Creating Ancient Languages for TV.

Gail Hairston, a PR person for the University for Kentucky, writes about a linguist with a great job:

Throughout Andrew Byrd’s successful career in academia, he has pushed to understand ancient languages to a depth no one has before. His goal was to understand how languages spoken thousands of years ago actually sounded. […]

He and his wife Brenna Byrd, also a linguist and assistant professor of German at UK, helped the video game creators Ubisoft bring to life “Far Cry Primal”, which involved warring tribes in 10,000 BCE. Ubisoft wanted realistic conversations among the prehistoric Homo sapiens, so the Byrds partnered their skills. While he created the words and their sounds, she relied on her award-winning skill at teaching foreign languages to teach the game’s actors to speak the languages realistically.

After the video game was manufactured, it wasn’t long before another opportunity was offered to Byrd.

This one came from television’s National Geographic Channel. Producers wanted the UK linguist to help create verbal languages for a new series to be called “Origins: The Journey of Humankind.” Futurist Jason Silva hosts this visually arresting new series that offers a twist on conventional historical documentaries as it explores the big question of how humans “got from there to here,” in the evolution from apes to astronauts. […]

Byrd remembered an early conversation with the producers, “First, they said they were fans of my work in ‘Far Cry Primal,’ then they asked me if I could create languages in different time frames, even different parts of the world.”

“What about 2,000 years ago, Europe?”

“Yes, sure,” Byrd answered.

“4,000 years ago?”

“Yes.”

“How about 14,000 years ago in Eurasia and northern Africa?”

“There’s a lot we don’t know about those regions, but yes, I can make an educated guess that will be reasonably close,” Byrd said.

“Southern France, 25,000 years ago? Australia, 50,000 years ago?”

“Those languages would be based on very little actual data, but I can make a good attempt,” he said.

The producers were more than happy with that answer, and now Byrd’s work can be found in virtually every episode of the new series.

Needless to say, I’m more than dubious about these sort-of-reconstructed languages, but hey, it’s only TV, and at least an actual linguist is involved. (I posted about Byrd’s website The *Bʰlog in 2014.) Thanks, Trevor!

Comments

  1. “How about 14,000 years ago in Eurasia and northern Africa?” “There’s a lot we don’t know about those regions, but yes, I can make an educated guess that will be reasonably close,” Byrd said.

    He may be a linguist, but in my view he has a strange definition of reasonable, unless he means that it sounds like any random human language instead of sounds that a different animal would make.

    It is indeed a great job!

  2. I think he meant “Hey, this sounds like a great job, and I’ll say whatever will get you to hire me!”

  3. At least nobody can ever prove you’re wrong.

  4. Trond Engen says:

    Í thought:

    2000 years ago in Europe? We have written sources for that. And much of what we don’t have sources for, we can reconstruct from living languages.

    4000 years ago? We can reconstruct a lot. Pick a branch of IE and try for Pre-Proto.

    14000 years ago in North Africa/Eurasia? Could well be Proto-AA. It’s not reconstructed with any accuracy, the urheimat could be anywhere in that general area, and this is in the upper end of reasonable age estimates, but it would still be a reasonable guess.

    Older than that? Nothing to base it on, so any conlang will do, but one might assume some general regional features in grammar or phonology.

  5. January First-of-May says:

    See also: Justin B. Rye’s Neanderthalese (which is explicitly labeled as a conlang, but, as the author noted, he’s trying his best not to say anything there’s solid evidence against; sadly, it’s probably too complicated, being based on completely different phonological structures (among other things), to ever appear in a TV show).

    EDIT:
    2000 years ago in Europe? We have written sources for that. And much of what we don’t have sources for, we can reconstruct from living languages.
    4000 years ago? We can reconstruct a lot. Pick a branch of IE and try for Pre-Proto.
    14000 years ago in North Africa/Eurasia? Could well be Proto-AA. It’s not reconstructed with any accuracy, the urheimat could be anywhere in that general area, and this is in the upper end of reasonable age estimates, but it would still be a reasonable guess.

