The idea of a crucial experiment that decisively confirms a model goes back at least to Francis Bacon (1561 – 1626) who used the term instantia cruci. Later, the term experimentum crucis was coined by Robert Hooke (1635 – 1703) and used by Isaac Newton (1642 – 1727), in particular with regard to his theory of light. Alternatively, Pierre Duhem (1861 – 1916) strongly disagreed with the possibility of crucial experiments. Somewhat in anticipation of Thomas Kuhn’s (1922–1996) paradigms, Duhem realized that scientific theories or models do not stand alone, but rather come coupled with auxiliary assumptions. Was what Galileo saw through the telescope features of the heavens, or only of his telescope, as some of his detractors claimed? One has to consider the combined heavens-telescope system to decide. When the detector is as complex as the ATLAS detector at CERN the question is even more apropos.
Karl Popper (1902 – 1994) refined the idea of the crucial experiment to one that falsifies a given model. But the Duhem-Quine hypothesis, a variation of Duhem’s idea, makes the point that falsification, at least in its naïve form, falls victim to same holistic argument: we can never test a single model in isolation. So is the idea of a crucial experiment just a will-o-the-wisp that vanishes on more careful evaluation?
We can think of many examples: Sir Arthur Eddington`s measurement of the bending of star light by the sun, the discovery of high-temperature superconductors, the measurement of the three degree microwave background, the Michelson–Morley experiment, and so on. Did none of these play a critical role in the history of science? I would suggest they did, but not in the simple manner suggested by Bacon or Popper.
Consider the Michelson–Morley experiment in 1887. Scientists did not do a Chicken Little impersonation and run around claiming the sky was falling or, in this case, that Newton (Newton`s laws of motion) and Maxwell (electromagnetism) were wrong. Rather, they started trying to understand what the explanation could be. This led to ideas like ether drag (the earth entraining the ether) or Lorentz-Fitzgerald contraction (the idea that objects shorten in the direction of motion). The latter idea was developed and expanded upon by Lorentz and Poincaré who developed the math for special relativity. Einstein claimed he was unaware of the Michelson–Morley experiment, but he was certainly aware of Lorentz`s early attempts to understand that experiment. Thus, the Michelson–Morley experiment started a chain of events that inexorably lead to special relativity, not in one easy step, but eventually and inevitably. If special relativity had been proposed thirty years sooner, it would have been treated as a curiosity like the Copernicus model when it was first proposed.
As another example, consider the measurement of the bending of light by the sun. The general theory of relativity and classical mechanics differ by a factor of two. Eddington`s 1919 experiment gave a result closer to general relativity and hence contributed to the early acceptance of general relativity (not that people are not still trying to test it; that is as it should be). A more striking example was the discovery of the three-degree kelvin cosmic microwave background. Before then, there were two models, both with strong support: the steady-state model and the big bang model. While the microwave background was a big boost for the big bang model, the solid state model did not give up without a struggle. There were various attempts to describe the microwave background in the solid state model but they were too little too late. Like the Michelson–Morley experiment, the discovery of the microwave background started a chain reaction that led to the acceptance of one model and the rejection of another.
Perhaps the best way of thinking of crucial experiments is not that they prove (that ugly word) one model better than another, but that they serve as a catalyst. Or perhaps, one can think of a super-cooled fluid that when slightly disturbed, suddenly solidifies. The same phenomenon is seen with people. A group are sitting at lunch and when one gets up to go and they all go, but only if the circumstances are right. Consider the discovery of the J/Ψ particle. The time was right and the background had been prepared so that when it was discovered, the particle physics community solidified around the quark model. Similarly, you can consider Galileo turning his telescope on the heavens as providing the catalyst for the acceptance of the heliocentric model.
Like models, experimental results do not exist in isolation. Rather, they build on each other and are given meaning by the prevailing models. The role of crucial experiments should be seen in relation to that milieu. They do not single handedly overturn or confirm the status quo, but rather, start chains of events that lead to or act as tipping points for the establishment of new paradigms. Thus, crucial experiments do exist but not in the naïve manner envisioned by Bacon, Hooke, or Popper.
Additional posts in this series will appear most Friday afternoons at 3:30 pm Vancouver time. To receive a reminder follow me on Twitter: @musquod.