How to Do Your Own Darwinian Evolution Experiments with the Random Mutation Generator

You don't have to take anyone else's word for it - you can use the Random Mutation Generator to find out for yourself how Darwinian evolution really works.

by Perry S. Marshall

Responses to skeptics and Frequently Asked Questions

Your Mission, should you choose to accept it, is to type a sentence into the Random Mutation Generator and then get it to say something else - something sensible, something more meaningful - without first producing a jumble of mis-spellings that drive your message into extinction.

According to the experts, Darwinian evolution is driven by a combination of Random Mutation and Natural Selection:

"Cumulative selection is the key to all our modern explanations of life. It strings a series of acceptably lucky events (random mutations) together in a nonrandom sequence so that, at the end of the sequence, the finished product carries the illusion of being very very lucky indeed." -Richard Dawkins, The Blind Watchmaker

It all starts with random mutation, from which natural selection can choose winners. Since DNA is a language based on a four-letter alphabet (A, C, T, G); since letters form words (Codons) and words form instructions (Genes) we can test the concept in Plain English.

The program gives you a piece of sample text "The quick brown fox jumped over the lazy dog." You can paste any text you want into the Random Mutation generator - computer programs, English, Spanish or whatever. And you have some choices:

1. How many mutations do you want at one time? (1, 2, 5, 10)

2. What kind of mutations do you want? Text mutations cause individual letters to be randomly substituted with other letters, numbers or punctuation marks. Binary mutations cause changes in the 1's and 0's that represent the letters.

How People Imagine Random Mutation Works

Most people have never seen a true simulation of random mutation and natural selection, and imagine that it works something like this:

You start with a sentence "The quick brown fox jumped over the lazy dog"

And then random mutation generates a bunch of variations:

The quick black fox jumped over the lazy dog

Too quick black fox jumped over the lazy dog

The jumped fox quicked black over the lazy dog

A quick brown flop jumped over the lazy dock

A quick brown dog flopped over the lazy clock

It would seem reasonable that is you substitute enough different words over a period of time, you might end up with a really great sentence like

The ferocious black fox leaped over the sleeping dog

Which your English teacher might feel is a much more colorful, more descriptive sentence.

...Now we add Natural Selection

With English, Natural Selection equates to people preferring one sentence over the other.

One of the best real-world examples of natural selection is advertising. Let's say you sell quick brown foxes and you want to advertise them on the Internet. You could write a paid Google advertisement (you see them running down the right side of the screen when you search) that says something like this:

Quick Brown Fox
Jumps Over Your Lazy Dog
Other Colors Available too - Black, White
www.QuickBrownFox.com

One of the cool things about Google is you can write several ads and they compete with each other for clicks. You could write a different version of the ad

Fast Black Foxes

Jump Over Lazy Dogs

Black, White, Brown - 50% off

www.QuickBrownFox.com

You run both ads simultaneously for a day or two.

Let's say the first one gets clicked on 1% of the time, the second one gets clicked on 3% of the time.

That means the second one is 3 times better than the first. The second one wins, the first one gets deleted. (This is exactly how it's done, by the way.) In fact Google is the ultimate Darwinian advertising machine! The winners rise and the losers fall.

And again, if we've never actually seen Random Mutation before, we might imagine that the Random Mutation Generator would be quite helpful in writing new ads.

See For Yourself The Random Mutation Generator is NOT Helpful AT ALL...

If you experiment with this yourself, you will quickly discover this doesn't work at all - because random mutation seems to only destroy your sentences:

The quick brown fox jumped over the lazy dog

after 5 mutations becomes

Qhe qu4ck brown fox jimpeX over.the lazy dog

and after 10 mutations becomes

Qhe qN4ck brown fox Vim3eX oeer.the lazy Iog

Or let's take these ads we've just written:

After 5 mutations:

Fast Black Foxes

Jump Ove6 Lazy Dogs

kylack, White, Brown - 50% off

whw.QuickBrownFox.coL

After 10 mutations:

Fast Black Foxes

Jump3Ove6 LaSy Dogs

kylack, White, Browf - y0% off

whw.QuickBrownFod.coL

If you play with the Random Mutation Generator for about 10 minutes, you begin to see that in order to get your sentence to evolve in any useful way at all, the mutations would HAVE to focus on individual words and leave the other words alone. But maddeningly, the mutation generator doesn't do that. It just randomly destroys the stuff that was already good. Random mutation is blind and has no respect for what is already working fine.

