How to Name Things: A Case Against Conlangs

I've been thinking about constructed languages a bunch since reading Manvir Singh's excellent New Yorker piece about Dune and conlangs; I rewatched Dune: Part Two over the weekend, too. I had thoughts. And I'm always thinking about worldbuilding and names, always scribbling notes for my current Stars Without Number campaign, my Lore24 project, my myriad half-baked fantasy settings, etc. Might as well try to make all this thinking useful to somebody else!

* * *

I see four major approaches to naming fictional people, places, and things:

1. Use real names. This is not only (usually) the easiest approach, it's a necessary one for certain genres. If you're writing realistic fiction set in the present day, you need contemporary names. Same with any genre that uses the real world as a foundation; most characters in an urban fantasy story, for instance, will probably have contemporary real-world names. Near-future science fiction will use all, or almost all, real names for its human characters; you just need to pay attention to demographics a bit (more than a third of Earth's population is either Chinese or Indian—don't skimp on Chinese and Indian surnames without good reason).
   
   On that note, using real names does involve research sometimes. If you're writing historical fiction, you need to make sure your names are appropriate to the period and culture you're representing. But there's never a time real names should be off-limits. Far into the future, it's reasonable to have contemporary names, particularly heritable surnames, keep appearing. Even in secondary-world fiction, real names sometimes work well (look at Disco Elysium).
   
   
2. Derive new names from existing names or from other words. Change some letters. Chop a few names apart and scramble the syllables. Grab common nouns from foreign languages. Make names out of existing words that aren't proper nouns, or move proper nouns across categories (use a company's name for a person's last name, or a person's first name for a country).
   
   Some people look down on this approach, but it has strong precedents. George R. R. Martin does this a lot in A Song of Ice and Fire: Eddard, Tyrion, Myrcella, Roose. George Lucas did it in Star Wars, too: Greedo, Wedge Antilles, Skywalker, Sidious. Your choices can be informed by a knowledge of linguistics and history, à la Martin, or they can be totally arbitrary, à la Lucas. Sure, Lucas came up with some duds, but he hit on some real winners too. Darth Vader! Han Solo!
   
   
3. Give names in plain English (or whatever language you're writing in) and explain that they're literal translations from a source language that you either never describe or perhaps relegate, in the form of a few instructive examples, to a footnote or an appendix.
   
   
4. Create a whole de novo conlang and derive names from that. Better yet, create multiple conlangs and then imagine an entire fictional history involving various fictional cultures. Share loanwords among your languages. Chart the progression of your fictional languages through vowel shifts and state-mandated standardizations. Simulate everything!
   
   Many people are drawn to this approach because it's an enjoyable challenge in and of itself. (That was Tolkien's motivation, famously: “The 'stories' were made rather to provide a world for the languages than the reverse.”) But many others go down this path because it seems like the most “realistic” or “authentic” way to generate names. Some think of it as the sine qua non of good worldbuilding.
   
   (We might say that there's a fifth approach, which is “make it up de novo, but in an unsystematic way,” but I think that you can pretty much always sort those into #2, if they really are unsystematic, or #4 otherwise: Either they're trying, if only unconsciously, to imitate some existing word or word structure, or they're making such an effort not to resemble anything that already exists that they're essentially a conlang, if perhaps a rather shoddy one.)

But enough taxonomy. What I want to do here is stick up for the non-conlang approaches. They have a lot to recommend them, starting with the fact that whereas your audience has to learn every single syllable of any conlang you show them, they already have a world of associations with existing words—even single syllables resonate with them. Maul? Sidious? Those are bad guys. Solo? He's a cool lone wolf. Skywalker? Probably a great pilot. Voldemort? You only need to know the last syllable: His name means death. These examples are all at the unsubtle end of the spectrum, but you know what? They work.

Examples from ASOIAF or Disco Elysium are a bit less literal, but still take advantage of associations and resonances with the familiar. Eddard Stark? Basically Edward Strong. Regal. Old and very English, which in fantasy usually means “us,” the familiar. Evrart Claire? First name is a slight twist on Evrard. Second means clear or understandable, which is almost certainly going to be ironic. Both are extremely French, which, at this point in the game, we've established means “here,” the familiar. One of us. And your audience doesn't actually need to know any of the trivia or etymology to just know, to have a gut feeling about the name. It resonates.

The more familiar a name is, the more your audience will identify with it. If you have people and places in your story who go by familiar names, or whose names are given as translations into common English, the audience will recognize them as “us.” Those with unfamiliar names, especially ones that feature unusual orthography and are difficult to pronounce, the audience will “other,” identifying them as…well, “them.” You should bear this in mind just as a gut check when naming stuff (anything with glottal-stop apostrophes in it is going to seem totally foreign to most English-speaking audiences, something I wish more fantasy writers understood), but you can also put it to work for you more purposefully. You can give your audience a lot of information about a person or place with just a couple syllables, and with a whole network of associations from multiple names, you can quickly sketch out an elaborate cultural geography.

