When we state something like "Every effective process is computable
... but there is a special emergent quality that transcends
effective processes." (taken at random from Jeffs clear-thinking
classification of possibliities), What do we intend as our meaning?
We may have a fairly absolutist view of truth, in which case we may
well argue in terms of absolute limits to computational process,
essential differences between types of causation, etc. This is a
typical philosphical approach, we seek to state true propositions,
which are true independent of viewpoint, language, time, frame of
reference, etc. In my view (only!) this tends to be unproductive
from the viewpoint of anything useful coming out (other than sheer
enjoyment of argument) - neither our means of distiguishing between
alternative propositions, our natural language, our formal systems,
our communication skills or even our brains seem to be up to such a
task!
I argue for a large degree of pragmatic criteria to be involved in
such a debate - truth is what is useful (in some long range sense).
I will not go inot the full philosophy of this (which is a real
mess!), but state some consequences:
1. Truth is sometimes relative to your frame of reference;
2. Purely formal arguments are, in themselves, insufficient to
establish such truth;
3. Sometimes aparently conflicting statements have to be accepted as
"equally true".
4. Establishing your *purpose* for statements, laws, concepts, etc.
is important in discussion.
5. This sort of truth is not ideal, but the best we can get (i.e.
others are, by definition, less good(=useful)).
6. You tend not to get absolute and unqualified distinctions (e.g.
something is either complex or not, alive or not, etc.).
Of course, this kind of truth is criticisable (otherwise it would be
a form of absolutist truth), self-referential (like other kinds of
truth, but more so (its ultimate justification is that it is a true
kind of truth because it is more useful than other versions of
truth!)), not completely defined (the starting point is
that truth is linked not to such dry definitions) etc. etc.
Both Ron Cottam's and my papers at E&M reflects this viewpoint
(maybe others as well).
Thus the question "will we be able to ever formally model life?"
rests MORE on practical questions (to which the answer is obviously
NO in the foreseeable future), rather than theoretical questions such
as limits to theoretical computational processes.
Thus my view on the possiblity of emergence of "life" inside an
essentially formal but environmentally rich system, I will explain:
The key argument (as I see it) goes something like this:
(0) From formal systems, you only get other formal systems. We are
starting with a formal system without life, therefore we won't ever get it
from a future one.
Other similar arguments have been proposed for various subject
matter, e.g.:
(1) A necessary condition for free-will is that the previous state of
mind is not determined, your first foetal state is genetically determined,
so there is no free-will.
(2) Life can only evolve from life, so if at some stage in the past there
was no life, then life can not have evolved.
Needless to say, I do not agree with these arguments, they all rest on
absolutist distinctions (formal/non-formal, free-will/dtermined,
life/inanimate), which in the context of the microscopic emergence of
these phenonema is (I feel) inappropriate.
Thus I reject these and other absolutist distinctions (closed vs. open
environments, closed to efficience cause/not closed to efficient cause,
complex/simple, model formally predicts/fails to predict etc. etc.).
Their application needs to be pragmatic to be useful!