Fog Creek Software
Discussion Board




Complexity

What separates Darwinism from any other theory of evolution is:
The variations which are the source of novelty were not generated for any purpose, and the mechanism for selecting among the variations also has no purpose. In other words, there is an absence of motivation or drive. Individual organisms have goals -- survival, pleasure, avoiding pain, etc. -- but species and ecosystems do not have goals.
In cultural evolution the existence of goals beyond the individual are acknowledged. In Darwinist evolution, which is the currently prevailing evolution theory, goals beyond the individual are not recognized.

According to systems theory (an earlier form of chaos or complexity theory) natural systems
evolve toward hierarchical organization and complexity. This is a built-in drive of natural systems. So complexity theorists should not be denying that species or ecosystems have goals.

This is an attempt to answer to some questions way down there in the Alyosha thread. (Can we consider this post #106?)

I would also like to suggest that talking about genetics is not off-topic on a software forum, since DNA can be considered a program, and since complexity is related to computer science.

The Real PC
Sunday, October 26, 2003

I use genetic algorithms to 'breed' computer programs; just for fun, of course.

I have a set goal for my population of programs to evolve towards.

However, the more often I run the simulation, the deeper I feel that the system as a whole has no designed-in direction.  I have to abort about 60% of my runs because they are not going in the direction I want, and thus wasting my cycles.

I look hard at my fitness function, at my parameters and everything.

i like i
Sunday, October 26, 2003

Wouldn't a species' goal under Darwinism simply be "to survive" - it's those who survive who can contribute their Selifsh Genes to the pool.

It's not that the species wants to become more complex, it's that by becoming more complex it's better able to survive. The complexity is a byproduct of survival.

=b~
Sunday, October 26, 2003

Complexity doesn't seem to aid survival. The simplest species are some of the best survivors. Before technology, humans had a low population. Gorillas, a highly intelligent species, never experience a population explosion, never over-ran other species. Sharks are great at surviving. Dolphins, another highly intelligent species with complex social systems, don't seem to be better at surviving than sharks which are far simpler.
And one-celled organisms are still doing fine after billions of years of evolution.

The Real PC
Sunday, October 26, 2003

This thread was totally offtopic about 50 posts ago.  Let it die already.

Matt Latourette
Sunday, October 26, 2003

If the this thread is fit enough then it will survive.

John Ridout
Sunday, October 26, 2003

Your last comment is a circular argument. And that objection is one of the more serious ones to be raised against Darwinism :)

Noevertheless, the Real PC's postings will spawn, despite their lack of any utility to the forum, exactly as junk DNA does.

INcidentally to talk about goals is anthropomorphism. And as I said in the other post, the species is a by-prodiuct of natural selection. The unit is the gene, and the vessel is the individual.

Why you get more complexity over time is simplest explained by Hawkins. And of course added complexity is a local phenomenum, in the context of the slide towards entropy.

Stephen Jones
Sunday, October 26, 2003

Darwinism is reductionist and bottom-up, systems theory is holistic and top-down.
The problems with reductionism are hard to understand and it took me a while to get it. Reading Koestler's Beyond Reductionism, many years ago, was difficult but very worthwhile. I also recommend Bateon's Steps to an Ecology of Mind, for understanding the complex information systems of nature.
I will try to explain it very briefly. I'm sure I'll leave something out, since there isn't time to explain every detail without leaving a single gap.

A standard statement of systems theory is "the whole is greater than the sum of its parts." For example, the information required to describe a molecule is not the sum of the information required to describe all of the atoms that make it up. The molecule exists at a higher level in the hierarchy. The relationships between the atoms is not specified within any of the atoms. Similarly, the relationships between the electrons, protons and neutrons within each of the atoms is not to be found within these particles, but in the atom they belong to.
Natural systems are, obviously, organized in hierarchies, similar to class hierarchies in Java (or some other o-o language). As you know, the relationships between your classes exist at a higher level of organization than the code within the classes. And the relationships between the methods in a particular class exist at a higher level than the code within any of the methods.
Darwinism, and modern genetics, are reductionist. There is a lack of understanding or concern with the hierarchical organization of systems, or with how the information describing this organization is encoded. For example, the behavior of genes is controlled by higher level codes that determine which genes will be switched on. Every cell of an organism has the same DNA, whether it's a nerve cell, muscle cell, etc. Cells with the same DNA have very different functions and structures. Reductionist theories fail to explain where or how this higher level information is encoded.
Sheldrake's book A New Science of Life gives extremely thoughtful and logical explanations of the failure of reductionist biology to account for the morphogenesis of embryos. Different genes are switched on or off in different cells according to some meta-program, which is not part of the DNA itself.

Holistic scientists have been addressing the failures of reductionist biology. But some people have an awfully hard time seeing that there are any problems with reductionism. They think that if you want to understand a brick building, all you need is a complete description of all the bricks that make it up. In reality, an understanding of the relationships between the bricks is just as important.

Artificial life (cellular automata) programs supposedly demonstrate the bottom-up emergence of simulated natural systems based on simple rules, and CA is the basis of digital physics. However these artificial systems obviously do have designers.

