Alife and Cyberanthropology

    ------------------------------------------------------------------------


Hacking Life Together - the beginnings of artificial life

Like "virtual reality", artificial life seems to refer to a cluster of
associated ideas. Some people think artificial life research is the attempt
to create lifelike robots, perhaps even at the nanoscale. Or that it is
another form of artificial intelligence (AI) research - AI being the effort
to design computers that think like human brains, either through emulation,
or, using neural networks and "biochips," imitation. Yet others seem to
think that it has something to do with biotechnology and the computer
sequencing of genetic information. But most often the term artificial life
is used to refer to something else - the modelling of biological processes
(sometimes called cellular automata ) on a computer.

The very first artificial life program was, of course, LIFE, created by
hackers at MIT in their heyday on a PDP-6 minicomputer. There are various
versions of this 30-year old program available today. It is a very
minimalist program, compared to complex simulations like SimLife which are
available for PCs today. Basically, the "cells" in Life are pixels which
either replicate, stay put, or disappear with each succeeding "generation",
depending on very simple algorithms which determine the result depending on
the number of "neighbors" (1-8) the pixel has. Since it's essentially a
"population" simulation, in Life, overcrowding results in mass death, and
so does isolation. But if you have the right number of "neighbors,"
surprising growth occurs. What was amazing to the creators of LIFE was that
cell combinations (organisms, if you will) of surprising complexity -
'floaters,' 'gliders,' 'shuttles,' and 'repeaters,' for example - would
appear from these relatively simple algorithms...

The early hackers, despite the obvious simplicity of their program, could
not help but think they had created a new life form on their computer...
after all, these 'organisms' displayed behavior that no one could have
predicted. But this was part of the hacking philosophy - with the hands-on
imperative, it was assumed that as each hacker layed on some sort of
modification or improvement to programming code, the emergent 'hacked'
together code would have capabilities not achievable by any one programmer.
The mentor of the hackers, Marvin Minsky, thought this was just how the
human mind worked - the 'ego' or 'consciousness' just being the 'hacking
together' of a bunch of subprograms which made up the mind... and that by
encouraging hacking, he was not surprised that their Mind Writ Large might
arrive at the type of surprising results found in LIFE.

People studying the hackers note that for many of them, LIFE was a way to
make life outside of the messy problems of biology, love, and relationships
(e.g. the 'real world.') Since many of them were essentially asexual, LIFE
was a means of bringing life into the world without having to surrender to
far messier, and less easy to algorithmize, reproductive processes... when
modern-day hackers like Hans Moravec call their creations "Mind Children,"
there is an implicit statement that they have found a novel way for the
human race to reproduce itself - an asexual and antiseptic way, with no
risk of communicable organic diseases, in which the resultant 'offspring'
might be more carefully modelled, and perhaps inevitably more 'fit' (metal
being more durable than flesh) than their predecessors...

Today, of course, artificial life researchers working in Santa Fe research
labs (or elsewhere) are using simulations far in advance of the original
LIFE. In these simulations, 'cellular automata' are actual 'simulated'
organisms, in 'simulated' environments within the computer which impose
'natural selection' on their destiny. These 'virtual environments' are
intended to model evolution itself - at a much faster timescale. The
'automata' contain 'DNA' of sorts - information for making copies of
themselves - which they pass onto descendants, through processes of 'sexual
reproduction,' e.g. they combine code to create their descendants... and
like real organisms, they contain the possibility for mutation and
adaptation. Mutation is usually introduced into this 'virtual world'
through random number generators that make sure it happens, as in the real
world, at random intervals...

The researchers working on these simulations have tried to make them as
complex (thus, in their opinion, close to the real world) as possible. Thus
their A-life creations contain algorithms for self-preservation,
competition for the 'virtual' resources in their environment (usually, CPU
time), and other behaviors to maximize their 'genes', e.g. replication of
their own programming code, in a Richard Dawkins-like way... and once
again, beginning with simple algorithms, they have found surprising results
and novel strategies. Their A-life organisms have resorted to parasitism,
symbiosis, cloning, cannibalism, and a host of behaviors both parallel to
the biological world, and often bizarrely different from it, in their
master program-imposed quest for reproduction.

