Astronomical Revisionism

In the course of arguing for school vouchers, Andrew Coulson tries out a little analogy:

Unaware that the planets orbit the sun, pre-Copernican astronomers tried to reproduce their trajectories with clever, intricate, but inevitably doomed geocentric models. In a similar vein, the Center struggled unsuccessfully to reproduce market incentives within their state-centric policy environment.

This is neither here nor there as far as education policy goes, but people are under some serious misapprehensions regarding Copernicus. The thing about pre-Copernican astronomy was that it was actually very good at calculating the orbits of the planets. It was also nicely integrated into a theory of gravity -- objects fall toward the center of the earth because solids have a natural tendency to direct themselves toward the center of the universe. Copernicus, by boldly casting aside the geocentric theory, managed to wreck this theory of motion (he had nothing with which to replace it), and was able to replace the previous astronomical tables with new, less accurate ones. Copernicus, you see, assumed that the planets moved around the suns in circles, which gives you the wrong results. Pre-Copernican astronomers, by contrast, were able to use epicycles to very closely match the theory with the observed data.

The virtue of Copernicus' system is that it was much simpler to do Coperican calculations than it was to factor all the epicycles in. People liked it because it was only somewhat less accurate than its predecessor but substantially easier to use. It took decades, however, for later theorists to work out elliptical orbits and the modern theory of gravity that gave "Copernican" astronomy the theoretical foundations and accurate results that it initially lacked.

Comments

And there is the whole parallelax thing: if the earth moves around the Sun, then the observed positions of stars should be sligthtly different now and six months from now -- just as if you are walking around a building with a largish radius, the positions of things on the horizon will change relative to you: a tree due north of the building will be to your left when you are east of the building and to your right when you are west. But (with the instruments avialable at the time) there is no observable parallelax change from 1/1 to 7/1 (or any other 6 month period), So the Copernican theory failed the test everyone agreed it should pass. (The stars are really far away compared to the radius of the orbit of the earth, by the way.)

Posted by David Margolies | January 2, 2007 1:18 PM

You know too much.

Posted by Ezra | January 2, 2007 1:44 PM

Okay, fine, but what was wrong with the writer's original invocation of Copernicus? The epicycles *were* doomed, because they were ad hoc and not predictive. They weren't correct, as MY implies, and simply more complicated -- they were like multiple asterisks added to a bogus theory. Once you've added, oh, 95 asterisks to your theory it's probably time to rethink it. Copernicus did. That was the -- correct -- point the author was making.

Posted by Cthomas | January 2, 2007 1:46 PM

My favorite of the pre-Copernican systems is the Tychonic system, http://en.wikipedia.org/wiki/Tychonic_system, and its last-ditch, head-in-the-sand effort to maintain the geocentric worldview.

Posted by PW | January 2, 2007 1:55 PM

Perhaps more importantly, however, it has been empirically verified that the earth really does go around the sun. I would think that the primary virtue of Copernican astronomy is that it more accurately represented the motions of the planets in space, even though it was initially less accurate at replicating the way these motions appear to an earthbound observer.

Pre-Copernican astronomy was a very nicely developed system for incorporating fudge factors to preserve a false assumption and uphold a misleading view of the universe. It was a good system for the astronomers and physicists of the era, but it would serve us very poorly in an era when we use our theory of gravity to place communications satellites into orbit.

As Matt suggests, Coulson resembles Copernicus in one important way, because he wants to wreck a useful, functioning system and replace it with one that does a poorer job of meeting society's current needs. However, unlike Copernicus, I don't believe the school voucher crowd is offering a revolution that will eventually lead to an education system that more accurately reflects the laws of the universe and better meets the needs of future generations. Don't flatter them with the comparison.

Posted by LaFollette Progressive | January 2, 2007 2:13 PM

How could you empirically verify that the earth goes around the sun? All motion is relative.

Posted by Martin Morgan | January 2, 2007 2:14 PM

Let's not get pedantic, Martin. The geocentric model has been experimentally disproven. The heliocentric model, to date, has not.

Posted by LaFollette Progressive | January 2, 2007 2:20 PM

If its pedanticism you want, neither sun orbits earth nor earth orbits sun, but both orbit their common center of gravity. In any event, that all motion is relative is hardly a pedantic quibble. It's one of the major achievements of 20th century science.

Posted by Martin Morgan | January 2, 2007 2:27 PM

I think what Matt is going for in complaining about "pre-Copernican" is that it's the wrong astronomer for the citation. You should say "pre-Copernican" with respect to geocentrism, "pre-Keplerian" with regard to circular orbits and epicycles vs. elliptical orbits, and "pre-Newtonian" with regard to the modern understanding of gravity (never say "post-Newtonian", however, to mean the years after publication of the Principia; the term is reserved for a relativistic framework viewed as a set of corrections to Newton's gravitational theory).

