The scientific method and quintessence analogs

[Gary A]

Wolfram Math World: A null hypothesis is a statistical hypothesis that is tested for possible rejection under the assumption that it is true (usually that observations are the result of chance). The concept was introduced by R. A. Fisher. The hypothesis contrary to the null hypothesis, usually that the observations are the result of a real effect, is known as the alternative hypothesis.

Most science knowledge is statistically validated. The scientific method requires scientists and all others who claim to think rationally to answer knotty questions by means of EXPERIMENT or careful direct observation.

In order to do this effectively, one must formulate a hypothesis, a statement of the putative principle that engages all the known implications (an alternative hypothesis). These implications must also be rather direct. Circumstantial consequences are just that - circumstantial and cannot PROVE a hypothesis. These better implications must suggest experiments that will verify them or not.

It is good if there are subservient implications of the alternative hypothesis that can be tested. It is better if a complete and utter negative statement of the hypothesis can be formulated as well that has testable implications (null hypothesis).

Then, the net sum of the original hypothesis and the negative hypothesis should be zero, a “zero sum” result in Game Theory terms. Logically, the negative completely cancels the positive hypothesis when the statements are algebraically added together. This negative hypothesis is called the "null" hypothesis not only because it would nullify the other if it proves to be true and it would otherwise "tend" to validate the positive or “alternative” hypothesis. Still, it would fail to PROVE alternate to be true, if Null was shown to be false in any way.

And then, if direct evidence can also be found that tends to corroborate the original positive alternative hypothesis, we can begin to regard it as a good logical beginning. AND THEN, if this confirmed positive hypothesis can be combined with statements that have already been proven and the combined implications can be verified experimentally or by observation, we have the beginnings of proof.

The key to this process is our ability to form an experimentally testable Null Hypothesis. The evidence FOR the positive alternative statement of the hypothesis is insufficient in itself because circumstances may combine to fool our little experiments. We are human. If an appropriate robust Null Hypothesis cannot be formulated, the original alternative hypothesis is unsuitable to merit the attention of the scientist.

This is relevant to the debates about Dark Energy, quintessence and Dark Matter.

To call one's self a scientist, one must respect the scientific method. It is not a scientist's dogma any more than it is dogma to be a fireman respecting the firefighters' code: one must acknowledge the power of fire - or else you die.

An unfalsifiable hypothesis (unable to be cast in terms of a null statement) has no business occupying the time of the scientist. Whole theories have often been constructed from unfalsifiable hypotheses. Such theories are often fun to think about, satisfying, even edifying and inspiring - but they are not science.


Cosmologists are always wrong, but never in doubt. – Lev Landau

[RichW9090]

Nothing ever PROVES a hypothesis.

[Gary A]

Nothing ever PROVES a hypothesis.

Well, a lot of scientists try to do so by applying the principles of falsifiability and the Null Hypothesis. Whether they succeed is another matter. Proof is very difficult to achieve. The closest we have come to proofs of any scientific hypothesis, principle or "law" is in regard to quantum theory and relativity. The theory of natural selection would be another that I would mention in the same breath. Maxwell's equations I would regard as just a description of reality, not so much as a theory.