If evolution is not a search, why is the term “evolutionary search” not an oxymoron?
Over at Uncommon Descent Elizabeth posted the following:
“…any “search” algorithm worthy of the name of “evolutionary search” comes with its own moderately smooth fitness landscape built in.”
So evolution is a search if it comes pre-built with its own moderately smooth fitness landscape built in?
Great. We agree. Now, recall that the comment of yours that I reacted to was at UD and said:
So I think we now see that what I said was not “just false”, because the “unless” condition held: Every process that occurs in Dembski, Ewert and Marks’s search space Ω and that gives a result can be represented as a probability distribution, and hence is, according to DEM a “search”, as they define the term.
Where Tom English and I went from there was to point out that when you required that there be genotypes which had fitnesses, the probability of getting a point in the Target was much higher than in all the “crazy” searches, such as ones that actively try to get worse fitnesses. Thus the presence of genotypes with fitnesses is enough to buy us some Active Information, and if the fitness surface is smooth, enough to buy us quite a lot.
Joe, I think you are focused on “even crazy ones” while I am focused on “every process.”
Here’s how I took what you wrote:
DEM call every process a “search”
So when I wrote:
Well now we know that’s just false.
I was referring to the portion of your statement that claimed that:
DEM call every process a “search”
But now I think we’re just covering again ground that was already covered.
If, when you wrote:
DEM call every process a “search”
What you actually intended to be understood was:
DEM call every process that they define as a search, a “search” …
Then how can I disagree with that?
As far as I am concerned. “even crazy ones” is totally irrelevant and not at issue.
I don’t know how many times I need to say this, but here it is again.
DEM DO NOT define every process as a search.
Of course, if every process can be be represented as a probability distribution, then it follows that every process is a search, as defined by DEM.
“all that we mean by search is a process that can be represented as a probability distribution.”
So apparently we’re right back to talking past one another.
Not at issue, Joe. Was never at issue.
Is EVERY PROCESS a process that occurs in DEM’s search space omega?
They have email addresses. Ask them?
Mung, you are right that not “every process” is “search” as defined by DEM. Mea culpa. Here are three processes that are not a search as defined by them:
1. A dog wagging its tail.
2. Me drinking orange juice.
3. My neighbor washing his car.
Here is something that is a “search” as defined by DEM:
Any process that results in an element from DEM’s search space Ω
You have said you agree with that last bit.
So now let’s take the argument one cricial step forward. This is your own thread,, on Evolutionary Search.
If we consider evolutionary processes, do they include ones that deliberately reduce fitness? No, they won’t. Does this mean that we already have substantial amounts of Active Information, just from that? That’s what Tom English and I argued.
It sounds like Mung agrees with that. So Mung agrees that Active Information is present as soon as we have results from reproduction of organisms that have fitnesses. No Designer needed once the organisms and the fitnesses are there.
Right?
Glad to find that Joe and I can actually have a meeting of the minds. Short lived, though.
I was not aware that evolutionary processes are deliberative. At this point, you’ve lost me.
Given substantial amounts of pre-existing Active Information, does it follow that no Designer is needed? No.
Joe F:
Mung:
Poor Mung.
Joe is talking about things like hill-descending, as opposed to hill-climbing, algorithms. Remember fitness landscapes?
Poor keiths.
What is the relevance of deliberately reducing fitness? Who or what is doing the deliberating?
Let’s rephrase for clarity.
1. If we [keiths, JoeF, TomE] consider evolutionary processes, do they [DEM] include ones that deliberately reduce fitness?
2. If we [keiths, JoeF, TomE] consider evolutionary processes, do they [processes defined as a search by DEM] include ones that deliberately reduce fitness?
3. There exist evolutionary processes that deliberately reduce fitness. Do they [DEM] include them [evolutionary processes that deliberately reduce fitness]?
Look into your crystal ball, keiths, and divine what Joe F meant to say.
Keep fighting, Mung. You can keep the light of realization unlit forever.
keiths, having failed to re-interpret Dawkins, also fails to re-interpret Joe Felsenstein.
But it’s not his fault. It’s never his fault.
Sorry to have missed this reply by Mung two weeks ago. Mung and I had apparently agreed as which processes constituted “searches” in Dembski, Ewert, and Marks’s argument:
Let me take those in turn. DEM do not raise the issue of where the “searches” come from, or whether they involve anyone doing any deliberating. They simply consider a space of all possible “searches”. They show that the average probability of reaching a “target” T is the same as a blind search that draws one element from the space Ω
Then they compute the quantity Active Information from how much more successful the actual “search” is than that.
Tom English and I, in our response to DEM, argued that just having organisms with fitnesses got us a fair amount of Active Information. No extra Designing needed to get that.
Have we proved that “no Designer is needed”? Of course not. What we have proved is that having some Active Information is not, in itself, evidence for there having been Design involved in ensuring the success of the organism. It can simply have genotypes that have different fitnesses, and that gets some noticeable Active Information.
So the issue of who or what is “deliberative” is not part of DEM’s argument, or of ours. And Tom and I make no pretensions that our argument eliminates any possibility of having a Designer. Just that the presence of “Active Information” does not establish that a Designer was involved.
So are we agreed then, that seeing Active Information may only establish that there were some organisms, that had fitnesses?
Let me add, since I was the one who used the phrase “deliberately reduce fitness”, that we can replace that by simply saying that one of DEM’s possible “searches” would be a process that had fitnesses for each genotype, but which at each step moved to the neighboring genotype that had the smallest fitness.
The issue of who or what is “deliberating” does not come in.
Such a process (one that has fitnesses, and ends up going downhill on the fitness surface) is one of DEM’s possible “searches”.