The recent discovery of a planet which has been [nick]named Xena, more distant and larger than Pluto, is stimulating a widespread discussion over the definition of a planet. I've suggested that Pluto should be demoted from the status of a planet, but I find that when I try to define a planet, there is no real need to demote Pluto. I'd like to describe why, but I want to emphasize two things. First I'd like to approach the problem by trying to defining "moon," which is a closely related concept. And I'd like to emphasize the question of time, namely how our categories need to accommodate not just change in our knowledge base as we explore and discover new things, but also dynamic change in the universe.
Let me note right at the start that there are many terms in astronomy which express old concepts but are no longer accurate. We speak of bodies rising and setting. We say that there is a sun and that there are stars, but only in relatively recent times have we started thinking of our sun as another star, or speaking poetically of the stars as other suns.
We speak of the moon, but now we know that our moon is just one among many. Jupiter has four moons, discovered by Galileo, and today we call them the Galilean satellites. When Sputnik launched, we called that satellite the first artificial moon. But increasingly we think of everything in the universe as either bound up gravitationally with other bodies or in actual orbits. The earth, Halley's comet, our sun, and perhaps even the Milky Way galaxy are all satellites of something larger.
Our common sense is that "moons" orbit planets, but since the latter term is ill-defined, I'm not sure we really know anymore what a moon is. But let's say that stars should not be classified as moons. Can we restrict the term "moon" to bodies where gases have never collapsed into nuclear fusion? Jupiter and Earth are two such bodies. So let's restrict the term further to bodies that do not orbit stars or former stars. Well, our Moon orbits the Sun. The latter's gravity acts on the Moon and is enough to keep it, along with Earth, in solar orbit. Astronomers sometimes speak of the Earth and Moon as a double planet system, because they orbit around a focal point located deep inside the Earth, close to the geometrical center of the Earth but not quite there. If the Moon were much larger, the focal point would lie between the surfaces of the two planets: then we would have to decide whether indeed to call them two planets, or to call the more massive one a planet while naming the other a moon. We need to restrict the definition further.
Let me suggest that a moon is a body which orbits a focal point inside the "surface" of at least one body that is non-stellar (by a stellar body I mean an object in which gases are, or once were, igniting in nuclear fusion). If the said focal point lies outside another body's surface or atmosphere (note that the latter's boundaries may have to be defined arbitrarily), then we may speak of a double-body system, rather than a body and its moon.
Suppose that a small body like a comet were to pass by our Moon and be captured into lunar orbit. We would call it a moon of our Moon. The new moon fits the above definition: it is orbiting a focal point inside at least one body that's not a star (it also happens to be in orbit around the earth, and in an orbit around the sun). The original moon retains its status as a moon, because it continues to fit the definition. A capture, unless it changes the original body in some drastic way, should not change the definition of that body. Now, the earth will have two moons: unless drastically changed, Earth will retain its definition, whatever that may be. But Earth will not be called a moon, because it does not orbit at least one non-stellar body. Its primary orbit is around a stellar body. Is that the definition of a planet? That it orbits primarily a stellar body? I think it is, so long as the orbiting body is not itself a stellar object. But I'd like to get at this by exploring one more issue: change.
A capture of an asteroid or comet seems like an unlikely event. But in the early solar system, such events probably occurred. Mars and the gas giants seem all to have captured smaller bodies at undetermined moments in solar system history. We hear today of asteroids passing "dangerously" close to the Earth-Moon system (we even think our moon formed from a collision of a former planet with earth), and though our aged and stable solar system may not be producing captures and collisions now, there is nothing in the law of physics preventing such a thing. There is much perhaps in our models of solar system formation that would argue against the probability of such a thing, but there are many solar systems in our galaxy, all of which will probably challenge our models of how solar systems form.
And even if we stick with our own, let's look at the big picture. In the early years, planets, moons, comets and all the rest were moving around in a riotous roullette; and when the sun begins expanding a few billion years from now, much of the current stability will again be disturbed. The histories of solar systems may contain, what is to our minds, long periods of unstable conditions in which it is not possible to say which bodies are the planets, how many moons they have, which bodies still retain large masses of gas and which ones are "terrestrial", etc. Orbits and masses are variable for much of the life of a solar system.
I think we'd be better off speaking of planets that are currently of a certain class or type: this planet over here currently has its gases and looks like a gas giant; that one over there retains gases but is in an elliptical orbit, and looks like a comet-type planet; another is a terrestrial-type planet with some gases; these others are irregularly shaped planets, of the asteroid-type; still others are Kuiper-type planets; and finally there are what we call Moon-type planets, orbiting not the star primarily but another planet. And for all of these we say, "currently", or "now."
The issue of change seems critical to me because our short lives, and our short history on the earth, makes the universe appear static to us, when actually it is not. It is alive with constant activity, much of it unstable or nascent -- and our stable conditions at this point in time, in this corner of space, make us still less aware of the universe's essentially dynamic character.
A planet much larger than Jupiter might live what we would consider a long time, several million years perhaps, before igniting into a sun. What we called its moons, if it had any, would cease to be moon-type planets (i.e., planets orbiting other planets), and would then become planets of another kind, depending on what the stellar fusion did to them (it could destroy them).
Other factors have restricted our vision of what planets are. Early in human history we saw only lights in the sky and called them stars, with some of them marked off as wandering stars, or planets. Modern science revealed that these wanderers were worlds of great size, comparable to our own world. But we have not seen them in their full glory -- or rather, we have only begun to see them as great, fascinating worlds. Until very recent human history, other planets, and even our Moon, have not pressed upon our consciousness as anything more than lights larger than the stars; they were reputed to be much more, but the details have been slow in coming. Earth-based photographs have started to change our impressions, as have visits by planetary probes; and critically, so have the photos from Hubble. Meanwhile, ground-based telescopes have gotten larger, and more accessible. Now we're starting to realize that all these bodies, even little asteroids and comets, are worlds of size and great interest. And when we visit them personally, our consciousness will regard them with greater equality, as non-stellar worlds.
The more I regard and study the Moon, the more I tend to think of it as a planet. It switches places with Earth twice every month as the third rock from the sun. It is sometimes called a failed planet, but it only failed to become an Earth-type planet. It is quite a massive world when seen in some of the best composites of Apollo photographs from the lunar surface; and it is an interesting world with the same planetary structure (core, mantle, crust) as Earth, and a similar historical process (up to a point). The Moon once had flowing lava and an atmosphere: at one time we may have had to include such a world in a different category of planet, closer to ours. Planets, even in a stable solar system, evolve enough to change categories -- or to push us to change our categories.
In sum, a moon is any object which orbits a non-stellar body; a planet is any object, but not a star, which orbits a stellar body.