A recent xkcd cartoon looked back to the time when Pluto was demoted from being called a planet to being called a dwarf planet (where dwarf planets don't count as planets):
We posted extensively here on various aspects of the story. Today I'm going to return to the status of the expression dwarf planet.
A few highlights from our earlier discussions of this expression: Ben Zimmer's posting, here, on the International Astronomical Union's declaration of dwarf planet; Ben's posting, here, on backlash against this declaration (in which Ben introduced the concept of hyponymy into the discussion); and my posting, here, on the semantics of noun-noun compounds.
The background for this discussion is the semantics of "N-headed composites", which I'll refer to just as "composites" in this posting. These composites come in several varieties, some of them N+N, some of them Adj+N. For my purposes here, the difference is not crucial, and this is a good thing, because (as I noted in my earlier posting) dwarf can serve as the first element of a composite either as an N or as an Adj, and often it's not clear what the right assignment is (N's in English often pick up Adj uses). Fortunately, we don't have to decide whether dwarf in dwarf planet is an N or an Adj or sometimes one and sometimes the other.
This is a side issue because both kinds of composites are normally "subsective" or "hyponymic". Using a prime for the denotation of an expression, in such composites:
(X+N)' is a subset of N'; that is, X+N is a hyponym of N.
A red apple is an apple (Adj+N), an electrical engineer is an engineer (a different Adj+N type), and a California university is a university (N+N).
Composites are usually interpreted this way when they're created "on the spot"; subsective/hyponymic interpretation is the default.
But conventionalized composites often have a looser interpretation for their heads than typical freely created composites do: an Irish elk isn't actually an elk (as Stephen Jay Gould liked to point out), though it resembles an elk, and a California is not actually a lilac (it's in the genus Ceonothus, not Syringa), though it resembles lilacs. (The connection of these beings to Ireland and California, respectively, is also complex, but that's a different topic, which I looked at most recently here.) That is, for this group of composites, (X+N)' is not a subset of N', though it is a subset of r(N'), the set of things that resemble N' in some specific way, different for each X+N combination.
I'll call such compositions "resembloid composites", distinguishing them from ordinary subsective composites. (I'm not claiming here that these are the only semantic subtypes of composites, only that these two cover a lot of the territory.)
[Side note: many times, resembloid composites have subsective counterparts. Irish elk could mean 'elk that is Irish' (perhaps one in an Irish zoo), and California lilac could mean any number of things, including an instance of the genus Syringa (rather than Ceonothus) growing in California.]
What's going on here is that resembloid composites are especially useful for NAMING TYPES, using the existing resources of the language (N+N and Adj+N combinations) — rather than metaphorizing some existing expression, borrowing a label from another language, using derivational morphology (as in plutoid), inventing a label from whole cloth, or what have you. The beauty of resembloid composites is that, although they're not entirely transparent, they lead you in the right direction.
They are enormously popular. Ordinary English is packed with resembloid composites, including many names of living beings. A 9/11/07 ADS-L posting by Larry Horn (talking about dwarf planet) cited several different N+N types, like peanut butter, sea horse, and phone sex (there are plenty of Adj+N examples as well). But technical vocabulary is also full of resembloid composites.
Back in the dwarf planet heyday, Mark Foskey wrote to me (on 8/27/06) about resembloid composites in mathematics. Among other things, he noted that a non-associative ring is not, in fact, a ring, but just sort-of-a-ring (almost a ring, in the specialized metaphorized sense of ring in abstract algebra, except for that associativity thing, which proper rings have). So non-associative ring is a resembloid composite, of a subtype with approximative semantics (a non-associative ring is almost a ring).
Life is complex: there are also plenty of subsective composites in mathematics (and other technical fields). An Abelian group is indeed a group (where group is another specialized metaphorization of ordinary language); it's a commutative group, a full-fledged group with the additional property of commutativity. (It could have been called an Abel group, using a N+N composite built directly on the name of the mathematician Abel, but algebrists fixed on the related Adj+N composite instead.)
I have considerable sympathy with the astronomers and other scientists who have objected to the resembloid composite dwarf planet handed down by the IAU. But I have to concede that resembloid composites are all over the place, in both ordinary and technical language.
A final note, about things that look like resembloid composites but might represent a folk taxonomy with categories that are broader than those of technical usage and most current common non-technical usage.
The background: unremarkable N+N resembloid composites, in the NOAD2 entry under fish:
used in names of invertebrate animals living wholly in water, e.g. cuttlefish, shellfish, jellyfish.
But of course this also reflects a folk view in which all sorts of sea creatures (whales and dolphins, for example) count as fish. This folk taxonomy isn't entirely dead, as we can see from Elisabeth Rosenthal's NYT piece of 8/30/08, "Swarms of Stinging Tentacles Offer Hint of Oceans' Decline", p. 9 (I've boldfaced the crucial bit):
Then there is pollution, which reduces oxygen levels and visibility in coastal waters. While other fish die in or avoid waters with low oxygen levels, many jellyfish can thrive in them. And while most fish have to see to catch their food, jellyfish, which filter food passively from the water, can dine in total darkness …
If the passage had gone "while fish die in …", it would not have conveyed that jellyfish are fish (and Rosenthal is clear on this point earlier in the article). But whether it came from Rosenthal or an editor, the "other" is there.