Tolkien gives us a textbook example of how to do this. The Mannish languages of Middle-Earth (i.e., those spoken in the Shire, in the Dale, in Rohan, etc.) have, as per #3 above, a “real” form (which he did not fully elaborate the way he did with Quenya and Sindarin, although he did give some examples), but they are “translated” into real-world languages. Westron, which is the language of our heroes the hobbits, is represented as modern English. Places like Rivendell, Brandywine, and Bag End, as well as people like Peregrin (or Pippin) and Merry, have “real” Westron names, but they are written as English in the stories. We know what “riven” means, we know what a “dell” is, we know “brandy,” “wine,” “peregrine,” and so forth. The Shire is familiar. It's home (and Rivendell is a safe, comfortable home away from home). The hobbits are “us.”

The farther we go from the Shire, the more foreign things become. Other Mannish languages are rendered as other Germanic languages: Rohanese (aka Rohirric) is Anglo-Saxon (i.e., Old English) and the language of the Dale is Old Norse. These are languages closely related to modern English, but not intelligible to English speakers. This tells us that these people are foreign to the hobbits, but with a fairly close cultural connection. The quasi-Welsh language of the grey elves suggests people more foreign than the Rohirrim, but still culturally connected to the Shire, whereas the quasi-Hebrew language of the dwarves and the quasi-Finnish language of the high elves are almost completely alien (English is distantly related to Welsh, but totally unrelated to Hebrew or Finnish—although of course Hebrew does suggest a spiritual connection, if not a direct cultural one; the dwarves may not be “us,” but they are like us). Tolkien uses our familiarity with real-world languages to establish a legible cultural geography. We don't need to be told that the high elves are more exotic to the hobbits than the grey elves are, and the grey elves in turn more exotic than the Rohirrim. We can just see it in the names.

Other fantasy and science fiction greats have used the same trick and similar ones. In Ursula Le Guin's Earthsea books, people's “use names,” though ostensibly belonging to the fictional Hardic language, are almost always given as plain English words, often those of animals or trees: Sparrowhawk, Otter, Dragonfly, Alder, Beech, Gamble. (So are some place names, like Ten Alders and Easthill.) True names, magic spells, and other words in the Old Speech (the language of dragons and wizards) come from a conlang. This others them, sharply separating them from use names and giving them an air of mystery. If Le Guin had come up with an entire Hardic conlang and used it for all the use names and common place names, the distinction would have been lost. To the reader, all the words would seem equally foreign.

Le Guin also used language associations to subvert readers' racial expectations. Her Archipelagan protagonists are dark-haired, dark-eyed, and dark-skinned, quite unlike the typical protagonists of young-adult literature in the 1960s and 70s. The Kargs, on the other hand, are pale-skinned, blue-eyed, and often fair-haired—much more akin to the assumed default in midcentury America. But the Archipelagans' Hardic speech is represented as English, whereas Kargish is represented by fictional words with a distinctly non-English phonology and orthography. Language helps a white reader see the Kargs as foreign barbarians and the Archipelagans as “us,” even if that reader is unaccustomed to identifying with nonwhite people.

Gene Wolfe, in The Book of the New Sun, also uses the translation conceit, explaining that his tale of the far future involves many words that don't exist yet in our time. To “translate” them, he describes or names many people and things with real but obscure or archaic words: Ascians, hipparchs, nenuphars, Pelerines. As with Le Guin's Old Speech, this lends them an air of mystery, albeit one that in this case communicates not an exotic origin but an ancient, forgotten one, appropriate for a setting in which there's literally nothing new under the sun, even if nobody can remember much of what is past.

Conversely, if you create everything from scratch, you lose all of the cultural and linguistic associations that real (or real-ish) words offer. This definitely has its merits in certain cases. When you want to create extraterrestrials or interdimensional demons that are genuinely alien, beings for which your audience has no frame of reference whatsoever, it's the only way to go—if you want those ETs or demons to be named in their own language.

But you don't actually need to do that. It might not even be plausible! If your totally alien ETs aren't humanoids, what are the odds their words can even be pronounced by human tongues? You can call them by exonyms derived from human language, as with many classic SF aliens: arachnids, xenomorphs, prawns, the Thing, Jean Jacket, good old Martians. Before you drive yourself crazy trying to whip up a whole world's worth of realistic conlangs, consider that even after Tolkien did just that, he still walked things back for his stories and relied on English and other real-world languages for their associations and resonances.