About 40 years later the French mathematician Pierre-Simon de Laplace came up with an alternative theory of enormous intrinsic beauty, whose sole flaw is that it doesn't actually work. Laplace thought that the Sun formed before the planets did, perhaps by some cosmic aggregation process like Kant's. However, that ancient Sun was much bigger than today's, because it hadn't fully collected together, and the outer fringes of its atmosphere extended well beyond what is now the orbit of Pluto. Like the wizards of Unseen University, Lapkce thought of the Sun as a gigantic fire whose fuel must be slowly burning away. As the Sun aged, it would cool down. Cool gas contracts, so the Sun would shrink.
Now comes a neat peculiarity of moving bodies, a consequence of another of Newton's laws, the Law(s) of Motion. Associated with any spinning body is a quantity called 'angular momentum', a combination of how much mass it contains, how fast it is spinning, and how far out from the centre the spinning takes place. According to Newton, angular momentum is conserved, it can be redistributed, but it neither goes away nor appears of its own accord. If a spinning body contracts, but the rate of spin doesn't change, angular momentum will be lost: therefore the rate of spin must increase to compensate. This is how ice skaters do rapid spins: they start with a slow spin, arms extended, and then bring their arms in close to their body. Moreover, spinning matter experiences a force, centrifugal force, which seems to pull it outwards, away from its centre.
Laplace wondered whether centrifugal force acting on a spinning gascloud might throw off a belt of gas round the equator. He calculated that this ought to happen whenever the gravitational force attracting that belt towards the centre was equal to the centrifugal force trying to fling it away. This process would happen not once, but several times, as the gas continued to contract, so the shrinking Sun would surround itself with a series of rings of material, all lying in the same plane as the Sun's equator. Now suppose that each belt coalesced into a single body ... Planets!
What Laplace's theory got right, but Kant's did not, was that the planets lie roughly in a plane and they all rotate round the Sun in the same direction that the Sun spins. As a bonus, something rather similar might have occurred while those belts were coalescing into planets, in which case the motion of satellites is explained as well.
It's not hard to combine the best features of Kant's and Laplace's theories, and this combination satisfied scientists for about a century. However, it slowly became clear that our solar system is far more unruly than either Kant or Laplace had recognized. Asteroids have wild orbits, and some satellites revolve the wrong way. The Sun contains 99% of the solar system's mass, but the planets possess 99% of its angular momentum: either the Sun is rotating too slowly or the planets are revolving too quickly.
As the twentieth century opened, these deficiencies of the Laplacian theory became too great for astronomers to bear, and several people independently came up with the idea that a star developed a solar system when it made a close encounter with another star. As the two stars whizzed past each other, the gravitational attraction from one of them was supposed to draw out a long cigar-shaped blob of matter from the other, which then condensed into planets. The advantage of the cigar shape was that it was thin at the ends and thick at the middle, just as the planets are small close to the Sun or out by Pluto, but big in the middle where Jupiter and Saturn live. Mind you, it was never entirely clear why the blob had to be cigar-shaped ...
One important feature of this theory was the implication that solar systems are rather uncommon, because stars are quite thinly scattered and seldom get close enough together to share a mutual cigar. If you were the sort of person who'd be comforted by the idea that human beings are unique in the universe, then this was a rather appealing suggestion: if planets were rare, then inhabited planets would be rarer still If you were the sort of person who preferred to think that the Earth isn't especially unusual, and neither are its life-forms, then the cigar theory definitely put a crimp on the imagination.
By the middle of the twentieth century, the shared-cigar theory had turned out to be even less likely than the Kant-Laplace theory. If you rip a lot of hot gas from the atmosphere of a star, it doesn't condense into planets, it disperses into the unfathomable depths of interstellar space like a drop of ink in a raging ocean. But by then, astronomers were getting a much clearer idea of how stars originated, and it was becoming clear that planets must be created by the same processes that produce the stars, A solar system is not a Sun that later acquires some tiny companions: it all comes as one package, right from the start. That package is a disc, the nearest thing in our universe (so far as we know) to Discworld. But the disc begins as a cloud and eventually turns into a lot of balls (Stibbons's Third Rule).
Before the disc formed, the solar system and the Sun started out as a random portion of a cloud of interstellar gas and dust. Random jigglings triggered a collapse of the dustcloud, with everything heading for roughly, but not exactly, the same central point. All it takes to start such a collapse is a concentration of matter somewhere, whose gravity then pulls more matter towards it: random jigglings will produce such a concentration if you wait long enough. Once the process has started, it is surprisingly rapid, taking about ten million years from start to finish. At first the collapsing cloud is roughly spherical. However, it is being carried along by the rotation of the entire galaxy, so its outer edge (relative to the centre of the galaxy) moves more slowly than its inner edge. Conservation of angular momentum tells us that as the cloud collapses it must start spinning, and the more it collapses, the faster it spins. As its rate of spin increases, the cloud flattens out into a rough disc.
More careful calculations show that near the middle this disc thickens out into a dense blob, and most of the matter ends up in the blob. The blob condenses further, its gravitational energy gets traded for heat energy, and its temperature goes up fast. When the temperature rises enough, nuclear reactions are ignited: the blob has become a star. While this is happening, the material in the disk undergoes random collisions, just as Kant imagined, and coalesces in a not terribly ordered way. Some clumps get shoved into wildly eccentric orbits, or swung out of the plane of the disc; most clumps, however, are better behaved and turn into decent, sensible planets. A miniature version of the self-same processes can equip most of those planets with satellites.
The chemistry fits, too. Near the Sun, those incipient planets get very hot, too hot for solid water to form. Further out, around the orbit of Jupiter for a dustcloud suitable for making our Sun and solar system, water can freeze into solid ice. This distinction is important for the chemical composition of the planets, and we can see the main outlines if we focus on just three elements: hydrogen, oxygen, and silicon. Hydrogen and oxygen happen to be the two most abundant elements in the universe, apart from helium which doesn't undergo chemical reactions. Silicon is less abundant but still common. When silicon and oxygen combine together, you get silicates, rocks. But even if the oxygen can mop up all the available silicon, some 96% of the oxygen is still unattached, and it combines with hydrogen to make water. There is so much hydrogen, a thousand times as much as oxygen, that virtually all of the oxygen that doesn't go into rocks gets locked away in water. So by far the most common compound in the condensing disc is water.