Feature Article - September 1998
by Do-While Jones

Information and Evolution

Darwin developed his theory of evolution before anyone ever thought of information theory. Although the study of information theory is still in its infancy, it is clearly posing some problems for the theory of evolution.

We now know that DNA contains the instructions for building various kinds of living cells. We even know how to measure that information in "bits."

But what really is information? Where does it come from? What does it have to do with evolution?

Let's answer the last question first. If information can be created by chance and natural selection, then DNA could have evolved the information necessary to build living cells without the help of an intelligent designer. But if information only comes from an intelligent source, then DNA had to be created by an intelligent source. It could not have evolved. So that is why it is important to figure out if information can come from a random process or not. We certainly can't figure out where information comes from if we don't know what it is, so that's where we must start.

What is information?

The nickname "Information Superhighway" is an overused cliché for the Internet. It is so overused that many people are sick and tired of hearing the term. But it does illustrate an important point. Information flows (in the form of text or images) between people over the Internet.

Let's talk about images first. Images are patterns of colored dots. These colored dots (called pixels) are formed by mixing various amounts of red, green, and blue light. It takes a certain number of bits to transmit each pixel. You can multiply the number of bits per pixel times the number of pixels per image, and that will tell you the maximum number of bits of information the picture can contain.

Images rarely contain that maximum number of bits of information. Images are generally "over-sampled," resulting in "redundant pixels." Since it takes time to send each bit over the Internet, engineers have worked hard to figure out ways of "compressing video" to remove these redundant bits. How this is actually done is irrelevant. The point is that ways exist to determine the amount of information in an image and the minimum number of bits needed to transmit the image. So, information is a quantity that is measured on a regular basis.

The amount of information in a message is inversely related to its probability. Suppose I could tell you which face would come up on the next roll of a die. Since the probability of a correct prediction is 1 in 6, that message would contain 2.585 bits of information.1 If I could tell you (within ½ mile) the latitude and longitude of the location where the next meteor will strike the Earth, that message would have 15.6 bits of information because the odds of predicting that location are about 50,000 to 1. A message containing the 6 numbers (out of 80) that will be drawn in a lottery would contain 28 bits of information because those odds are about 300 million to 1.

Information by Chance?

It is possible to write a computer program that generates random characters (letters, numbers spaces, punctuation marks) with the same statistical distribution as found in a typical English document. If we let that program generate random characters, and run that gibberish through a spelling checker, most of the "words" will be misspelled. But there will be a few character sequences that happen to be correctly spelled English words. A mathematician can calculate the expected number of correctly spelled words in randomly-generated text. If he does the calculations correctly, then the number of randomly produced words will be close to the prediction. If the number of words spelled correctly matches the expected value, then there is no information in the randomly-generated text. No surprise means no information.

If you select a text file at random from somebody's floppy disk, and run it through a spelling checker program, nearly all the words will be recognized by the spelling checker. (There might be some technical terms or geographical names that the spelling checker doesn't recognize.) Since the number of correctly spelled words is not even close to the expected number for random text, it is a clear indication that the text file was produced by an intelligent agent rather than a random process.

A thorough analysis of a typical text file will show some patterns. The letter "q" will often be followed by the letter "u". Certain sequences of letters, like "t-h-e" will appear more frequently than would be produced by a random process. So, the existence of patterns suggest the presence of information. (The same is true of images. Images that convey information have a pattern in them. But if you disconnect your TV antenna you just get pattern-less "snow" or "confetti" on your screen.) Truly random data does not contain any patterns.

There is more to English text than just patterns of characters that form words. One could write a program that randomly selects English words from a dictionary and builds sentences from them. If you pass such a file through a spelling checker program, it would find many correctly spelled words. This is an indication of an intelligent source. In fact, the process wasn't completely random. An intelligent programmer selected words from a dictionary that had been created by someone with a knowledge of correct English spelling. So, the surprisingly high number of correctly spelled words is a valid indication that an intelligent source was somehow involved in producing the file.

A file produced this way might contain, "Ate sound John." This sequence of words is not a sentence because it is syntactically incorrect. English sentences don't take the form verb-object-subject. Words have to be in a valid order to make a sentence.

"John ate sound" has the subject, verb, and direct object in a valid order, so this "sentence" is syntactically correct. Even so, it is semantically incorrect. In other words, it has no meaning. One cannot eat sound.

Text files produced by intelligent humans contain words that are spelled correctly, arranged in syntactically (that is, grammatically) correct sentences, that have semantic content (that is, they make sense). This gives them the ability to convey information from one person to another.

The probability that a random character generator (even if it is filtered by a spelling checker) would produce the sentence, "Amelia Earhart's airplane crashed into the ocean at 5 degrees, 8 minutes, 24 seconds south latitude, 172 degrees, 54 minutes, 3 seconds east longitude." is essentially zero. But just for the sake of discussion, suppose it did. The chances that the wreckage of her airplane would be found within one-half miles of the specified latitude and longitude is about 1 in 50,000.

Suppose you received a message from an unknown source that correctly specified the location of the wreckage of Amelia Earhart's airplane. A random character program could not have generated that message. Someone who knew where she crashed must have written the message. Random messages don't contain that much information.

