Konrad Völkel
Many many times I talked about this topic in personal communication. I repeated myself and I developed my presentation of arguments over the years. The development will continue, but now I present you the current status. I hope you enjoy reading this and give some comments!
Stories serve a purpose. I'm pretty sure, if a story didn't serve a purpose - why would you tell it anybody? This purpose may be, to get some events and possible outcomes into our mind, so we realise what could happen in future. This is good to determine strategies what to do or to avoid in order to achieve a certain goal. The purpose may as well be to transport social norms, that are vital in the survival of the group, where the story is told. It is also possible, that this once was the case and nowadays the story serves other purposes. Some stories match our need in answering difficult questions, such as "where do we come from". The bible does a pretty good job in containing examples for these different types of stories and (of course) many more.
Theories are stories that make predictions, like "when you perform task A, then afterwards B will happen" or "B happened because A happened before, so it's more likely to observe the event B after the event A instead of not observing the event B after the event A". A strange but correct example of a theory is found in the bible: "if you don't honour your parents, god will punish you" (the phrasing was my choice, of course, and I'm simplifying here). This is a very clear statement, that "not honouring the parents" (A) implies "god punishes you" (B).
Scientific theories are theories whose predictions are all falsifiable, that means the theory contains a precise description how to test the prediction in a way such that it could actually be wrong. That excludes the statement "god loves every human" from scientific theories, because there is no test available that could possibly falsify this statement (and the same is true for "god does not love every human"). I want to call those theories that fail to match my "scientific" criterion esoteric theories.
Note that existence statements like "there are quantum particles" are not part of any scientific theory, since "quantum particle" is just a word, whose existence cannot be tested. The statement to be found in a scientific theory would be "there are things that behave in this-and-that way and we might refer to them as 'quantum particles' in this text...". There you can test that there really are things that behave "this-and-that" way.
Surprisingly, the famous "String Theory" is an example of an esoteric theory, because there is not yet any way to test it's predictions. Mass media ignore this, of course, because science fiction sells.
More important than existence statements are generalisations, like "all dogs are mammals". These statements can never be true with absolute certainty, unless they are trivially true. Either we leave open the possibility of finding some day a (genetic engineered?) dog which isn't a mammal, or we define dogs to be mammals with certain properties (and I guess the latter is common). It's funny to imagine children who might call some strange looking cats "dog" and refuse to call some strange looking dogs "dog", so this is really an issue of words. To get a little bit more scientific, consider "the speed of light is a constant, smaller than infinity, and nothing travels faster than that". Of course, up to today, nobody encountered something that travels faster than light. But who knows - some day, we might be able to produce such a thing and therefore we can never consider a general statement to be true.
A direct conclusion is, that almost all (= those who include general predictions rather than only existence predictions) scientific theories aren't true. They are just good "approximations of truth" if you really want to talk about "truth" here. This is a common misconception of science in the media that has already caused much damage.
Imagine you live in the times of Newtonian physics, where Einstein and Minkowski didn't revolutionize gravity yet. One day you wake up, think "how good that those scientists have found out the truth about reality", get the newspaper and - Bam! - you read about Einstein's discovery and that they were wrong! How is that possible? Science must be an unreliable thing .. you should better not believe those scientists any longer.
Sadly, many scientists have not understood that this is a big issue, so they continue talking about "true" theories in the media.
If you still want to think about "true" and "wrong" theories, maybe calling every theory "wrong" which has been falsified, then I have to inform you, that Einstein's general and special relativity are falsified (in experiments involving high energy). Keep in mind that the GPS (global positioning system) relies on calculations that come out of Einsteins theory, but at the same time of phenomena that come out of quantum physics - so if you don't "believe" in either of these theories, you can't believe in GPS (which is strange, since GPS is obviously working). Sadly, we know of no scientific theory to replace our current theories of gravitation and quantum physics. This is the major question in today's physics.
So I strongly advice you to get this into your mind: Science is not about finding out the true nature of our reality. It can't do this, and it doesn't try to do so. It merely provides models that are modelled upon the obstructions we encounter in performing actions in our reality. Let me explain this difficult sentence with an analogy:
Imagine a blind man put into a forest, who is supposed to find his way out of the forest. You tell him, that he should find the way out as fast as possible and that he will have to do that multiple times. Of course, this blind man will try to figure out the shortest path in the forest, around the trees. So he will try to memorize where he found a passage and where he found obstacles (the trees). Afterwards, he could provide a map of the forest, where the trees are indicated by, e.g. black circles on white paper. But he will never know (and doesn't need to know) about the green leaves that are hanging on the trees above him - unless someone tells him.
