In last week’s Creationism seminar, Donald James Batten, B.Sc.Agr. (Hons 1), Ph.D. ended his talk with a calculation showing how unlikely it is that a particular protein would form via random chance alone. I did not verify all the numbers in his calculation. I see no need to, because his calculation was completely unrelated to evolution. By implying it had anything to do with evolution, he merely illustrated his own lack of understanding of the subject matter. And yet, many people are credulously listening to his rhetoric, under the impression that he does understand evolution…
Dr Batten’s Calculation
He started out by calculating an upper bound on the number of “experiments” that could be performed, making many extreme assumptions in favour of extremely many experiments. Using (I assume) “secular” scientific estimates for the number of atoms in the universe and the age of the universe (14 billion years), he calculated the number of experiments assuming you could run a thousand experiments per second, for every atom in the universe. Cute. Or rather, absurd, and intentionally so, as he is trying to ridicule evolution.
Next he calculated what the odds are that selecting 300 amino acids randomly from the 20 “standard” amino acids would produce a very specific protein (consisting of 300 amino acids). This produces a number of the order 10390, which is incomprehensibly larger than the number of experiments in used for the upper bound in the previous paragraph.
He is suggesting that every possible combination of amino acids would need to be tested (with an “experiment”) in order to find the correct one, and illustrates (correctly) that such an exhaustive search through every possible combination would indeed be impossible.
Dr Batten’s Errors
There are two very significant errors in this calculation, if it is supposed to have anything to do with evolution.
- The first is the error of characterising evolution as a random process. If a particular protein could have developed through a sequence of mutations, with each step in this sequence being useful in its own right, an exhaustive search through every possible combination of amino acids is unnecessary. You do not have to look everywhere to find the highest peak on your local hill, you only need to make a sequence of successive steps, each taking you a little bit higher, never having to look further than twenty meters in any direction. Or stated differently, you can climb the hill from the east or from the west, there is no need to climb it from all sides.
In my opinion, the second error is the greater one. It is the assumption that any particular protein must develop. This mistake has a lot in common with the kind of mistake that makes us believe our planet is the centre of the universe, with the sun orbiting it. Just because we have a particular protein, does not mean that intelligent life needs that particular protein. Life could have developed with a completely different set of proteins, or even without protein as we know it.
To explain this with the hill-climbing metaphor: the significant mistake in Dr Batten’s calculation is to assume that we are at the highest peak of the highest mountain of the evolutionary landscape. In order to reach such perfection, a search through all possible “proteins or equivalent” would indeed have to be much more thorough. With our twenty-meters short-sightedness, we would effectively have to visit every possible place in the evolutionary landscape, simply to see if it isn’t at a higher altitude than our own local hill. Evolution and genetic algorithms only find local optimums. And we did only climb the nearest hill. There are very many other hills that could also have been climbed, evolving intelligent life not based on any of the proteins we know, just like there are animals with much more advanced eyes than our own (for example the squid).
If you throw a ball in a random direction with a random velocity, the probability of it landing in any specific place is very small. However, the probability of it landing somewhere is 100% (in our simplified model). Once it has landed somewhere, arguing “wow, that could not have been random chance, the probability of landing specifically there is so minutely small!” would be incorrect.
For the record, to make our hill-climbing metaphor a more complete metaphor for evolution, the evolutionary landscape is not stable. The hills and mountains are constantly changing, with evolution constantly climbing up the new nearest hill. The forces that shape this evolutionary landscape are fascinating, including the likes of predator-prey interaction, environmental concerns, food shortages etc.
Are there any questions at this point about the statistics involved? Any further questions about Dr Batten’s calculation, or why it is incorrect?