Posted by jfaberuiuc | January 2, 2007 2:32 PM

I do not believe cthomas is correct when he says:

"The epicycles *were* doomed, because they were ad hoc and not predictive. They weren't correct, as MY implies, and simply more complicated"

Whatever set of astronomical data you have, there is an epicycle model that fits it to well within observation error (that is, fits it as well as any model fits it). As one of my professors said at graduate school (in Math): epicycles are dense.
(Mathematicians will understand. Others, it repeats in mathspeak what I said just above.)

The epicycle model was, therefore, (1) more accurate that the copernican model, (2) came with a consistent and commonsensical theory of gravity, and (3) did not have a parallelax problem. The question is then, how did it end up on top? (See Kuhn for a suggested answer.)

Posted by David Margolies | January 2, 2007 2:35 PM

(a) I don't see where Coulson is guilty of the 'serious misapprehension' to which you refer.

(b) If the general point about Copernicus is supposed to be: the received view that he replaced a worse system with a better one is wrong -- well, yes and no. Copernicus is predictively/descriptively worse, but is better wrt to other metrics such as simplicity. Also, the whole issue of the theory's relation to physical mechanisms of planetary motion is more complicated than you are letting on. For a long time before Copernicus, the explicit goal of astronomy was not to provide a physical model of how it works for predictive, technological purposes, but to "save the phenomena" in terms of uniform circular motion, for something more like aesthetic reasons (circular motion is the most perfect kind and thus the only kind worthy of heavenly bodies; hence the whole system of epicycles, deferents, etc.), without worrying about what the underlying mechanism is. One of the interesting things about Copernicus is that he is really a transitional figure here; e.g. he retains the idea that a satisfying account has to be in terms of circular motion, and a physical mechanism doesn't play a role in his story. To that extent, Copernicus still isn't fully 'scientific' in a modern sense. It's not until Kepler, taken together with Newton, that you get the real tying of the theory to a physical mechanism, and the abandoning of the older aesthetic in favor of a different idea of what a satisfying explanation is.

Anyone interested in these issues, and especially the idea that the history of science is not one of monotonically increasing rationality, is encouraged to consult Feyerabend's Against Method, which I suspect is in the background of MY's post.

Posted by live | January 2, 2007 2:35 PM

While we're being pedantic, all non-accelerated motion may be relative, but accelerated motion is not, and we can demonstrate that in the solar system (neglecting the acceleration in the galaxy) the non-accelerating frame is centered very near the sun (the barycenter is primarily determined by the sun, with the largest correction by at least an order of magnitude coming from Jupiter, not the Earth).

Posted by jfaberuiuc | January 2, 2007 2:36 PM

The sun and earth orbit a common center of gravity that is located beneath the surface of the sun. As I said later in the same paragraph, "the primary virtue of Copernican astronomy is that it more accurately represented the motions of the planets in space, even though it was initially less accurate at replicating the way these motions appear to an earthbound observer."

Don't be an ass.

Posted by LaFollette Progressive | January 2, 2007 2:38 PM

Martin - Einstein says the no frame of reference is privileged - you can use the earth as your fixed frame of reference if you like. But, as Copernicus noticed, its much easier to do the math if you use the heliocentric one as your frame. The question of whether the earth revolves around the sun or vice versa is really pedantry.

The issue in question is how do you verify that the planets don't revolve around the earth?

Posted by Lee | January 2, 2007 2:40 PM

If its pedanticism you want, neither sun orbits earth nor earth orbits sun, but both orbit their common center of gravity.

Which is within the sun. Hence heliocentrism.

Posted by Al | January 2, 2007 2:49 PM

Actually, I think I just said the same thing as Matt, only more turgidly. Well, that's why he's the pro. Sorry about that, chief!

Posted by live | January 2, 2007 2:59 PM

Copernicus was engaged in using Occam's Razor -- his heliocentric system required fewer epicycles than Ptolemy's geocentric system. Kepler then simplified it farther by getting rid of the requirement of orbits being perfect circles.

It's amusing to read this and then read the comments on the "Discrimination" post below, in which readers concoct some awfully Ptolemaic elaborations to maintain their faith, against all evidence, in the preconception that males and females are exactly the same except for how they are socialized!

Posted by Steve Sailer | January 2, 2007 3:04 PM

Look I'm not some ec-centric. But the fact is helio/geo- centrism are not absolutes.

It's like saying you can empirically verify the North Pole is north of the South Pole. It's all dependent on the initial frame of reference.

Posted by Martin | January 2, 2007 3:14 PM

Martin, there is an important empirical difference between the geocentric and heliocentric frames, namely the heliocentric frame is much, much closer to being an inertial reference frame.

