Can mutation and selection increase information?

It is natural selection that is goal directed, not random mutation. Random mutation increases information, selection reduces it.--Percy

All you have is a semantic argument, as usual. What would it look like if random mutations were not goal directed? How would they differ from goal directed random mutations? Can you show us a single peer reviewed paper that uses those terms in those ways?

Only if you're talking about Shannon Information and even Shannon acknowledged that had nothing to do with meaning. Yes.

Yes, of course. More like he stated it right up from in his seminal paper:

quote:

---

The fundamental problem of communication is that of reproducing at one point either exactly or approximately a message selected at another point. 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.

---

So, yes, random information increases Shannon information. Selection reduces it.--Percy

Gosh, you never got around to an operational definition of the kind of information you are talking about.

Think of it this way.
I have a copy of Origin of Species which contains information.
I buy another copy. Do I now have twice as much information? Or 2 copies of the same information?
I tear a page out of one copy and glue in a page of random numbers. That's a deletion and an insertion. Does that give me more information?Werner Gitt in In the Beginning Was Information distinguishes 5 levels of information of which Statistical (Shannon Information) is the lowest level.

Think of it this way.You're studying the population of a certain organism. One of the offspring experiences a mutation that gives it two copies of a gene that produces a certain protein. The organism now produces twice as much of that protein as before resulting in differences from others in the population. So of course gene duplication increases information.Mathematically, the number of bits required to encode the number of genes of an organism is log₂(number-of-genes). Naturally log₂(number-of-genes + 1) is a larger number and therefore more information. Werner Gitt confuses information with meaning.But repeating the answer to your original question about whether mutation and selection can increase Shannon information, random mutation increases information, selection reduces it.--Percy

Depends on your definition of information. What's yours? Depends on your definition of information. What's yours?

No, he just uses a different definition of information than Shannon; must people do. Shannon's definition is very restrictive but suited for the purpose for which he was using it. Random mutation degrades existing information. Take a page from a book and start randomly mutating the letters and pretty soon it is unreadable and any information on that page is lost.

You're confusing information with meaning. In order to measure information mathematically you must use Shannon information.Here's an example using a single gene of a population. Imagine that in this population the gene has 4 alleles. The amount of information for this gene across the population is: Now imagine that one of the offspring in this population experiences a mutation in this gene to form an allele that is different from the existing 4 alleles. Now there are 5 alleles for this gene in the population, and the amount of information is: So the amount of information in the population for this gene has increased from 2 to 2.32.If you have some other way of measuring information, let's hear it.--Percy

You could sort of use Kolmogorov complexity.But we all know that CRR will never provide an operational definition of what he's talking about. He doesn't have one and if he did it'd be certain that under some circumstances it increases.

How does that apply to biology?Can a gene duplication change phenotype? If yes, does that count as an increase in information by your definition?

But is that definition applicable to biology? Then you would define every difference between the chimp and human genomes as a loss in information. In other words, your definition is not relevant to biology and evolution since evolution can proceed without needing to gain information as you define it.

How can you know whether the information is "degraded" before it is filtered by selection? What if a mutation makes an organism fitter in a new environment while the unmutated individuals remain fitter in the old environment? Isn't that an effective increase in information?

Which has more information a pig or a cow?