Our scientific theories are like the blind man's map. They serve the purpose of allowing us to predict ways to perform certain tasks and they predict the obstacles. But we will never know of anything we can't measure, so the theories are really modelled after that what we can see. Of course, the blind man could imagine purple animals living above the trees, but that would be an esoteric theory and clearly won't help him in finding the shortest way out!
We can compare two theories on at least two axes: Expressibility and simplicity. By expressibility I mean that we consider the theory more useful, thus better, which provides more predictions. For example, a theory that predicts the behaviour of water between 10 and 90 degrees (Celsius) is not as good as a theory that predicts the behaviour of water in any temperature. By simplicity I mean, that from two theories we would prefer the simpler one over the more complicated one. It is just a matter of usability that we want a theory with the least assumptions and the most easily checkable predictions possible. This principle is historically often called Occam's Razor, since it "cuts away unnecessary parts with a sharp razor-blade".
What to do if a theory was falsified? Think of our theory of gravitation - now that we know it's wrong, shouldn't we abandon it? Well, it would be nice if we could do that, but there is nothing else that makes that many unfalsified predictions. Newtonian physics is falsified in much more cases, so we have to stick to our current theory. A simple way to save the theory is, to include some assumption as "In the cases, where we already falsified the theory, we don't provide predictions any longer". The new theory, that includes this statement, is now a non-falsified theory. This may seem a little bit odd at the beginning, but if this "saved" theory still provides more predictions than every other theory, then it's clearly our preferred one (and this is the case with our current theory of gravitation).
So we talked about scientific theories and what's great about them (we can compare them in a reasonable fashion and they predict useful stuff, like GPS). But I want to state, that esoteric theories and more general, stories, are useful, too. Any normal person doesn't know about quantum physics and general relativity, but instead some simple Newtonian-like physics - so he or she wouldn't try to spontaneously fly into the sky, because the theory implicitly used predicts that this is impossible. Since humans want to solve riddles and have their questions answered, they think a lot about stories that are able to answer questions like "Where do we come from" - and it's a very good thing to have a story that gives an answer to those questions. As in the physics example I have just given, stories are more useful than the latest scientific theory because they can over-simplify things and include statements that are falsified in general (but good approximations in everyday life).
So, is there any way to compare which story to use? We can still use the criterion of falsifiable expressibility, that is, one story is better than another if it provides more falsifiable predictions. But it's also important that the story is not to complicated, so we have to use another criterion of simplicity here, which depends heavily on your personal needs. Let's look at an example of story comparison:
Imagine an ill patient, who wants to decide what treatment to choose. There are several classical treatments in a hospital as well as some esoteric treatments, involving religious rituals or some other non-falsifiable mechanism. I think it is very clear that the patient should prefer the treatment where the probability of a positive outcome is maximized (in reality, there are often several "positive outcomes" that are hard to compare, but this choice is the patients job, not ours). The esoteric theories (related to the esoteric treatments) are not falsifiable, so there was never an experiment to see if the treatment ever worked better than other treatments, so there is just no probability of success. The scientific theories (related to the "classical" treatments) are falsifiable, so there have been experiments to find out the probability of success compared to a variety of other treatments. The patient can only choose from the classical treatments in any reasonable way, since with them he knows that they could work, while with the others it is known that they don't work in any reproducible way.
Sub-cases of our patient example include the usage of homoeopathy, but there is also the Placebo effect, so I don't want to discuss these issues here. Let me just stress that the choice of an esoteric treatment can kill a patient, so this is a good example of a problematic story. Other examples of problematic stories include those used by some group of humans to control or maintain power over another group of humans. The undeniable historical evidence for these examples is found in the history of slavery, the history of women rights and the history of inquisition. Men have told stories to "explain" (in an esoteric, non-falsifiable way) that slaves, women and certain individuals called "witches" are inferior or have to be killed. I think it's very good that we abandoned these stories and punish everybody who still supports them.
There is a way to abandon the belief in an esoteric theory. Just try to replace the words in the unfalsifiable sentences by something funny. In the sentence "aliens are secretly controlling everything, therefore the climate changes" you can replace the unfalsifiable part "aliens are secretly controlling everything" by another unfalsifiable, similar part, thus producing a sentence like "a big, fat mushroom is secretly controlling everything" - and that makes obviously no sense at all.
To come to an end, let me just remark that my usage of words differs heavily from Wikipedia's article about the scientific method. I think this article has many errors, but to discuss that I would have to require you to read some texts from Popper, Lakatos and Feyerabend before (and I think this is a little bit too much for a blog post). I think it's still safe to say, that a theory chosen from all known scientific theories (for a specific phenomenon) under the criteria of simplicity and expressibility, is a theory that satisfies the criteria from the "scientific method" article in Wikipedia. So my opinion is not so far from mainstream at all.
This blog post grew out of attempts to write other blog posts, so there might be follow-ups on the following subjects: "homoeopathy, justifying doing mathematics, climategate".