This interesting paper deserves to be known more widely: “And Yet It Moves: The Observability of the Rotation of the Earth” by Peter Kosso, Northern Arizona University, published in Foundations of Science, (2010) 15:213–225. Excerpts:
Abstract A central point of controversy in the time of the Copernican Revolution was the motion, or not, of the earth. We now take it for granted that Copernicus and Galileo were right; the earth really does move. But to what extent is this conclusion based on observation? This paper explores the meaning and observability of the rotation of the earth and shows that the phenomenon was not observable at the time of Galileo, and it is not observable now.
In our own time there are lots of outstanding scientific questions regarding objects and events that cannot be observed, but few rise to the intensity of genuine controversy. One that does is the issue of evolution. We cannot observe the origins of life – here it is, after all, well along – and we cannot observe the long process described by evolutionary biology. The controversy, the run-in with creationism and intelligent design, arises from the combination of this inability to observe and the challenges the theory presents to significant cultural ideals. It is exactly this pairing of unobservablity and cultural challenge that made the movement of the earth more than merely an academic detail. p.214
We could use the starry background as the reference system, and say that the earth rotates relative to the stars. That is a precise concept and a determinate claim, but it encounters no disagreement. The x-moves-relative-to-y relation is symmetric; it could just as easily, and just as accurately, be put as y moves relative to x. Had Galileo emphasized, All I’m saying is that the earth rotates relative to the stars, Aristotelians and the Church would most likely have responded, Oh, well, when you put it that way, there’s no problem. p.217
Consider the possible perspectives from which to do the observing. There are three, terrestrial, celestial, and external, that is, observing things on the earth from the earth, observing things in space from the earth, and observing the earth from space. p.217
The key difference between observation and evidence is the necessary role of inference in the latter. It is usually a causal inference, from effect to cause. The effect is observed, and we infer what must have, or is likely to have, been the cause. p.217
To summarize the analysis of the terrestrial perspective, the rotation of the earth is not observable, neither directly nor indirectly. That was clear from the start, and the work has been to determine whether there could be evidence of the rotation. That depends on our understanding of the dynamic connection between what is observed, the tides or the bulge of the equator, and its cause. Using the interpretive framework of Newtonian dynamics, the bulge is good evidence of rotation, and it provides some justification for that aspect of the Copernican model. Using Machian dynamics, the data provide no such evidence. p.220
The important point here is that the celestial data amount to evidence of the earth’s rotation only with the interpretive help of some theoretical understanding of motion. Whatever is being observed, it is not the rotation of the earth. Whatever is being observed is, by way of some interpretive principles, evidence of rotation. The celestial perspective does not make the rotation observable. p.221
The point is that the splendid video of the rotating earth, as filmed by the Galileo spacecraft, requires some understanding of forces and the causes of motion in order to count as evidence that it is indeed the earth that is rotating. It is, in other words, evidence but not observation. There is no perspective from which the rotation of the earth is observable. p.222
The underlying reason that the rotation of the earth is unobservable can be clarified using the modern physicists’ distinction between kinematics and dynamics. p.222
This distinction allows for a very clear summary of the difference and dispute between the two chief world systems. They are kinematically equivalent. p.223
The difference between Tycho and Copernicus is in the dynamics. Once we understand the dynamics of the situation, in particular the real nature of gravity, there is simply no way the Tychonic system could work. It violates laws of physics, that is, laws of dynamics. Following Newton, the laws of dynamics single out a group of reference frames, the inertial reference frames, as being those in which the laws such as F = ma are true. In a spinning reference frame, or an accelerating frame, fictitious forces show up. On a carousel or in a fast car rounding a bend you feel a force to the outside. But this is no force; there is no F causing any a. It is the result of being in a noninertial reference frame. This dynamical principle can be used to specify the appropriate reference frames for measuring properties of nature, appropriate in the sense of avoiding fictitious causes and effects. These are the inertial reference frames, those in which Newtonian dynamics work. p.223
Before we celebrate, full disclosure requires noting that there are alternatives to the Newtonian laws of dynamics, and since the dynamics direct the inference, there are alternative conclusions regarding the rotation of the earth. Mach, again. Mach’s Principle claims that the determination of inertial reference frames is fundamentally in reference to the aggregate masses of the universe, all the stars and galaxies. This restores the relativity of rotation. That force you feel on the carousel is not fictitious at all; it is a real force caused by the relative motion between you and the distant stars. There are two aspects of gravity, by Mach’s dynamics, the normal force of attraction described by Newton and an additional force that arises when there is relative acceleration between the masses. … Under the influence of Machian dynamics, the rotation of the earth is not only not observable, it is not properly defined. The property is only determinate with respect to some other actual object. p.223
Observing the motions of things on the earth or in the sky is an act of kinematics, and it cannot uniquely determine the nature of the cause of the motion. Kinematics does not determine dynamics. On the other hand, once the dynamics is known, the kinematics can be inferred. If you know the forces, you can predict the motion. But from the motion you cannot infer the forces. The observable details of rotation are matters of kinematics. They indicate what is happening but not why. Observation tells us how things move but not what moves them. p.224
Back to the Copernican Revolution, then, there is no observable difference between the two chief world systems. It takes a dynamical theory to interpret the data. Aristotle, whose dynamics described natural motion as earthen matter seeking its natural place at the center of the universe, would naturally put the earth at the center of all things, though this does not secure its being motionless. Galileo, who participated when the Revolution was a work in progress, contributed an early version of the concept of inertia that led to Newton’s first law of dynamics. But it was not until Newton himself put the dynamics in order that the difference between the world systems was determined, both conceptually and empirically. This means that belief in the Copernican system before Newton was somewhat premature, since the dynamics needed to interpret the evidence as supporting the rotation of the earth was yet to be written. p.224