And it doesn't matter how many times you hit the reset button and start over. The most you can ever get from this is a minor spelling change from a few mutations. But once you've attained some kind of very modest change, the random mutations continue to destroy what you've built. After 20-30 mutations it's not even recognizable as English anymore!

There's an Even Bigger Problem

Let's say you want the word Brown to evolve into the word Black. Shouldn't be too hard, should it? Only four letters need to change after all. But even if you could get the mutations to concentrate just on those four letters, you'd still get a mis-spelled word, which natural selection would eliminate before it ever evolved into the correct spelling.

Let's try it, one mutation at a time:

Brown > Brorn > Brorb > BrorW > qrorW > qKorW > qKoJW > qKoyW > qFoyW > qjoyW > qjTyW

After 10 mutations we didn't have a single letter remaining.

Maybe we need to do more than 1 mutation at a time? If I do 5 at a time maybe I can make the leap in one step:

Brown > Vr17n

Remember: In real life, mis-spelled words will cause our ads to go extinct. A mis-spelled ad with garbage characters can't compete with a correctly spelled one - nobody will respond to it. Mis-spellings are not what we want and they're NOT good!

Our only hope is some kind of "punctuated equilibrium" where big jumps happen all at once. What's the chances of evolving Brown > Black in one step?

You can easilycalculate the statistical chances of this. For each letter there are 52 possible letters (including lower case and caps) and 10 numbers, plus a few punctuation marks. There's a total of about 65 possible characters per letter.

So the chances of evolving Brown to Black in one step are one chance in 65 to the power of 5, which is one chance in 1,160,290,625 (just over one in a billion).

One in a billion odds against evolving just ONE WORD - that's pretty remote.

...And it only gets worse

Remember, that's just ONE five-letter word. For a whole sentence like

The quick brown fox jumped over the lazy dog

which has 44 characters, there are 5.8639153496314421699960747595891e+79 possible combinations.

That's about the same as the number of particles in the entire universe.

Now obviously there are many possible sentences that can be constructed with 44 characters, maybe billions. Maybe even trillions. But even with billions or trillions - compared to 10 to the power of 79, all you're doing is knocking off a dozen zeros. Whether you do this experimentally or statistically, you quickly realize - this is hopeless!

The problem is this: It's impossible to re-write a sentence one or two letters at a time without getting fatal spelling errors in between - which cause your sentence to become extinct.

Oh, and there's one other problem: The longer the text, the worse it gets. It's fairly easy to produce a few real words with the Random Mutation Generator if you start with only five or six characters. But a complete sentence, like The quick brown fox jumped over the lazy dog, is nearly hopeless, as we've already seen. Every single letter you add to your sentence makes your odds worse by a factor of 65. Those 65's multiply real, real fast. An entire paragraph or page of text - that's a disaster!

Your own DNA contains as much information as a stack of encyclopedias. Would you even dream of trying to edit an encyclopedia with a random mutation generator? (Anyone who wants to is welcome to try.)

Short Easy Sentences vs. Real DNA, Real Biological Systems

Again, we've just played with a 44 character sentence. But what about real life?

The simplest known micro-organism is a parasite called Nanoarchaeum. It's got 490,885 base pairs. In other words its DNA has a string of almost half a million characters, each of which is the letter A, C, G or T. A simpler living organism is not known to exist.

Our sentence The quick brown fox jumped over the lazy dog requires 308 bits (1's and 0's) to represent in binary computer language (ASCII). Nanoarchaeum's DNA, by comparison, requires 981,770 bits to represent in binary computer language. Nanoarchaeum contains 3000 times more information than our simple sentence.

