|
INTELLIGENCE:
LOGICAL OR BIOLOGICAL?
A
personal view[1] of the
unintelligent artifice that is Artificial Intelligence
Originally published in
Communications of the A.C.M., 36, No.7, July 1993, 15-16
& 119.
Still
as relevant today as over a decade ago! This material has been recycled in
an attack on knowledge management set out in the next article 'The
Knowledge Scam.'
"
Oh! how I just hanker after those long lost days
before gullible governments and organizations were assaulted by an
unremitting onslaught from the proselytes of
artificial intelligence. How
I remember those sunny days when the acronym AI stood for artificial
insemination. I say, I say, I say; do you know the difference between
artificial insemination and artificial intelligence? There isn't one.
Some poor dumb animal is getting 'screwed,' and only the tool-waving
professionals are enjoying it!
Why is it that this low humour doesn't help? Why
is it that I cannot even think about artificial intelligence, expert
systems, neural networks, cognitive science and the like, without feeling
extremely irritated? - irrationally irritated. Is this nervous response
of sarcastic humour merely covering up my own inadequacy, my fear of the
coming intelligent machine that "will keep us as pets"[2]? Marxism is dead - with automation, capital has
overcome labour; is intellect next for machine domination? I had hoped
that AI was dead too; even the United States Government has finally 'wised up' and stopped throwing money at expert systems[3]. But AI is no 'straw man'; it won't go away
that easily. For they are still out there, those hard core of fanatics who
insist that intelligence can be discerned through the application of logic
and mathematical models, those 'would-be gravy-trainers' who absolve
past failures with 'we haven't got the models right yet.' Do those
AI people really think they can capture meaning with a logico-mathematical
analysis that merely manipulates inert tree structures and networks? It
seems they do!
So, my indignation is not inadequacy. My annoyance
is with the arrogant stupidity of these pseudo-scientists who persist in
spreading the malignant lie, that intelligence can be disembodied and then
automated[4]. My aggravation is with those who believe this
present age will have all the answers[5]. My exasperation is with the ignorant who take it
upon themselves to pontificate on rationality, knowledge, intelligence - wisdom even. My irritation is with those who do not realise that the
closer we look at knowledge, intelligence or wisdom, the less we
understand them. These words, any words, are where confusion begins. "As
if every word were not a pocket into which now this, now that, now several
things at once have been put!"[6] Even Humpty Dumpty[7] knew that a word "means just what I choose it to
mean - neither more nor less." "Every
word is prejudice."
Ah! So my annoyance isn't with AI after all, it is
with AI people and with their faith in abstract data structures. Why
can't these people see that our human intelligence, our knowledge, our
reason, our rationality, is interpretation of 'reality' according to a
coarse systematization that we cannot escape. It is what we do, how we
think, how we cope - it is essentially us. It is a systematization of
sense data and biochemical input, that is our inheritance from distant
ancestors, even from those amoebae in the primaeval slime. It is a
systematization that prejudices us in the belief that our sensations show
us the truth about things, about reality. Those amoebae developed reactive
interpretations of their environment, enabling them to survive and
prosper. Humanity also. At each stage in the feedback of our development,
and this includes the present, a logic was formed that was sufficient for
the effective interpretation of sense data. "Not
to know but to schematize - to impose upon chaos as much regularity and
form as our practical needs require."
"Just as
certain human organs recall the stage of evolution of the fish, so there
must also be in our brain
grooves and convolutions that correspond to that cast of mind: but these
grooves and convolutions are no longer the riverbed along which the stream
of our sensibility runs."
|
|
Did those amoebae know the truth about things? Of
course not! Was there some magic moment when, on the tree of evolution,
our progenitors (and why pre-humanity alone?) were given the keys to the
kingdom of knowledge? Has intelligence been perfected into some 'state
of grace'? Or rather, are we trapped; forced to build on a framework set
down before the dawn of intelligence? Have we merely developed upon the
amoebae's-eye view of the world, and cultivated a more sophisticated
schema? Our progress(?) to understanding(?) is entangled in an
evolutionary spiral, genetically restricted and limited by "the
curriculum of an earlier mankind." What if[8] our means of thought was never its original
function? What if intelligence is mere epiphenomenon, mere side-effect of
survival? What if "logic was
intended as facilitation; as a means of expression and not as truth; only
later did it acquire the effect of truth"? What if
"our
intellect is a consequence of the conditions of existence," a
culmination of feedback in this species' successful (to date) quest for
survival? Then human intelligence is tied to life; then it is biological,
genealogical rather than logical. Then human intelligence cannot be
separated from what is the human total. Then synthetic intelligence, mere
mimicry of changing patterns of neurons (themselves a fantasy of the
side-effect), cannot be intelligence, because there is no biological 'side' to be
'effect'ed.
But worse, what if "error is a condition of observation in general." What if we
possess merely "a convincing
criterion of reality in order to misunderstand reality in a shrewd and
advantageous manner." What if the concept of comparison that
pervades all our thinking, "the
presupposition that there are identical things, that the same thing is
identical at different points of time," the idea of 'sameness,'
the seed of equality and enumeration, and thus of logic, was merely appropriate
in our evolution? Then the logic of mathematics is just idiosyncrasy. What
if it has become mere self-indulgent over-sophistication? What if it is no
longer appropriate? Was it ever universally appropriate? Then artificial
intelligence is just error compounded on error.
