A Conclusive Definition of New Genetic Information


Anti-evolutionists often point out that new genetic information is needed for evolutionary change to occur. But what exactly is new information? The definition is absolutely critical because it dictates how we distinguish evolutionary change from simple variation. And so we have found it necessary to provide a conclusive definition. But first, information itself must be defined. NOTE: in this article, ‘evolution’ refers to the theory of evolution which claims all life is descended from a common ancestor.

Shannon Information Theory

Genetic information is usually described by evolutionists in terms of Shannon’s Information Theory invented by Claude E. Shannon. This theory basically defines information as a reduction in the amount of uncertainty of the state of a system. According with Shannon Theory, to obtain information you must increase the receiver’s certainty about a particular thing (usually the probability of occurrence of a symbol). But Shannon Information has little to do with genetic information because it was invented for telecommunications — not biology or genetics.

Following are some examples showing why Shannon Theory cannot describe genetic information adequately.

1. Imagine little Johnny asking his father whether the Moon is made of cheese. His father mischievously replied that it is. Johnny’s uncertainty prior to the answer is now gone since he is fully certain the Moon is made of cheese. According to Shannon Theory, Johnny would have received new information since his certainty has been increased — he is certain that the Moon is made of cheese. But of course, the Moon isn’t made of cheese! The ‘new information’ Johnny received is false.

2. Imagine you were challenged to explain something about the following: ‘CAT’ or ‘100110010100’. Every answer to the challenge must presuppose a prior knowledge of what the word means. ‘CAT’ could refer to an animal or to a CAT scan. The string of 1s and 0s could be part of the ASCII binary code used in computers or it could be a secret code used by spies. If we use Shannon Theory, the phrases do not conduct meaningful information by themselves.

3. Consider a perfectly cut beam of wood. It wouldn’t take much to describe it — only a description of the material and the position of each edge or corner. However, if it is shredded into a pile, an enormous amount of Shannon information would be required to describe the position and shape of each woodchip.

So if we apply Shannon Theory to biology, an increase in information could become lethal or increase randomness (just as more information is required to describe random woodchips than is needed to describe a perfectly cut beam).

Shannon himself even admitted his theory does not recognize meaning:

“Frequently the messages have meaning; that is they refer to or are correlated according to some system with certain physical or conceptual entities. These semantic aspects of communication are irrelevant to the engineering problem.”[1]

Shannon’s theory also tells us that the information content increases with the number of symbols used. Thus a very concise sentence would have less information than a sentence which uses many unnecessary words and does not adequately explain or state anything meaningful.

Shannon Theory does not discern between meaningless and meaningful, purposeless and purposeful, or incorrect and correct information since it relies heavily on the certainty of the receiver.

As Werner Gitt stated in his paper Information, Science and Biology, Shannon Information ‘completely ignores … semantic aspect [meaning of a word or sentence], [and] is wholly unsuitable for the evaluation of chains of symbols conveying a meaning.’ (Emphasis added)

A more concise definition for genetic information is therefore needed.

Gitt Information

While Shannon’s information theory has only one dimension and does not recognize meaning, the theory proposed by Werner Gitt has five dimensions and does recognize meaning. Gitt’s theory may be summarized like this: “Information is an encoded, symbolically represented message conveying expected action and intended purpose and is only present when the five hierarchical levels are observed in a system: statistics, syntax, semantics, pragmatics, and apobetics.” What exactly are these five dimensions, or hierarchical levels of Gitt information?


Statistics is concerned with the actual letters of a ‘word’. For instance, ‘cat’, ‘bat’, and ‘sat’ all have equal statistical information since they all have three letters. In genetics, the English letters are replaced with nucleotide bases (G, C, A, and T).


Syntax describes the ordering of the letters and words within a sentence, whether the word is spelled correctly, whether the rules of a language are observed, and whether the correct language is used. ‘I went shopping’ has more syntax information than ‘I w∀n tshappi∑g’. The first sentence contains correctly spelled words, the words are ordered, and proper English letters are used; the second sentence has misspelled words, contains spaces where they shouldn’t be, and non-English symbols. Therefore, the second sentence has less syntax information. Whether a sentence makes sense is not included in syntax information.

The same rules hold true of genetics, except the letters are replaced with the four nucleotide bases.


Semantics involves the meaning of a word or sentence. For instance, ‘bat’ refers to ‘a small dark coloured mammal that only comes out at night’. ‘Bat’ has more semantic information than ‘bjt’ because ‘bjt’ does not convey meaning to the receiver while ‘bat’ does. The language, size of the words or sentence, number of letters, and method of transmission do not have effect on semantic information — it is only concerned with the conveyed meaning.


Every information transfer is performed with the intention of an action being performed by the receiver — whether it is to make the receiver watch an advertisement in full or to start designing a new building. Pragmatics therefore concerns itself with expected action. An example could be the advertisement ‘buy our new product for only $15!’ — it expects the receiver to come into the store (i.e. perform action).

The pragmatic level exists in genetics since a gene usually ‘commands’ a certain protein to perform action.


Apobetics is the most important aspect of the Gitt theory and describes the result or purpose of the information itself. The sentence ‘milk is white’ has a purpose of making the receiver understand that milk is white. The sentence ‘buy our new product for only $15!’ requires the intended result of going to the store in conjunction with actually buying the product (the intended result of the advertisement).

