Functional Information (again)

Before the math.

Nope. “Number of functional sequences that still work” seems to be saying “number of functional sequences that are functional”. Perhaps colewd thinks that in all such calculations we have, for each possible sequence, a 0/1 according to whether it works (1) or does not (0). If so he is completely wrong, as Hazen and Szostak consider cases where each sequence has a number indicating how well it works, and that number can be anything from 0 on up, not just 0 and 1. It is clear that colewd does not understand what Functional Information is. Nevertheless he wants to straighten us all out on what it is.


I means exactly what is in the explanation in the paper.

fraction of all possible configurations of the system that possess a degree of function ≥ Ex .

So we need to identify the number of working sequences as all possible sequences is a straight forward calculations.

What I meant to say is will work vs still work. Thanks for the correction but perhaps next time you could do better then posting in the manner that Gunter Bechly accused you of.


Empirical research on frequency of function in sequence space for the gene/proteins being measured.

fraction of all possible configurations of the system that possess a degree of function ≥ Ex .

So we need to identify the number of working sequences as all possible sequences is a straight forward calculations.

No, we need to find the number of sequences that work as well as the given sequence or better. That’s what the quote from Hazen and Szostak says. Not just “working”.

So colewd is wrong. And apparently I’m being “vitriolic” when I say that.


Well, ID Creationism has very high standards of polite behavior. It’s best never to point out how wrong people are, as this is the worst of the worst sort of vitriol. Far better to make video of them making fart noises, like Dembski did with Judge Jones. That’s well within the range of polite discourse, so if you’d done that rather than simply pointing out the error, nobody could possibly object.


You are being exceptionally obtuse, especially since the answers were given to you (comments #54,56,57). The first thing is to select a system and level of function, then you can do the math.
FI will be dependent on the function chosen (if there is a choice, as JH notes in #55/58. If the observer switches interest to a different (possibly related) function of the same system, then the FI they measure will be different.

None of this has any bearing on the probability that a system will evolve, it is describing properties of systems that already exist.

No one is saying the paper is wrong. Everyone is saying that you do not understand the paper. It doesn’t mean what you seem to think it means.


You simply need the quantity of sequences that fits the function Ex. Why are you complicating this?

This is what I stated several posts ago with one exception. You do not need to select the level of function. This is what is empirically observed by the number of functioning sequences and dividing that by the total sequence space.

Functional information is very problematic to evolutionary theory or universal common descent as the theory cannot explain the estimated quantity calculated (conservatively 20 million bits for human exons) between the simplest species to the most complex. It is another challenge to the simple to complex model.


Except that your own favored definition of Functional Information (Hazen and Szostak’s) does not measure information in bits – it does so as the negative log of a fraction.

So what, Bill Cole, is the actual measurement of FI, according to the Hazen and Szostak definition, of “the simplest species”, “the most complex” species and a “human” respectively?

It is becoming blatantly obvious to me (as it has to others on this thread) that you have absolutely no idea how Hazen and Szostak’s definition works in practice.


That doesn’t matter. The rhetorical strategy here has never, in the hands of any of the IDC proponents, been expected to make actual sense. It consists entirely in the shouting of the word “information!” over and over again, interspersed with vehement, but completely fact-free, assertions that this new elan vital, infermayshun, is something evolution can’t explain.

This does seem to fool a lot of people. But the constant silly assertion that somehow the mathematics of information theory demonstrates the impossibility of evolution is just another example of railing at the bees: prove, via math, that bees can’t fly, and then, when the bees keep proving you wrong by simply flying, rail at the bees instead of evaluating what’s wrong with your math.

Here, of course, the actual math has been dispensed with, which makes it even simpler.


I think the legerdemain works better when they use their own special ‘magic pixie dust’ definition of “information”. When they use one created by actual scientists, it tends to come with baggage like actual meaning, actual constraints, and an expectation from the better-informed members of your audience that you should understand how the whole thing works. That can get in the way of shouting and gesticulating effectively.


Because Szostak and Hazen complicated this, and you say you are using their definition. They didn’t say “fits the function”, they said

possess a degree of function ≥ Ex .

Why are you oversimplifying this?


Perhaps this is an opportune time to remind everyone of one of the pivotal moments in the history ID’s attempt to demonstrate that information proves Goddidit. I remember when this thread was first created and watching with astonishment as the replies grew into the hundreds, without even an attempt to provide an answer.


What is the difference in these statements in your opinion? How does it affect calculating FI?

What is the difference in these statements in your opinion? How does it affect calculating FI?

My understanding of H&S’s calculation of FI is that

  1. You decide what “function” you’re going to measure,
  2. You calculate it for each possible sequence
  3. To calculate the FI for one sequence you consider its function value f(x)
  4. You determine what fraction P of all sequences have values >= f(x)
  5. The Functional Information for sequence x is then -log2(P)

I have the distinct feeling you’re doing something else when you ask who “fits the function”. Am I right about that?


I agree with your process 1 thru 5. I apologize for the sloppy language.

1 Like

What is the difference in these statements in your opinion? How does it affect calculating FI?

Bill, if you really want to understand,then you should be trying to explain these to the rest of us.

A single sequence will have a single level of function, but a system may have multiple levels of function. There is no point in "the proportion of sequences with function \geq Ex" if there is only one level to consider. Thus those nifty graphs with level-of-function on the Y-axis.


As I said I agree with the process Joe articulated. I understand his statements to be the process of estimating functional information.

I am not sure the point you are trying to make here.

You will note that there’s nothing in there about complexity. Note also that FI can easily increase if the current sequence is not at the maximum.