A couple of days ago I read and linked to an interesting article over on io9, called How to write the next great alien language. The article compares the Elvish languages (Sindarin and Quenya) and the Klingon language, and opens as follows:
Constructing an entire alien language is the most challenging task in all of speculative fiction, and there are two examples that tower above the rest: J.R.R. Tolkien's Elvish and Marc Okrand's Klingon. We'll show you how to outdo even them.
When it comes to world-building, there's no finer way to capture an alien culture than to give it a language that seems utterly strange to human ears. It's obviously a challenging task, and one that requires a decent working knowledge of linguistics. So while we have to leave the nitty-gritty of language construction to a textbook, what we can do is examine the very different overarching approaches used in constructing the two most iconic alien languages - Elvish and Klingon - and then explain how you could create a language that combines the best of both.
When they talk about overarching approaches, they are effectively referring to the fact that Tolkien based his Elvish tongues in part on a pair of human languages (Finnish and Welsh) and extensively developed the history of changes these languages, while Marc Okrand used existing names and a couple of lines of dialogue as a basis for creating a Klingon language system with an extensive vocabulary that could cover all kinds of topics (Klingon Hamlet, anyone?) but didn't delve into history of Klingon at all.
Let me start by saying that I'm terribly impressed with both Tolkien and Okrand for their achievements. The io9 article didn't mention Paul Frommer or Na'vi, but as I've understood it from independent research, Frommer designed the Na'vi language of Avatar in a somewhat similar way to Klingon (i.e. by approach from a contemporary linguistic perspective, though without the snippets of previous language use as a basis).
I wondered about a couple of things in this article. First, there's a general idea that one should give an alien culture a language that sounds "utterly strange" to human ears. How do you define "utterly strange?" I suppose I'd say that an alien language must have two major features: it must be unintelligible by speakers of any existing human language, and it must have no etymological connection with any existing human language. If it did have such connections, it wouldn't be "alien" by definition. Beyond that it's a matter of personal taste.
Marc Okrand had to work with an existing description of Klingon as "guttural," which affected his choice of the language's sounds. He also tried to have patterns of sounds in the sound system that didn't occur in existing world languages - his example was having a sound for V without having one for F. In the area of grammar he chose to use object-verb-subject word order, which is the most unusual in world languages. Overall, I would summarize by concluding that he was trying to defeat the human language "universals" in order to make the language very alien.
Paul Frommer was freer in his choice of sounds because he got to start from scratch, and in articles about his creation of Na'vi he has said he deliberately chose language sounds for his sound system that came from rare Earth languages and thus would be hard for most humans to recognize. Na'vi is in some sense the opposite of Klingon, in that it has no V - instead it has P and an ejective, written "px". Verb conjugation is by infixes rather than prefixes or suffixes - another deliberate choice to make the language unusual. You can read more about it at Wikipedia or at Learn Na'vi.
One of the key features of both Klingon and Na'vi, though, was that they had to be pronounceable by humans so that they could be used in a movie - and as it turns out, Sindarin Elvish is also pronounceable by humans; its vocabulary has been developed for use in the Lord of the Rings films.
That brings me to the question of how one might "outdo" Tolkien and Okrand - one of the proposals of the io9 article. They imply that the next great alien language should involve both an extensive contemporary vocabulary and a sense of historical development. Seems logical, at least inasmuch as that would "fill in the blanks" of what the developers of these existing languages didn't cover.
But my commenter Megs had a different point of view: "I think one of the most fundamental things about creating an ALIEN language isn't so much to use things that are rare in human languages, but to not BASE them on human languages. Just use logic to create a system that communicates. But maybe that's just me."
Okay, so Megs would want to create a system of communication that's not based on human languages, just on logic. One could argue that the logic required has some human basis too, but I see her point. What about extremely different alien physiology? What about scent language, or visual language? Alien is alien, right?
In order to understand the parameters here, I think we have to ask one critical question:
Why am I creating this language?
