Return To 12/95 Prototype Edition Index
Can it be defined? Could an amalgamation of metal legs and circuit boards, or a pattern of activated silicon ever be classified with dolphins, daisies and diplodoci? However, this is not a new question. Philosophers have been losing sleep over it since human consciousness itself emerged from our primate lobes as Mathematician Colm Massey explains....
To the uninitiated, the concept Artificial Life (AL) probably triggers off images of hard disk trashing computer viruses, or even of Frankensteins monster. Those responses are not far of the mark. But to be fair, AL is a vast spectrum of ideas and inventions that ingenious ecologists, mathematicians and computer programmers have been conjuring up since the industrial revolution.
Computer viruses are one of the very few destructive examples from this spectrum and although we should not take the Frankenstein scenario too lightly it shouldn't keep us from exploring and enjoying what this expanding field has to offer. The vast majority of the work which can be classified as AL has been benign, exciting and occasionally very useful! Some of these ideas have been used to help ecologists solve real biological conundrums. (Click for more info)
Some have taught computer programmers how to evolve code rather than design it(Click for more info) , with results often far better than existing human created versions, and some have inspired a few clever games. But how could a computer program or a clever robot ever be given the august label Life? Life is spontaneous, unpredictable, adaptable. How could something preprogrammed by a human ever do more than mimic those characteristics? This is where it starts to get interesting.
Its just that the revolution in computer technology and the quality of science fiction writing has given it greater urgency. Before getting lost in a Zen koan, its important to bare in mind that a full definition of life is more than likely a holy Grail. As members of one family of lifeforms, can we possibly dream up a definition to incorporate all possible lifeforms?(Click for more info). We will probably have to settle for a soft definition and not worry too much if some alien intelligence appears out of the woodwork (or silicon) which doesn't quite fit.
To start, we shouldn't be picky about the substance from which our universal lifeform is made. Terran organic life is primarily composed of carbon, hydrogen and oxygen, but to limit ourselves to those building blocks would be to blinker ourselves immensely. We should try and think in terms of patterns, patterns of molecules, patterns of energy, patterns of information. Once we do this, whole vistas of potential ecologies open up to us. So, what characteristics should these patterns have, before we can label them as lifeforms. Once again to think like this is to hobble ourselves. We should never say this is life, but this is not, we must see it as a sliding scale.
Does a seed have as much life as the tree it grows into? I have heard many convincing arguments saying that the flu virus is merely a complex protein. If we see life as a sliding scale, that problem disappears. Intuitively, a virus has less life than a bacteria, and a diamond has less life than a virus. So now our task is to find the qualities which are more abundant in a bacteria than in a virus and if we have been general enough we should be able to apply this template to our computer viruses and our clever robots and measure how life-like they are.
One lesser known but very powerful quality which could form the template for a definition of life is autopoiesis : The ability of a lifeform (or pattern) to maintain its integrity when interfered with by forces external to it. All organic life satisfies this criterion. How autopoiesis is achieved varies from species to species. Some focus on rapid self-replication, so that their essential form survives even if individuals die.
As we move up the scale of complexity, replication gives way to self-defense mechanisms and on up to what we now call intelligence. Most other definitions rely on self-replication as an axiom, but this rules out the possibility of the biosphere itself being a lifeform. We need at least one other quality, as a diamond has high autopoeisis, simply by being physically robust, but intuitively we would say it is lifeless. One obvious factor is complexity.
Although complexity is very difficult to define without getting into some surreal mathematics we all have a good intuitive understanding of it. So, we have the life of an entity as a function of its complexity and its autopoietic skill. One without the other is no good, as the diamond example illustrates. Similarly a combustion engine is very complex, but tinker with a few screws and it won't work. Both qualities are required.
Now, lets look at the most infamous example of so called AL , the computer virus, and see how well we can lay our complexity / autopoieses template upon it. They, like their protein counterparts achieve high levels of autopoeisis through rapid replication. The more stealthy ones surviving better than others. So, without a doubt computer viruses satisfy the autopoiesis requirement. However they are still vastly simpler than their protein brethren, so this lowers their life rating considerably.
There is no rush to make a Bill of Rights for silicon, but someday that science fiction nightmare may come to pass. The acceleration in computer speeds and memory is unrelenting, and AL gurus are learning new tricks every day.
Within a few years we may have artificial lifeforms as complex as bacteria, another ten years and it may have reached the level of simple multicellular creatures like protozoa. Unless some unforeseen glitch appears the only barrier to this evolution is the technology on whose back these lifeforms will live, and the ingenuity of the new breed of Dr. Frankensteins.
Colm Massey is a mathematician and software engineer at Southampton University Return To Text
More info: Artificial Life has really taken root in behavioral ecology. AL forms excellent models for the interactions of social insects. Return To Text
More Info:The main trick used by AL designers is to immitate the gene shuffling in the sexual reproduction of organic life. The DNA of an AL lifeform is usually made from instruction set, that the computer or robot can understand. The designer can either play God and impose criterion upon these lifeforms, or allow them to compete with each other and let the survival criterion emerge for the virtual world s/he has created. Return To Text
More Info: One lifeform trying to define all life is
very similar to the Zen riddles about thinking about thinking about thinking...
If this gives you a headache, try to grab your left hand with your left hand! They
are both futile activities for identical reasons. Return To
Text
