Crazy Idea Versus Scientific Theory

How does one go about deciding if an idea has merit as a scientific theory? Robert Ehrlich, in his book Nine Crazy Ideas in Science (Princeton University Press, 2001, ISBN: 0-691-09495-0), gives nine questions to consider. I have reorganized those questions into questions about the theory and questions about the proposer. I have also restated them and added one other consideration that I feel should be explicitly stated. These are the criteria for determining whether an idea has merit as a scientific theory that will be used for this course.

Each idea should be evaluated in terms of these eight criteria. For our purposes assume that an idea has merit as a scientific theory if at least 50% of the conditions are met. Those ideas that satisfy less than 50% would be considered to not have merit as a scientific theory.

The theory needs to be coherent and internally (logically) consistent.

To be coherent and internally consistent means that all aspects of the theory itself are consistent. One part of the theory doesn't contradict another part of the theory.

The theory needs to be consistent with fundamental principles.

Fundamental principles have held true throughout all of the history of humankind. A theory that depends on something that contradicts a fundamental principle will be very suspicious. Claims of perpetual motion machines which function without the input of energy are always viewed as suspicious. No such claim has been shown to be true. No example exists where the machine can produce enough energy to replace the energy lost to friction or other heat generating actions.

The theory needs to explain observations and experimental results.

The purpose of science is to explain physical observations and experimental results. These explanations are in the form of theories and so any scientific theory should explain the observations and experimental results.

The theory needs to be applicable to an appropriate range of material and have an appropriate number of free parameters.

When a theory purports to explain everything about the universe, a flag should go up and the theory should be considered with suspicion. Most theories have somewhat limited application and only attempt to explain a specific set of observations.

Free parameters are associated with experimental variables. If the theory has a large number of variables that can be adjusted to fit any situation, it is again cause for caution. It has been said that anything can be modeled if there are enough variables to twiddle. As an example, most any curve can be modeled using a polynomial, if you have enough terms in the expression.

The proposer needs to be known as competent in the field of the theory and open to new data and critique by others in the field.

When someone is competent in a field, they have generally studied and published in that field. There are a few scientists that are competent in several areas, but most of the time competence is restricted to one or two closely related areas. Just because someone is extremely competent in one area (maybe even won a Nobel prize), it doesn't mean that their views in a new area are better than someone who has studied and worked in the new area.

Competent scientists are also open to new data and critique by their peers in that field. If a peer questions or gives an alternative view of some aspect of the theory, the proposer must be willing to listen and either agree with or give evidence to counter the alternative view.

The proposer needs to refer to the work of others in the field and openly disclose all data and methods used to develop the theory.

Scientists don't work in isolation. All of the observations and experiments, even those done by other researchers, have to be explained by the theory in order for the theory to persist. A working knowledge of what other scientists in the field are doing is critical to advancing any theory. In addition, all methods of doing experiments and the results, or data, that come from the experiments should be made known to all scientists that are working in the field. This also allows other researchers to repeat the experiments and helps to ensure that nothing is overlooked.

The proposer needs to use statistics and other math tools appropriately.

It is often said that statistics can be used to explain anything. There are well-established ways of using statistics and testing the statistical significance of data. Statistics or any math tools should be used in these well-established ways and the purpose for using them should be clear.

The proposer needs to not have an outside agenda.

We all know people who are opinionated to the point that it is better to just stay away from the subject. If there is this kind of behavior about a scientific theory it is reason to be cautious. Or, if the person can benefit financially from the theory, it is also a reason to be cautious. Politics and other considerations may also cause someone to propose and vigorously support a particular theory.

These criteria can be used as premises in logical arguments. But that is getting ahead of ourselves. The next section will introduce some simple logic and show how to argue that an idea does (or doesn't) have merit as a scientific theory.