If a short sentence like The quick brown fox jumped over the lazy dog can't evolve through random mutation, then even "simple" microorganisms that are thousands of times more complex won't be able evolve through random mutation either. Random mutation does not create new information, it only destroys information.

Evolution through Random Mutation:

Possibly the Greatest Myth in Modern Science

The idea that random mutation creates biological diversity fails computer simulations*; it fails if you compute the statistics; and it fails biologically. This observation was confirmed by Theodosius Dobzhansky's fruit fly radiation experiments, Goldschmidt's gypsy moth experiments, and others. Decades of research were conducted in the early 20th century, bombarding fruit flies and moths with radiation in hope of mutating their DNA and producing improved creatures.  These experiments were a total failure – there were no observed improvements – only weak, sickly, deformed fruit flies. 

A bit of experimentation with the Random Mutation generator makes it clear why:

  • Language, plans and instructions do not evolve from the "bottom up."
  • Language, plans and instructions do not evolve in microscopic increments, one or two letters at a time. They evolve in increments of entire words, phrases, sentences and paragraphs.
  • Language, plans and instructions evolve as an expression of ideas which come first.

When you write or speak, you begin with intent which becomes a sentence made of words which are made of letters. When you translate from English to French, for example, you cannot translate one letter at a time. And you can't really translate one word at a time either, that's a disaster. You have to translate idea for idea. The same is true when you edit a document that you or someone else has written - the letters are subordinate to the words, which are subordinate to the ideas. Ideas come first.

Evolution via Random Mutation is nothing more than an urban legend. Why does this superstition persist? I believe it's because 99.9% of the people who believe it and talk about it have never conducted an experiment like this to see with their own eyes that it just doesn't work.

Randomness vs. Science

The entire enterprise of scientific inquiry has always been the assumption and discovery of underlying order. NOT disorder! From Copernicus, Galileo and Newton right up to the present day, science has always succeeded by assuming in advance that there are specific undiscovered causes for the behavior of the physical world. Scientists have always been motivated by a belief that these causes could be discovered. In fact science itself was born from a philosophical worldview that believed (for mostly theological reasons) that the universe was governed by an unchanging, predictable, discoverable set of laws.

Could life and DNA have risen randomly? There is perhaps some remote chance that they did. However I contend that such an explanation does not even qualify as a scientific explanation at all, simply based on the dictionary definition of science itself:

sci·ence (sī'ens) n. The observation, identification, description, experimental investigation, and theoretical explanation of phenomena.

Chance or accident as a proposed origin of anything defies observation, identification, description and experimental investigation. So "Randomness" per se not only doesn't explain how anything operates, it is categorically not even an explanation. 

It doesn't produce a testable hypothesis. It doesn't give anyone the ability to reproduce what happened in the past.

Whenever randomness is said to be an explanation of any particular process, then that explanation by definition can never be reproduced or tested. It only evades the question and ties the powerful hands of science behind its back.

"Randomness" as a theory of biological diversity is not merely bad science, it's a wholesale avoidance of scientific inquiry. It leads to theories and terms like 'Junk DNA.' (An October 2004 article in Scientific American described the Junk DNA hypothesis as "one of the biggest mistakes in the history of molecular biology.") There is no junk in DNA, and to assume there is just eliminates the possibility of making new, important discoveries.

A Newer, 21st Century View of Evolution

G.K. Chesterton, the well-known 20th century intellectual, said "The Christian is quite free to believe that there is a considerable amount of settled order and inevitable development in the universe. But the materialist is not allowed to admit into his spotless machine the slightest speck of spiritualism or miracle." Chesterton is suggesting that battle over 'creation vs. evolution' is, to some extent, a false dichotomy.

The real issue, both scientifically and philosophically, is naturalism vs. design.

Random Mutations cause birth defects, tumors, cancer, death and extinction; mutations are helpful in only the rarest of circumstances and are completely incapable of introducing significant improvements.

The current dogma which says random mutations drive evolution is almost completely false.

So the faster we can discard this wrongful notion, the faster we can get on with real research and understanding.

"Is the Ability to Evolve Pre-Programmed?"

