Uncommon Descent Serving The Intelligent Design Community
evo

The hole of the SLoT

Share
Facebook
Twitter
LinkedIn
Flipboard
Print
Email

Definition of the 2nd law of thermodynamics (SLoT).

This law (in its statistical mechanics sense) states that an isolated system goes towards its more probable states (those more numerous). Since the disordered states are countless, while the ordered/organized ones are few, a closed system spontaneously goes towards disorder/disorganization (related to entropy).

Difference between order and organization.

Increase of order implies decrease of entropy. Examples of order in nature are crystals; soap bubbles and raindrops are examples of naturally ordered quasi-spheres. Examples of order in human artefacts are the pattern of wood in a fence and the configuration of seats in a cinema.
Organization also implies persistent decrease of entropy, but is far more and far higher than order. Organization is qualitatively different from order. Organization always involves functional hierarchies and complex specified information (CSI). Examples of organization in nature are cells and organisms. Examples of organization in human technology are engines and computers.

A key point: the relation between organization and entropy is non symmetrical. (Intuitive example of non symmetrical relation: rain implies decrease of dryness, but decrease of dryness does not imply rain – decrease of dryness may well have other causes.) While organization implies persistent decrease of entropy, a decrease of entropy alone does not imply organization. Put differently: while entropy destroys organization, its opposite – neghentropy – doesn’t create organization. While it is true that what decreases order destroys also organization, it is false that what increases order creates also organization.

By increasing order we don’t get organization, like by increasing numbers we don’t get elephants or spaceships, like by increasing a rectangle we don’t get a circle or a fractal. Organization is not at all the limit which order tends to. Between increasing order and organization there is a deep discontinuity, a “hole”.

Graphical representation of the 2nd law.

See this picture:

close

Where the organized state (red zone) is one, the ordered states (yellow zone) are some and the disordered states (green zone) are countless. Since the disordered states are far more numerous than the other states they are more probable (leading to the continue tendency for disorder stated by the 2nd law). In the picture the 2nd law tendency is symbolized by the gravity force applied to the red ball. The red ball always tends towards the bottom, towards the disordered states. The discontinuity between organization and order – the “hole” – is represented by the tunnel between the red zone and the yellow zone. The ball never reaches the red zone of organization because, also if it climbs the mountain, it falls in the hole and crosses the tunnel.

Biological unguided evolution.

Evolution supposes that all the biological organization on Earth arose spontaneously (naturalistic origin of life + naturalistic origin of species).

Corollary of the 2nd law.

In an isolated system, organization never increases spontaneously. Hence the 2nd law refutes evolution. The absurdity of evolution is illustrated in the following picture:

evo

Evolution would involve countless scenarios where the red balls stay permanently on the top of the peaks. Consequently the 2nd law disproves evolution because evolution would represent a set of events practically impossible.

Evolutionist “compensation argument”.

To rebut the above corollary, usually evolutionists resort to this argument. Since the Earth is not isolated, the 2nd law does not forbid a local (on Earth) decreases in entropy (which is all biological organisms represent, and no more than evolution is posited to do), gained at the cost of increased entropy in the surroundings (the solar system) (or, as long as the system exports a sufficient amount of entropy to its surroundings). So evolution can happen on Earth.

Refutation of the “compensation argument”.

The main counter-point is that, no, decrease in entropy is not “all biological organisms represent”. Organisms eminently represent organization. They are even ultra-complex systems. As said above, simple decrease in entropy is not organization. Evolutionists use “entropy” as a “free lunch” for evolution: entropy increases there, so entropy decrease here and organisms arise here at zero cost, while the 2nd law is safe. Too good to be true. Since entropy is related to disorder, then I cause a big mess (easy task) there to get organization (difficult task) here? Do you see the nonsense?

Second, call A the open system and B its surroundings. “Increased entropy in the surroundings” means that B has increased its disorder, going towards a more disordered state. This additional disorder in B becomes (in the mind of evolutionists) sort of “money” to pay the organization in A. Just this concept appears paradoxical: to pay organization by means of disorder. It is like to say: a disease in my wife 🙁 increases my health :).

Third, the reasoning is also absurd when we speak of probability. “Increased entropy in the surroundings” means that in B happened events more probable than the events happened before. These more probable events become (in the mind of evolutionists) sort of “magic” that creates organization in A. In turn, this organization in A is events with low probability that happen. So the whole reasoning is: probable events happened in B cause improbable events in A. It is like to say: the shopping expenses of my wife 🙁 cause my winning the lottery :).

