hard logic built into computers has been the sort
.from this zero/one, yes/no, true/false, black/white, all/none
tradition. As the physical embodiment of the logic glided across
vacuum tubes and transistors and various electronic forms, there
was no questioning of the binary mapping of zero/one to true/false.
Facts became primarily true or false. When a fudge was needed,
any uncertainty became a probability, and without much satisfaction.
Yet another probability is then required: a measure of confidence
that the result was indeed correct!
Alternative logics have existed in philosophy and mathematics
for many, many decades. Only recently are these other choices
coming into fashion with those closer to technology. These logics
have various for but all strive to represent values in shades
of grey.
One direction to move is to the more primitive: "In the
beginning was the distinction." The distinction is said
to be more primitive than yes or no, 0 or 1. Breaking down any
complex structure can be performed by carving more and more distinctions
until no more can be drawn with the available data. Indeed any
division along a continuum.(another means of providing multi-values),
or the convolving of many values along many dimensions, can be
modeled as distinctions, and vice versa.
These alternatives have been invoked to handle belief systems.
Contradiction and conflict, ambiguity, context, commands and
questions all survive in representations that are not true/false
based. Ancient and modern paradoxes are not solved but are resolved
in formulations that provide multiple values for any given variable.
Modal logics have been found by some researchers (prominently
the Japanese) to hold greater promise for representing complex
problems.
Recently an algebra has been proposed which embodies an "ethical"
structure. It has been used to explain differences in the morality
of whole cultures. If validated (and already the evidence is
strong) such an "algebra of conscience" would be a
candidate for implementing the core of Asimov's Three Rules of
Robotics.
Mathematics can also express the structure of existence. The
topology of Systems would be an odd place to look for an explanation
of laughter and tears, but it has been done. The dimension of
event that is required to be wise can also be derived, and it
corresponds closely to a lifetime.
15 The Nexus
There is a danger in the convergence of so many strong fields
on a single domain: divergence of descriptions, competition for
dominance, in general the view that one view should take general
precedence because of its scope, aesthetics, or (heaven help
us in a subjective domain) predictive power. But, like in an
imaginary science fiction, there is also available a single (I
hesitate to use the world "unifying") framework of
description in which to couch the disparate approaches of these
many fields. Cybernetics, the most misunderstood field of the
era, is this framework.
Lest it immediately appear that a single icon is being invoked
to prevent the dominance of conventional science, some perspective
is necessary. Cybernetics is an epistemology, rather than a scientific
discipline defined by a domain of application. Not concerned
initially with embodiment, it has no entrenched interests, as
physics, biology, mathematics, or computer science invariably
do. Though not from mainstream science, cybernetics is periodically
adopted by those fields to provide answers to otherwise unanswerable
questions.
Cybernetics is concerned with how we know what can be known;
therefore it is about uncertainty of the ultimate kind. It is
not control theory, or neural modeling, or robotics. It may be
about all of these things, but only if they are modeled in particular
ways. it is near to mathematics, because it models the trajectories
of systems. It 18 near biology because it models a system's ontology.
It is like physics because it seeks to encompass the macro and
micro, in a unity of physical and mental worlds. And it encompasses
AI and computer science because It can combine their objective
needs with the subjective acts of inventing distinction and having
purpose,
The above comments must simplify things somewhat. We similarly
simplify conventional biology to be concerned with living things,
AI with intelligent machines, physics with the large and small
of the universe, and mathematics with representation in symbol
systems. All of these are human endeavors, which is their unifying
core. Each Involve the observation of the interaction of systems.
Cybernetics, as an epistemology, is concerned with the observation
of interacting systems; its primitive is the interaction.
One major product of this enquiry is a theory of conversations
that 18 not restricted to verbal exchange between humans. It
has the scope to model exchange among the tiniest particles of
physics, and the largest concepts of mind. It can model consciousness,
time and the specious present. In a view of existence as solely
interaction (as that is all that we know)* the use of the word
"conversation" is not just a brilliant metaphor, it
is the unification of subjective and objective, scientific and
humanistic.
This is the contribution that cybernetics makes to the fields
outlined above: it brings an image of synthesis, not to replace
but to enhance and connect. In the process, many a paradigm is
challenged and revised. And this is how we will complete our
whole story, with one additional contribution from cybernetics
to the nature of machine intelligence, summarized as follows:
If interaction is primary, then distinct individuals must exist
in order to interact. The observer of the interaction posits
their existence --- and at the same time, posits his or her own
existence, by acting as observer of one's own conversation with
the interaction under observation. All knowing is relative to
the ability to interact, and hence it is relative to the existence
of the distinction of the individual. If knowledge cannot exist
without a knower, making machine intelligence means making an
individual --- not in the sense of robots, but in the ability
to draw distinctions, of oneself and others. To succeed at this
is to create, in a single stroke, self and other, the here and
now.
16 The Mirror
Recognizing oneself ("one as a self"), then, becomes
central to intelligence. The mental places we inhabit provide
distinctions we can make (our culture), and create a context
for further distinctions that we invent (our evolution). To view
ourselves on the outside, we model the cosmos. To view ourselves
on the inside, we model intelligence. Everywhere between these
extremes, our language and society attempt to unify experience,
to smooth any edges and to create a seamless flow of description
from inside to outside, macro to micro, within to among. After
centuries of specialization and particularization, the movements
of science intertwine and converge: physics becomes computation,
mathematics becomes reason, and conversation becomes epistemology.
And if science allows biology to become cybernetic (remember
synaptic gaps and the dance of agreement), intelligence can be
embodied in new forms.
Our story, then, begins and ends with a reflection of ourselves.
Our insides project into the world, and the world reflects how
we conceive. We project into ourselves, and our mind reflects
back the universe. The nested reflections of science and culture,
our world and ourselves, are the cycle of human endeavor that
our story is about. At the center is what we name as intelligence.
Our naming makes it ours. At some time, machine intelligence
will also do naming, drawing the distinctions of the world in
its own language, just as the old story says Man and Woman were
commanded to do.
- end -
© Copyright Paul Pangaro 1994 - 2000. All Rights Reserved.