There are a lot of misconceptions about science.

There are ideas that Scientific Laws are somehow authoritative or that they can’t be broken. There are ideas that Scientific Theories are no more than hypothesis, lacking any significant supporting evidences. And there are no end of misconceptions about the findings of science, the “Scientists discover moon made of cheese,” or “Science discovers Black Holes leak,” and such like misleading headlines.

There’s too many to do justice to them in a single post, so they’ll come up from time to time in this blog.

Today I’ll just deal with one of the most common ones.

Science Proves Things

This is one of the major misconceptions. It seems that ‘proof’ is a difficult concept for many people to understand. There’s a lot about this in The Oracle.

It is possible to mathematically ‘prove’ things in certain axiomatic systems.

In the real world one cannot ‘prove’ anything absolutely, one can only provide supporting evidence for things. What sort of things? Things like statements, observations, opinions, theories, explanations, hypothesise, ideas, conjectures, and very importantly accusations.

The legal profession has been concerned with the issue of proof for accusations for at least three millennia. In most courts the standard required is that the evidence presented reaches a level that precludes ‘reasonable doubt.’ While evidencing something to the level where it appears to exceed a test of ‘reasonable doubt’ in the eyes of a judge, a group of elders, a jury of one’s peers, or twelve good men and true, may be referred to as having ‘proven’ someone’s guilt, it should not be confused with ‘absolute proof’. There are more than enough people who have been convicted of something with proof ‘beyond reasonable doubt’ where subsequent DNA evidence has convincingly shown that they were clearly innocent of the accusations. Link

The fact that something has been ‘proven’ in a court of law does not mean that it has been ‘proven’ in the philosophical or scientific sense of that term. It only means it has been evidenced beyond reasonable doubt, and sometimes that is clearly wrong.

The main evidence for many ‘things’ is observations by Human witnesses. We all know that Human Beings do not always make good witnesses. Unfortunately for many things Human witnesses are the only evidence. For example, in a test of the effectiveness of a new drug to assist people struggling with depression often the only evidence concerning the effectiveness of the drug is the patient’s own witness concerning their sense of depression and well being. There is no other way to ‘prove’ whether the drug is effective or not.

The other major evidence is confirmations. Confirmations may be observations of something else.

So - if one has a conjecture that all ravens are black, one can start gathering confirmations for this. One wants observations from trained observers (who can tell the difference between a Raven, a Crow and a Rook) who have no bias, preferable confirmed by two observers at once.

Records are gathered of time, place, observation conditions (spotting ravens at night leads to observation errors), the identity of the observer, the distance of the observation, etc. etc.

One may then write this down, or record it on tape or in a computer, which creates records of the observation.

And thus one gathers two and a half thousand records of observations of black ravens.

Does this prove the conjecture that ‘all ravens are black’?

No. It simply gives us two and a half thousand confirming instances for the conjecture.

What happens if a single person then reports seeing a white raven?

We have already encountered one example with the problem of confirming instances as evidence as shown in CK’s Blog for the 28th October 2018. Two and a half thousand confirming instances for Goldbach’s Conjecture (2) can be found, but that doesn’t prove it’s true. In fact it isn’t true. At least two exceptions are known to exist.

And this is the problem for science. Many scientific theories have many millions of records of observations all confirming the theory is right, but THAT DOESN’T PROVE IT IS RIGHT. In science, beyond reasonable doubt doesn’t cut it.

And that’s why science has Theories, which may be as well evidenced as it is possible to evidence something, but they still are not considered to be proven.

And science also has ideas, concepts, hypothesise, and conjectures, which may be supported with some observations, some maths, or nothing at all. Which doesn’t mean they aren’t right; it just means we are a long way from confirming them to anything approaching a convincing level.

So - science doesn’t prove things.

In a court of law CSI scientists may provide evidence sufficient to convince a court beyond reasonable doubt. That may be referred to as scientists ‘proving’ the accused is guilty. But science has been known to be wrong even with things such as DNA evidence. We’ll look at that later.

Science doesn’t prove things. It evidences them.

Next week I’ll look at the difference between ‘Science’ and ‘The Scientific Method.’