If random mutations are not the source of adaptation, could there still be an adaptability mechanism pre-programmed in biological systems? Could DNA re-write itself? As you can see here, random processes simply are not the answer.

When I discovered this for myself, I wondered if evolution could be engineered to happen.

Microbiologist James A. Shapiro of the University of Chicago has done extensive research on this very question and has published a number of papers confirming that yes, there is an adaptation mechanism in DNA that is marvelously sophisticated. Mr. Shapiro has discovered that a protozoa, subjected to extreme environmental stress, can splice its own DNA into over 100,000 pieces (!), re-arrange them in a highly systematic fashion, producing new protozoa that inherit a new set of characteristics.

In other words DNA can be likened to a computer program that re-writes itself on the fly. In fact this is how our own immune systems adapt to a nearly infinite range of possible threats and ward off attackers - through this kind of sophisticated adaptation.

Mr. Shapiro published a fascinating paper in the Journal of Biological Physics called "A 21st Century View of Evolution." In this paper, the failed "Random Mutation" view is discarded and replaced with a much more sophisticated, engineered process. Instead of seeing DNA as a static database that is acted on by outside random forces, it is now seen as being more like a intelligent operating system that repairs files, corrects errors, and adapts to changing circumstances. It's nothing short of amazing.

Shapiro's work clearly shows that evolution, however it may be understood, is anything but a random process.

Further Reading: The Case for Intelligent Evolution

At my other website you can read or listen to my talk "Language, DNA and the Case for Intelligent Evolution" in which I describe what natural processes are capable of producing and what they are not capable of producing - and a view of biological evolution through the lens of information theory.

Go back to the Random Mutation Generator

"Where Did the Universe Come From?"

In this 5-day email series you'll discover:

  • The mistake Einstein later called “the biggest blunder of my career” – and a dangerous assumption that nearly blinded him to the greatest discovery of the 20th century
  • “Bird droppings on my telescope” – a strange piece of radio data that was almost attributed to… well, birds – and how this Nobel Prize-winning experiment now shapes our understanding of time itself (Day 2)
  • How “one extra atom” at the birth of the universe could have wiped out entire galaxies, or even the whole cosmos (Day 3)
  • A riddle: So simple, any child can understand; so complex, no atheist can solve (Day 4)
    The Big Bang and new implications for science, philosophy, and beliefs about God

The 5-part series starts today. Just enter your name and email address and click the button. I promise not to sell or rent your email address to anyone (that only creates hard feelings anyway), and you can un-subscribe at any time.

Enjoy...

Perry Marshall

*Random Mutation fails computer simulations: Some readers will object to this statement, as there are many evolutionary programming algorithms available - for example Richard Dawkins' methinks it is like a weasel program, Thomas Schneider's Ev program, and Cal Tech's Avida. But in every case, the computer program does NOT evolve the same way Darwinian Evolution allegedly evolves. All of these programs either 1) randomly mutate carefully selected portions of the code while keeping everything else intact (hey, that's design!), or 2) are designed to converge towards some pre-determined (pre-designed) state. In the case of Schneider's Ev, "evolution" is merely defined as creating more bits of information, but the information itself is totally meaningless. I'm not sure Schneider's program demonstrates much of anything at all. Every successful evolution simulation I'm aware of is, ironically, an example of intelligent design.

Frequently Asked Questions

Q: The Random Mutation Generator doesn't simulate evolution because it doesn't have natural selection.

A: The Random Mutation Generator has a button labeled SELECT, and a button called REVERT TO SELECTED TEXT. You can select any version of the message you want and revert back to it later with the push of a button.  And you're certainly welcome to save multiple copies of your message and mutate them all in tandem.

The Random Mutation Generator does not have a pre-programmed "natural selection" function because if it did, the selection would be designed, not natural.  It would be cheating with a pre-DESIGNED goal. 

In the case of English letters and words, the only selection that would be truly natural is real English in a real situation with real people.

