up a level
post article
search
admin
about
rdf
Karl's Fiction Writing Site
|
Posted by Karl on Monday October 25, @09:47AM
from the dept.
As reported in this month's Wired magazine, the old spectre of so-called "intelligent design" is rearing its head again. This blatant case of Lysenkoist pseudo-science has as one of its advocates George Gilder, who is given space to provide his justification of the idea in Wired--very civilized of the magazine.
It's interesting when Gilder says, up front, that "the concept precedes the concrete". He's identifying himself as a philosophical Idealist; he also identifies proponents of evolutionary theory as Materialists. He claims to be basing his views on "the largely mathematical facts of 21st-century science." Readers of this blog should know by now that it is precisely the old 17th century ideas of materialism and idealism that are being superceded today. Gilder's argument does not come from the Age of Information; it comes from the Age of Steam.
So "the concept precedes the concrete," eh? As always, this notion conveniently lets us off the hook for having to actually investigate how real-world things work. If the concept precedes the concrete, then the concrete need not participate in its own explanation. We can turn to dialectic, argument, or the Word of God to know the physical world; at best, the physical processes themselves are obstructionary, full of inexplicable behaviors, noise and contingency. Better to forget about the physical, and concentrate on the Ideal.
*Yawn.* The 17th century view of the world that led us to separate the material from the conceptual is gone, folks. Gilder is playing on a false dichotomy so his argument doesn't even get off the ground. He's trying to invoke the notion of the universe as dumb matter infused with a motivating Spirit. If you want to see the union of matter and concept, though, just run any one of Stephen Wolfram's cellular automata--and then consider that the developing embryo of a human being works the same way. As Wolfram has shown in exhaustive detail, mind-boggling complexity can be produced by fundamentally simple processes. The basic claim of the "intelligent design" folks that natural selection alone can't explain the complexity of organic life might be trivially true--but natural selection operating in a universe where Wolfram's cellular automata are possible does explain it. And, a curious thing about Wolfram's automata is that as physical processes they are in a sense stateless--they are their state, in other words they store information about themselves and their environment but not in any abstraction separable from themselves. They are simultaneously information systems and physical systems and the one can't be teased apart from the other. If it could, it would be possible to predict the behavior of all automata but there are automata that have to be run in order to determine their behavior--their result cannot be abstracted from their operation even in principle. Ergo, if Gilder's identifying information with the Concept (as he pretty clearly is) then in many systems in Nature, the concept does not precede the concrete.
The defense of "intelligent design" is in any case a great example of the relevance of this blog. The ideas I'm talking about are foundational and political when they become clothed in such guise as school curricula. Examine Gilder's position and you see the old system that I'm claiming we've moved past. And if you're wondering about the importance of these ideas, consider the political context of that position.
<
|
>
|
|
The Fine Print: The following comments
are owned by whoever posted them.
( Reply )
|
Re: Unintelligent Design
by R von Neumann on Wednesday December 01, @04:09PM
|
You said:
"They are simultaneously information systems and physical systems and the one can't be teased apart from the other. If it could, it would be possible to predict the behavior of all automata but there are automata that have to be run in order to determine their behavior--their result cannot be abstracted from their operation even in principle."
I did not read Gilder. I only read the piece of yours from which this quote was taken. I am too impatient to just make this point and perhaps should have read what you were responding to. …but I see something here that I would like to consider with you.
With regard to cellular automata (CA), I wonder how large a universe is necessary such that the ‘natural selection’ will kick-in on its own without the programmer’s ‘artificial selection’ being a piece of the formula. While CA produces some interesting machines, how would we mechanize a means that would save the ‘interesting’ ones. …and in so doing, would not this automated selection process have to be designed by we the designers? How do we synthesize neo-darwinistic methods – methods that are clean and without our prejudice. It appears to me that the ‘mutation’ part (random seeding) of the CA universe, may be trivial, but the selection part, the subtractive part, cannot be designed, it must somehow just ‘be’, in the world-view of a neo-darwinist.
How would our synthetic universe model ‘selection’? In the natural world, how does natural selection ‘bootstrap’ itself. Natural selection is, by definition, non-teleological. Looking back at your quote above, there are behaviors that cannot be predicted, not even in principle. So the universe ‘computes’ over time, the selection criteria and that criteria is always changing. We do not know what it will be until it finishes computing (so to speak).