    Older than that? Nothing to base it on, so any conlang will do, but one might assume some general regional features in grammar or phonology.

    Basically what I was thinking of too (though IIRC Proto-AA is closer to 8000 years ago than 14000).

  6. 4,000 years is enough time for the linguistic diversity of an area to be wiped clean and started anew. The Mediterranean of 4,000 BP had a number of barely recorded phyla, now all gone, and who knows what else.
    So even if we think that at 10,000 BP there was Proto-AA somewhere (horn of Africa?) the situation at 14,000 BP (pre-glaciation!) around the Mediterranean is nothing I can imagine. The languages we think of as ancient and autochthonous—Sumerian, Hattic, Etruscan, Proto-Kartvelian—were many millenia away in the future, in locations unknown.

  7. Well, you’re not getting a job at the National Geographic Channel.

  8. Trond Engen says:

    Sure. And even several of the IE languages we know of from 2000 years ago are barely known beyond their mention in contemporary sources. But exact reconstruction of all historical and geographical detail of the continent was not part of the job description. Neither was the language at a certain location at certain time. We (1pp exclusive) do know enough to come up with at least one European language from 2000 years ago. We also know enough to be able to approximate one from 4000 years ago. We have hardly any idea of what was spoken where 14 000 years ago, but within a general area as wide as North-Africa/Eurasia, one might use the dim light from the distant but conveniently placed lamppost of AA.

  9. The main problem the further you go back in time is that most humans would have lived in small groups of a few dozen individuals, meaning that language fragmentation would have been rampant. Several bands that split recently and live close enough together to have regular contacts may share the same language, but the centrifugal tendencies would overwhelm any language cohesion beyond, say, a few thousand people (not sure if that is a good estimate, but I’ll stick to it—in any case nowhere near as large as the typical language communities we’re familiar with today).

    It wouldn’t have been until after developments such as plant cultivation and animal husbandry led to many more humans living much closer together that larger language communities would have arisen.

    Around 14,000 years ago in North Africa and Eurasia, there would have been a rather diverse constellation of languages spoken, at least as diverse as what we find in New Guinea or in North America before European contact. A possible exception might be in the area of the Early Natufian culture, which seems to have developed at least a semi-sedentary lifestyle predating the advent of agriculture. It is intriguing that Proto-Afroasiatic is estimated to have been spoken at roughly that time period, meaning that it likely had a significant language community already (i.e. not just a few thousand individuals) to account for its later spread. But this would have been an exception, and its contemporaries would probably have had much fewer speakers.

    A typical language you might have heard in Southern France 25,000 years ago is likely not to have left any trace of its existence, let alone any descendant that is still spoken somewhere.

  10. Trond Engen says:

    We (1pp exclusive)

    I meant exclusive myself, not anyone else here.

  11. We (1pp exclusive) do know enough to come up with at least one European language from 2000 years ago.

    Indeed, no reconstruction is needed: it suffices to use Latin or Greek, both widely spoken in the Europe of 17 CE.

    The trouble with AA reconstruction is that there are two dictionaries of AA, both recent, agreeing in almost nothing.

  12. SFReader says:

    Circa 14000 years ago, people were concentrated in a few small Ice Age refugia. Wouldn’t be surprised if all of Europe had less than a dozen languages.

  13. David Marjanović says:

    Around 14,000 years ago in North Africa and Eurasia, there would have been a rather diverse constellation of languages spoken, at least as diverse as what we find in New Guinea or in North America before European contact.

    Europe, though, was apparently cleansed of humans (both Neandertal and Neander) by the Last Glacial Maximum some 25,000 through 14,000 years ago, and then repopulated from the east by people related to the Ancient North Eurasians, so its mesolithic languages probably weren’t that diverse after all. Funnily enough, they might even all have been Nostratic in the wide sense (which includes Afro-Asiatic).

    The trouble with AA reconstruction is that there are two dictionaries of AA, both recent, agreeing in almost nothing.

