The following is a collection of materials I have produced on the question of defining the word "Planet". This is a controversial issue, and the articles here primarily represent my own opinion on the topic. Opinion in the community has always been divided, but has drifted towards the position I've advocated for since 1999, to the point that the current IAU resolution activity almost converged on it. Obviously, others have also suggested similar schemes. In any case, you will find that the material below contains or refers to the most comprehensive and penetrating analyses of this whole topic. The image above is from the 2003 Mercury article.
A brief commentary on the IAU August 2006 General Assembly resolutions establishing nomenclature for Solar System planets. I applaud the basic decision to make "roundness" the basic criterion for establishing the lower limit of the planetary mass domain. I'm sorry the IAU didn't actually take on the task of defining "planet", but restricted the discussion to our Solar System (actually, considering how things went, maybe I'm not that sorry!). I agree that Pluto, Ceres, and the other round KBOs deserve their own classification, since they really are smaller than the other planets and do orbit in belts of numerous even smaller objects. I strenuously object to the proposition that they are not planets - in fact the IAU calls them "dwarf planets", which sure still sounds like planets to me. I have suggested - and the public reaction certainly bears out my reasons for doing so - that one should consider "dwarf" an adjective instead of a noun. Then Pluto retains its planetary status, but we can distinguish it and its ilk from the 8 major planets in the Solar System. Some feel that public reaction is irrelevant to a scientific definition (and the argument over parts of speech does suggest a certain lack of substance to the whole thing). I actually provide a scientific definition (with a slightly different word) in what I propose; it avoids all concerns beyond the physically-motivated setting of the planetary mass range between roundness and the onset of thermonuclear fusion. To pretend that the definition of the word "planet" has no cultural significance, however, is to invite culture to bite back, as it is doing. It would have been easy to do the right thing both culturally and scientifically, and I am disappointed that my colleagues did not take the chance to do so.
Now, the supporting material...
This is the one scientists or informed laypersons should read. It is primarily a detailed discussion of all the issues governing the general definition of "planet" (for both small and large planets). It notes that the debate on defining "planet" really takes place in four disparate arenas: Characteristics (physical properites), Circumstances (all the orbital considerations), Cosmogony (how objects are formed), and Culture (people's preconceptions and historical inertia). Most of the arguments and fun-poking this debate generates are due to the advocates not realizing they are operating in the different arenas, or insisting that theirs is the right one. Mike and I offer different solutions; mine more closely resembles the one originally up for consideration by the IAU (but is more generally applicable). Mike's contains much of the dominant opposing position, with the stronger insistence on dynamical dominance that was finally adopted by the IAU for excluding "dwarf planets" as "planets".
What is a "Planet"? (Mercury 2003 article)
This is the one by me that most lay people should read. It provides the rationale for the IAU decisions, and why I say what I say at the top of this page.
IAU Final Resolution: Solar System Planet Definitions
This is what finally passed at the IAU General Assembly on Aug. 24, 2006
It set the lower limit for planetary masses to be gravitationally-induced roundness. It makes a couple of political mistakes. It defines "classical planet" and "dwarf planet" separately. Both have roundness as a lower mass limit, but planets clear their orbits of other bodies, while dwarf planets don't. This is a very circumstantial definition: if one took Mars and put it into Xena's orbit, it would be demoted to a dwarf planet, since dyamical dominance depends both on your distance from the star and on the presence of neighboring large planets. It also establishes the term "plutonian object" to distinguish dwarf planets beyond Neptune from Ceres (which is the only dwarf planet I know of that is not plutonian).
IAU Working Group on Extrasolar Planets : Large Planet Definition
This sets the upper limit for planetary masses to be the fusion boundary
My proposal for defining planets splits the astrophysical issue from the others (see Ann. Rev. article below) by first defining "planetary mass objects" or "planemos". This is a purely mass-based definition, and the IAU has also now proposed the same mass limits: gravitational roundness at the lower end and fusion cutting off the upper end. Objects more massive than 13 jupiters I call "fusors" (including brown dwarfs and stars) because they all at least fuse deuterium into helium (the essence of a "star's" distinction, and something planets should not do). [Frankly, I'd be happy to just call fusors "stars", and make brown dwarfs a special kind of star, but that is off the subject...] This led to the following definitions:
Planemos: round non-fusors (roundness refers to the surface closely following a gravitational equipotential)
Planets: Planemos orbiting fusors (orbiting refers to the primary orbit - with the closest center of mass)
This is a very succinct and easy-to-understand scheme. I believe it is preferable (although consistent with) what has been decided by the IAU. It does make for a number (tens at least) of overall planets in our Solar System, but that is easily dealt with. The definition of "dwarf planets" passed by the IAU should have been presented as the definition of a new subclass of planets, with "dwarf" as an adjective (as with "gas giant planets", for example). That way, Pluto would officially still be a planet, albeit a particular sort of planet. Schoolchildren would learn about the classical 8 planets first, and about the dwarf planets in a general way (with a couple of examples, like Pluto and Ceres). The Solar System would be understood to have a good many dwarf planets, but only 8 major ones. We also have dwarf stars and giant stars, but nobody questions whether any of them qualify as stars. The particular treatment of Pluto is a political error that I hope can be rectified somehow.
