I want to talk about an approach to science called Instrumentalism that avoids interpretation in favor of calculation. Theories are not educated guesses, nor are they explanations backed by evidence. Rather, they are predictive models that may have no basis in reality. If someone observes a phenomena, instrumentalism states that one is not to speculate on it, but simply produce a model that can take initial conditions and predict the results. If one has not done so, then has not done science.
For example, if you have a description for what happens when a coin is tossed, it’s only scientific if you can use it to predict the results of a toss. If you can do this, then you have a theory. This theory says nothing about is “really” happening, nor should it. Rather, it’s basically a program where you enter the observations and get predictions. Furthermore, like software, it’s upgraded when something better comes along. That’s right, theories are always provisional because — contrary to popular belief — science was never about the truth.
Theory Y is better than theory X if its predictions are more accurate, it takes new observations into account that X couldn’t, and/or it unifies multiple fields.
Not only does instrumentalism say nothing about the truth of theories, it says nothing about the truth of non-theories. When something is non-scientific, it doesn’t mean it’s not true; it only means that it cannot produce a predictive model. There is nothing pejorative about saying something is not scientific. To say X is not scientific is as neutral as saying X is not mathematical.
Instrumentalism keeps science on solid ground by ensuring its results are useful, restricting the phenomena it can study and providing a criteria by which the quality of theories can be judged. Competing narratives can’t be judged, but competing predictions can by comparing success rates. What’s more, since instrumentalism avoids narratives, it avoids ideological debates.
As an example of an instrumentalist approach, take genetic engineering. Some people conflate it with evolution, but it’s irrelevant. Genetic engineering requires only a correlation between genes and traits, and this can be done while believing in evolution or intelligent design. Speculating on where these genes came from, or the implications of shared genes doesn’t affect the practice at all. Furthermore, what historical accidents or narratives led to genetic engineering are equally irrelevant; one doesn’t reflect on the accident that led to penicillin when practicing medicine, nor should one reflect on what may have enabled genetic engineering.
Simple, deterministic, quantifiable, useful, non-ideological. That is instrumentalism and it is the way science should be.