    Yes, and in one of them the way Arabic dictionaries are organized shines through, to mention the one issue I remember. It doesn’t help that lots of Chadic and Cushitic languages remain mostly unknown to science and that therefore the protolanguages of these branches haven’t been reconstructed yet… assuming Cushitic is even a single branch.

  14. Would it be OK to just borrow some languages from, say, New Guinea — if the producers made a sufficiently large donation to the originators’ cultural and material wellbeing?

  15. January First-of-May says:

    Would it be OK to just borrow some languages from, say, New Guinea — if the producers made a sufficiently large donation to the originators’ cultural and material wellbeing?

    For the 50,000 BP Australian option, it’s probably the best fit one can get, anyway (at least, assuming any humans lived in 50,000 BP Australia at all… which they apparently did).

  16. There are good reasons not to just borrow some languages from New Guinea. It puts living speakers of those languages in the same basket as long-dead peoples now known to us only through fragments of bone and reconstructions in books and museums—that is, specimens to be viewed for our edification, rather than real, living human beings on the same planet with us at this present moment. It implicitly places those speakers outside the viewership of the program, which—this being a documentary for the BBC—would traditionally have been “all of humanity, in theory.” Maybe a sufficiently large donation could outweigh these for the speakers in question, but I’m not sure you could manage the logistics well enough to ensure that no-one ended up denouncing the BBC as arrogant colonialists on Twitter.

    All this aside, there are practical reasons not to do it. Lots of people are going to be at least curious about a “reconstructed” language supposed to be 14,000 years old. Much fewer are going to be interested in living languages from New Guinea (alas), and there’s also going to be a significant segment of the viewership who gets confused and ends up believing that the languages of New Guinea are 14,000-year-old fossils frozen in time, or some nonsense like that. So from the outreach + education + ratings point of view, it wouldn’t be a good idea.

    And given that the language crew is just 2 people, I doubt it would even be any cheaper!

    Instead, the BBC should make this documentary with conlangs, and then invest big in a 25-part documentary on the linguistic situation in insular SE Asia.

  17. January First-of-May says:

    Instead, the BBC should make this documentary with conlangs, and then invest big in a 25-part documentary on the linguistic situation in insular SE Asia.

    Well, conlangs where they are needed (i.e. when the relevant population won’t be actually speaking Latin, Greek, Sumerian, or some other reasonably well attested language).

    I agree about the rest of your comment, however.

    EDIT: though sometimes you do need a good linguist. The guys who invented coins would’ve been speaking a lightly-attested branch of Anatolian…

  18. @SFReader, thanks for reminding me about the Last Glacial Maximum. By 14,000 years ago, however, deglaciation had progressed enough for the Magdalenian culture to be quite widespread through Europe, including most of present-day France and Germany. This would have been the result of a recent expansion from the Ice Age refugia, meaning that the picture was possibly of just a few shallow language families with wide geographic distribution. In time, the fragmentation would have been rapid, though, since Europeans of this era were hunter-gatherers without dense settlements.

    When I wrote the comment, I had the rough idea that the Sahara would have been green at this time, but I was wrong—the Neolithic Subpluvial in Northern Africa only started 9500–9000 years ago and lasted until 5500–5000 years ago. 14,000 years ago, the Sahara would have been as dry as it is today.

  19. Trond Engen says:

    Hunter-gatherers in open landscapes often move over large areas and maintain distant contacts. The dynamics would have been different before the introduction of trade with differently specialized societies, but I’ll postulate* that it’s sedentarianism and local self-sufficiency (in vital goods and marriage partners) that allow rapid development of mutually unintelligible dialects.

    *) Not original to me in any way, but I didn’t bother to find support.

  20. “Europe, though, was apparently cleansed of humans (both Neandertal and Neander) by the Last Glacial Maximum some 25,000 through 14,000 years ago, and then repopulated from the east by people related to the Ancient North Eurasians, so its mesolithic languages probably weren’t that diverse after all.”

    I probably wouldn’t use the term ‘related to’ in this case, as their genetic split could have been around 20,000 years before the last glacial maximum. And they probably didn’t spend the last glacial maximum in the east, but rather in southern Europe, or close to Anatolia, as the earliest tested Anatolian Farmers already had very old admixture from this population.

