Science as process and product

Science is about finding stuff that works. No more, no less.

Science is a process that has given rise to a body of knowledge. Science is not just that body of knowledge.

That body of knowledge is often venerated as "science" and used to beat the credulous over the head. This is often factually correct, but usually just turns them off what they think is science. Nobody likes to be told they are stupid by an arrogant nerd.

It is easy to use "science" as buzzword, a holy cow. To set it up as an icon to be worshipped, now that we've killed god.

Science, in essence, is just stuff that works. If it doesn't work, it's not science. This is the much-venerated scientific method. it is well explained in other writeups, but here is my take on it:

  • First, observe what happens. Try to do so in as impartial and repeatable manner as possible. If it's not repeatable, it's not science.
  • Come up with a theory that explains the facts that you have observed. Don't get overly attached to your theory, you may have to abandon it.
  • Use your theory to make predictions about things you haven't tried yet, then try them. If your theory's predictions match what you observe then you may be on to something.
  • Think of something that if it happens, would disprove your theory. Try to make that thing happen.
  • If your theory does not explain the facts, throw it out and get a new one that does. Repeat.
  • If your theory explains all the facts that other theories in the same area do, and facts that other theories don't, then you are definitely onto something.
  • Tell other people about what you have done so that they can try to replicate your observations, spot flaws in your observations and theory and generally build on your acomplishments.

Any theory that lasts is going to end up with a good fit to reality. In short, it will work.

If you want to be strict about the way that you refer to a good theory, the theory is never true. The theory is not reality, it is just marks on paper. It is not the truth, It's only a model A theory is either more accurate in it's predictions of reality than another theory, or less so. It is simple or complex. It is elegant or not-so-elegant. It either has known flaws, or doesn't. It has been superceded or not. But it is never the absolute truth, the last word. All theories are accepted on a provisional basis, as they can be discarded if something better comes along.

Science as an evolutionary process towards a defined endpoint

Some say in the postmodern fashion that "all explanations of the world are just narratives, and all are equal". I agree with the first part. Yes, the theories that come from the scientific method are just narratives. But not all narratives are equal. They can be judged according to their fitness to their purpose. A kind of natural selection applies.

For instance, a hunter-gatherer jungle tribe might have a explanation that the world was created by being hatched a cosmic egg laid by the celestial duck. This story can be judged according to its fitness to its purposes: to give an answer to kids who ask "why?" to everything which gives them a sense of wonder, to reinforce the power base of the shamans, and so on.

A scientific theory is one that is judged according to criteria that strive towards some kind of complete truth, i.e. a 100% accurate description of how the real world works. I'm not claiming either that we can or can't have a perfectly accurate description of the world, just that we can and do strive toward it.

A scientific theory is judged good if it does not conflict with any known fact about reality: It must not be provably false, and it is better if it makes testable predictions about things yet unobserved. It must be falsifiable and the experiments must be repeatable by others.

Science works because in part it harnesses the human ego: if you can disprove someone else's theory or come up with a better one, you will be noticed. It is a free market of ideas. It is competitive. It has natural selection.

Science and non-science

There is no such thing as a supernatural occurrence. If something occurs, then it is by definition natural, even if we don't know the hows and whys of it.

If something occurs, then there by definition is a scientific explanation for it. To take an extreme example, we haven't come to believe the theory that our reality is controlled by a big guy with a long beard, but if there was strong evidence that this was a better theory than what we find in modern physics, then it would be a scientific one.

For instance, what is alternative medicine? If a treatment was proven to have an effect, even if we had no idea why, it would be assimilated into scientific medicine. Of course, the effect would have to be consistent and not just the placebo effect. And naturally, there would be intense interest in finding out why it works because of the predictions (ie new medicines) that could be made from knowing. Understanding is control. Knowledge is power.

Yes, there may be delays in assimilating new ideas due to egos and vested interests. But in the long run, it is inevitable. Alternative medicine is alternative to what works.

Hence the purveyor of any alternative treatment will try to market it as something that is going to soon be noticed by the mainstream. For instance, homeopathy is just about to gain mainstream credibility, and always will be.

The body of knowledge generated by science gets complex. That's because reality is complex, and people even more so. But the concept of science is an astoundingly simple one: It is just stuff that we know works.

Dismal sciences

What about economics, psychology, social science? Are these sciences? They do not have the repeatability that I talked about above.

First off, these studies are important and useful independent of if they fit the definition of science or not.

Secondly these studies all have one thing in common – they study the actions of human beings. As physical systems go, human beings are by far the most unpredictable, contrariwise, variable, self-referential systems you will ever encounter. How do you account for all variables that affect a person's actions? Getting them to do exactly the same thing twice is very difficult. How do you repeat a circumstance? If you are using the same persons as subjects, it's different, as they have done it before. If you use other people, then that's a difference.

It could be that simply we are studying the wrong level here, that a complete scientific understanding of say, economics (which itself requires some theory of mob psychology) is like trying to model the earth's weather with quark physics. Imagine trying to replicate your experiments with that model!

Thirdly if a result is only statistical, then that statistic is itself a result. If you can only produce a given result half the time, then others should also be able to verify that 50% mark.

Thank you for staying with me through this rambling rant.