Isaac Asimov - The Relativity of Wrong (1989)

Isaac Asimov - The Relativity of Wrong (1989)
chem.tufts.edu

However, I think he has misclassified the error of the English Lit correspondent in this essay. Since I have witnessed this misunderstanding many times (and I want to postpone writing my C++ thread pool code), let me elaborate. There are basically three areas of intellectual endeavor, classified by how they go about investigating the truth of the statements: One is scientific, where questions are handled using the scientific approach, so well described by Asimov. The other is group is axiomatic systems, like mathematics (and heraldry), where statements are mostly proven rather than checked through experiments. Given a statement, e.g. "there are infinitely many primes", one does not go about making measurements in the real world to check its truth (although sometimes real-world models can help, pioneered by Archimedes in the "Sand Reckoner").

Then, there's a third category of questions where these two approaches fail. In fact, when asked a question of this type, one does not immediately know how to go about proceeding to investigate it, e.g. "What is truth, Is there a God", etc. This, of course, is the realm of philosophy with somewhat different techniques of tackling problems then the previous two.

With this (rather simplistic) classification, we can see the impedance mismatch between Asimov and the Lit guy. Asimov is the scientist par excellence and is talking about the scientific method. However, the Lit guy's letter (truth is relative, etc.) is the culmination of a certain philosophy that is still the accepted standard in liberal arts school and may be summed up as moral relativism. The guy's obvious error is to apply this approach to the scientific realm. But mind you, this is very common with people of this background. If you don't believe me, just find a liberal arts major and talk for 10mins. He/she will be dumbstruck that you think otherwise. A common example with such people is how Newton's theory was "wrong" and Einstein's theory "corrected" it, clearly showing there's no absolute truth. Years ago, my professor in graduate level Syntax course gave this very example with the same conclusion that the Lit guy arrived.

But apart from its misapplication to scientific matters (such as the shape of the earth), is the moral relativistic approach a useful one, despite its acceptance as an axiom in a lot of universities? Now, that is a long debate.

Science consists of both theory and experiment. The old theory does a reasonable job of explaining the old data that came from the old experiments. Along come new experiments producing new data that the old theory cannot explain.

To understand science you must contrast it with folk-wisdom. Folk-wisdom comes up with a three part explanation of the new data. It is conservative in that it keeps the old theory, as its explanation of the old data. It comes up with a new theory that explains the new data, but not the old data. Thirdly it offers an ad hoc guidance as to which theory applies.

Science is also conservative, but with a different vision of what this means. Scientific conservatism lies in the high esteem granted to the old experiments. The new theory cannot merely explain the new data, it must do pretty well on the old data. Experimental error is important and we must at first cut the new theory some slack as experimenters are prone to optimism about the accuracy of their experiments especially when they agree with established theory. But not for very long. Old experiments can be redone with greater accuracy, and the new theory is expected to win both on the new experiments and on the most accurate versions of the old experiments.

Notice my gradual build up to the central dilemma. A good theory has a degree of inevitability about it. Newton's Law of Gravity is an inverse square law. Distance raised to the power of 2. Perhaps we can get a better fit to the orbit of Mercury with 2.00036, but that would be silly. Inevitability implies brittleness. If new data contradicts the theory there may be no way to adjust the theory.

If the old theory is a good one, but new experiments contradict it, folk wisdom can accommodate this straight forwardly by keeping the good old theory as the folk account of the old experiments. Scientists however have a serious problem. The good old theory breaks when they try to bend it to fit the new data, and the new theory has to agree with the old data, even though the old data appear to force one to believe in the old theory as the unique good fit.

Scientists are caught in a trap. Their conservatism with respect to old experiments forces them to be radical with respect to new theory. The new theory cannot just beat the old theory on the new data, it has to beat the old theory on the old data, the very data that gave rise to the old theory! This requires radical new ideas. As [Lampedusa says](http://en.wikipedia.org/wiki/The_Leopard) "If we want things to stay as they are, things will have to change."

Hence the 20th century phenomenon of lay people reading popular accounts of Quantum Mechanics and of Relativity and coming away with the impression that scientific truth is prone to dramatic changes. Is there any antidote? I recommend A Quantum Mechanics Primer, by D.T. Gillespie. This slim, clear book reaches Ehrenfest's equations on page 111, allowing the interested layman to see how Quantum Mechanics leads to F=ma and beats Newtonian dynamics on its home ground. But the book requires single variable calculus and a good enough grasp of three dimensional vector spaces to realise that nothing goes too horribly wrong when we add more dimensions. Is it really accessible to a layman?

Ehrenfest's equations symbolism a new paradigm of total backwards compatibility through radical change. Asimov ably covers the data side of this, but wimps out of trying to explain the implications for theory.