    As for the region around North Africa and southern Eurasia before agriculture boosted certain languages, they could have been speaking anything. Every time a new genetic study investigates a genome that old, they find a new population that was very highly divergent.

    There will be some new ancient DNA studies from India coming soon, which should be interesting.

  21. David Marjanović says:

    14,000 years ago, the Sahara would have been as dry as it is today.

    Just slightly shifted to the south.

    their genetic split could have been around 20,000 years before the last glacial maximum.

    True; I remember an old date of 51,000 years ago…

    And they probably didn’t spend the last glacial maximum in the east, but rather in southern Europe, or close to Anatolia, as the earliest tested Anatolian Farmers already had very old admixture from this population.

    Apparently not in southern Europe, which is why we don’t have more Neandertal admixture.

    There will be some new ancient DNA studies from India coming soon, which should be interesting.

    Awesome!

  22. @Trond Engen, thanks for that point about hunter-gatherers maintaining distant contacts over large areas, and that sedentary and locally self-sufficient societies may be more prone to dialect and language divergence.

    I wish we knew more about the language dynamics of such societies as they existed before contact with sedentary populations. We know that even before the introduction of agriculture some areas were productive enough to support at least a semi-sedentary lifestyle and populations that were greater than your average hunter-gatherer band.

  23. SFReader says:

    Some sedentary hunter-gatherer societies probably had some form of smallscale supplementary agriculture without supplying main nutrition needs.

    Beringian population – ancestors of Amerindians – apparently was engaged in cultivation of bottle gourd plant used for making containers.

  24. Even if hunter-gatherers maintained distant contacts over large areas, that does not require that they spoke the same primary language. As you can see in a recent post here, many Native American tribes that spanned a huge area, with quite seperate languages, shared a sign language lingua franca/trade language, probably for thousands of years.

    Humans are quite good at developing secondary bridge languages whenever necessary, and pretty good at not mixing it up with their primary language, unless new items and ideas come from the contact with outsiders, which requires new words.

  25. I know Andy Byrd personally, so I can testify that he’s perfectly at home in reconstructed Proto-Indo-European. As for pop applications of historical lingusitics, he has produced (and recorded) a very plausible modern version of Schleicher’s fable (The Sheep and the Horses). I’m also pretty sure he would agree that 14 kya is beyond the applicability of comparative reconstruction as we know it. Knowing something about actual protolanguages can be more of a hindrance than help if your job is to design a realistic conlang. Anthony Burgess openly admitted that the Ulam language he created for Quest for Fire was biased towards IE, with familiar-looking roots and inflections: “If I see the fire it is atrom, if the fire sees me it is atra, if I am surrounded by many fires they are atrois. Very Indo-European — or shall we say very unChinese, very unMalay.” Words and morphemes recognisable today used 80 kya in Palaeolithic Europe by not-quite-modern humans? Did they say udra for ‘otter’ and mamota for ‘mammoth’?

  26. Piotr! Great to see you around these parts again; don’t be a stranger!

  27. Matt: There are good reasons not to just borrow some languages from New Guinea. It puts living speakers of those languages in the same basket as long-dead peoples now known to us only through fragments of bone and reconstructions in books and museums—that is, specimens to be viewed for our edification, rather than real, living human beings on the same planet with us at this present moment.

    I agree. Since I mentioned Quest for Fire — Inuit and Cree were used (or shall I say exploited) there in this way.

  28. Hat, of course I always lurk around. I’ve been a bit busy recently, hence my low profile, but I’m no stranger.

  29. Marja Erwin says:

    ““Far Cry Primal”, which involved warring tribes in 10,000 BCE.”

    In addition to reconstructing languages, there’s also the problem of reconstructing conflicts. Modern war relies on newer social structures: states/armies, militias, hired mercenaries, and bombing cities, stealing cattle, burning crops, taking prisoners as slaves, in some cases freeing slaves, and so on.

  30. Trond Engen says:

    Good point.

Speak Your Mind

*