Disney’s Atlantis: The Lost Empire is unusual in having a linguist hero, and less unusual in having rather silly linguistics. The linguist is mainly a translator, and in the first scene of the movie announces he’s found a mistranslation in a “Viking” text which is English spelled out in futhark. (This isn’t a little detail in the background for specialists to freeze-frame and roll their eyes at; Milo Thatch has written out the futhark on a blackboard, with an English letter under each character, helpfully spelling out an English phrase.)
However, Thatch’s grandfather went to Georgetown, so I’m going to be generous with the movie (Hoya Saxa!) and focus on what’s more interesting: the conlang Marc Okrand created for the Atlantean culture.
Okrand is better known for creating Klingon, which he deliberately designed to not resemble human language. In many respects, his Atlantean is the opposite – in sound, vocabulary, and grammar, it’s meant to be a very plausible human language.
When Thatch and his companions reach Atlantis, they find the Atlanteans can understand English – and French, and German, and Hebrew, and Chinese – because all later languages are descended from Atlantean. This is of course silly – could someone who would say
To ðam het se gumena baldor ealle ða yldestan ðegnas; hie ðæt ofstum miclum ræfndon, rondwiggende, comon to ðam rican þeodne feran, folces ræswan. þæt wæs þy feorðan dogore þæs ðe Iudith hyne, gleaw on geðonce, ides ælfscinu, ærest gesohte.
automatically understand someone who said
To that feast, the warrior called his chief retainers. They came with great haste; the troop of warriors traveled to that strong prince; leaders of that folk came forth on the fourth day since Judith, that shining lady, first sought him, and the wicked warrior welcomed that wise maid.
? And modern English is only eight hundred years older than Old English; Atlantis, we’re told, sank more than eight thousand years ago. (For a fuller comparison of Old and Modern English, here’s the complete Old English poem “Judith” and a modern translation.)Still, we’re being generous, and Atlantis does break away from the usual SF tropes where either everyone speaks English, or there’s some sort of magical spell or perfect machine that translates everything automatically. And it’s true that several of the languages we see the Atlanteans understanding – Latin, German, Russian, French, Italian, Greek, and Russian – do indeed come from a common source, called Proto-Indo-European. (It’s generally abbreviated to PIE.)
Which brings us to Okrand’s contribution. Okrand wanted Atlantean to be a plausible ancestor of all the world’s languages. It’s obviously not possible to deduce what such a language actually was, or someone would have done it already; there’s no known relationship between, for example, Chinese and the Indo-European languages, or between Arabic and the Iroquoian languages.
Okrand started by building on the existing reconstruction of PIE. Now, PIE was never written down, so we have no direct evidence of what its vocabulary was, or how it was pronounced. But people had noticed that many words were similar across multiple languages in Europe and India, suggesting they came from a common root.
Here are a few examples from Kathleen Hubbard at the University of California:
|bharami||phero||fero||baira||bear (i.e., “carry”, not “big furry animal”)||bher-|
You’ll notice how regular the sound changes are. A “p” at the beginning of the word becomes “f” in English and Gothic, and “p” in Latin, Greek, and Sanskrit. “Bh” at the beginning of a word becomes “bh” in Sanskrit, “ph” in Greek (that’s a “p” with a puff of air, not an “f” sound), “f” in Latin, and “b” in English and Gothic. Reconstruction isn’t about squinting at a pair of words and saying “yeah, I could see how those are kinda similar”; it’s about finding regular patterns of correspondences in numerous pairs of words. (If this sounds intriguing, here’s a website that explores more sets of words that descend from common PIE roots.) (Also, Hubbard explains in more detail why we believe that those particular sounds are most likely for PIE – the reconstruction draws on what types of sound changes are known to be common or uncommon.)
As Andras Rajki shows at the Atlantean Language Institute, Okrand drew heavily on PIE roots, along with a smattering of roots from both Indo-European languages (such as Irish, Russian, and Greek) and non-Indo-European languages (such as Cantonese, Hebrew, and Japanese.) I made a PIE chart – DEAL WITH IT – to show the proportions.
As you can see, the vast majority of roots come from Indo-European languages. What’s more, if we look at the non-European roots, we see Japanese, Indonesian (including one word borrowed from Arabic), Cantonese, and Hebrew. Nothing from North or South America, or from Australia, which is disappointing. (Rajki’s etymologies don’t cover the entire Atlantean vocabulary, so there might be other languages represented that we don’t see here – still, this is about a tenth of the total lexicon, so I’d be surprised if the overall proportions are strikingly different from what we’d see if we looked at everything.)
Still, while non-Indo-European languages got somewhat short shrift in the lexicon department, they’re better-represented when we get to sounds and syntax – as we’ll see in Part 2.