Consider the DNA molecules in your body. Your DNA is a sequence of about 3 billion base pairs that contains the blueprints for all the cells in your body. The information coded in this molecule tells how to make blood cells, skin cells, brain cells, bone cells, and optic nerves. It contains instructions far more detailed than directions to Amelia Earhart's missing airplane.

If a tree falls …

Philosophers sometimes debate the question, "If a tree falls in the forest, and nobody hears it, does it make a sound?" We may not be able to answer that question, but we can answer a more pertinent one. "If someone sends a message and nobody receives it, is information transferred?" Certainly the answer to that question is, "No."

The information in a DNA molecule would be useless if the cell didn't know how to process it. Consider this actual fragment of DNA.

ctgcaggaaa ctttatttcc tacttctgca taccaagttt ctacctctag atctgtttgg ttcagttgct gagaagcctg acataccagg actgcctgag acaagccaca agctggtgag ttgtaggcat tttttccatt actttctgat tcataggctc aacgcacctc aaagctggaa atgccgggtc tgggtacacc ctggggaact gcaaagcctg cacacttggg gggaatgatc aagatgagag gcaggggtgg ggatggcatg tgcaccagga gatgttagag aaacctgagg aagagcagag tgcagcaggt gatgggggag agtgggcagc aagcgaggcc aggacagcca ctctgctcag tcaccagtcc acacacccag gggctcactc tgcccctctg agcacccaag gacgttaaag agctggaact gttagtctaa atataggacc atccaagctc tgaaccaaaa tgtgtccctt gcctcaactc aggagatcca cagaggcaga agtaaggaat ttattttctg aaagatagat ttctatcagt tctgggtgac atgttctgac act

Does it tell you anything? If it does, you can pick up your Nobel prize. But your cells know it is a human monocyte chemoattractant protein 1 receptor gene, 5' region. Your cells know how to read it, even if you don't. Did the cell's ability to read DNA evolve by chance? Do computers figure out how to read information from the Internet without someone programming them to do it? Certainly not.

The Evolutionists' Dilemma

Evolutionists must explain where the genetic information in DNA came from. They can't do it. Here is how one evolutionist tries to dance around the problem:

It is simply not possible to change a hemoglobin gene into an antibody gene in one step. … To understand how evolution really works, we have to abandon the notion that such mutations can happen. Instead we must think of mutations as small changes affecting the functions of preexisting genes that already have long and complex histories. Usually, new mutations tend to damage genes in which they occur because they upset their precise functioning and their finely honed interactions with other genes. But sometimes they change them in ways that increase the fitness of their carriers, or might increase the fitness of their carriers farther down the line if the environment should alter in a particular way.2 [italics in the original]

Dobhzhansky's view [is] that much of the variation needed to accomplish the transition was already present in the gene pool … 3 [italics in the original]

There are two fallacies in this argument. The first is that random changes in existing information can create new information. Random changes to a computer program will not make it do more useful things. It doesn't matter if you make all the changes at once, or make one change at a time. It will never happen. Yet an evolutionist tells us that if one makes random changes to a hemoglobin gene that after many steps it will turn into an antibody gene. That's just plain wrong.

Information does not gradually increase through small, incremental random changes. There is no scientific evidence to even suggest that it does. But many of us are painfully aware that just one bit error in Windows 95 can cause our computer to die! Information certainly can be lost by a small, incremental change.

The second fallacy is the assumption that information "was already present in the gene pool" just waiting to be changed. Where did that previous information come from? Presumably it came from modifying other existing information. But where did that existing information come from?

Remember, a message telling the 6 winning lottery numbers (a 300 million to 1 shot) contains 28 bits of information. Just the first 50 letters of the monocyte chemoattractant protein 1 receptor gene printed above contains about 100 bits of information. That whole gene contains about 1166 bits of information. But that gene represents just 583 of the 3 billion base pairs in a single human DNA molecule.

Information Isn't Just Structure

At our last Fourth Friday Free Film showing, our favorite critic argued that hydrogen and oxygen naturally combine to form water, which is a more complex structure, and therefore contains more information than hydrogen and oxygen alone. (We admit, he didn't say it quite that succinctly. He never says anything succinctly.) He confused structure with information.

The carbon atoms in a diamond are arranged in a neat crystal structure. The carbon atoms in a lump of coal are not. The carbon atoms in a diamond don't have any more information than the carbon atoms in coal. Information is coded in patterns that have meaning to the receiver. Water molecules and diamonds don't contain patterns that represent information. Yes, they have structure, but no, they don't have information.

Information Needs a Source

If you find a message that contains information (such as the location of Amelia Earhart's airplane), someone had to write it. Random chance does not produce information. If there is information coded in a DNA molecule (and there certainly is), an intelligent source must have put it there. There is no scientific evidence that even a small amount of information can be generated by chance. There is scientific evidence that random changes to a message can remove information. Mutations might remove information, but they will never create it.

To believe that a DNA molecule evolved by chance, you have to reject science.

Quick links to
Science Against Evolution
Home page
Back issues of
(our newsletter)
Web Site
of the Month
Topical Index


1 The amount of information is log2(1/probability).
2 Christopher Wills, The Runaway Brain, 1993, page 166 (Ev)
3 ibid. page 64