In the Part 1 of this series, I wrote that the two most important philosophical innovations that led to the incredible success of science is that science will seek to create only naturalistic explanations about natural phenomena. And science will determine the truth of a scientific explanation by appealing directly and only to nature itself. In other words we will ask nature to explain nature and we will use nature as the final authority about the truth of that explanation. So how do we go about asking nature about nature? And how do we go about asking nature about truth?
Hypothesis: Make a Claim!
The first thing to know is that no one gets anywhere in this process without making a truth claim. Just like no one goes to criminal trial without some prosecutor somewhere making some kind of claim against the defendent. In both cases, the claim is called a hypothesis. Webster gives an excellent definition for this use of the term hypothesis:
a tentative assumption made in order to draw out and test its logical or empirical consequences.
So a hypothesis is a tentative assumption. Initially you don’t know if it is true or false yet. And like the criminal trial prosecutor’s hypothesis, which we assume is wrong until proven right, in science we assume the same thing. In fact, you could say that in a criminal trial, it is not the defendent that is on trial so much as the charge against him is on trial. The burden of proof for the guilt of the defendent lies with the prosecutor.
In science, we proceed the same way. We make a truth claim about nature in the form of a hypothesis which we then put on trial. And we must formulate our hypothesis to allow us to “test its logical or empirical consequences.”, (as Webster so aptly puts it) so that nature can be the judge and jury.
Prediction: Make a Testable Claim
The way we do that is to make a special kind of truth claim that says something definitive about nature that we can easily test. For example, suppose we offer this hypothesis:
Swan Hypothesis: All swans are birds.
As you can see, there are quite a few consequence to this claim about swans. If we state the hypothesis within the body of a logical proof, you can see how these consequences appear as inescapable conclusions that follow from the first two premises:
- Assume the accepted definition of birds.
- All swans are birds.
Therefore, (because of 1 and 2)
- all swans must have wings.
- all swans must be warm blooded.
- all swans must have beaks.
- no swans can have gills.
- no swans can live underwater indefinitely.
- no swans can live in the vacuum of outer space.
Notice that 1 is a premise that we have assumed and 2 is a premise that is our swan hypothesis. And 3 through 8 are some of many conclusions that follow as a logical necessity from 1 and 2. Because the conclusions are logically connected to our hypothesis, the hypothesis can only be true if the conclusions about swans are true.
In science we would say that conclusions 3 through 8 are predictions about swans that come about as consequences of our hypothesis. They are consequent predictions. Notice that these predictions are not just our opinion about swans, but rather they are generated by combining our swan hypothesis with the other premise. It is these logically generated predictions that allow us to test our hypothesis against nature. All we have to do now is to observe a lot of swans to see if the predictions are accurate or if they fail.
How do we ask nature about the truth of a claim? We make a hypothesis, combine it with other premises and coax out testable consequences in the form of predictions. Then we test the heck out of the predictions (but more on the testing later.)
Falsifiability: Make a refutable claim.
Whether our swan hypothesis is true or not, it would be considered scientific because it makes highly testable predictions about some aspect of nature. The precise scientific term for “testable” is “falsifiable”. Notice the irony here that the bolder the claim our hypothesis makes, the easier it can be falsified when compared against observations in nature if it were wrong. This is a good thing, however, because if the predictions from the hypothesis manage to keep from being falsified by repeated observations, our confidence in this risky and vulnerable truth claim about nature keeps going up.
When we say “all swans…” our predictions have to apply to every swan we see. If our hypothesis were wrong, it would appear pretty readily because one or more of our predictions would be falsified. But if we say “some swans…”, everything suddenly gets loose and fuzzy because there is no way to be certain for which swans our predictions would have to apply. If we run across a swan with gills, our “some swan” hypothesis will not be falsified because we don’t know if that swan is one of the “some swans” or not. A “some swans” hypothesis would be unfalsifiable because of that. More importantly, we can’t really build much confidence in the “some swan” hypothesis if there is no way to falsify it.
However, if you supplied some selection criteria in the hypothesis you can make it falsifiable again, such as “In Medina County, Ohio, 95% of the swan population are birds.” To test this, you would collect a large enough number of swans in Medina County and test if the predictions we listed above applied were accurate for 95% of them.
So you see why the term “falsifiable” is used rather than testable. We can build a lot of confidence in a hypothesis if it could easily be falsified if it were wrong, but fails to be falsified under repeated testing.
To summarize where we are so far, science seeks strictly natural explanations for natural phenomena. And the authority for the truth of those explanations is nature itself. We can begin to ascertain that truth by formulating our explanations in the form of a hypothesis that when combined with other premises produces logically consequent predictions about nature. Then we test those predictions against nature.
Has anyone notice a problem yet? If I say all swans are birds, how can I prove this without testing our predictions against every single swan in the world? Isn’t this a big problem? Well, yes and no. In part three of this series I will discuss the epistemology of “proof” when it comes to scientific theories that make universal claims about nature.
This next step is very important, because using all this, we are going to prove that we can play billiards on the planet Mars without learning new rules, and we are going to prove that all life on the planet evolved from a common ancestor organism over 3 billion years ago.