Re: "Algorithms" in Molecular Biology?

Guy Hoelzer wrote:
Hi John,

in article e4srsq$i6j$1@xxxxxxxxxxxxxxxxxxx, John Wilkins at
john@xxxxxxxxxxxxx wrote on 5/22/06 10:20 AM:

dougwedel@xxxxxxxxxxxxx wrote:
"Glen M. Sizemore" <gmsizemore2@xxxxxxxxx> wrote

The unusually-brilliant moderator of this group will probably censor my
post, but I have a comment: asking whether molecules "use algorithms" is
the
same thing as asking "does the Moon calculate its orbit?" It is
troublesome
that the latter (as well as the former) are not more widely recognized as
stupid.
I tried to hint at the scientific respectability of my inquiry by citing
Sydney Brenner, but Brenner is not the only Nobel Prize winning scientist
who has expressed the view that the future of science, and in particular
biological science, lies with understanding nature's ALGORITHMS. Manfred
Eigen said that we must look for the algorithms that generate biological
information. If you look at the deepest thoughts on these subjects of
mathematicians and physicists like Stanislaw Ulam, Alan Turing, John von
Neumann, and Erwin Schrodinger, you will find that they were all looking at
life this way, and indeed, the idea goes back to Liebnitz and Descartes. So
"stupid" seems...a bit unkind, frankly, and not indicative of broad
scientific learning.


Even respected scientists, and in my opinion physicists more often than
others, have a problem distinguishing abstractions from realities, models from
the systems modelled, and equations from the dynamics they described.

Agreed.

Biological processes do not do calculations (unless they are biological
processes in a suitably complex nervous system, ranging from mathematicians to
corvids. I will not say which I think the more complex). We do not calculate
the trajectory of an object we throw. We do not work out the stress equations
in our skeletal growth. Phyllotactic plants do not work out the Fibonacci
sequence. These equations, and the underlying algorithms, are *abstractions*
that we have developed in order to model and predict behaviours. It happens
that the relevant causal factors can be mapped using these successful
equations. But we have to be careful to tell the difference between the
abstract idea of things and the things themselves.

I don't know of anybody who has argued against the examples you give here.
Indeed, your comments suggest to me that you do not understand the essence
of the "computational universe" argument. The claim is that the dynamics of
the universe is one, big, unfolding computation manifested by the
interactions among its components. It is not the thrower who calculates the
trajectory of the ball they are about to throw. It is a system of
interacting agents in which the thrower is only one of them. In fact, I
would argue that all biological processes are manifested calculations; not
that they do calculations. This perspective remains rather philosophical as
I don't see how it can be empirically or logically tested at this time.


There are versions of this. One is called, boringly, the Matrix Ontology. We
can ignore that. But suppose we do treat the universe not in the way John
Wheeler ("it from bit") wants to do but the way you present here. So analogue
processes can be seen as computations of arbitrarily high precision of...
what? Themselves?

There are no algorithms in play when physical processes play out. I don't
exemplify the John Wilkins algorithm. I just am John Wilkins. That's why it
makes no sense to think of me being a bat, or Guy Hoelzer - these aren't
algorithms the hardware can "run", they are just states of the physical world
that I happen not to be.

Nothing is gained by calling ordinary causality "information". So there are
mathematical equations (notably, thermodynamic ones) that can be used to
explain the ways those states occur, so what? We hope this is true when we
apply formal explanations to physical systems. But the explanation isn't the
system.

The causal processes involved in throwing a ball or catching it (much harder,
computationally, which means my dog is a better mathematician than I am; which
is true, and I have no dog) are best described by some set of equations. But
when I catch a ball, I don't *do* those calculations. I have a set of
responses learned by trial and error over years, I have observations to anchor
those responses in this case, and I have success some of the time (so we
ignore the "bad calculations" I did when I *missed*). In saying "the processes
are manifested calculations", all you are saying is that we can describe the
processes with calculations. Which is what I said to start with.

To go back on-topic, I am reminded here of Michael Ghiselin's perspicuous comment:

"We must not take the term 'strategy' too literally. The only real strategists
here are evolutionary theorists. Carrots do not ponder differential
reproductive success. - Organisms play the game because, and only because,
their ancestors did not lose." (1974, The economy of nature and the evolution
of sex. Berkeley: University of California Press, p22)

Likewise, DNA is not an information bearer in the usual sense.

Indeed. :-)

Guy




--
John S. Wilkins, Postdoctoral Research Fellow, Biohumanities Project
University of Queensland - Blog: evolvethought.blogspot.com
"He used... sarcasm. He knew all the tricks, dramatic irony, metaphor, bathos,
puns, parody, litotes and... satire. He was vicious."

.

Quantcast