A-Life and Evolution: The Post-Darwinian Synthesis

Not surprisingly, A-life computer wizards, in modelling Darwinian
evolution, have been led to revise some of the current theories in biology
as to how it occurs...this resulting in a post-Darwinian synthesis which
incorporates the 50s synthesis of Mendelian genetics, Darwinism, and
molecular biology, while perhaps inviting through the back door forbidden
notions of Lamarckism or even "Wallaceism" (e.g. teleology or
purposiveness) back into evolutionary theory, along with some of the more
modern modifications, such as "genetic drift" and "punctuated equilibria."
Hardly creationists, the A-life theorists still feel there are great
inadequacies in the existing evolutionary model, which offers no other
guiding force to evolution than the "Blind Watchmaker" of natural
selection.

In standard neo-Darwinian theory, the environment has no effect on
organisms. They may make organic changes to deal with a changed environment
(e.g. using a certain muscle more than before), but because of the
"Weismann barrier," these changes are not passed to the germ plasm. The
only motor for genetic change is random mutation - scrambling of the genome
through random blasts of cosmic rays. And the only thing that in any way
constrains and shapes that change is natural selection. If the mutation is
not favorable to allowing the organism to survive or reproduce in its
current environment (and 99% of mutations result in this outcome), e.g.
increasing its 'inclusive fitness,' then the mutated gene and its carriers
die. All the myriad adaptations in organisms (eyes, wings, fingers, etc.)
we see today are the result of countless numbers of random mutations and
the ruthless slaughter of those mutations which reduce fitness...

A-life organisms, however, are modified directly by their "environment,"
since ultimately their environment is made up of the same machine language
that composes them... and A-life mutations, despite being driven by an
essentially 'random engine,' still show strange coupling to the
environment. Many A-life theorists also note that their mutations look
curiously purposive , seeming to employ "heuristic search strategies" much
in the same way as chess-playing programs do, i.e. they explore all
possible alternatives (their "possibility space" in mathematical terms) and
make directed self-modifications based on evaluations of those
alternatives. The point A-life theorists make is that their cellular
automata appear to be employing strategies for maximizing reproduction -
which is, after all, what one form of conscious 'real world' life on this
planet, homo sapiens sapiens does already.

Some people see this reintroduction of teleology into evolution as just
latter day Wallace-style mystification, an attempt to get some form of god
back into the machine. A-life theorists say no, it's entirely possible that
all life was governed by the natural selection 'algorithm' originally (1.
MUTATE RANDOMLY 2. IF YOU'RE NOT FIT, DIE, AND DON"T REPRODUCE.) , but
perhaps somewhere along the way, through this process, another 'algorithm'
was created which superseded (more properly, worked alongside to) natural
selection. We know that the genome is more complex than originally thought
- there are McClintock's "jumping genes" and some genes which activate or
deactivate or modify the action of others. It might be possible that some
mutations are not just "copying errors," caused by a nasty dose of
radiation, but are instead self-directed 'metaprograms' for
gene-modification, e.g. produced according to certain conditions already
'coded' within the genome. (These conditions could include response to
changes in the environment.)

Biologists and naturalists cry foul at computer scientists having the gall
to step into their territory. They insist such conclusions are not valid,
because the simulations of the A-life researchers are not life and cannot
be governed by the same biological laws. Any phenomena they observe which
are discordant from the existing principles of biology are simply the
results of their failure to program their simulations to conform to
'real-world' biological processes. Nonetheless, the A-life researchers'
work have come to the attention of philosophers like Varela and Maturana,
who do see definite implications for things such as cognition, evolution,
and reflexivity... they understand that organisms are coupled to a larger
system (Gaia, or Sunflower World, if you will), which in fact constrains
the ways in which they evolve, beyond just random blind change.