In short, the evolutionist “compensation argument” is something like “non-X causes X”. It helps exactly zero the case for evolution, and doesn’t save evolution against the 2nd law.

The bottom line is: improbable events related to organization in a system remain improbable independently from the fact that we consider the system closed or open. Unless evolutionists are able to prove that some external cause is really able to reduce somehow such improbabilities, by injecting CSI to create organization. So far evolutionists have not succeeded in such task, their “compensation argument” is laughable. While IDers have a name for an organizational cause: intelligence.

Comments
F/N: Mung is of course referring to Maxwell's Demon who, by knowing he state of molecules, can allow faster ones to go through to one side and/or slower ones to the other, of a container, through a barrier with a trap door. Thus, he creates a disequilibration and undoing of diffusion based on intelligent and purposeful action. Szilard's point is that he does not come by that info on the cheap if he is material. This is of course the first step to the long chain of points that shows that constructive work resulting in FSCO/I -- complex, functional organisation according to a plan -- is most credibly explained on information and intelligence and is not a plausible product of spontaneous processes like diffusion. Where also, entropy can be analysed as the average missing info to specify microstate on observing macrostate, which if high forces us to treat the entity as highly disordered and/or random, e.g. we try to harvest work based on heat. EL et al by fixating on reduction in such uncertainty on heat transfer out of an entity [as opposed to the Darwin's warm little pond relevant rise in same on importing energy without coupling it to a conversion mechanism performing shaft work issuing in constructive work . . . cf. here on as has been repeatedly linked and cited . . . ], distract attention (especially their own) from the problem of the statistics involved in getting to FSCO/I by diffusion and the like. Those statistics make it effectively unobservable on the gamut of solar system or observed cosmos, to see that happening, just as for essentially similar reasons, we will reliably not see by chance 500 coins, tossed, and showing all H's. Empirical impossibility by overwhelming statistics is just as effective as logical impossibility, in the real world. KFkairosfocus
July 21, 2013
July
07
Jul
21
21
2013
06:47 PM
6
06
47
PM
PDT
If course there is a hole in the slot. Else the demon could not allow things to pass through.Mung
July 21, 2013
July
07
Jul
21
21
2013
05:36 PM
5
05
36
PM
PDT
LoL! You don't have anything to argue, Elizabeth. And you are easily refuted. So of course you have given up arguing with me.Joe
July 11, 2013
July
07
Jul
11
11
2013
07:52 AM
7
07
52
AM
PDT
No, just given up arguing with you, Joe.Elizabeth B Liddle
July 11, 2013
July
07
Jul
11
11
2013
07:40 AM
7
07
40
AM
PDT
Is Elizabeth abandoning her diatribe wrt EAs? Or does she still think, albeit contrary to the evidence, that they model darwinian evolution?Joe
July 11, 2013
July
07
Jul
11
11
2013
06:59 AM
6
06
59
AM
PDT
Elizabeth B Liddle
The argument that the 2nd Law is a bar to evolution is simply wrong. If it were, we wouldn’t have weather.
Strange woman. I would have thought that you said "if it were, we wouldn’t have evolution". Weather, meteorology is based on a very difficult field of physics: fluid dynamics. A field that is even hard to mathematically model and hard to simulate on computer. Its laws explain all weather phenomena, tornadoes included (which you so frequently speak of). I don't see as the SLoT-bars-evolution argument could bar also weather. The SLoT bars spontaneous evolution of organisms. It doesn't bar spontaneous evolution of clouds, rain, snow, hurricanes... The latter have nothing to do with the former. The Earth's weather, in spite of its huge complexity, is not an organized system. In fact we should carefully distinguish between the mere quantitative complexity of chaos and the real qualitative complexity of organization. The error of evolutionism is indeed to pretend that the latter can arise from the former gratis, without the need of intelligence.niwrad
July 10, 2013
July
07
Jul
10
10
2013
10:54 PM
10
10
54
PM
PDT
Ben W:
“Eliminate the weak and deficient” is a selection criteria, and so long as there’s any population pressure, it will result in “good” mutations being favored.
Well "good" is relative. More like the non-detrimental will be favored. And taht isn't a creative force.
All that you need for an EA is a selection criteria, a genetic mixing algorithm (preferable, but not strictly necessary), and some starting bits of code/DNA. How is that different from nature?
1- Nature doesn't seem to be capable of producing reproducers 2- Nature doesn't seem to be capable of producing neither DNA nor a code 3- Nature doesn't do selection criteria- eliminative criteria, yes 4- Nature doesn't have a goal- EAs do. Why do you keep ignoring that?Joe
