FRIDAY, APRIL 25, 2008
Prototype: Pacman lives on screen, pacman is sufficiently advanced to discover fundamental laws of universe:
Power pellet grants invulnerability for time t, inversely proportional to level of advancement.
I travel at a constant speed, can stop only head on into a wall, etc.
Now, this fictitious pacman must be an early success of A.I, if it can analyze its own universe and deduce natural laws. However, I ask, is it possible for the pac creature to further deduce how it came into being, what programming language it is written in, which country the hardware is housed in, or even what a country is? If it is capable of growing in intellect and knowledge should it someday be able to accurately describe the motivations that drove some human programmer to give it life?
I’ll depart momentarily to discuss the problem with induction in the foundations of science, as I see it.
Science is a game of bouncing back and forth from deduction to induction and back.
Observe. Abstract. Deduce. Predict. Observe.
The laws that are abstracted and function to predict and explain are discovered, and then reinforced, by observation. If a theory fails, it is modified or replaced. Then science never makes the statement “prediction P MUST be true,” but rather “if P is not true, our theory must be altered or replaced.” So questions of “why” can never fully be answered. Answers to “why” follow from deduction: a woman sitting in a cardboard box, left with nothing but canned food, water, and a few axioms of set theory, emerges after some months and surprisingly has theorems in mind that coincide with ours on the outside world. She follows deductive reasoning, concludes *[your favorite theorem]* MUST be true. In the process she has answered the “why,” as well.
But science cannot be practiced in this way. Each “why” demands another level of mystery, and until things become fairly reliable, really anything can happen. We need to observe in order to induce. But then anything that is beyond our ability to observe is beyond our induction, even if the phenomenon is causally relevant to that which we can observe.
For example? Pacman! It may very well be the case that the reason pacman exists is that a programmer was obsessed with mazes, early video games, and artificial intelligence and therefore created the pondering pacman. Pacman’s story of his own origins might include this fact, except for the fact that pacman was made in an entirely formal universe, which might have had a million different origins, each ridiculously distinct. So really the best pacman can do is say “space is black, the pellets are white, there are never more than 150 small pellets on a board,…,and there may be many things outside this universe, but we will never know.” As useless as that last statement is it is certainly more profoundly true than “…, and that is the entire grand unified theory of the universe and everything.”
At this point I may have a reader who is concerned I am working up to an argument for Intelligent Design. This would be unjustified guilt by association. Instead the reader may safely infer I think staunch atheists are fools, and flaunting a serious lack of imagination is something I find irritating if not just boring. I do insist that any reasonable philosopher can entertain ideas of creation without falling into religion’s old clichés , but that is not my point to make today.
[to be continued…]