Saturday, January 10, 2009

Fragmented Thoughts on Algorithms


An algorithm is simply a finite set of unambiguous instructions performed in a prescribed sequence to achieve a goal. There are three key factors to algorithms that must be satisfied in order for the process to be legitimate. These are:

1. Substrate Neutrality.
The result of the algorithm is dependent upon the logical structure of the steps, not by the materials used to complete the process.

2. Underlying Mindlessness.
The series of steps must be utterly simply, in that no judgments, decisions, or intuitions are required for steps to be completed correctly.

3. Guaranteed Results.
If all of the steps are followed accurately, the algorithm must do whatever it is that it does. (This seems redundant, but is important.)

The point of interest is this: whatever it is a given algorithm does, it doesn't have to have a conceivable use or value. People are accustomed to considering such systems from a perspective of interest or utility, and many don't realize that there are some algorithms that are so irregular or pointless that there is no way possible way to articulate just what it is they are for. As Daniel Dennett says, "they just do what they do and they do it every time." For the majority of people, it's simply not worth the effort to identify and understand an algorithm that has no apparent purpose.

And so: it is interesting to wonder how many of the phenomena we experience in the world are controlled by such seemingly useless algorithms. How many useless processes does it take working in concert to produce something that does have apparent value and/or purpose? And would each useless algorithm simply constitute a single step in a larger one? As I said, it would be difficult to even recognize such systems, as human thought patterns generally aren't likely to devote much power into handling inherently pointless processes, especially if the series of steps required to produce the useless (or more accurately, the apparently useless) result is staggeringly large.

I initially encountered the discussion of algorithms in Daniel Dennett's book on evolution, Darwin's Dangerous Idea. Everything above ties into evolution and natural selection as follows: the mistaken assumption is that natural selection is an algorithm with the intended result of producing us. For opponents of evolution, it is difficult to view the theory of evolution outside of its relevance to the origins of human design, and thus they falsely classify the process of natural selection solely as an alternative explanation for the existence of human beings. While natural selection may occur as the result of a series of algorithms, it is erroneous to assume that those algorithms must specifically be for something, that there is a certain meaningful goal to their process.

No comments: