Bruce Edmonds writes:
> This is not true, they *can* play such a role. Techniques of
> self-modifying assembly language are now standard, not to mention
> compilers that are used to compile themselves, etc. You could use
> your argument by analogy - "that organisms organise themselves is an
> illusion, they play no part in the iteractions between atoms"! These
> arguments are false at the relevent level - DNA does not alter the way
> atoms interact but does control the context in which they do -
> software does not alter the way the electronics operates but does
> control the context in which they do. Life is no illusion.
> ....
> Yes, a turing machine has no 'final causation', but why can't real
> computers be an environment where such causation could emerge?
I equate final cause as "being closed to efficient casue"
In organisms,
material cause = material substance (substrates of chemical reactions)
origionally from the environment
efficinet cause = enzymes that convert substrates to products
formal cause = genome that modulates the efficient cause information
final cause is the idea that the system is closed to efficient cause
(all enzymes are made by other enzymes)
Unentailed (from within the system) natural laws of physics and chemistry
provide the explaination for why potential reactions so organized within the
system take place.
In Computers
Material cause = data (numbers) origionally from the environment
Efficient casue = individual program instructions that opperate on data
formal cause = the algorythm (information) that governs which instruction
step will next be executed.
final cause = (would be that all instructions come from the output of other
instructions) This is what I think is imposible.
Unentailed (from within the system) physical hardware provides the
explaination for why potential computations so organized within the system
take place. Additionally, a partial explaination of the operation of the
physical hardware comes from the natural laws of physics and chemistry.
In self-modifying code, one part of an algorythm will be responsible for
over-writing some other part of the algorythm. If you try to entail every
program instruction this way, I believe you will get an infinite regress
of needing more and more statements to do the entailing. Even if you
could get it down to just one instruction being the efficient cause for
a computation that results in a number that can be used as another
instruction, there has to be a second instruction to place that result
in the appropriate place in the computer memory to be considered as an
instruction. This, I believe, guarantees the infinite regress for
computers (all formal systems). It takes the additional instruction
to interpret the result of the computation. Therefor no finite number
of instructions can be self-producing.
In biology, there isn't the need for "the extra interpretation
instruction". Whether a product molecule acts as an enzyme (efficient
cause) is based on just the laws of physics and chemistry and what the
molecule IS. Therefor, it is at least conceptually possible for a non-formal
system to be completely closed to efficient cause (have final cause within the
system). This is because the "meaning" of the stuff is retained.
In a formal system (in which meaning has been stripped), you can never
get efficient cause to close...you always need extra unentailed
structure (code) to compensate for the stripped meaning.
Jeff Prideaux