The existence of people who vehemently reject the theory of evolution is no big news. They have essentially been around for as long as the scientific notion has, and they have been fairly consistent in their objections to it, as well as in their motives. For the most part, this rejection is religiously motivated and largely an inherited stance, ironically enough. They have, many through no fault of their own, been taught to a prioristically reject the notion of evolution, while at the same time steering clear of any opportunity to learn more about it. Among the many claims they have been taught to rote learn and blurt out in defence of their denialist position, there is one that states that a process driven by random mutations can only decrease and degrade the genetic information contained in an organism's cells, and cannot, therefore, account for the complexity of life in nature. Essentially that mutations can only be deleterious and cripple an organism. This claim strikes at the heart of how the process of evolution works, and it is both factually wrong and logically absurd. Hopefully, by the end of this article, you will see why. You will also see how one can only have a rational, adult debate with someone if definitions of the terms being used are agreed upon. That is one crucial aspect of rational conversation that still eludes many, be they creationists trying to attack a well-established scientific theory simply because they don't like it, or your average politician trying to sway voters. One simply cannot make the claim that mutations only degrade genetic information, and yet refuse to provide an operational definition of what "information" means to them.
Now, creationists are particularly fond of a little routine they do. Whenever they are presented with evidence of a new trait arising in an organism as a result of a mutation - say, a bacterium evolving a modified version of a particular enzyme that allows it to feed of a previously untapped environmental resource - they have a way of construing it as a loss of the original trait. Et voila, they have proven that mutations can only destroy pre-existing information. Or so they think. All they have really proven is that they are truly desperate. By the same logic the evolution of a reptile into a bird would also inherently require the loss of many somatic traits, such as teeth. I'm certain your average hawk would resent the accusation that it is nothing but a degraded lizard. Fortunately for our feathered friend, that's not the case.
Let's talk briefly about information theory. Information theory is a branch of applied mathematics that is concerned with the quantification of information content. Wait, don't run away! No equations, I promise! You see, there is one big construct in this field, called Shannon's information theory and it's involved in communication over noisy channels. Imagine you're on the phone with a friend who's inviting you to a party. He tries to give you the address, but the signal is rubbish, and so you end up on the other side of town at a furry convention. Been there? Well, blame entropy. In Shannon's information theory, entropy is the measure of randomness or distortion applied to the information that is being conveyed by the noise that is on the channel that is being used. It's important to note that this particular model is concerned with the faithful communication of a specific token of information between transmitter and receiver. In this view, nothing can increase the information content of the message, because it's content is determined by the transmitter - much like we cannot add to the content of a letter once we have posted it. All that can happen is that the envelope is lost, delivered to the wrong address, shredded by the huge dog chasing the mailman, or who knows what. Shannon's theory, however, is not the be all and end all of information theory. Creationists have simply twisted it to mean that mutations cannot produce new information, even though it simply does not apply to evolution. Let's think back to our telephone analogy. In it, you and your friend are trying to communicate using a common code, a language. You are using this language to convey a meaning that is established prior to its communication. Your friend needs to convery an invitation, and therefore produces the utterances that represent an invitation to a party in the particular language being used. DNA is often referred to as the language of life, but that is a deeply flawed analogy. You could actually say that the genome itself is not really information. DNA is just a long molecule. While language is used in communication between more or less conscious and intelligent beings, DNA is a long molecule, used by another set of molecules to produce further molecules. At no time during the transcription of DNA into RNA into proteins do the enzymes involved have any understanding of the protein structure that will eventually result. As such, we cannot think of DNA as a text being written with any intent, and since there is no original DNA message that needs to be faithfully conveyed from a transmitter to a receiver, we also cannot apply Shannon's information theory to measure the information content of an organism's genome.
One could certainly claim that errors in DNA replication and transcription can have disastrous consequences, and that's entirely correct. That is not the issue. The issue is that while noise in communication will invariably degrade the original message, the noise produced by mutations has no original message to degrade! Furthemore, any qualitative impact of a mutation cannot be independently judged simply by looking at the genetic sequence it affects. Let's not forget that the process by which mutations generate new information - and I use the term with all the caveats I have just introduced - has no bearing over its quality, over whether or not the new information will be useful for the organism. If we wish to push the language analogy to the limit of its usefulness, the whole process is rather more similar to one person generating random strings of letters that may or may not represent "information" depending on the reader. Let me give you an example: take a string of letters, something like "vervoerdersaansprakelijkheidsverzekering." Now, I will assume it will strike you as meaningless gibberish. However to a Dutch speaker it should mean something like "carriers' liability insurance." Meaning - and therefore information - is a function of the interpreter being used. Better said, information is only really tangible upon interpretation. In the case of genetics, the genetic code is only discernible information when it is interpreted, when it transforms from genotype to phenotype, from complex, three-dimensional molecule into the resulting organism in its entirety and in its ability to interact with the outside world. The quality of the genetic cipher can only be assessed relative to the environment the organism lives in. While many mutations will prove deleterious, others - as a matter of fact the majority of them - may have no effect whatsoever on the protein product. A few may offer an advantage. There is no way to tell just by reading the sequence. Their protein "meaning" must be observed in action, and the ever-changing environments most organisms have to contend with will always have the last word.