Richard Dawkin's big blunder

[Esquilax]

First, I would like to ask you if you can demonstrate cumulative selection without utilizing a target? Has this ever been done?

It's actually kinda begging the question to presume that any natural selection did have a target without demonstrating one...

Second, Do you agree that for any selection criterion, there will exist some set of targets which evolution will home in on?

Yes and no: yes, there are certain ideal traits that reoccur, no they do not constitute a "target" that is being shot for. You're still approaching this backwards: rather than being the end goal, these traits just bear out to be the most useful in a set of given circumstances, and so those mutations that lead to them get favored.

Imagine it like a river, that branches of periodically. Some branches have rocks blocking the way forward, some don't, and looking at the whole you can see that those pathways that don't get blocked off end up having more smaller branches off of them. Would you say then that the river is "targeting" non-blocked paths, or would you just say that this is physics acting on environmental factors?

Third, specifying the precise phenotype is just a lazy way of programming a selection criterion. Suppose the target sentence was "I am". He could write a selection criterion that homed in on this sentence just as well as it homed in on his precisely stated phenotype. For example, His program could favor 4 character sentences(I'm including the space character). His program could favor more vowels than consonants. His program could favor the characters "I", " ", "a", "m". His program could favor sentences where vowels proceed consonants. His program could favor sentences in which the vowels are in reverse alphabetical order....so on and so forth. Instead he wrote a selection criterion that favored one specific sentence because it was easier.

That's why he called it a cheat, precisely because it was simple and easy, yet did not reflect the true process.