Mechanistic versus Electronic Life

The adherence of many biologists to the old Darwinian synthesis is, in
part, due to the fact that they still adhere to the basically mechanistic
view of life advanced by such molecular biologists as Jacques Monod. It was
a physicist, Erwin Schrodinger, who first suggested that life might be
based on quantum-mechanical properties that are non-deterministic and non-
mechanistic - based on the same subatomic processes which make possible the
transistor and much of existing electronic technology. Descartes thought
that animals were basically automata, but men were not because they had
souls; LaMettrie went further, suggesting that even humans were machines -
driven by the same mechanical processes that governed the simple engines of
his day. Today, we are starting to find out that humans might well be
machines - but they are not steam engines; instead they are
digital-electronic machines, possessed of the same strange emergent
properties as Alife... including a capacity for unexpected and
unpredictable behavior and novel strategies for reproducing their kind.

Recent studies in the bioelectric and electromagnetic basis of life have
shown just how close this view may come. Besides the electric potentials
carried by the nervous system, researchers have found other electromagnetic
signals mediating bioinformation throughout the body - such as the "current
of injury" found in the healing process. There have been all kinds of
speculations about the implications derivable from the electromagnetic
basis of life; perhaps it is a verification of the Oriental view that the
body is full of currents of life-energy-order (ch'i ), or a vindication of
the Sheldrakean hypothesis that there might be "morphogenetic fields"
directing the specialization of cells and the form-maintenance of
organisms. The field of 'electrobiology' is still controversial, but it is
apparent that EM fields do have far-ranging impacts on people's health,
their circadian rhythms, and perhaps even their cognition... and that
humans and other life forms may be as electronic in basis as their
computers.

This may explain the curious reaction that people, and especially children,
have to computers. Their immediate reaction is to look at themselves, note
sociologists like Tracy Kidder and Sherry Turkle. Rather than asking
themselves the questions of the AI researcher - am I a computer? - they ask
about the computer: is it alive? (Does it think? Does it have a soul? Can
it see me? If it can do all this, but it isn't alive, what does that mean
about me? Could people do all these things and still not be alive?) Coming
to the conclusion that it is not, they inevitably come to the "deep"
questions that AI researchers have argued about - what is it about me that
is different, that makes me alive? Turkle in particular is amazed at the
creativity and experimentation that the computer brings about in children -
almost as if they have discovered a playmate of sorts.

Since it forces us to confront that age-old question - what separates life
from nonlife - the computer is a remarkable catalyst. One thing that has
been found - once again, through computer modelling of chemical processes
such as the Zhabotinsky Reaction discussed by Ilya Progigine - is that life
is negentropic, an open system, and self-organizing. Through essentially
"chaotic" (nonlinear) processes, life reverses the normal entropic
processes toward decreasing order (but not permanently - not at the
macrocellular level, certainly, since complex organisms still age and die),
mainly by incorporating energy and information from the environment. The
Second Law of Thermodynamics is thwarted, locally, within the organism,
because it increases the disorder of its environment, rather than itself.
The organism maintains coherence through a constant "dialogue" of self and
not-self (the essential working of the immune system.)

Schrodinger suggested that at the quantum level, the thermodynamic arrow of
time might not be manifest. If life - whether in the computer or in the
real world - is based on processes occurring at this level, it might be
able to reverse macroscale physical processes leading toward entropy. He
thought that life was essentially different from nonliving matter because
it was the result of emergence from properties of phenomena at the
subatomic/quantum level, such as indeterminancy, complementarity, or
nonlocality. The gene was at so small a scale that it was likely to be
influenced by quantum effects - just as the microprocessor is. As we
miniaturize the nanocircuitry for our computers more and more, approaching
ever closer to the threshold for quantum mechanics to "take over," we may
be surprised to find that we are on the verge of observing the sudden
"phase transition" that may have produced life from nonlife aeons ago...