July 10, 2013
July
07
Jul
10
10
2013
06:46 PM
6
06
46
PM
PDT
No, there is a selection coefficient actively selecting which will (continue to) reproduce and which will not.
No. The selection coefficient changes all the time.
Reference please. I know in Dawkins' weasel it doesn't change. I know in the antenna EA it doesn't change. I know in all the EAs presented in SciAm's "Evolving Inventions" it doesn't change. So please tell us how we can solve a problem that keeps changing? It selects those which are deemed better suited to solve the problem.
No. Those that solve the presented problem better (not “are better suited to solve the problem”)...
LoL! Liz, the problem isn't solved with just one iteration. The antenna EA didn't design the antenna after one interation. The evolving inventions didn't come about by only one iteration. Dawkins' weasel confirms what I said.
.. have a greater chance of surviving to reproduce, just as real life creatures who solve the problems presented by their environment do.
Real life creatures who wait for a just-so accidental mutation are in for trouble. And yes that means it actively eliminates those which it deems are not suited to solve the problem.
No, it doesn’t.
Yes, it does. Again see Dawkins' weasel.
Well, some may, I guess, but not the ones I do.
Whatever, you don't model natural selection. And perhaps you don't even write/ use EAs. Heck you thought AVIDA was an EA. Oh well. Also your idea of NS runs contrary to Mayr's. You lose in that debate.
The most effective ones make sure that the poorer ones also have a non-zero chance of reproducing. Just as in life.
In life we see plenty of cooperation. Those that would normally have a non-zero chance are often given more than one.Joe
July 10, 2013
July
07
Jul
10
10
2013
06:41 PM
6
06
41
PM
PDT
re mine at 117: The most effective ones EAs make sure that the poorer ones also have a non-zero chance of reproducing. Just as in life.Elizabeth B Liddle
July 10, 2013
July
07
Jul
10
10
2013
03:08 PM
3
03
08
PM
PDT
niwrad
This is a clear sign that, among IDers, the awareness of the power of this ID argument is increasing.
Not necessarily, niwrad :) The observation is sometimes made that if you want to boost your citation index, the best way is to publish something that is wrong. It will be copiously cited by everyone who subsequently gets it right. The argument that the 2nd Law is a bar to evolution is simply wrong. If it were, we wouldn't have weather. There may be bars to evolution, but I suggest that the number of responses to these threads means that this one is not one that helps your case!Elizabeth B Liddle
July 10, 2013
July
07
Jul
10
10
2013
03:07 PM
3
03
07
PM
PDT
Joe
No, there is a selection coefficient actively selecting which will (continue to) reproduce and which will not.
No. The selection coefficient changes all the time.
It selects those which are deemed better suited to solve the problem.
No. Those that solve the presented problem better (not "are better suited to solve the problem") have a greater chance of surviving to reproduce, just as real life creatures who solve the problems presented by their environment do.
And yes that means it actively eliminates those which it deems are not suited to solve the problem.
No, it doesn't. Well, some may, I guess, but not the ones I do. It simply gives those that solve the problem better a better chance of reproducing. The most effective ones make sure that the poorer ones also have a non-zero chance of reproducing. Just as in life.Elizabeth B Liddle
July 10, 2013
July
07
Jul
10
10
2013
02:49 PM
2
02
49
PM
PDT
There isn’t anything in nature that says natural selection can do anything beyond eliminate the weak and deficient. That means there isn’t any evidence that it is a designer mimic. And if you remove the goal and selection coefficient from the EA then nothing will happen.
"Eliminate the weak and deficient" is a selection criteria, and so long as there's any population pressure, it will result in "good" mutations being favored. All that you need for an EA is a selection criteria, a genetic mixing algorithm (preferable, but not strictly necessary), and some starting bits of code/DNA. How is that different from nature?Ben W
July 10, 2013
July
07
Jul
10
10
2013
02:00 PM
2
02
00
PM
PDT
If a program is designed to do something, and it does it, then it did so by design, not by natural selection.
But… that’s how all experiments work.
No, it isn't. Lenski's long running experiment alone refutes you. Then there are all the natural experiments we conduct during eclipses.
By this argument, an experiment designed to, say, test gravity, is not a good representation of the real world.
Just cuz you say so- NOT.
Yes, experiments are designed to reproduce real-world phenomena in a controlled environment.
