Why I Don’t Believe in the Butterfly Effect, Part 9
Part 9 of 12: How I Would Interpret Lorenz’s Observations
Given how much the idea of the butterfly effect has become a trope in science fiction movies and YouTube videos, it seems strange to remember that it started out as a serious scientific idea. As explained in Part 1 of this series, the butterfly effect was proposed in the 1960s by the meteorologist Edward Lorenz, who noticed that when he changed the initial conditions of his weather model by a tiny amount, the resulting weather would be completely different after a few weeks’ time. He popularized this phenomenon, and over time, it became known as the butterfly effect.
Weather models are far more advanced now than they were in the 1960s, but I believe that some of them, mostly the deterministic ones, still show the extreme sensitivity to initial conditions that Lorenz found. (I could be wrong about that.) And meteorologists still believe in the idea of the butterfly effect, at least to some extent.
I have spent this whole series of essays arguing that the butterfly effect is not true. This begs the question: if I think that the butterfly effect isn’t true, then how would I respond to Lorenz? How would I interpret the extreme sensitivity to initial conditions that scientists (like Lorenz) have observed in their models?
My answer is twofold: 1. My problem with Lorenz’s ideas is not with his results; it’s with how he interpreted those results, 2. My arguments against the butterfly effect were based on philosophy and logic, not on science.
It is true that if you’re using Edward Lorenz’s weather model, and you change the initial conditions by a very small amount, the resulting weather will end up being very different. But that does not imply that “a butterfly flapping its wings in Brazil creates a tornado in Texas.” That’s because that is assuming that there is only one node in world history. Lorenz’s results are true, but the fallacy lies in taking those results and turning them into a causal statement in which there is assumed to be only one node in world history. This the fallacy that I explained and debunked in Part 4 of this series.
To make this point clearer, I will use another example of the butterfly effect in pop culture. Some years ago, I remember reading a strip of the comic strip FoxTrot in which the character of Jason Fox, an extremely nerdy ten-year-old boy, promotes the idea of the butterfly effect at the breakfast table, much to the chagrin of his sister. Jason declares, “I could wave my hand like this, and it might rain next month in Malaysia. I could wiggle my tongue like this, and it might be sunny there instead ………… it even makes a difference which way I stir my special blue oatmeal.”
No, Jason, you’re wrong. Your decision to wave your hand is not the only node in world history. The universe either is or is not deterministic. If it is deterministic, then it was pre-determined that you would wave your hand, and Malaysia’s weather next month was also pre-determined. There was never any alternative. Meanwhile, if it is not deterministic, then your decision to wave your hand is just one of millions of nodes in world history, such that you waving your hand does not singlehandedly determine the weather in Malaysia.
As I explained in Part 5, the trick to think in terms of probability over time. The probability that it will rain next month in Malaysia varies over time. By the time we get to next month, it will have reached either 0% (it rains) or 100% (it doesn’t rain). But Jason’s decision to wave his hand only changed this probability by an extremely small amount.
Exactly the same argument applies to the butterfly and the tornado. The statement, “a butterfly flapping its wings in Brazil creates a tornado in Texas” is fallacious because it assumes that the butterfly’s activity in that moment is the only node in world history. It is not. If the universe is deterministic, then there are zero nodes in world history, and everything was pre-determined. The butterfly’s actions were pre-determined, and the future weather was also pre-determined. There was never any alternative. Meanwhile, if the universe is not deterministic, then there are millions of nodes in world history, such that the butterfly flapping its wings still leaves open millions of possibilities.
Again, I do not dispute the results of weather models. I do not dispute that if you’re using Edward Lorenz’s model, and you add a butterfly flapping its wings into the initial conditions, then you might end up with a tornado in Texas, where there had not been one before. But it’s fallacious to conclude that the butterfly caused the tornado or that if the butterfly hadn’t flapped its wings, then we wouldn’t have had that tornado. If you say that, you are making the logical fallacy of assuming that there was only one node in history.
Rather, it makes much more sense to make a graph of the probability that there would be a tornado in Texas as a function of time. And when you make that graph, you can see that the butterfly in Brazil had only an infinitesimal effect on that probability. And therefore, the butterfly doesn’t matter very much.
Other parts of this series:
Part 2: The Butterfly Effect in Pop Culture
Part 3: The Wrong Way to Disprove It
Part 5: Responding to Arguments in Favor of the Butterfly Effect
Part 6: Exceptions (And Why They Aren’t Really Exceptions)
Part 7: Three Wrong Ways to Discuss Alternative History
Part 8: The Right Way to Discuss Alternative History
Part 10: The Butterfly Effect and the Slippery Slope