In the quantum domain (just as within computer simulations), as Feynman
diagrams demonstrate, processes are completely time-reversible and involve
essentially the exchange of information. The electron does not smash into
another electron like a Newtonian billiard ball. Instead, since each is
more of a "smeared" out wave than a compact sphere, what is really going on
during, for example, a jump from ground state, is the exchange of a photon
from one to the other. The photon - and other 'messenger' particles like
the Z or the graviton - is really basically a messenger, carrying the
information: "Hey. Move over." Moving from Newtonian to quantum mechanics,
we also move from the physics of matter and locality to the physics of
information and probability... the electron on which we and our A-life is
based is neither here nor there, but smeared out all over a probability
space between here and there.

Viruses, Information, Life

In dealing with that age-question, what is life?, we cannot help but turn
to that boundary-straddling organism, the virus. Biological viruses are
generally just strands of DNA surrounded by a crystalline protein sheath.
They are clearly far more 'primitive' than the eukaryotic cell - yet
viruses reproduce by invading the cell and 'hijacking' its chromosomal
machinery for making copies of themselves, thereby killing the cell.
Viruses can be dormant for incredibly long periods of time, often not
needing many of the things (oxygen, light, water, etc.) that other living
things do. They do not grow. And they are incapable of reproducing on their
own. Indeed, if you look at one under a microscope, it really appears quite
'dead' - just a microscopic crystal sitting there doing nothing. But
introduce it to a living cell, and quite amazing (and deadly) things start
happening. That infinitesimal little crystal can easily kill things as big
as us.

The computer virus is so-named because it does exactly what biological
viruses do. The computer virus enters a host (computer), "hijacks" its
programming for its own purposes, and it makes copies of itself, often
ending up overwhelming or erasing all the data in the host. Through
communications links or disk exchange, the virus' replicants then propagate
to other computers. Like the biological virus, the computer virus is a
clever parasite. Well-designed computer viruses periodically 'mutate' to
resist antiviral software, just as biological viruses occasionally appear
in strains resistant to earlier vaccines... further, they are capable of
"hiding" in parts of the RAM that make them "invisible" to users, of
altering their replication rate, or of only "activating" the destructive
parts of their code at a specific time, so that they can remain in the host
for longer periods. Like biological viruses, computer viruses display novel
strategies for assuring their reproduction. You can see where I'm going...
from this viewpoint, we can say that the virus (both variants) is indeed
alive.

Indeed, Richard Dawkins provides us with the meme, a mental parasite rather
akin to the computer virus. Dawkins says that the 'selfish' gene uses the
'vehicle' of the physical organism for making copies of itself. Likewise
the 'meme' (we can call it a 'belief' or 'idea' or 'mental self-replicating
proposition,' but I prefer to look at it as 'programming code which
contains instructions for its own replication,' e.g. a computer virus, if
we take the metaphor of the brain as a computer) uses the vehicle of
"culture" (acquired/learned behavior) as its mode of propagation. If this
form of exploitation were not bad enough (genes making organisms run about
for their selfish purposes), humans invented culture a few thousand years
ago, and allowed memetics to take over genetics - the memes started running
the show, wresting control from the genes. Biological/genetic evolution
braked for the human race, because a new mode of adaptation, cultural
evolution, took over, which was quicker and more effective...

The memes can't survive without the genes, since humans are needed to
spread them to each other through communication. (Suicide and celibacy
memes quickly disappear, unless supplemented with recruitment memes.) but
basically, memes can be seen, like computer viruses and the selfish gene,
to be exploiting their hosts (human minds) for their own purposes - to
spread themselves as far as possible, as fast as possible. Indeed,
so-called memetics experts try to adapt the gene metaphor to memes - when
and how do memes mutate? (clearly they do; ideas, despite idealists,
change.) Under what conditions do they propagate? Which memes are the most
'fit'? The ones that enhance the biological survival of their hosts, or
not? What cultural and environmental processes govern memetic selection
(when do memes not take root in a new cerebrum?) Etc.