Genes also conform to the apobetic level since they ‘command’ proteins to create a particular function or trait (i.e. requiring an intended result). Apobetics differs to pragmatics in that pragmatics only requires action from the receiver (going to the store, or proteins acting at the command of genes), while apobetics requires actions intended by the transmitter to be performed (buying the product, or proteins performing the proper function as required by the genes).

Information in DNA is used to translate RNA into proteins. This would not happen without meaningful information

Information in DNA is used to translate RNA into proteins. This would not happen without meaningful information.
Image from Wikimedia Commons and is in Public Domain

Conclusion of Gitt Information

Now for information (defined as “an encoded, symbolically represented message conveying expected action and intended purpose”) to be present in a system, the five levels of information must be observed in that system. Genetics fulfils these requirements since DNA (1) is made up of ‘letters’ called nucleotide bases, (2) nucleotide bases can be properly ordered into ‘words’ (codons) conforming to rules, and is made up of a ‘language’ which the cell understands, (3) genes have meaning and code for amino acids/proteins, (4) genes ‘expect’ proteins to act on their commands, and (5) genes expect proteins to fulfil the intended function.

Gitt Theory is therefore the best information theory that can be used in genetics.

Definition of New Information

Keeping all this in mind, we can confidently define new information as:

“The arising of a new encoded, fully useful, symbolically represented message conveying expected action and intended purpose in where all of the five hierarchical levels of the message — statistics, syntax, semantics, pragmatics, and apobetics — are altered so that they have entirely new, useful, and functional symbols/meaning than what they had originally.”

This means that new (genetic) information only comes about when the five dimensions of information have been meaningfully changed to something different — this includes the word and how many letters it contains (syntax) all the way to a change in the intention of the information (apobetics). So for instance, if a gene is duplicated, no new information has come about because the duplication has the same function as the original (the apobetic level hasn’t changed).

Evolution and New Information

So what has all this got to do with evolution? Can microbes evolve into man by information increase? No. When genes are recombined, new information can arise (but information is also lost in the process) — yet evolution cannot happen by recombination alone, because it requires many thousands of genetic ‘letters’ to be added. So microbes-to-man evolution would actually need a net gain in new genetic information for it to begin. A net gain happens when new information is added without the loss of previously existing information.

Possible Examples of New Information?

Evolutionists are convinced they can provide examples of new information being added to a creature’s genome, or at least they may provide a mechanism by which it might happen. Keeping in mind that a net gain in genetic information is needed, let us see if these examples and mechanisms stand up to scrutiny.


Gene Duplication

Gene duplication is the alleged mechanism allowing an organism to acquire a net gain in information. It is believed that a gene will duplicate which would change meaning as it mutates over time. Technically, this would be classified as a net gain, given that all levels of information (statistics, syntax etc.) were changed to something completely different from the original. But there are some valid problems when we extrapolate this process to one which provides all the new genes and information needed for evolution. Dr. Robert Carter puts it best:

“Invariably, the people who use this as an argument never tell us the rate of duplication necessary, nor how many duplicated but silenced genes we would expect to see in a given genome, nor the needed rate of turning on and off, nor the likelihood of a new function arising in the silenced gene, nor how this new function will be integrated into the already complex genome of the organism, nor the rate at which the silenced ‘junk’ DNA would be expected to be lost at random (genetic drift) or through natural selection.”[3]

The main problem with duplication as a mechanism for evolution is that the duplicated gene must be ‘turned off’ for it to not harm the organism (if it is ‘turned on’, it will almost indefinitely be harmful as such duplications cause cancer and diseases). But when the gene is turned off, there is nothing to keep it through the generations! In other words, a turned off duplication is not beneficial to the organism so it will probably be lost. Indeed, even helpful genes are commonly lost[4] — so how much more will this occur to a useless turned off gene!

If deletions of functional genes occur quite commonly, why should we believe that non-functional duplicated genes would last for hundreds of thousands of years while they acquired new meaning?


Recombination of DNA could conceivably produce new information, but it doesn’t result in a net gain of information. An analogy is in order to demonstrate the point: What amount of modifying and shuffling of letters of a small book will transform it into the British Library? You might get a different type of book, but never the British Library. Recombination is not sufficient to produce a net gain in information and therefore, it isn’t a sufficient process whereby microbes-to-man evolution might occur.


This article can be summarized with these four points: (1) Shannon theory is insufficient to explain biological information since it ignores meaning, intended purpose, truth, and quality. An increase in Shannon information could also be lethal. (2) Werner Gitt’s definition is the best definition that describes biological information. (3) There have been no examples of new information arising (where the statistics, syntax, semantics, pragmatics, and apobetics levels must be changed) in any organism, including the nylonase adaption. (4) Evolution requires a net gain in information — which likewise has never been observed.


  1. Shannon, C.E., A mathematical theory of communication, p. 1, reprint with corrections from Bell System Technical Journal 27:379–423, 623–656, July, October, 1948, at <http://cm.bell-labs.com/cm/ms/what/shannonday/shannon1948.pdf> Back to text
  2. Yomo, T., Urabe, I. and Okada, H., “No stop codons in the antisense strands of the genes for nylon oligomer degradation,” Proceedings of the National Academy of Sciences USA 89:3780–3784, 1992. Back to text
  3. Carter, Robert W., “Can mutations create new information?” Journal of Creation 25(2):92–98, August 2011. Back to text
  4. The 1000 Genomes Project Consortium, A map of human genome variation from population-scale sequencing, Nature 467:1061–1073. Back to text
comments powered by Disqus