| Scientists confirm a theory as they amass more and more evidence that fails to refute any of its laws. These laws hold for any member of the stated class of events. Such classes are open, i.e., one cannot enumerate their members. Karl Popper pointed out that since empirical laws are generalizations about open classes, the scientist cannot verify them directly. Instead, the most one can hope is to falsify them. Simply demonstrate the existence of some counter example. Popper’s claim suggested to some that the task of science was to find a member of the class of events defined in the law for which the law fails to hold. This is hardly a full characterization of the scientific method, which is much broader. |
Lett’s Laws.Over the years Professor James Lett (1990) has developed six rules intended to sharpen students’s ability to think critically. He characterizes these rules as his own understanding of the scientific method. Some variations of these rules have since made their way into many science curricula. The scientific approach of the linguist Stanley Starosta (2000) includes the most important of these criteria. He also points out the importance of their being ordered, each one finding meaning when based on success in the one previous. In essence according to these rules the proposed scientific theory:
|
| The last of Lett’s rules relates to the investigator avoiding self-deception. This is easy to fall prey to when the investigator fails to consider the work of others in his field. The present author finds it impossible to keep abreast of much of the current work in linguistics and the kind of evidence that might come from a knowledge of numerous languages. |
Formality.There is much to say about a theory being formal and explicit, as the present work attempts to be. It is tempting and all too easy to take shortcuts in this area. For example, Lett himself has run just a little rough-shod over logic. His division of logical cases into valid/invalid arguments and sound/unsound arguments is useful even though his examples are faulty. In any case, if the argument is invalid or unsound, the argument can be falsified. First, every valid argument has a true conclusion. The conclusion must be either true on its own, or follow directly from the premises. Suppose we can accept the following two premises: “All dogs have fleas” and “Xavier has fleas.” The conclusion that claims “Xavier is a dog” does not follow from these premises. It is irrelevant to validity that “Xavier” could refer to anyone or anything, with or without fleas. We could fix its meaning definitely on a flea-carrying thing and not make the argument logical. It is the form of the argument that is illogical. A form of argument that is accepted as valid takes the following as premises: “All dogs have fleas” and “Xavier is a dog,” and the following as the conclusion: “Xavier has fleas.” |
| For Lett an example of a sound argument is one that actually conforms to reality. In the first example above, one of the premises, “All dogs have fleas,” can be easily falsified. The argument is unsound, even though its form is logical. |
Repeatability.In the case where the evidence for a claim is based on experimental results, or the evidence supporting the claim could be explained as coincidental, then the evidence must be repeated in subsequent experiments or trials, i.e. the evidence has repeatability. It is possible to fake results or get results that are coincidentally as expected, and not for the theoretical reasons. These possiblities make it the more important for anyone to be able to repeat the experiment and get the same results, no matter how many times they try. The absence of repeatability, as in palaeontology, slows progress and makes confirmation of theory difficult. In such cases simulation by means of computer software has been invaluable. Care must be taken that the arguments thus built avoid circularity. |
Falsifiability.Lett follows Popper in claiming that in order for a theory to be true, it must be falsifiable. This is the case for any proposed law. If a claim is false, the evidence will prove it to be; if a claim is true, the evidence will not disprove it. Falsifiability is essential because: if nothing conceivable could ever disprove the claim, then any counter evidence that exists would not matter; there would be no point in examining any possible evidence to the contrary; the conclusion is already known and the claim is invulnerable. This cannot mean, however, that the claim is true scientifically; what it means is that the claim is either tautologically true, e.g., “This sentence contains five words,” or otherwise unscientific. It is logically impossible for any scientific claim to be true no matter what. It should always be possible to conceive of evidence that would disprove the claim. |
Sufficiency.The evidence supporting a claim must be adequate to establish its truth, with the following stipulations: 1) the burden of proof rests on the claimant, 2) extraordinary claims demand extraordinary evidence, 3) an appeal to authorities cannot support extraordinary claims. One cannot argue, “No one has proved me wrong.” The logical possibilty of a claim is not enough to establish reality, nor are by themselves witness reports, no matter how sincere, and expertese, no matter the length of the authority’s postnomial. Even the honorable Dr. Einstein made blunders. It is the theory itself that exhibits sufficiency. |
Comprehensiveness.As time goes by, counter evidence may accumulate for some theories. Sometimes, even then, it is possible to save the basic structure of a falsified theory. The defender simply modifies certain laws slightly or gives some of the laws the necessarily more restrictive interpretation. One makes out certain special cases to be exceptions. As this sort of thing happens more and more, however, the theory will appeal to scientists less and less, and eventually they will feel obliged to apply Occam’s razor (simplify it) and replace it with some other more esthetically pleasing formulation. |