Intelligent Design and the Information Argument

One of the biggest arguments the Intelligent Design (ID) movement use revolves around information. At its simplest, they claim that a process without intelligence - such as evolution - cannot produce information, and therefore cannot lead to the rich diversity of life we see around us.

I have never been that clear on how IDists actually measure this quantity, despite it being a big part of their arguments. They seem very sure evolution cannot create information, but how can they say that without a clear way to determine what it is?

How can we discuss information in regards to evolution if we have no idea what it really is?

Even the Discovery Institute are clear that we need a way to measure information:

How does our definition of information apply to biology, and to science more generally? To render information a useful concept for science we need to do two things: first, show how to measure information; second, introduce a crucial distinction–the distinction between specified and unspecified information. ​

https://www.discovery.org/a/118/

In my view, it is really up to ID to provide an answer. Different ways of measuring information can give very different results, and a claim made using one system may not be true in another. If they are saying a natural system cannot create information, they need to be clear what they mean by information. What I want to do here is to highlight some issues and hurdles that must be considered.

As I starting point, suppose I have a sheet of paper with text on it. Let us suppose it has 500 words on it, or 3000 characters. It was printed from a Word document that is 17 kB long.

How much information is that?

The Data​

Does the information include:

• the dimensions of the paper, spatial coordinates and orientation (important if this was a clue in a video game)
• the font family and size
• the page layout
• the composition of the paper and ink

With regards to the text, are we measuring the information assuming a seven bit character set or UTF-8 or UTF-32? If we compress the text file, has the amount of information changed? Does it depend on the compression algorithm?

The Eye of The Beholder​

If the text is just random characters does it have less information than English prose? If I later realise actually it is in Polish, does it now have more information, even if I cannot read it? Suppose I hand it to someone fluent in Polish, does it now have more information? Or did it always have that information?

Does a work of non-fiction have more information than a fiction book; that is, are the facts in the text the information, or do they add to the information? If so, can we measure the information to see how accurate it is? Or do we need to verify the claims to determine the information content?

Double The Information?​

If I have two identical sheets of paper, do I have twice the information? Or is the information the same if it is merely a copy?

If I now realise there is a paragraph that was accidentally repeated in the text, does that mean there was less information than I originally thought?

Actually, I can now see that the word "deoxyribonucleic" is repeated eight times? Does that reduce how much information? The word "and" is used nineteen times, does that have an impact?

By the way, this is very important to the issue of evolution because an important type of mutation replicates sections of DNA. Does that lead to there being more information? How about when the copy is altered?

A Revised Copy​

Suppose the text was edited slightly correcting some typos and updating the factual information, so now I have the original, and a revised version. Is that twice as much information as either one on its own, or is it the same?

Does the revised version have more information that the first now the information is more accurate? Does correcting a spelling mistake increase the information or decrease it?

This leads us into...

Information Content vs Communication​

There is a problem here that a lot of discussion about information is actually about the communication of information. There is a difference. The sheet of paper above contains information, but in the above discussion there is no sense of communication of the information.

A major point of difference is that in the communication of information you start with a message at point A and want to get the same message to point B. Any deviation in what arrives at B is - in most situations - considered a bad thing. As a simple example, if a sentence is sent from one person to another, and random noise on the line happens to correct a typo in the original, this would still be seen as a loss of fidelity; a deterioration of the information.

This is certainly the case for Shannon's view of information.

The nature of the system is that information can never increase when it is communicated because by definition any change is considered bad. This is quite different to a genetic mutation, which can, at least in theory, be good or bad.

This can lead to the erroneous belief that evolution cannot create information - and is why Shannon's information is so popular with IDists. But that does not follow because evolution is not in the business of faithfully transmitting information. This is why a consideration of information communication is not appropriate to a discussion about evolution.

Think about the revised copy discussed in the previous section. According to Shannon, the correction of a typo and the updating of the factual information in the revised copy represents a deterioration in the information transmitted. To Shannon, information has been lost in the revised copy. And yet by any normal standard the revised copy is superior to the original.

I am not saying Shannon is wrong, but I am saying his approach is not always applicable and is not applicable to evolution.

The DI Approach​

The Discovery Institute, in the article linked above, offer a probabilistic approach that they use with poker hands. However, this is a very contrived situation; a poker hand is nothing like a sequence of bases in DNA! I appreciate they are citing Shannon's work, but that is about the communication of information, not the information intrinsic to the thing.

That is, we are not concerned with how successfully an organism transmits its genetic information to the next generation, we are concerned with how much information is in the organism's genome.

I do not see how this can be applied more generally. Which brings us to...

Mount Rushmore​

This is a great example of something that we would expect to be able to measure the information of... if the ID argument has any validity. The Discovery Institute even cite Mount Rushmore as an example of something that is designed, so this is a great example to look at.

Specifically, we should be able to measure the information across a number of mountains and determine that Mount Rushmore has significantly more information.

How can the probabilistic approach the Discovery Institute showed us with the poker hand be applied to Mount Rushmore? Can another approach be used?

If we cannot do that, that would pretty much destroy the ID argument - especially given is an example IDists bring up themselves.

So where is the calculation?​

Of course, IDists have never tried to calculate the information in Mount Rushmore.

I think on some level they all know it is pure hokum. It is just a façade - something that looks reasonable to a casual glance, but has no substance.