What happens if, instead, we admit to being in the
trap of evolution? What if we call ourselves human, and accept 'sameness,' and hence
'number,' as a practical choice with
circumscribed appropriateness, while at the same time denying its
validity? This is only a problem to those who insist on the logic of 'false opposites,' grounded in sameness. This is only a problem to
those who fail to see that logic contrives to make simple, to make
consistent, that which is not. The scientist's humble pride of 'standing on the shoulders of giants' merely perpetuates
past-advantageous fallacies of logic and numbers. It is only the dominance
of the authority of logic in our education system (our conditioning), that
makes it unassailable, not logic itself.
But just because formal mathematical logic is
illogical, doesn't mean that it can't be useful. After all, its
domination over our thinking is based on the authority of past usefulness.
"What convinces is not necessarily
true, it is merely convincing." So just what do we humans do? We
describe things, and we describe them well; we describe them in a way that
is useful. We designate. We create descriptions that stand the tests of
time, empirical tests. But those descriptions don't need to be true to
be useful. Do they even need to be convincing? For what convinced us there
and then, need not convince us here and now. What remains for us to decide
is whether, as individuals, we find it necessary to convince ourselves of
the validity of logic merely in order to apply it appropriately. After all
according to this thinking, logic came into being because it was useful,
not because it was true. Oh no! Does this mean that some of the products
of artificial intelligence may be useful after all. It seems so - but
despite the theory and not because of it. For AI remains mere
unintelligent artifice!
But can the products of this logical intelligence
actually control what is 'out there,' if there is an 'out there'?
Can an amoeba control? It is sobering to consider that we share a common
ancestry of interpretation with that single cell! That human evolution has
spawned logic, mathematics and technology from this base, does not mean
that we can control the "flux of
becoming" within reality, any more than the amoebae can. Perhaps
what is 'out there' is far more wonderful than our drab shadow world
of perception can ever appreciate. What if it is a flux more complex than
can be perceived through the naive inventions of our presently evolved
state?
Critics of this attitude would say that it is
paradoxical, because it must ultimately describe itself within a schema
that itself admits to being a false logic. The underlying concepts of
evolution and feedback are therefore themselves products of a
self-confessed misinterpretation, negating any 'logical validity' in
the conclusions it may draw. But does that matter?
For at the same time this approach implies that 'true' and 'false' are themselves merely artificial concepts born in a successful
feedback, themselves sometime-appropriate misinterpretations.
Inevitably these self-indulgent mind games are
futile, "how should a tool (our
intellect) be able to criticize itself when it has only itself for the
critique," "how can we look
around our own corner?" Such arguments must ultimately undermine
themselves, and every other theory for that matter. Absurd! Or is it?
Darwin's contemporaries were only too aware that his evolution theories
held this nihilistic message. They knew that the real argument was far
more profound than the side-show of our descent from the apes. They knew
that evolution held a reductio ad
absurdum rejection of any formal understanding of intelligence. Why
has our own age forgotten this? Perhaps it is appropriate that we choose
not to know that there is no knowing about knowing?
We must just get on with it; decide to play a
different game. Why should we play by the rules of the logicians? Why
shouldn't we play with our own individual self-decided designations of
appropriate and inappropriate? Accepting complexities and ambiguities in
the unknown and unknowable is what we 'humans' do all the time.
Ambiguity is only a problem when we think about it within the cage of
logic; only then does ambiguity make us rigid and we cease to function.
But ambiguity is part of life[9]. People who need certainties, do so because they
have no joy in life.
Ours is an island of reason, of rationality, of
logic, created in a sea of unreason. In the madness of our technological
age, Canute-like, the futile kings of Artificial Intelligence sit on the
beach ordering the tide back. They and their followers fail to see that
there is no knowing these alien waters, and that there is no escape. This
is an island of our own construction, an island 'built of sand.' "Only very naive people are capable of believing that the nature of man
could be transformed into a purely logical one." "The world is logical because we made it logical." We made it
logical in the eons of feedback that is life on earth. The "ideological
thuggery" of logic forces us either to ignore the illogical, or to
treat it as a logic yet to be perceived.
But we humans know better, we can transcend logic -
only the drab obey logic without question. Why am I irritated by
Artificial Intelligence people? You know, I don't feel that way any
longer. Writing this essay has got it out of my system. Catharsis! I can
happily leave Artificial Intelligence to bathe in the sea of its own
futility and ... utility!
"
[1]
I intend that this polemical essay be read aloud. I am using the
rhetorical device of "talking to myself" here, because rhetoric is
far more appropriate than logic for the ideas expressed.
[2]
Edward Fredkin, 'Horizon,' BBC Television, 1983.
[3]
Artificial Intelligence, Daedalus,
Proceedings of the American Academy of Arts and Sciences, Vol.117,
No.1, Winter 1988.
[4]
Every age has its myths of intelligent life through technology: Adam
and Eve, Coppelia, Frankenstein: mud technology, clockwork and
electricity respectively.
[5]
It is for this reason that this essay uses no modern arguments. The
most modern ideas here date from the last century, and some go back to
Plato (c.428-347 B.C.) and earlier. Even the style of presentation is
the Ancient Greek form of reading aloud.
[6]
All the italicized quotations in this essay are taken from
translations by W. Kaufmann and by R.J. Hollingdale of the works of
Friedrich Nietzsche (1844-1900).
[7]
Through the Looking Glass by
Lewis Carroll
[8]
I cannot resist using 'what if scenarios' to argue against the
expert systems proselytes who propound 'what if scenarios.'
[9]
The joke at the very beginning of this essay depends on ambiguity, the
double meanings of the words 'screw' and 'tool,' and the
coupling (sic) of 'gullible governments and organizations' with 'dumb animals.'
|
