Scientific Method

We use a particular method of thinking in this blog, which we call the "scientific method". In the following paragraphs we'll attempt to define what we mean by "scientific method", acknowledge that other definitions of the same phrase exist, and then reject those other definitions on the grounds that they are tedious.

In our scientific method, we begin by thinking about the world and dreaming up some hypothesis about how it works. We might conclude from some calculation that natural selection gave rise to evolution. We might conclude that the practice of witchcraft causes crop failure. Maybe we think that electric charge is quantized into tiny little particles that we are going to call "electrons". Maybe we conclude that human carbon dioxide emissions are going to cause catastrophic global warming. A hypothesis is any such unproven claim about the world. Our hypothesis serves an important purpose: it motivates us to perform experiments, observe the natural world, and study existing recorded observations, because we are wondering if our calculations could be correct, and we are even hoping it will be correct, or maybe we will be so surprized to find that it is incorrect that we are intensely curious as to how it could possibly be incorrect. The hypothesis is not a scientific theory. It is not proven. Its only purpose is to motivate us to study the world.

The next step in our scientific method is to adopt the "null hypothesis". The null hypothesis is the assumption that there is no causal relationship between any two phenomenon until proven otherwise. We assume natural selection has no effect on evolution, witchcraft has no effect on crops, electric charge is not quantized, and human carbon dioxide emissions have no effect on the climate. Now we look for observations that contradict our null hypothesis. No matter how compelling our calculation, no matter how attractive our hypothesis, it remains untrue until the null hypothesis is disproved.

When it comes to disproving the null hypothesis, we are not allowed to assume that our own hypothesis is true. We cannot, for example, assume that our carbon dioxide emissions will cause catastrophic global warming, build a computer simulation upon this assumption, and then claim that the catastrophic warming predictions of this simulation disprove the null hypothesis. That would be circular. The predictions of the null hypothesis must itself be contradicted by observations of nature. It is not sufficient for the null hypothesis to be in disagreement with our own hypothesis. It's obvious that the calculations that lead to our hypothesis disagree with the null hypothesis. They must disagree, or else we would have no reason to start testing the null hypothesis. No disagreement between our own hypothesis and the null hypothesis absolves us of our commitment to the null hypothesis.

When the null hypothesis is not consistent with observations, we abandon it and adopt a new hypothesis. The new hypothesis must contain as few causal agencies as possible, and it will become our new null hypothesis. For example, we observe that natural selection causes the evolution of antibiotic resistant bacteria in a petri dish. Our null hypothesis that natural selection has no effect on evolution is disproved, because our null hypothesis stated that evolution was random, and yet we are able to reproduce this particular type of evolution at will through the application of natural selection. Now we adopt a new null hypothesis: natural selection causes evolution, but nothing else causes evolution. That's how we arrive at the theory of natural selection: by rejecting all other causes of evolution other than the one we have observed directly. Any further argument about how incredible it is that we could have evolved by natural selection, or how some external agency makes it all much easier to understand, is rejected out of hand by our particular version of scientific method. So far as we are concerned, natural selection is the only cause of evolution. Using the same process, we conclude that natural causes are the only causes of climate change, because the climate has done nothing that cannot be explained by natural variation.

That's our version of scientific method. But this is not the method being taught in the schools in my state. The official definition of the scientific method in the school system is, "build a model of your hypothesis, and if it is consistent with past observations, it has been validated and is therefore true." The fact that there are an infinite number of hypotheses that fit past observations, all of which make conflicting preditions about the future, presents a serious problem for the school's definition of scientific method, but they resolve this problem by referring to a "consensus among scientists" to determine which of the conflicting hypotheses is the one that is going to be the "accepted theory".

Any definition of scientific method that relies upon "consensus" is boring. Who wants to debate with someone who is going to fall back on "consensus" as the reason for their being correct? Nobody is going to be motivated to examine the world in detail when the result of their examination can be dismissed by a bunch of people saying, "you're wrong because we say so". There's no excitement to be had in the persuit of consensus, so we reject the consensus scientific method on the grounds that it is so tedious that nobody will want to study the world if they have to follow it, and so it will defeat itself. We'll stick with our null hypothesis and keep going our own way because it is more fun.