Furthermore, the IAU has definitely not defined the word "planet". Instead it has provided nomenclature for planets (classical and dwarf) that applies only to our planetary system. Planemos are what planetary scientists and astronomers actually study: they include moons and free-floating 10 jupiter mass objects along with everything else in this mass range, without prejudice as to where they are found, what they are doing, or how they were born. Planets as I define them are a particular instance of planemos, consistent with the cultural use of the term. These terms apply not only to the Solar System, but in general. "Planemo" avoids all questions about surrounding populations of bodies, orbital circumstances, modes of formation, definition of "moons", etc. I encourage liberal use of adjectives (including "dwarf") as needed to describe more about the planet you are discussing.
Objection 1) There would be too many planets in our Solar System.
Answer 1) I suggested dealing with that in the Ann. Rev. article by labeling them "mini-planets" to distinguish them from the 8 major planets. It is a fact that our Solar System has a lot of such bodies; the question is what to call them. Schoolkids should only have to learn the 8 major planets, and know that there are many dwarf planets. The most important of these are Ceres, Pluto and [Xena] right now, but the list is likely to change. I understand the cultural inertia suggesting that 10 planets are OK but 50 are not, but since the science is not so neat we just have to choose our words and teaching carefully. The nice thing about adjectives is that they can be applied somewhat inconsistently and subjectively without creating the heat that the planet debate generates. I view the IAU resolution as flawed only in that "dwarf" should be an adjective modifying "planet". While this is an entirely semantic distinction, I'd venture to say that the public would have a very different reaction to it!
Objection 2) Planets should be dynamically dominant, and not found in belts of objects with too similar orbits and masses.
Answer 2) That sounds at first like a reasonable proposal, but is very hard to apply and vague in general. The issue of dynamical dominance depends on the relative masses of the objects, the presence or absence of sufficiently larger objects in sufficiently nearby orbits, the distance of the belt from the star, and other factors. In our system it turns out to be relatively easy to apply this criterion, but it is likely to be problematic in general. It suffers from the arbitrary question of when the number of companion objects rises to the status of a "belt" or "population" (Trojan asteroids don't count, a certain number of orbit-crossing objects don't count), all the planets were in a "belt" (the circumstellar planetesimal disk) at sufficiently early times, and so on. The part of the Resolution dealing with the "neighbourhood " is therefore a little more problematic - just exactly what does that mean and what constitutes "cleared"? For example, how much evidence is there that Xena has not cleared its neighborhood (or at least lives in a very uncrowded neighborhood); it certainly does not orbit in the classical Kuiper Belt. One can easily imagine ambiguous extrasolar systems. However, so long as one is arguing about the adjective "dwarf" and whether it applies to a particular object, the fundamental definition of planet need not be disturbed. My main complaint is that the "definition" is really just aimed at our Solar System, as is clear from the use of "Sun" rather than "star" throughout.
Objection 3) There is no real difference between a non-fusing object of 10 jupiter masses which forms by direct gravitational collapse and a 14 jupiter mass object that forms the same way. They should both be called "brown dwarfs" or the smaller one should be called a "sub-brown dwarf".
Answer 3) This is something of a matter of taste. The question is whether to go with fusion as opposed to direct gravitational collapse as a more basic property of stars. The issue of "forming in disks" is complicated; one could argue that close binary stars form in a (circumbinary) disk, and it is possible that planets as small as Jupiter form quickly by gravitational instability in the disk as well (show me the core!). I prefer the obvious qualitative difference of thermonuclear fusion as being well understood and observationally verifiable. At the upper end of planetary masses, I would help out my friends by calling anything above 2 jupiters a "superplanet" (because degeneracy pressure support begins to set in). This has the advantage of relegating all the objects of questionable cosmogony to that category (separating them from "normal" planets).The substantial contingent who prefer to use cosmogony to define planets should use adjectives (like "agglomerated" or "direct collapse") to make their points, rather than trying to define "planet" itself using a hard-to-observe and not always firm theoretical basis with a rather striking Solar System bias built in. It is obvious by now that we have a lot more to learn about planets and planetary system formation, architecture, and evolution around other stars, and undoubtedly face further surprises.
The issue of "dwarf planets", and "double planets" vs. "moons" (these are all planemos).
The question of sub-brown dwarfs vs. superplanets vs. free-floating planets (they are all planemos, and brown dwarfs are not).
(image from my Scientific American article on brown dwarfs, April 2000)
Other Reading Material:
What is a Planet? Universe in the Classroom link (ASP)
This is one that teachers should read and maybe use.
What is a Planet? Slides for a talk (PowerPoint 8Mb)
An Orb by Any Other Name (2003 Berkeleyan article by Robert Sanders)
A campus article on the Pluto controversy.
Commentary in the San Francisco Chronicle, Friday Aug. 25, 2006
A local article on the IAU decision.