Scientific Guesses Aren't for Amateurs

by Richard Schowen
Professor emeritus of chemistry, molecular
biosciences and pharmaceutical chemistry

There is probably no modern myth so entrenched as the image of working scientists as cool, systematic thinkers and doers who follow, like automatons, (desperate with boredom, one might think), the rigid protocol of "the scientific method." In real scientific life, progress toward research goals occurs rather less neatly and with much more excitement and interest. The "intuitive leap" is what allows expert ornithologists to glance at a video image and, in a flash, feel the bird they are looking at is a pileated woodpecker and not an ivory-billed woodpecker.

Many good scientists tend to examine the available data with great care and find they have one or two pretty clear ideas about how to explain the findings. Usually, these ideas have not been laboriously drawn from the data but instead have "sneaked up" on the scientists, and they feel less like hypotheses than like good guesses.

Linus Pauling, the greatest chemist and one of the greatest scientists of the 20th century, who, among many other achievements worked out the nature of the chemical bond and founded molecular biology, was famous for his intuitive leaps. Some criticized him, but the Pauling style is the common way that real science is actually practiced.

The idea that thoroughly informed persons ("experts") are capable of instant judgment about questions in their fields — guessing the answer — and that their guesses are often very good indeed is the subject of Malcolm Gladwell's recent book Blink: the Power of Thinking without Thinking (New York: Little Brown, 2005).

The more general idea that human knowledge, including scientific knowledge, can be acquired by an informed guess, followed by a rigorous evaluation, goes back much further. C.S. Peirce, a 19th-century American philosopher considered the father of the theory of signs, or semiotics, suggested that knowledge could be acquired not only by induction and deduction, but also by "abduction," by which he meant an informed guess followed by careful investigation.

John Platt, a spectroscopist and philosopher of science, in 1964 made much the same point using the term "strong inference." He describes how "[c]ertain systematic methods of scientific thinking may produce more rapid progress than others." He suggests devising several hypotheses, then devising a crucial experiment to exclude at least one of the hypotheses by showing it to be wrong, and then cycling back to devise new hypotheses and new crucial experiments. The hypotheses are of course highly informed guesses, and indeed Peirce considered "abduction" a synonym of "hypothesis." An enormously entertaining 1983 article by Thomas A. Sebeok and Jean Umiker-Sebeok explains that the same point applies to the investigation of crimes and to medical diagnosis.

So guessing, followed by checking whether the guess is right, is the way good science usually works, and that is much more enjoyable than turning the crank of a methodological machine. It is also much more efficient, as Platt showed. The only thing to remember is that, to avoid foolish guesses, the guesser must be an expert — and that takes years of education and more years of experience.

Sources

"[Pauling] made great intuitive leaps and was frequently criticized for the conclusions he drew from what some felt was too little experimentation, often outside of Pauling's area of expertise." Gail Thompson and R. Andrew Viruleg. In The History of Chemistry, 1992 Woodrow Wilson Summer Institute." Linus Pauling, A Biography." See also http://www.paulingexhibit.org/exhibit/process.html.

The Sign of Three: Dupin, Holmes, Peirce. ed. Umberto Eco and Thomas A. Sebeok, Bloomington, Indiana University Press 1983: Ch 2, "You Know My Method," Thomas A. Sebeok and Jean Umiker-Sebeok.

John R. Platt, Strong Inference, Science (1964) 146: 347-353.