Re: Maximally constrained

[Judith Rosen <***>]

Howard, it seems to me that you keep changing the point you say you're
making.

In your first post, you said that the reference of my father's was about
ORGANISMS as "having the maximum number of non-holonomic constraints".
Instead, what he was talking about in that reference was the only kind of
machine that could mimic behaviors of seeds described by somebody named
Morowitz, whose comment is not included in this excerpt.

Secondly, you are saying that you believe living systems are minimally
constrained, and he said his discussions with you put a maximally
constrained system as one way to describe a MACHINE-- not a living system.
Why are you arguing with that? It seems to me that he is not disagreeing on
that point.

Thirdly, The quote of my father's as given: >(RR) "The results of our
analysis above suggest that, whatever else may be true of relatively simple
but nevertheless highly evolved contemporary cells, they must at least
behave physically like maximally (nonholonomically) constrained,
programmable mechanical systems." has nothing whatsoever to do with your
corollary: (HP) "the physical principle I proposed that cells behave like
minimally constrained systems." Here's the reasoning:

He is saying that the simplest living systems must AT LEAST behave PYSICALLY
(appear to an observer)  LIKE (as if they were) maximally nonholonomically
constrained, programmable mechanical systems. Putting it another way; Even
the most sophistocated programmable, informationally and contextually
constrained machine will only be able to mimic the simplest organism. This
statement is on a completely different subject from your proposed physical
principle.

Fourth: There's a difference between "context-dependent constraints" and
"constraints". It's the difference between nonholonomic and holonomic
constraints. Context-dependent, or nonholonomic, constraints actually would
allow degrees of freedom in an organism that would otherwise not exist in an
unconstrained system. So, your point: " the physical principle I proposed
that cells behave like minimally constrained systems." is one I suspect my
father actually WOULD disagree with. I thought that was your original point
and I have already made my arguments giving the reasons why that logic
doesn't hold. If he was disagreeing with you, it wasn't because he
misinterpreted you, it was because he thought you were wrong.

Minimally constrained systems would exhibit very little autonomous activity;
they certainly would not be "context-dependent" which all living organisms
are. How are living systems context-dependent? Because the context (both
internal and external) is in constant interaction with (constraining) the
system-- it does so in direct ways but also in INDIRECT ways, via the
aspects of the context encoded in the internal predictive models that are
part of the organization itself of all living systems.

Judith

Judith


----- Original Message -----
From: "Howard Pattee" <***>
To: <***>
Sent: Sunday, October 31, 2004 10:38 PM
Subject: Re: [ROSEN] Maximally constrained


> Tim and Judith,
>
> Both your responses go way beyond the issue I was addressing. I have no
disagreement with most of what you both say. To return to the issue I
raised: My comment was that Rosen, although referring to discussions with
me, did not understand my principle point (at least in this paper). I still
don't see any interpretation of Rosen's 1986 paper that would suggest the
principle I have proposed, i.e., that life harnesses natural laws as much as
possible, and uses the minimum number of informational constraints. He
appears to express the opposite. Rosen says, "The point we wish to make is
that such a maximally constrained system is the only kind of mechanical
system for which Morowitz's comments [about seeds] holds. It is also,
according to the many discussions of Howard Pattee, one way of defining the
concept of a machine as a material system."
>
> Rosen says further (p. 122):
> "The results of our analysis above suggest that, whatever else may be true
of relatively simple but nevertheless highly evolved contemporary cells,
they must at least behave physically like maximally (nonholonomically)
constrained, programmable mechanical systems." Neither of you has managed to
convert this statement (Tim's half-full, half-empty logic notwithstanding)
to the physical principle I proposed that cells behave like minimally
constrained systems.
>
> Rosen's example of what he means by a maximally constrained is the special
purpose computer; that is, a machine that computes one function, like a
particular Turing machine. Even a programmable computer is maximally
constrained because its sequence of states has maximal dependence on the
program constraints and minimal dependence on laws. Laws depend on energy,
time, and rates that do not play essential roles in computing a function.
Also, no new parts are constructed. Rosen (in this paper) does not
distinguish life in these terms but by the inequivalence of programs
(simulations) and causal structures.
>
> In my view computers are nothing like life. Unlike computers, all the
functions of life depend on energy, time, and rates. Folding and
self-assembly processes are minimally constrained energy minimizations or
relaxation processes. Metabolism and replication depend on specific rate
control. Enzymes are the universal rate-controllers. It is a simple fact
that the genetic sequences constrain only a tiny fraction of the enormous
number of variables (degrees of freedom) of biopolymers. New parts are
continually constructed.
>
> As I said, this is a good paper. It is technical enough to be tested, and
it is full of stimulating ideas. I agree with much of it, but that does not
mean it can be defended as the final word. For example, Rosen's larger
crucial point, namely, "that most system descriptions do not contain
sufficient 'information' to allow realization at all" is certainly true.
That is to say, most sequences do not construct useful enzymes. However,
Peter Schuster's models have shown that because of symbolic and physical
redundancy far more sequences than we ever imagined can construct the same
function. This makes effective evolutionary search much more probable, and
destroys a classical criticism of Darwinism.
>
> Enormous progress has also been made in enzyme design. Rosen's statement:
"Thus, if I am given the primary sequence of a protein, I can in general
deduce nothing about its functional or kinetic properties," would not be
accepted today. Whole industries are doing just that.
>
> Howard
>