People often have hostile reactions to the notion of memetics - you're
making it sound like ideas are alive, they often say. Well, A-life
theorists are greeted with the same skepticism when they suggest that
computer viruses are alive too - that viruses are A-life organisms that do
theirs one better, because they are not stuck in one computer simulation on
one computer... many virus writers do see their creations as lifeforms. The
problem is that the virus has the potential to get out of hand, just like
the Internet Worm did in 1988. In that case, the virus represents a danger
- like a weed, it may replicate itself so rapidly that everything around it
is killed. Certainly, this is one of the few worries that roboticists raise
about their creations, especially if they are von Neumann machines.

It is certainly feasible to construct a robot that can build another robot
exactly identical to itself. Such von Neumann machines might be perfect for
exploring intergalactic space, making copies of themselves when they've
started to wear out, some people have suggested. Nanotechnologists like K.
Eric Drexler see nanorobots doing this at the microscale - indeed, they
have a process which they can model/emulate, the replication of the cell.
Of course, the inevitable risks in creating such constructs is that their
replication could easily get out of control, killing off their creators.
Strangely, some artificial life gurus don't seem to be bothered by this;
they assume that they are creating a form of life superior to organic life,
and that it's only 'natural' that 'artificial' life supplant its organic
predecessor someday...

The Metaphysics of Alife: the Quantum-Computational Cosmos

Even the hackers who hacked together LIFE couldn't help but feel the
gnawing presence of a Big Question in what they were doing. What if they
themselves were programs or automata in a simulation called the "universe?"
(If so, where was the programmer?) Thinking on the fact that since he was
not omnipotent, he was also an imperfect programmer, Gosper wondered openly
if the LIFE organisms might notice the imperfections in their environment,
come to the 'realization' that they were automata, and ask their 'creator'
to give them the code that governed their existence... sure, it was a
stupid, way-out thought. The lives of the hackers revolved around the
computer, so they could be excused for stupid musings like "Where does the
computer end, and real life begin?" Or "what if the programmer for Life
(the real world version) left some Easter Eggs in the program for us to
find?"

But physicists like Heinz Pagels were and are searching for the "Cosmic
Code" - the "Master Control Program," if you will - even if they've put
aside the search for a programmer. Many of them are convinced that we will
eventually find GUTs - the unifying grand mathematical theorems that "code"
for all the particles and forces in the universe. Many are now starting to
advance Gosper's proposition - the universe is basically computational or
digital in nature, composed not fundamentally of matter or energy but
instead information. The "Quantum-Computational Cosmological Model" (QCCM,
for short) is also called by physicist John Wheeler "the It from Bit"
hypothesis. It suggests that the universe is itself made up of cellular
automata, so to speak, on/off bits which are constantly turning each other
(billions of times a nanosecond) on and off, this flux being the basis of
the most basic things we can observe - subatomic particles such as the
photon or electron.

Computer scientists like Vallee who advance such a 'cybernetic' model of
the universe point out that it also suggests that events may be linked as
much by information, association, or analogy, as they are by location in
space and time. Further, there are those who suppose that if the universe
is a computer program, the emergence of conscious life for feedback loops
(you need conscious observers to collapse the quantum wave function in the
Copenhagen interpretation) was built into the code from the beginning - the
Cosmic Anthropic Principle. Likewise, one would also expect that such a
program would need a 'control system,' a sort of thermostat or CPU to keep
various subroutines from overrunning the main program - make of that what
you will.

The possibility that we inhabit an infoverse suggests startling new
possibilities for consciousness, as physicists like Fred Alan Wolf have
realized. It turns out that consciousness may be a fundamental emergent
property, necessary for the maintenance of the entire system, rather than
an epiphenomenal accident from matter, which some behaviorists saw as
"ultimately not providing much value for the organism." In an infoverse,
the emergence of beings capable of processing information, especially about
themselves (self-consciousness) would be vital. The hackers were creatures
of technique, more interested in engineering solutions to problems than
philosophizing, but LIFE, that strange program running on a mechanical
device, came to threaten the mechanical view of the universe. To think that
all of this came out of some dancing pixels on a monochrome screen...

Steve Mizrach, aka Seeker1