Me Thinks it Looks Like a WeaselIn the last example, the computer program written to test evolution was obviously flawed. After 200,000 tries, the code executes a line that calls the experiment, claims it is successful and displays the result of a successful experiment. Usually, the flaws in computer simulations designed to test evolution are much more subtle. The above demonstration will illustrate this. Its inspiration comes from the famous Richard Dawkins "me think it looks like a weasel " applet. Dawkins is a best selling author whose unwavering conviction to evolution and atheism has probably done quite a bit to promote intelligent design. But here I want to concentrate on some flawed computer simulations he did many years ago. He like many scientists assumes that chance is not a dominant factor because natural selection guides everything. In his simulations, he could type in any target phrase (for example: me thinks it looks like a weasel), and watch a very different starting phrase evolve towards the desired target. The target phrase was always found and it was found very quickly. At one point in time there was a Java applet available online that would duplicate these results. The above demonstration duplicates the logic behind the Dawkin's program. The program breeds strings of characters and creates mutations (it changes letters). The program is fast because it only considers offspring with mutations. In other words, while the program only creates 5000 offspring with the create offspring button, all of these offspring have mutations (some as many as 11 mutations), so the population of animal needed to create these mutations would have to be very large and may require a hundred thousand generations to create this many mutations.
|
Directions: 1) Click Create Offspring. This will create 5000 offspring based on the random letters designated "here are your letters." Some of these offspring will have 1 mutation (1 changed letter) others may have up to 11 mutations. 2) Score each mutant with the three score buttons. Pick the one with the highest score. This will change the starting point to the selected string of characters, so click ok to accept the new string.. (If two or three scores are equal just pick one it does not matter which one). When each score button is pressed, 500 mutants are evaluated and the score and characters of the one that is most like the target phrase are displayed. The target phrase is not shown so try to guess it as you go along ( this is important). 3) After picking the highest scoring mutant, click create offspring again. Score each mutants as before, and pick the mutant with the highest score. The score will gradually work its way up until you reach 18. At a score of 18, the letters will be very close to the target, so the program will reveal the secret answer in the last box. This take about 5 minutes to find the right answer. Thus, it is a very powerful demonstration of evolution and natural selection. Almost all computer programs written to test evolution follow this pattern So does this program really model evolution? If you as an intelligent reader cannot figure out the target phrase when
the score is less than 14, then how can natural selection do so? The answer is that it
cannot. Natural selection cannot guide gibberish and help it to evolve into something
useful. This is the role of chance. |