Critique of evolution has, it seems to me, mostly dwelled on the mutation side. Example: take a strand of DNA, so many base-pairs long, and then turn the whole computing power of the universe into a parallel processor that does nothing but try all commutations of base pairs, to randomly produce this exact strand. For a strand that is fairly short, the result is ‘not enough time’, and that is a hyper-understatement.
I have seen where many still hand-wave around this problem – maybe that’s what you do. But I think I remember Stewart Kauffman stating in his Origins of Order that the mutation side is a problem. That’s why he tried to synthesize a model using templates.
But what about the selection side of the equation? Hasn’t CA always had human selectors? With no human observer, or human designed observer, would not an occasional jewel that instantiated, get boiled away? How does a universe with no teleology create the selection landscape? Try bootstrapping it from nothing – instead of presuming it already exists. How does it start and how does it advance?
This past century with the Goedelian and self-referential theorems, some camps would say, that Goedel’s Theorem speaks on this. The problem I am trying to point out is that the state of the system (CA or natural) might instantiate an occasional new thing, a thing that potentially could be useful in a new level system that does not yet exist. If the selector is the landscape of objects that currently exist, how do they (how does it) select an object of the next order set?
What about this gedanken experiment:
We humans start receiving a 40GHz data stream from ‘out there’. We figure out how to read the data and it turns out that the data represents a book on how to build a transporter, or some other alien artifact that is far beyond our current state of science. As soon as the book is sent to us, another one is sent. Then another. After five years we have received trillions of books and we have determined that the aliens have made a game out of this. They tell us that only one book will actually build a good working transporter. All the other books are clever fakes. They are fakes that we cannot determine by principle, that they are. Without building the machine from the instruction book, that they are indeed fakes, is unknown.
How can we ‘select’ the proper book. This alien information uses construction techniques and materials that we already possess. Without knowing the working principles, how can we do it? We would have to build each one and then we cannot remove the teleology from this scenario. We have been given a statement of purpose from the aliens – there is a ‘transporter’ design in this huge stream of data. So we will build and test, build and test. The test criterion has purpose.
This scenario seems to me to illustrate the problem of non-teleological-selection. That there is a better version of “Hamlet” buried in ASCII-code, somewhere deep in the digits of pi, is a statement of faith until we find it. I do not have a problem accepting it as fact. Again, the mutation side is trivial – potential information can be easily produced. How do you find it? How do you select for it without assigning some sort of purpose to the selection process?
If we say that natural selection is not intelligent, and is non-teleological, that statement is merely axiomatic. But if natural selection is not intelligent, then what IS intelligence. What IS purpose? What is the creative power we humans hold in our minds? We do not know. We do not know what intelligence is, but so many will argue that it only exists in the mind of humans, and maybe some other yet to be discovered and proven places (aliens, chimps, etc).
If we do not know what intelligence is, other than that is it exists and we humans possess it, then how can we say it has not played a part in the instantiation of ‘what is’ in our physical world? You say that Gilder speaks from the ‘steam age’ but I think you speak from the ‘cyber-age’ – from the 1950-1960 era. There seems to me to be a post-materialism aspect of science, trying to emerge now. Behind the quantum curtain, there is a bunch of stuff we still do not know anything about. Maybe universal intelligence and teleology lies behind that curtain.
R. von Neumann
|
[
Reply to this ] |
Re: Unintelligent Design
by Karl on Thursday February 24, @01:57PM
|
|
Well, Stephen J. Gould talks about the problem you describe in his book "Full House." There's no need for teleology in natural selection because there is no direction to it. What appears to be direction is the result of selection operating in a situation of having a "wall" on one side but not the other: organisms have a minimum possible complexity but not (so far as we've seen) a maximum. Random fluctuations in complexity in such a system--what statisticians call a "drunkard's walk" result in the complexity meandering slowly away from the wall, which translates into complexity increasing with time. But it's not increasing because of any built-in teleology. It's purely a result of a random walk with one limiting wall.
There is no direction to evolution. So questions of purpose or teleology don't apply.
As to cellular automata, they require no programming nor any "external environment" in which to develop. They are entirely creatures of their own initial conditions--unless they are start to interact with other entities which could also be CAs. Within and between themselves they exchange data and in fact can be complete Turing machines--computers. So even the simplest CA is capable of calculation, memory functions, simulation and information-gathering. All without an external programmer.
Natural selection really does account successfully for the appearance of complexity, including complex organs, organisms etc. It also accounts for the creation of 'purpose' from non-purpose.
|
[
Reply to this ]
|
|
The Fine Print: The following
comments are owned by whoever posted them.
( Reply )
|
|