Some are, but that is moot as it misses my point. There isn't anything in nature that says natural selection can do anything beyond eliminate the weak and deficient. That means there isn't any evidence that it is a designer mimic. And if you remove the goal and selection coefficient from the EA then nothing will happen.Joe
July 10, 2013
July
07
Jul
10
10
2013
01:39 PM
1
01
39
PM
PDT
Elizabeth:
We communicate by defining our terms with excruciating care.
LoL! That alone requires information of the type IDists are talking about. IOW you lose, again. Nice job.Joe
July 10, 2013
July
07
Jul
10
10
2013
01:33 PM
1
01
33
PM
PDT
Natural selection is eliminative. The only reward for the survivors is that they may get the chance to reproduce. That isn’t how EAs operate.
It is exactly how they operate.
No, it isn't. No, there is a selection coefficient actively selecting which will (continue to) reproduce and which will not. It selects those which are deemed better suited to solve the problem. And yes that means it actively eliminates those which it deems are not suited to solve the problem. and If a program is designed to do something, and it does it, then it did so by design, not by natural selection.Joe
July 10, 2013
July
07
Jul
10
10
2013
01:32 PM
1
01
32
PM
PDT
Ben W said: What we scientists mean when they talk about “information” is different from what Dembski means when he talks about it. Really? How the heck do you guys communicate? What are all those words in your peer-reviewed papers? What are all of those words in your science textbooks? Those words are not information to you and your alleged scientists?
What we mean by "information" is bits. So, no, letters on a piece of paper are not by themselves the same as the "information" that we use to communicate. By the physical definition of 'information', the random string of letters "ajwi" has the same amount of information as the string "blue", since both are four letters. Both contain the same amount of bits. That may seem non-intuitive, but it's how we define information in thermodynamics and physics. For the scientific definition, the semantic meaning of a string of letters or its 'functionality' is completely irrelevant. What matters is just the length of the string.
If a program is designed to do something, and it does it, then it did so by design, not by natural selection.
But... that's how all experiments work. By this argument, an experiment designed to, say, test gravity, is not a good representation of the real world. "You designed the experiment so that the ball would fall to the ground", you'd say. Yes, experiments are designed to reproduce real-world phenomena in a controlled environment. That doesn't mean they don't tell us anything about the real world; quite the opposite, in fact. We can take the same forces and systems that we see in Nature, and apply them in a lab or in a model system, and come to understand them better. And we can take something quite like natural selection, and put it in a computer program, and see what it does with strings of bits or abstract structures. Can it form more complex structures and functions out of simple ones? The answer: "yes".Ben W
July 10, 2013
July
07
Jul
10
10
2013
01:17 PM
1
01
17
PM
PDT
Elizabeth B Liddle
There may be perfectly good reasons why computers cannot arise spontaneously.
There are, Elizabeth, there are good reasons. See here .niwrad
July 10, 2013
July
07
Jul
10
10
2013
01:14 PM
1
01
14
PM
PDT
Nobody denies this, that “systems [in some circumstances] spontaneously go towards *order*”. What I deny, based on the SLoT, is that systems spontaneously go towards *organization*.
Oh, good. It's strange to say, but I'm glad I misunderstood you. Unfortunately, the 2LOT says nothing at all about organization, and as such, the compensation argument also has nothing to do with organization, since the compensation argument only addresses the 2LOT. When you say things like this:
It is the evolutionist “compensation argument” that deals with entropy AND organization, by saying that the latter simply comes from a decrease of the former, while this is impossible.
, it might appear that you don't understand the evolutionists' argument. Biologists do not say that organization comes simply from a decrease in entropy. Really, no one thinks that a decrease in entropy is sufficient by itself to cause organization or evolution, otherwise Mars would be teeming with life. We see local decreases in entropy all the time - does the freezing of an ice cube cause organization to emerge? No. A decrease in entropy is just one requirement -- among many -- that is required for evolution and life as we know it. No one thinks that this meeting this requirement is sufficient, by itself, to produce life. But failing this requirement would be sufficient to debunk evolution. However, as fate would have it, evolution passes 2LOT with flying colors. The amount of entropy decrease caused by life and evolution on Earth is less - in fact, much, much less - than the increase in entropy elsewhere.Ben W
