Wednesday, February 12, 2014

Science Doesn't Prove Things

Science provides hypotheses and models of how things work. The value of a model is measured by its ability to explain past observations and predict future ones.

A new scientific theory is accepted into the mainstream canon when there is consensus that it provides better, simpler, more accurate, or more detailed explanations than existing theories. New theories often meet with great resistance unrelated to their merit or utility, for a variety of reasons outlined in my previous post.

Today’s breakthrough theories will become tomorrow’s mainstream theories. And sooner or later, many of today's best and brightest theories will be displaced by newer theories.

Does this mean that theories that are “right” today somehow become “wrong” in 50 years? Assuming that the underlying assumptions don’t change, can something that is “true” today become “false” somewhere down the road? Can things that have been “proven” later be “disproven”?

The laws of statistics do not change. But as time goes on, our abilities to measure things precisely do change. And perhaps more significantly, our cultural biases, conceptual frameworks, and descriptive and abstractive abilities change.

For centuries, Newtonian mechanics were sufficient to explain the motion of matter to almost everyone's satisfaction. Then experiments with light led to the theories of relativity, which showed that Newton's laws were subtly but consistently inaccurate (also known as “wrong”). Then quantum mechanics revealed that the world of light and matter was a world of likelihood rather than laws.

So what?

So: Just because we are told that the experts have “proven” something doesn’t mean that it will stay proven forever. To me, this means that nothing has been proven at all.

Does this mean that I am “anti-science”? No. It simply means that I do not worship false idols.

When discussing science, it is crucial to keep in mind that it deals in ever-evolving models, not in truths. If this is the case in the field of physics, then it is even more the case when it comes to anything having to do with humans, health, food, and medicine. In these areas, our abilities to measure things, or even to decide what to measure, are evolving rapidly. On top of this, humans are terrible experimental subjects, especially when it comes to long-term health and diet studies, and especially when they are expected to self-report: subjects tell researchers what they think they're expecting to hear; subjects lie about what they eat; and sometimes subjects simply forget things or get confused.

In my next post, I’ll explore some embarrassing moments in the history of mainstream medicine, and I’ll talk about why the alternative health world is can sometimes be years ahead of the mainstream world.

1 comment:

Alex Lewin said...

A cool perspective on some of the same questions: http://www.clickz.com/clickz/column/2287609/all-models-are-wrong-but-some-are-useful