Wednesday, May 13, 2015

What Makes a Good Theory

Using as a backdrop the recent report from Seymour Hersh on the bin Laden raid, Michael Rozeff presents an excellent discussion of what makes for a good theory. Below.-RW

A Brief Foray Into Sorting Out Theories
By Michael S. Rozeff

Our colleague Paul Craig Roberts has proposed that bin Laden died years before the raid in Abbottabad in which he is said to have been shot to death. One can assemble a number of supporting facts or near-facts or interpretations of facts or near-facts in order to support this hypothesis or theory. A good many people no doubt assign a high probability of truth to this theory. I do not.

Under this hypothesis, the raid is taken to be a complete hoax. How does that explain the surviving family members making comments about the raid? Under that theory, the sources that Hersh used have to be liars and he has to be wrong in thinking that they are valid. This theory must then explain these implications. The theory must also explain why, if bin Laden were dead, the U.S. government would have denied it. It has to explain why the government chose the time it did to perpetrate a risky hoax, especially since the uncovering of the hoax would be damaging. What great benefits would such a hoax bring. and could not such benefits be obtained in other less risky ways? And if such a hoax were attempted, would it not have been better planned with the cover stories all pre-arranged?

Now we have an alternative conspiracy theory, which is the one proposed by Hersh. This theory must answer to fewer questions and fewer serious challenges. It is internally consistent and explains a good many facts and ambiguities, some of which have existed from the outset and gibe with earlier news reports. I never bought into Roberts’ theory in the first place, but I didn’t express my opinion about it. I thought it was implausible even though we knew he was ill. Bin Laden was born in 1957. He was 44 years old in 2001. An early demise from his ailments seemed unlikely. He could afford good medical care. Had he died of natural causes, the news would probably have leaked out. There were some reports of his activity after the date of his supposed death. All in all, his death in 2001 or thereabouts seemed unlikely to me because of considerations like these and others. Hersh’s article makes this theory even less plausible.

We constantly devise theories and choose among them. This is required in this case. This comment is nowhere near a complete explanation of what criteria we should or do use in choosing or in assigning probabilities to alternative theories. Each criterion requires a ceteris paribus proviso — other things equal. Occam’s Razor is frequently invoked as a criterion: choose the simplest or most parsimonious explanation, one that requires the fewest and least problematic assumptions. Of course, choose the theory that explains more of the phenomena. A better theory will have some implications not previously thought of that when looked at are true. It is said to be richer. One wants a theory to build upon assumptions or postulates that are correct, if at all possible. This is not always known for sure. The fewer the leaps of faith in the postulates, the better. Simplicity, explanatory power, richness, and correctness of assumptions are a few criteria. Another is lack of ambiguity, or that the theory is clear or strong in some sense. People have other ideas about theories, such as here and here.

There is controversy over the subject of what makes a good theory. I mean only to open the door, if it needs to be open for some lay readers who haven’t thought about this before or been exposed to how scientists think about and use theories.

The internet is chock full of theories. Everyone needs ways of choosing among them.

The above originally appeared at

No comments:

Post a Comment