A good place to test real English in real situations with real people is Google ads, the most prevalent form of advertising on the Internet.  Google's system has built-in natural selection (popular ads move up and are shown more, unpopular ads get eliminated).  If we use random mutation to mutate Google ads which are then favored or eliminated by natural selection, we will have something very close to Darwinian evolution.

Darwinists can put their money where their mouth is.  I have yet to meet one proponent of Darwinian Evolution who is interested in spending their money on this experiment.  (I ain't spending my money on it.)  It's clearly an impossible proposition.

Q: [From PZ Myers' Blog on Pharyngula]: It ignores fitness functions, synonymy, ranges of functionality, multiple functions, etc., etc., etc., all in the name of slavishly and crudely mapping English to protein evolution.

A:  You can't have a fitness function in a computer program without designing one.  That's Intelligent Design.  Any ad on Google is in a real-world environment with a natural fitness function; if you can randomly mutate text and write successful Google ads then you have successfully demonstrated that Random Mutation + Natural Selection is a viable design mechanism. 

Synonymy implies synonyms which do exist in English as we all know.  The quick brown fox jumps over the lazy dog and The fast brown fox jumps over the lazy dog are synonymous, but getting from one sentence to the other via random mutation requires a minor statistical miracle (one chance in billions).  And that's just with one sentence.  

(That's why 8-10 letters is more than enough security for most passwords on the Internet.  An 8 digit password with upper and lowercase letters and numbers has 218 trillion possible combinations, which means if a hacker has to randomly guess your password, it's gonna take him a long, long time.)

If we were starting with a whole paragraph or page of text, getting synonyms without having your paragraph destroyed first would require a major statistical miracle.

Ranges of functionality and multiple functions: Any useful English sentence might be considered good; there's a vast variety of acceptable outcomes.  But random mutation won't produce any of them, because random mutation doesn't obey the laws of the English language.  Nor does it obey the laws of the genetic code.

Q: [From PZ Myers' Blog on Pharyngula]:  Oh, but it is so much better than Avida or Tierra.

Neither Avida nor Tierra actually represent Darwinian evolution.  Both are intelligently designed computer programs that seek pre-programmed goals.  

The About Avida page says: "The population adapts to the combination of an intrinsic fitness landscape (self-reproduction) and an externally imposed (extrinsic) fitness function provided by the researcher."  

What is an externally imposed fitness function?  It is a pre-programmed goal.  The Darwinian version of evolution has no pre-programmed goals and it is not possible to create a simulation of evolution without designing one.

The What Tierra Is page says: "The Tierra C source code creates a virtual computer and its Darwinian operating system, whose architecture has been designed in such a way that the executable machine codes are evolvable...the machine code can be mutated (by flipping bits at random) or recombined (by swapping segments of code between algorithms)"  

Tierra only works because it has pre-designed goals and fitness functions.  Notice that swapping segments of code between algorithms is not random, it's designed.

All Genetic Algorithms, without exception, are examples of Intelligent Design.  And the more cleverly designed they are, the more useful the results that they produce.

Q: [From PZ Myers' Blog on Pharyngula]: This looks like something a sixth grader would slap together. 

A: That's becasue a sixth grader could slap this together.  After all, the essential concept of Darwinian Evolution is Sixth-Grade Simple, is it not? According to Darwinists, evolution is an all-pervasive principle, that random variation combined with the immense powers of natural selection cause more and more sophisticated organisms to emerge. But if evolution can only be successfully simulated by programs designed by (there's that dreaded "D" word again) people with PhD's and computer science degrees, then... isn't that a vote for design?

Thank you, Mr. Myers.  Your "sixth grader" accusation eloquently proves my point.  And, Mr. Myers, if you have more to say, I am happy to debate this further with you in public on the Infidels Discussion Board.  Or anyone else who wants to challenge me on this.  I've been successfully debating information theory, evolution and the origin of life there for well over a year. I welcome all comers.

Q: So are you saying the Random Mutation Generator disproves Evolution in general?

