I think I know what I'm talking about. Please refrain from assuming I'm basing my argument on some misconceptions unless you clearly see that. NS is a process

I think I know what I'm talking about. Please refrain from assuming I'm basing my argument on some misconceptions unless you clearly see that. NS is a process for selecting changes. It doesn't introduce any change. NS does not produce variations. If NS produced variations then it would be sufficient on its own, and no one would go looking for other causes of variation, such as mutations. NS is not the guts of evolution, it's only a part of it, and by itself it is useless. This is actually a testable statement. If NS by itself is sufficient for producing variations and selecting them, no one would go looking for some other source of variation. From what I understand, Darwin himself didn't think too much of explaining where variations come from; they seemed to occur naturally, so it doesn't matter where they come from. This doesn't work anymore, so people had to actually think and fill this gap. The guts of evolution, really, is producing variations. For evolution is nothing but the progression of species by having small variations accumulate over a long time. NS without variations to choose from will do absolutely nothing to advance the cause of evolving any species. The patching/test analogy works great, I don't know why it didn't make sense to you. NS by itself doesn't do anything, for the same reason that testing by itself doesn't actually develop anything. Suppose you have two teams, team A and team B. Team B does nothing but testing. They test and test and test, and they only accept something of good quality, and no code makes it into production unless it's passed the testing by team B. Team A do the actual development, they use a decentralized version control system, and every change is made into its own branch, and it's then sent to team B to be tested. Team B doesn't by itself make anything; they only accept or reject changes done by team A. If team A doesn't send any changes to team B, the software will not evolve. Similarly, NS can only select variations, it doesn't produce them. Is that analogy really hard to understand? There has to be some other forces that create these variations, and the proposed idea is that these variations are caused by random mutations. What tends to happen in evolutionary literature is, the role of NS is magnified, and the role of mutations is marginalized, as if it's the selection process that matters, not the variation-producing process. The mere occurrence of mutations is not the subject of my argument. The part that is often overlooked, even though it is so crucial, is the lack of evidence for whether mutations have the ability to actually produce any useful set of variations for NS to act on. It's assumed that eventually mutations will always produce something useful, but this is a claim that lacks evidence. This is the heart of evolution; the mechanism that produce these small changes that eventually accumulate. This is ten times more significant that NS. Darwin focused on NS so much, but that's because he assumed that NS is the force that drives the variations. It's as if competition would motivate individual animals to better adapt themselves to their environment, by extending their necks, and things like that. He even talks a lot about the competition! If you really think about it, he's making an analogy with market competition: market competition drives people to provide better services/products. The fact that people will make better product is not so important, because they always can, but who's to force them? If a crappy product sells, you won't be motivated to improve it. Competition forces you to improve your products. Darwin must have thought the same about natural selection: it doesn't matter where variations come from, they just naturally occur. This is where neo-darwinism comes in. It says, no no, that doesn't really work. Variations come from mutations. Once you say that, NS stops becoming so important, and the mechanism that produces the variations becomes much more important. Natural variations in the gene pool (like hair color) cannot contribute to evolution, because they always revolve around a closed circle. Suppose I have 3 options for hair, each option having 5 possible choices: * color: white, black, red, yellow, blue * type: straight1, straight2, curly1, curly2, curly3 * length: really-short, short, medium, long, really-long That's 5^3 possible variations right there. It's really easy to imagine having more such options. You can have 20 trivial options, each having 2 choices. That would be about a million possible variation (2^20). You can try to mix and match any number of configuration for these options, but it would never result in a new species even after a billion years. The only mechanism for actually creating new kinds of variations that can lead to the evolution of new species is mutation. As for your self-replicating pixel-drawing program that can evolve to a web browser -- you are more than welcome to have a go at it. I can tell you from now that it will fail. There's a simpler experiment, that will also fail: http://www.randommutation.com/darwinianevolution.htm http://www.randommutation.com/darwinianevolution.htm > You seem to be suggesting that it is unlikely for mutation to create useful variations; this is also a fallacy because mutation is random. So despite being unlikely it is not impossible Yet another instance of hand waving. You might think it is possible, but until you prove it, I will claim that this is an unscientific claim. > remember we are not talking about a mutation that suddenly means the offspring grows a lung - that is a common misconception. I never suggested that a single mutation must produce a lung. This is a common technique. I might call it "The argument from you don't really understand". I think you're the one who's not understanding my argument. Suppose complex feature X requires 100 steps to come about. Each step is a tiny change in something that already exists, but after a 100 steps, it will be basically transformed into X. Each such step must be producible by a mutation. Not only that, but each step must be advantageous enough to be selected by NS. (Notice how NS here actually makes things more difficult). A small example: the word "yellow" can be transformed to "hello" by two steps: yellow -> hellow -> hello The problem here is, "hellow" is not a valid word, and therefore NS would fail it, and this chain of variations will not work. Of course this example doesn't exactly apply to genes, but it clearly illustrates two points: - Each "tiny step" must be producible by a random mutation - Each such step must pass NS. Now, from reading the brain-simulation thread the other day, I sort of conclude that no one really knows what these steps might even look like, because genetic engineering is really hard, and the way genes translate to features is still largely a mystery. > The theory of evolution has survived this challenge a number of times without being proven fallacious. Not really. Of course, you will ignore everything I'm saying, because instead of focusing on evidence and logic, you're fixated on the notion that I'm an idiot (or misinformed) and therefore my argument is obviously false. I mean, how dare I question evolution? Right?