Physicists frequently stray into the field of philosophy; notable examples include Thomas Kuhn (1922 –1996) and Henri Poincaré (1854 – 1912). This is perhaps because physicists frequently work in areas far removed from everyday experiences and, in order to be successful in communicating their ideas, underlying assumptions must be dealt with explicitly. Although less well known today than Kuhn and Poincaré, Percy Bridgman (1882 – 1961) also falls into this category. In physics, he is noted for his work on high-pressure physics, winning the Nobel Prize in 1946. In philosophy, he is credited with coining the term OPERATIONAL DEFINITION and promoting the idea of operationalism. These ideas are laid out in the 1927 book: THE LOGIC OF MODERN PHYSICS. If nothing else, it shows the folly of using MODERN in book titles. None-the-less, it is an interesting little book and, in its time, quite influential.
In his book, Bridgman introduces several interesting ideas, for example, that when one explores new areas in science, one should not be surprised that the supporting concepts have to change. Hence we should not be surprised when classical concepts fail in the relativistic or quantum domains. This illustrates why interpretations of quantum mechanics, explaining it terms of classical concepts, are poorly motivated. A related idea is that an explanation is the description of a given phenomenon in terms of familiar concepts.. Of course, with this definition, what qualifies as a valid explanation depends on what the explainee is familiar with. If one cannot succeed using established concepts, one must explain the new idea using familiar, albeit far removed, concepts But what happens when even this does not work? Bridgeman suggests that the solution is to introduce new concepts and become familiar with them. Seems reasonable to me. Thus quantum mechanics can be explained in terms of the, familiar to me, concept of the wave function; no need for many worlds and the like.
While it is natural to think of high speed (relativity) or small size (quantum mechanics) as new areas of science, Bridgman includes increased precision as well. He talks about the penumbra of uncertainty that surrounds all measurements and that is penetrated by increasing the precision of the measurements. Thus the idea of the distinct high-energy and precision frontiers, commonly discussed in modern particle physics planning exercises, goes back at least to 1927.
Bridgman was also a phenomenologist to the core. He did not believe that a priori knowledge could constrain what could happen; in his words: Experience is determined only by experience. C.I. Lewis (1983 – 1964) in his 1929 book Mind and the World-Order agrees. The similar ideas, in books of about the same time, indicate the concerns of that age.
Despite these interesting sidelights, the main idea in THE LOGIC OF MODERN PHYSICS is that concepts are defined by how they are measured; that is by the measurement operation, hence the term operationalism. So why was he interested in operational definitions? It was to avoid the problem in classical mechanics where concepts like distance and time were taken for granted. It then came as shock when the concepts proved to be rather complex when special relativity was invented. To avoid such shocks in the future, Bridgman proposed the idea of operational definitions. For example, to measure length you go down to the local Canadian Tire® store (in the USA it would be Walmart®), buy a tape measure and use it measure length. Thus the concept of length is defined by Canadian Tire®, oops, I mean by a tape measure. What if I measure length by surveying techniques that make use of tranquilization? Bridgman claimed that that is a distinctly different concept and is covered by the same term only for convenience. Here at TRIUMF, distance and location are also measured using laser tacking. This is again a different concept than the original concept of length. Things get even more complicated when we talk about the distance to stars which use again a different operation. Bridgman suggests that length loses it meaning at lengths less than the size the electron because such lengths cannot be measured. Today we would say they can be measured but length in that case is simply a parameter, in a mathematical formula, describing the scattering of particles. Hence we do not have one concept of length or distance but many, although they are the same numerically in regions where the techniques overlap.
Bridgman then goes on to consider various other concepts and how they might be defined operationally. He seems to have been very much influenced by Albert Einstein (1879 – 1955) and Einstein’s discussion of the synchronization of clocks (which actually goes back to Poincaré). The possible operational definitions of velocity are particularly interesting. In contradistinction to the definition given by Einstein based on clocks synchronized and distances measured in a fixed inertial frame, Bridgman suggests that the velocity of a car could also be defined by counting mileposts that the car passes to determine distance and using the clock on the car dashboard to measure time. This velocity can become infinite and would be useful to a person going to a distant solar system who is interested in how many of his years it takes to get there. For most purposes Einstein’s definition is more convenient and hence it is the one in textbooks though other definitions remain possible.
And on it goes. In some cases the definitions seemed quite forced. Never-the-less, three groups of people picked up on the idea of operational definitions. One group was the logical positivists. They tried to avoid theory and were pleased when a physicist gave definitions directly in terms of observables. The second group was the phycologists, who wanted a more secure foundation for their subject. The third group was in quality control and business management where Walter Shewhart (1891 – 1967) and Edwards Demming (1900 – 1993) adopted the idea.
However the concept, as the end all and be all of meaning, had its problems. Like logical positivism, it missed the idea that the meaning is in the model. While we may have different ways to measure length there is common idea behind them all. We can consider this common idea to be an abstraction from the different operational defined concepts or we can take the operational definitions as approximations to the abstract idea. One could say that operationally there is no difference between the two approaches.
Ultimately, operational definitions are useful. They tie concepts tightly to observations where they are less likely to be dislodged by future discoveries or new models. They also help eliminate fuzzy thinking. A lot of the concepts that do not have operational definitions are, in general, poorly defined. Who knows, I might even take the concept of scientific realism seriously if someone gave me an operational definition of it.
To receive a notice of my future posts and my pending book, In Defense of Scientism, follow me on Twitter: @musquod.