The answer to that question will determine everything about the language you create. Tolkien set out to create a history of mythology, and thus spent a lot of time on the history of his languages; that was a natural consequence. Okrand and Frommer set out to create languages that could be used by human actors on film, which required extensive vocabulary and contemporary usage, but did not require that they trace the history of the languages - and heavens forbid they should create something that was unpronounceable!
So let's get back to basics. The purpose of most science fiction and fantasy languages is to function in a story. The nature of that story is what determines the features and qualities of the language.
Here are some examples of languages that I've created, what parts of them I've created, and why:
Gariniya was based on the idea of a canon-based language - a language that consisted primarily of oblique references to a set of stories that all speakers of the language knew. To my knowledge, this language concept originated with Star Trek: The Next Generation's episode Darmok. My own contribution was to the culture surrounding it, in particular to finding ways for the language to have evolved and ways that the language would be learned and passed on (which I had felt were missing in the Star Trek episode; see my entry Darmok and Me). I also needed culture surrounding it, dictating the conditions of its use. For the purposes of the story I did NOT need it to have either an extensive vocabulary or a fully realized grammar. I designed phonology pretty thoroughly, constructed a few words to use in the story, and that was sufficient for my purposes.
Aurrel & Khachee
Both Aurrel, from "Cold Words," and Khachee, from my forthcoming story "At Cross Purposes," are entirely original in concept and execution. Similarly to Gariniya, though, they have well-developed phonology but don't have a big vocabulary because I only needed to use a few names and concepts in the stories. As I've said before, the more of an alien language you use, the more you alienate the reader, so if you're working from an insider perspective, it's good to use as little of the alien language as possible. Both of these languages differ from Gariniya in that I did much more design of their grammar. I started with one single key feature that was going to cause the most trouble in the story - status dialects in Aurrel and discourse/turn-taking structure in Khachee - and then developed from there. Another reason that grammar knowledge was important was because I tried to alter my use of English so I could sound like a native-speaker-of-alien-language-speaking-English when in the alien point of view. Aurrel also had a language evolutionary angle, and a historical angle, because these were directly relevant to the plot.
The Varinn language (from Varin, as you've probably guessed) is the most extensive language I've created. I did intend it to be pronounceable - and relatively easily pronounceable - by humans, since the Varini (people of Varin) themselves are humans. It has phonology and syntax and morphology, and I've delved far into cultural issues surrounding its use by various social groups. It also has a rudimentary history as a language, including sound changes and changes in verb conjugations. This is because I have other stories that I've designed in which this history becomes important. I've also developed more vocabulary for this one because I wanted to have a Varinn translation of certain songs and oaths that appear in the story. However, I hardly ever use it in the stories themselves. All of the stories are told from an insider perspective, and thus need to be effectively "in translation" whenever possible so the language won't distract from the story.
So when it comes right down to it, what makes a really great invented language?
The io9 article, and some comments that have followed it, seem to suggest that it would be better to include both concurrent vocabulary and language history - an additive sort of solution based on their analysis of Elvish and Klingon. Even if you had both, though, the language wouldn't have all the richness, irregularity, cultural grounding, manners, etc. of a natural language.
I'm going to argue, though, that the next "great alien language" is not going to be much like a natural language, for one single reason: without a great story to rest upon and guide its form, a language would have no reason to exist. It might be quite comprehensive and possibly quite alien (if that's what its creators were after), but nobody would care about it. A language needs a world, and characters, and a story, to make it compelling and worthwhile. Once it's in the hands of dedicated fans and learners, anything is possible. After all, creole languages are natural languages, vibrant, with functional grammar, and these languages develop naturally in locations where adults have been using awkward pidgin. Once you've put the accepted usage into the hands of learners, especially child learners, language tends to take on a life of its own.
So if you're creating a language, be careful of losing yourself in the conlang process, and make sure to keep an eye out for the needs of the story you want to tell. I encourage linguistic research, exacting standards, and lots of hard work - if that's what you're into. But for those of you who don't have tons of time to devote to language development, keep in mind that an effective and functional alien tongue doesn't need to be extensive to work. It just needs to be systematic and serve the needs of the story you want to tell.