Posted by Ben V-L | January 2, 2007 4:10 PM

Weren't you just bashing Friedman for standing up for the flat world model?

Posted by aleks | January 2, 2007 4:15 PM

Foucault's Pendulum was, as I recall, the first bit of empirical evidence for the heliocentric model (in the late 1800's!).

It's worth noting that there's a pretty wild difference between the earth-orbiting-the-sun of the heliocentric model and the sun-orbiting-the-earth of the geocentric model -- the latter explains the year, while the former explains the day. So Foucault's pendulum provides evidence that the earth is, in fact, rotating on its axis, which would be unnecessary in the geocentric model.

The geocentric model would call for some crazy speeds, now that we know the true orbital distances.

Posted by Michael Sullivan | January 2, 2007 4:38 PM

Actually, now that I think more about it, my earlier post may be inaccurate. Certainly, Gallileo's discovery of the phases of Venus constitutes considerably earlier empirical evidence for the heliocentric model.

And I'm not 100% sure that the contemporary-to-Copernicus model explained the day via solar motion.

Posted by Michael Sullivan | January 2, 2007 4:46 PM

I wonder if pre-Copernican orbit forecasters had had access to high speed, automated computation for their epicycles, would the Copernican model have caught on? If the intial advantages of the Copernican theory weren't substantial, when would we have moved to the heliocentric or Newtonian model? Presumably, not until the simper, Newtonian model allowed forecasts that were simpler, faster and MORE accurate. As an exercise in an alternative history of science, when would this have occurred?

Does our access to machine calculation encourage us to follow more finely approximated, but ulitmately less revealing, lines of research?

Posted by Fred | January 2, 2007 4:48 PM

Well, as an analog, we have lots of ideologies today that require huge numbers of moving parts to describe the world even halfway accurately. The intellectuals who can operate these absurd mental models take great pride in their brainpower that allows them to keep such unwieldy contraptions.

For example, it doesn't take all that much intelligence to notice that people differ in intelligence, and that that has all sorts of ramifications. It takes a lot more smarts, in contrast, to construct quasi-plausible denials of these obvious facts, which, in turn, demonstrates the superior IQ of the IQ-deniers!

And that's mostly what people want from espousing ideologies: prestige and self-esteem.

Posted by Steve Sailer | January 2, 2007 6:50 PM

There is a critical difference between Copernicus' conception of nature and that which you described. Copernicus had no sense of any notion of gravity. Had you explained "gravity" to him he would have rejected it. He would have preffered and indeed subscribed, like so many of his peers, to what we know as an Aristotelian conception of nature whereby things had a natural tendency to move towards their "place" and that natural tendency was from a principle within itself, or rather by virtue of itself. There is no outside cause. Left to itself a thing would move to its "place", and there could be no such thing as gravity.

Posted by Charles B | January 3, 2007 12:54 AM

Michael Sullivan: the transit of venus is not proof of heliocentrism, just that Venus may be between the earth and sun (which the geocentric models allowed anyway). you may be thinking of observations of the changing phases of venus, established by Galileo. That disproved the Ptolemaic model of the solar system, but does not actually prove that the earth moves about the sun, just that venus must move about the sun. the earliest definitive proof that the earth moved around the sun was the observation of the parallax reflex motion of the nearest stars. This was long expected to be the definitive proof, but it was so hard to measure, that by the time astronomers got there (1838 for 61 Cygni), everyone was already completely convinced. The observations of the transit of Venus was an important milestone in setting the true scale of the solar system. Until then, all orbits could be determined relative to the earth-sun distance, but that distance was poorly measured.

Fred and others: the epicycle concept was essentially a fourier decomposition of the elliptical orbits, and can be a perfect representation of the orbits, as long as enough terms are used. Moving from the geocentric to the heliocentric version removed the first big term from all orbits and added one to the earth (actually, probably added several). The result was an improved simplicity for the model, but as MY says, no improvement in accuracy for orbits, but an explanation for the changing brightnesses. It is still not uncommon for numerical models of orbits to use the equivalent fourier decomposition to speed up calculations of orbits, and in fact the fourier analysis technique was invented for exactly that reason. It turns out that, even though the elliptical orbit can be simply written, solving for position as a function of time can be a real hairy problem. Many of the famous math techniques used for solving equations were developed in the late 1700s and early 1800s to model the motions of objects in the solar system.

Posted by eam | January 3, 2007 1:38 PM

blfq ucgxbda imyvwrno ljvxk mgforwcun axmjbzstp npxc

Posted by nxmgsbr dtwgcnbj | June 11, 2007 4:24 PM

pewn dkergixj viasntqpu whvpbofr klmc cbxmvjpu knsmrzpyd http://www.uoryvnjcl.ngckuodqs.com

Posted by ieyq aesbzolut | June 11, 2007 4:24 PM