July 10, 2013
July
07
Jul
10
10
2013
12:57 PM
12
12
57
PM
PDT
Folks, In few days at UD even 4 threads have been dedicated to the anti-evolution ID argument from the SLoT (Sewell's, kairosfocus' (comments off), Sheldon's and mine). Overall so far these articles have got 600+ comments. This is a clear sign that, among IDers, the awareness of the power of this ID argument is increasing. And, in the same time, among evolutionists, the awareness of the weakness of evolution - before the SLoT - is increasing too. Evolutionists continue to deny all, it is their tactics. But they are really afraid. They know that we know that "the evolution-king is naked", also from the SLoT point of view, beyond all others.niwrad
July 10, 2013
July
07
Jul
10
10
2013
12:00 PM
12
12
00
PM
PDT
Joe:
Really? How the heck do you guys communicate?
We communicate by defining our terms with excruciating care.Elizabeth B Liddle
July 10, 2013
July
07
Jul
10
10
2013
11:44 AM
11
11
44
AM
PDT
niwrad
As Prof. Sewell says: if in a closed system a computer arises, it is insufficient that you says that order can arise.
Of course it is insufficient. Nobody is saying it is sufficient. It is however necessary. Therefore Granville's claim that it can't happen, and therefore evolution can't happen is incorrect. There may be perfectly good reasons why computers cannot arise spontaneously, i.e. be designed and built by evolved organisms that arose spontaneously, but it's not because order can't arise spontaneously. It can.Elizabeth B Liddle
July 10, 2013
July
07
Jul
10
10
2013
11:43 AM
11
11
43
AM
PDT
Natural selection is eliminative. The only reward for the survivors is that they may get the chance to reproduce. That isn’t how EAs operate.
It is exactly how they operate.Elizabeth B Liddle
July 10, 2013
July
07
Jul
10
10
2013
11:38 AM
11
11
38
AM
PDT
Ben W:
What we scientists mean when they talk about “information” is different from what Dembski means when he talks about it.
Really? How the heck do you guys communicate? What are all those words in your peer-reviewed papers? What are all of those words in your science textbooks? Those words are not information to you and your alleged scientists? We use the word information the same way that everyone uses it- the same way scientists use it when they communicate their ideas. It's nonsense like Ben's that demonstrates why people don't take their side seriously.Joe
July 10, 2013
July
07
Jul
10
10
2013
11:32 AM
11
11
32
AM
PDT
Ben W
Imagine what happens to every Earth-like planet that is cooling down from its early, molten state. It goes from molten rock (high disorder) to solid rock (low disorder). The system spontaneously goes towards order. How can it do that? Because it’s an open system.
Nobody denies this, that "systems [in some circumstances] spontaneously go towards *order*". What I deny, based on the SLoT, is that systems spontaneously go towards *organization*. I explained the abysmal difference between organization and order, and even I symbolized such difference in my pictures as an insurmountable "hole".
compensation doesn’t “cause” evolution [...] compensation allows for evolution..
Compensation allows for order increasing, which has nothing to do with organization creation. As Prof. Sewell says: if in a closed system a computer arises, it is insufficient that you say that order can arise. Your order is ZERO compared to organization. Sewell continues: you must prove that the CSI of the computer gets in the system somehow. And your compensation doesn't provide such proof.
If your understanding of a physical law is contradicted by reality, then you’re not understanding the physical law correctly.
I assure you that my understanding of the SLoT is NEVER contradicted by reality. It is the understanding of the SLoT by evolutionists that is poor: in fact they continue to dream, also if the SLoT evidences should awake them.niwrad
July 10, 2013
July
07
Jul
10
10
2013
11:29 AM
11
11
29
AM
PDT
Ben W:
Well, what organisms (besides humans) can even understand the idea of evolution or “goals”?
If James Shapiro is correct, all of them.
And it’s trivially obvious that a piece of code is not self-aware, or aware of a “goal”. It’s not sentient.
The programmer is. And the entire program is designed for a purpose. If a program is designed to do something, and it does it, then it did so by design, not by natural selection.
What I mean to say here is that evolution and EAs work regardless of whether the thing that’s being “tested” is aware of it.
EAs work because they were designed to. And according to ID, organisms were designed to evolve and eveolved by design.
That’s how EAs work, too. They eliminate the pieces of code that don’t work (or equivalently, these pieces of code don’t “reproduce”),...