Connectionist neural networks were also supposed to demonstrate the bottom-up organization of cognition. Even with designers, these systems have not resulted in real cognition.

The organization of complex natural systems must be at least partly top-down.

The Real PC
Sunday, October 26, 2003

This thread is no longer a fit thread.

John Ridout
Monday, October 27, 2003

Cellular automata obviously have designers?  How exactly do you mean that?

K
Monday, October 27, 2003

You know, the Nazis had pretty strong opinions about evolution and darwinism, too.

Godwin
Monday, October 27, 2003

"Noevertheless, the Real PC's postings will spawn, despite their lack of any utility to the forum, exactly as junk DNA does."

Do forums have goals, or only individuals posting in forums?

Jim Rankin
Monday, October 27, 2003

The thread is now over!

Bobo
Monday, October 27, 2003

Jim: your suggestion that the forum is lacking a teleology is a by-product of your outdated "reductionist" worldview.  If you were to see the JOS Forum *properly* from a holistic standpoint you would see that these forums have achieved a sentience beyond the combined intelligences of the contributers, that the sum is now greater than its parts; and hence, it is proper to say that the Forum can have a goal which it actively pursues.

Alyosha`
Monday, October 27, 2003

"natural systems
evolve toward hierarchical organization and complexity."
WRONG. Natural systems do not evolve toward complexity. That assumption is wrong. Natural systems favor diversity.
Many different animals is the best thing when the meteor hits. It means that some will have a way to survive. Monoculture is bad. (We can debate why in an M$ vs Linux thread.)
Because there is a lower limit on what we call life (virus, bacteria), the only way to be diverse is to become more complex.
If complexity were favored, wouldn't there be more people than bacteria? We are just a blip on the radar, the microscopic stuff has been and always will be the main event on this little planet.

Doug Withau
Monday, October 27, 2003

Alyosha`, your attempt at some humorous sarcasm reveals your own reductionist mindset when you say the "forums have achieved".  Complexity theory is an attempt to understand how complex behaviour can emerge from the interaction of simple interacting agents, and thus make existing life forms a plausible result of Darwinism.  OTOH, "system theory" reveals that the "organization ... must be at least partly top-down."  Saying the "forums ... achieve" leaves out this essential top-down imposition of the organization.  How this is accomplished, or even what it means, is left as an exercise for the reader.

Godwin, you, of all people, should know that intentionally using Godwin's Law to stop a thread is doomed to failure, along with Darwinism, complexity theory and reductionism.  Perhaps noting that Lamarckism was promoted by the Communists will do it.


"The thread is now over! "  Not likely!  But for me it is.  This isn't going anywhere and just isn't interesting anymore.

Z
Monday, October 27, 2003

Been away, but I see this discussion hasn't done much mean time.

"Natural systems favor diversity. Many different animals is the best thing when the meteor hits."

Yeah right. One bacterial species looks at their neighbour, and thinks: "Uh-oh. We're really much too similar. This ecosystem won't stand a chance when the meteor hits. Quick, we'd better evolve into something different..."

If you're going to spread utter nonsense about evolutionary processes, a plausible mechanism is least you could provide.

Real PC's comments about complexity are almost as misguided, but I'm not getting into that again - I said my bit. If you want a sensible overview of the perils of reductionism that's a bit more in touch with reality, try Ian Stewart and Jack Cohen (no, not The Science of Discworld, some of their other books). A good antidote to Dawkins' arch-reductionism.

Dave Hallett
Tuesday, October 28, 2003

Dave,
Well at least you can see that reductionism has some problems. Therefore you must have some understanding of why I'm skeptical.
If reductionism has problems, the answer is not to stubbornly repeat, like the pseudo-skeptics, that science has always been this way, look at all the progress that's been made, we don't want to re-write the laws of nature, we don't want to throw away all that progress, etc.
So many people miss the point. The goal is not to throw away science and start over. The goal is to acknowledge certain problems and limitations and look for ways to overcome them and go forward.
Yes maybe science can continue to progress while hanging on to reductionism. But in certain areas progress is blocked and will remain blocked -- unless mainstream science starts to bend and let the non-reductionist scientists at least get their opinions heard and their evidence seen.

I honestly feel that some people are addicted to authority and certainty. Questioning and doubting what it says in the textbooks is just not part of their nature.
I can certainly appreciate skepticism and I realize the world is full of deception,  ignorance and credulity. I realize that not every dowser or fortune-teller is genuine. Balancing doubt and belief is difficult and requires thinking, which is hard work. It's easier to believe every weird thing you hear, or else to disbelieve anything that was not in your eighth grade science books.

The Real PC
Tuesday, October 28, 2003

Real PC -

You clearly have some disdain for folks who disbelieve your particular set of evolutionary hypotheses, yet you return to this topic with a kind of religious zeal.
Has it not once occurred to you that other people may have considered these ideas and then rejected them for any of the reasons expounded upon in the many, many responses to your various threads?

Devil's Advocate
Tuesday, October 28, 2003

The reasons are always appeals to authority and convention. They never give logical reasons or evidence.

The Real PC
Tuesday, October 28, 2003

*  Recent Topics

*  Fog Creek Home