A: Nope.  What I am saying is that the Random Mutation Generator (and all evolutionary and genetic programs) demonstrate that evolution is an engineered process, not a random process.  Evolution works the same way Genetic Algorithms do: by seeking pre-defined goals. The ONLY way to get English to evolve is with a program that follows the rules of English.  Likewise the only way to get DNA to evolve is with a program that follows the rules of the Genome.  

The point of all this is very simple: Random mutations are equivalent to noise, and noise only destroys information and degrades functionality.  I'm a communications engineer, and in the world of digital communication there is no example anywhere of adding noise to a signal and getting a better signal that contains more information and more functionality.  The Random Muation Generator is a simple tool that illustrates this with English, but the results are the same regardless of what kind of code you mutate.  You could use French or HTML or C Code or DNA code, and the result will always be the same.

Random Mutation is noise and in engineering (where people design things for a living, and if the designs don't work, they get fired) noise is always the villain never the hero.  Ever.

Q: But nearly all biologists agree that Random Mutation of DNA, combined with Natural Selection, account for the evolution of all living things.

A: There is no doubt that many people believe this.  And in biological literature there are thousands and thousands of books and papers that say Random Mutation is the source of biological diversity.  But... NOWHERE in the biological literature is this ever demonstrated or proven. Nowhere.  

Note: I didn't say evolution hasn't been demonstrated, in some ways it has.  But no one in any experiment or scientific journal has ever demonstrated that the source was RANDOM mutations. (If you disagree, I challenge you to locate even one example.)  Relevant research:
  • Theodosius Dobzhansky tried for decades to mutate Fruit Flies with radiation and get them to evolve, and his experiments were an utter failure.  He did not get one single instance of improved fruit flies.  In fact the results were pretty much identical to what you get from the Random Mutation Generator: Sickly, deformed fruit flies with legs growing out of their mouths and so on.
  • Richard Goldschmidt conducted similar experiments with moths, also for many years.  Same dismal results.
  • Barbara McClintock discovered in the 1940's that damaged chromosomes re-arrange themselves in an impressively sophisticated pattern of internal genetic engineering.  Far from random, this process is remarkably similar to what a magazine editor does with an article: She makes judicious word substitutions, re-arranges entire sentences and paragraphs to improve a document.  The genome does the same thing with segments of DNA.  McClintock won the Nobel Prize in 1983 for her work on Mobile Genetic Elements.  Genetic information is modular, just like English or computer programs, and the only way to successfully modify it is to respect its structure.  Random mutations have no such respect.
  • James A. Shapiro of the University of Chicago has continued McClintock's work, noting that the Genome has sophisticated error correction mechanisms that actively prevent random mutations; and that organisms under stress can re-arrange their own DNA in an immensely sophisticated system of cellular genetic engineering.  Shapiro defines a "Third Way" (an alternative to both Darwinism and Creationism) and a "21st century view of evolution" which sees evolution not as a product of random mutation; far from it.  Evolution is an engineered process.

Q: Is your theory about Random Mutation (that it does not contribute to evolutionary progress at all) Falsifiable?

A:  Yes. Here's how to falsify it:

Show me just one paper, book or experiment anywhere in the history of biology that empirically demonstrates and proves that random mutation of DNA produces novel adaptive features (eyes, wings, legs, functional organs); and that the mutations that produced those features were in fact random. (And not Mobile Genetic Elements or some other systematic process.)

Again I repeat: In 150 years of research on this subject, there is not a single peer-reviewed paper, book or experiment that demonstrates this to be true. Not one.

No one who considers themselves a skeptic or a scientifically literate person can believe this to be true and still be consistent with skepticism and scientific literacy.

Q: So what is your aim with the Random Mutation Generator?

A: To demonstrate as clearly as possible that the notion of Random Mutation as a source of evolutionary progress is utterly false and absurd.  Evolution by Random Mutation is nothing more than a fairy tale. It is the most widely circulated myth in 500 years of science.  The dogma has been repeated ad nauseum but is always assumed, seldom examined, and never demonstrated by anyone to actually be true.

The evolution of information is ALWAYS a top-down process, not bottom-up.  The top-down structure of information itself makes the strongest possible inference to design in biology.

©2006-2007 Perry S. Marshall