Complex Specified Information
|The divine comedy|
Complex Specified Information (CSI) is a concept created by William Dembski of the Discovery Institute. This concept is used specifically to support the intelligent design hypothesis. CSI is a measure of the quantity of information contained in an item that meets the requirement of being both complex and specified, which shows whether the item is either designed or naturally occurring.
There is no scientific evidence that this concept holds in practice, and where Dembski and company have attempted to study it they have insisted on doing it in computer science rather than biology.
One easy illustration of Dembski's idea is if you randomly draw out 5 cards from a deck of 52 cards, the sequence of cards you have is highly improbable, but as long as the cards do not fit a preconceived pattern, it can be explained by chance. But if the cards fit a pattern (such as a royal flush), then it must have been designed and couldn't have happened by chance. Dembski, of course, never explains how the fact that both a "random" sequence of cards and a "royal flush" both having the same probability allows us to say one must be designed but the other cannot be. Nor does he address the more complex issues of evolution, such as scaffolding or co-option of function.
 Irreducible Complexity
An integral part of the concept of CSI is the probability of a particular configuration. The calculation of this probability is complicated substantially if it is made up of sub-parts that can arise separately; CSI is therefore most valid when applied to a configuration that is irreducibly complex. A commonly presented example of alleged irreducible complexity is the bacterial flagellum. Note that this claim has been substantially refuted. A computer program that demonstrates evolution has been found to produce complex specified information during the process of mutation and selection.
Under no circumstances should these be considered examples of CSI:
- The pattern of relationships among all living things known as the "tree of life". Even though the same nested hierarchy shows up repeatedly wherever one looks, and in every form of life, whether living or extinct, in comparative anatomy, biochemistry, DNA, distribution in space and time, embryology, and who-knows-what-next. Even though this is much more complex than the bacterial flagellum, and actually makes predictions (which some might think of as a real standard for being "specified" in advance).
- In particular, the pattern of similarities between the human body and the bodies of chimps and other apes, and to lesser degrees, other mammals, other vertebrates, and other living things. This is a really sensitive topic.
- The various systems of dating of things to more than several thousand years, whether by radioactivity, chemistry, astronomy, geology, or anything else that anybody has ever thought of. While many of the advocates of CSI would probably personally not be averse to "old life on an older earth in an older universe", it's probably not advisable to make too much of it.
The issue of falsifiability is unclear from the hypothesis. High values of CSI show design, but is it possible to show a natural occurrence since design might be imposed in the genome of a living creature or the laws of physics that control for example chemical interactions?
 Begging the question
Classifying something as "CSI" requires one to calculate how probable that it would arise "naturally". Thus, saying that there are examples of CSI in biology is simply a fancy way of saying that evolution is very unlikely. Dembski's calculations as of the probabilities basically assume that evolution is false. Furthermore, the treatment of design as being separate from "natural" processes pre-supposes a dualistic point of view. For example, he gives Shakespeare's sonnets as examples of CSI. But if evolution is true, then Shakespeare, and hence his sonnets, are the result of natural processes.