No, there is a selection coefficient actively selecting which will (continue to) reproduce and which will not. It selects those which are deemed better suited to solve the problem. And yes that means it actively eliminates those which it deems are not suited to solve the problem. To sum it all up- EAs use reproduction and variation to achieve a goal. Natural selection doesn’t have any goals and a chance to reproduce is all that awaits the survivors, whatever they are.Joe
July 10, 2013
July
07
Jul
10
10
2013
11:28 AM
11
11
28
AM
PDT
Ben W, I can’t make heads or tails out of anything you said, save to gather you do not want to be reasonable to the evidence.
BornAgain, I'm basically saying I can't see how your conclusions follow from the evidence you presented. I also have no problem with the classical/quantum information connection. To a physicist, the distinction between the two of these is irrelevant for this conversation. Neither of them have anything whatsoever to do with the idea of "organization", CSI, or what Dembski and Behe claim about evolution. And that's the problem. What we scientists mean when they talk about "information" is different from what Dembski means when he talks about it. We can measure physical information (whether classical or quantum), but you can't measure 'specified complexity'; heck, you can't even really rigorously define it. That's on reason many of us see it as nonsense.Ben W
July 10, 2013
July
07
Jul
10
10
2013
11:26 AM
11
11
26
AM
PDT
Ben said: An important part of both is that the organism or code does not know about the “goal”,… On what evidence do you make that claim?
Well, what organisms (besides humans) can even understand the idea of evolution or "goals"? And it's trivially obvious that a piece of code is not self-aware, or aware of a "goal". It's not sentient. What I mean to say here is that evolution and EAs work regardless of whether the thing that's being "tested" is aware of it. The "goal" isn't held by the organism, it's held by the entire system (and even then, we shouldn't really call it a "goal", as that implies some kind of sentience. Does gravity "try" to pull things downward? Is that its "goal"? No, it just does what it does.)
Obvioulsy you don’t understand natural selection. As Mayr said it is eliminative, not selective. And BTW biological fitness is an after-the-fact assessment.
Surprise! That's how EAs work, too. They eliminate the pieces of code that don't work (or equivalently, these pieces of code don't "reproduce"), and we do this after they've been created. We don't create pieces of code that we think will work well, instead, we combine other pieces of code that work, and we see how they do, and then the ones that do well are allowed to "reproduce". Honestly, it's extremely analogous to evolution by natural selection.Ben W
July 10, 2013
July
07
Jul
10
10
2013
11:11 AM
11
11
11
AM
PDT
Ben W, I can't make heads or tails out of anything you said, save to gather you do not want to be reasonable to the evidence. Good luck with all that! I'm out of this one. Welcome to Joe's world! :)bornagain77
July 10, 2013
July
07
Jul
10
10
2013
11:09 AM
11
11
09
AM
PDT
Ben W:
On the contrary, “to reproduce” is as much of a goal as any other problem.
Well Lizzie's position cannot explain reproduction.
An important part of both is that the organism or code does not know about the “goal”,...
On what evidence do you make that claim?
In life, nature judges an organism; in EAs, the programmer has provided some criteria.
I doubt that nature judges and the programmer provides everything the program needs to solve the problem.
If we were to design a real-life “program” to solve the problem of “make an animal as fit as it can”, we couldn’t do better than evolution by natural selection.
Obvioulsy you don't understand natural selection. As Mayr said it is eliminative, not selective. And BTW biological fitness is an after-the-fact assessment. To sum it all up- EAs use reproduction and variation to achieve a goal. Natural selection doesn't have any goals and a chance to reproduce is all that awaits the survivors.Joe
July 10, 2013
July
07
Jul
10
10
2013
10:30 AM
10
10
30
AM
PDT
And you cannot demonstrate that “EAs are an exact replication of natural selection in silico” and I can demonstrate they are design mechanisms, ie tools designed by us to solve specific problems (they have goals).
On the contrary, "to reproduce" is as much of a goal as any other problem. An important part of both is that the organism or code does not know about the "goal", it is simply judged by how well it meets it. In life, nature judges an organism; in EAs, the programmer has provided some criteria. But neither the organism nor the piece of code being judged needs to have a goal; they're simply judged by an external source, and the 'successful' ones reproduce. If we were to design a real-life "program" to solve the problem of "make an animal as fit as it can", we couldn't do better than evolution by natural selection.Ben W
July 10, 2013
July
07
Jul
10
10
2013
10:21 AM
10
10
21
AM
PDT
1 2 3 5

Leave a Reply