On the Revolutions of the Heavenly Spheres

Overview

On the Revolutions is five hundred and fifty pages, and roughly twenty of them are the ones anyone reads. Book I is a general account of the universe, written in plain Latin for an educated reader. Books II to VI are technical astronomy: spherical trigonometry, a star catalogue, tables, and geometrical models for the motion of the moon and each planet in longitude and latitude. That part is a working replacement for Ptolemy’s Almagest and is very hard going without training.

Copernicus finished it decades before it appeared. He had circulated a short sketch, the Commentariolus, around 1514, and then sat on the full work for thirty years. What broke the deadlock was a young Lutheran mathematician, Georg Joachim Rheticus, who travelled to Catholic Warmia in 1539, stayed two years, published a summary of the theory himself to test the reaction, and finally persuaded the old man to print. The book came out in Nuremberg in 1543, and Copernicus is said to have been handed a copy on the day he died.

The printing produced the book’s strangest feature. Rheticus had to leave partway through, and the job passed to a Lutheran theologian, Andreas Osiander, who added an unsigned note to the reader explaining that astronomical hypotheses need not be true and are merely devices for calculation. It was not Copernicus’ view, it was inserted without his consent, and it was not publicly exposed as Osiander’s work until Kepler said so sixty years later. Whether it protected the book or muffled it is still argued.

The argument

The dedication. Copernicus writes to Pope Paul III, and the letter is a piece of rhetorical management. He explains that he hesitated for years, that he expects to be laughed at, and that he was pushed into publishing by churchmen. His stated motive is not new data but a scandal: the astronomers cannot agree, their models are inconsistent with one another, and the whole enterprise is like a painter assembling a figure from hands and feet and a head taken from different men, each part well drawn and the result a monster.

Book I, the case. The universe is spherical, the earth is spherical, celestial motion is uniform and circular or compounded of circles. So far this is Aristotle. Then the break: the earth is not at rest. It turns daily on its axis, it travels annually around the sun, and its axis holds its direction.

Copernicus argues this the way an ancient would. Against the objection that a spinning earth would fly apart, he asks why the same argument does not apply with far more force to the enormously larger sphere of the stars, which on the old model has to complete a circuit every day. Against the objection that we would feel it, he uses the ship: a passenger sees the shore move and not the vessel. Against the objection that thrown objects would be left behind, he argues that the air and everything in it share the earth’s motion.

Then the payoff, in chapters 9 and 10, which is the part of the book that changed things. Once the earth is a planet, the order of the planets is no longer a matter of choice. Ptolemy could arrange Mercury and Venus more or less as he liked, because nothing in his system fixed them. Copernicus can determine the order from the periods, Mercury fastest and innermost, then Venus, earth, Mars, Jupiter, Saturn, and he can calculate each planet’s distance from the sun in terms of the earth’s. Retrograde motion, which Ptolemy explained with a dedicated epicycle for every planet, becomes a single consequence of watching a slower body from a faster one. And the fact that Mercury and Venus never appear far from the sun stops being a peculiar fact requiring its own rule.

There is a cost, and Copernicus takes it. If the earth moves, the fixed stars should shift against one another over a year, and no such parallax could be measured. His answer is that the stars must be so far away that the earth’s whole orbit is negligible by comparison. That is correct, and in 1543 it meant accepting a universe far larger and emptier than anyone had been asking for.

Chapter 10 ends with the passage everybody quotes, the sun enthroned in the middle of a beautiful temple, governing the family of planets that circles it.

Books II to VI, the machinery. Spherical astronomy and a star catalogue largely taken over from Ptolemy. Precession of the equinoxes, which Copernicus handles as a wobble of the earth’s axis rather than a drift of the stars. The moon, where his model is a real improvement, since Ptolemy’s had required the moon to double in apparent size over a month and it plainly does not. Then the planets, in longitude and then in latitude.

This is where a modern reader’s expectations are most likely to be wrong. Copernicus did not sweep away the epicycles. He was committed to uniform circular motion as a principle, so he needed circles upon circles to fit the observations, and by some counts he used as many as Ptolemy. What he removed was the equant, Ptolemy’s device for letting a body move uniformly as seen from a point other than the centre of its circle, which Copernicus regarded as a violation of the principle rather than a useful fiction. The system is therefore not simpler in practice and not more accurate, and its tables were adopted quickly by working astronomers for reasons of convenience rather than conviction.

How to read it

Read the dedication and Book I. That is around fifty pages and it contains every argument a non-specialist needs. Then stop, unless you have trigonometry and a reason.

Read Osiander’s preface, and then read the dedication, and notice that they are making incompatible claims about what the book is. The question of whether a scientific model is a description of the world or a calculating device is opened on the first two pages of the first modern scientific book, by two people who disagreed about it.

Do not expect new evidence. Copernicus had essentially the same observations Ptolemy had, many of them literally Ptolemy’s. The argument is about what arrangement makes sense of them, and it is worth noticing how much of it is an appeal to proportion, order and fitness.

Read Kuhn if the question of why anyone accepted it interests you. A theory that was no more accurate and no simpler won because it made a set of unrelated facts into consequences of one fact, and that is the clearest available case study in how scientific persuasion actually works.

The common misreading is that the church suppressed the book on publication. It was dedicated to a pope, circulated freely, and was used by Catholic astronomers for seventy years. It went onto the Index in 1616, after Galileo had made it a claim about the world rather than a calculation, and it was suspended pending correction rather than banned outright.

Questions it raises

• Copernicus offers no new observations. What kind of argument is he making, and should it have persuaded anyone?
• He kept uniform circular motion and paid for it with epicycles. Was that conservatism, or the reasonable defence of a principle he had good grounds for?
• Osiander says the model need only save the appearances. Is there a difference between a theory that predicts correctly and one that is true?
• No stellar parallax could be found, so the stars had to be almost inconceivably distant. When is it legitimate to save a theory by enlarging the universe?
• The strongest argument in Book I is that the new arrangement is better proportioned. How much should elegance count as evidence?
• If the earth is a planet like the others, what else in the old picture has to go? Copernicus does not follow that through. Should he have seen how far it went?