When scientists talk about how forces act on objects, all of us can envision what happens when we push tables around a room. No doubt that we are talking about real motions and objects, and we can all agree on their form.
Not so true of the unseen objects and forces that science hypothesizes on very small scales. Are there, in fact, atoms and nuclear forces? Or are they imaginary objects that simply embody the ideas of current theories?
I would have said, as a scientist, that unseen objects described by scientists are, in fact, real, a doctrine known as scientific realism. The case for and against realism is nicely laid out in a recent podcast of Philosophy Bites, interviewing David Papineau. The arguments also apply to the interplay between data (observation and intelligence) and the hypotheses and theories that try to explain and extrapolate from them, for example in controversial areas of ecology and economics.
I didn’t realize that Newton understood science only to be descriptive, not explanatory. He realized that his Law of Gravitation captured the form of gravitational attraction, but did not relate to what gravity was or why it existed. It wasn’t until the birth of atomic theory, when so much evidence pointed to the existence of discrete atoms, that science again had confidence to state that their unseen objects were, in fact, real things.
Papineau describes two way in which we can be skeptical about the existence of unseen things that science asserts to be real. One is that the data is usually sparse enough that multiple explanations can fit the observations. We can think of some as being more unlikely than the others (God has arranged the universe so as to produce this misleading result, contrary to the real objects involved), paradigm shifts toppling accepted models still happen with some regularity. Atoms are now envisioned as wave packets rather than billiard balls.
This leads to the second source of skepticism: scientists have been wrong in the past, so they are probably wrong now. But look at the process: when two theories compete to explain the same data, someone performs a decisive experiment and one theory fails. Predictions are made from the surviving theory, data is collected, some is not explained, and, again, the hypotheses arguments are challenged and some are discarded. Science makes progress, one theory ultimately survives, and many are discarded. This doesn’t make science more often wrong than right: it says that is approaches truth iteratively.
I really liked this program: it goes to the nub of what beliefs I can trust, and which I should question. It reminds me that it’s easy to confuse reality and appearance, and that I should scratch the surface now and again before asserting that something is true. And it’s endless fun to think up alternatives that explain my data.