World’s languages traced back to single African mother tongue: scientists.
New Zealand researchers have traced every human language — from English to Mandarin — back to an ancestral language spoken in Africa 50,000 to 70,000 years ago.
Scientists say they have traced the world’s 6,000 modern languages — from English to Mandarin — back to a single “mother tongue,” an ancestral language spoken in Africa 50,000 to 70,000 years ago.
New research, published in the journal Science, suggests this single ancient language resulted in human civilization — a Diaspora — as well as advances in art and hunting tool technology, and laid the groundwork for all the world’s cultures.
The research, by Quentin Atkinson from the University of Auckland in New Zealand, also found that speech evolved far earlier than previously thought. And the findings implied, though did not prove, that modern language originated only once, an issue of controversy among linguists, according to the New York Times.
Before Atkinson came up with the evidence for a single African origin of language, some scientists had argued that language evolved independently in different parts of the world.
Atkinson found that the first populations migrating from Africa laid the groundwork for all the world’s cultures by taking their single language with them. “It was the catalyst that spurred the human expansion that we all are a product of,” Atkinson said, the Wall Street Journal reported.
Atkinson traced the number distinct sounds, or phonemes — consonants, vowels and tones — in 504 world languages, finding compelling evidence that they can be traced back to a long-forgotten dialect spoken by our Stone Age ancestors, according to the Daily Mail.
Atkinson also hypothesized that languages with the most sounds would be the oldest, while those spoken by smaller breakaway groups would utilize fewer sounds as variation and complexity diminished.
The study found that some of the click-using languages of Africa have more than 100 phonemes, or sounds, whereas Hawaiian, toward the far end of the human migration route out of Africa, has only 13, the Times reported. English has about 45 phonemes.
The phoneme pattern mirrors the pattern of human genetic diversity as humans spread across the globe from sub-Saharan Africa around 70,000 years ago.
This gives me LIFE from people who insist all languages (ALL no matter what) derive from latin bases.
Reblogging this for three reasons:
1) It’s awesome and worth knowing
2) It makes sense when you think about, you know, the whole history of human development (from a NOT white supremacist perspective at least)
3) To add that if anyone ever tries to say that all languages are derived from Latin [insert choked sound of disbelief and anger] you can inform their ignorant (probably racist) asses of this: Latin, as far as languages go, is an INFANT. It’s part of a subset of Indo-European languages and MOTHERFUCKER EVEN ENGLISH ISN’T ONE OF ITS DERIVATIVES. (French, Spanish, Italian, Romanian, and Portuguese are, as well as lots of their related languages and dialects, that’s it.) Latin isn’t even remotely old enough to be a mother language. It’s like saying alpacas were the original dinosaurs or some bullshit.
HUH! You dont FUCKING SAY?
WILL YOU LOOK THE FUCK AT THAT!
Not surprising, but this is cool information!
This is not only wrong, it’s also really bad science. And I know every field has problems with misreporting of new research discoveries, but I think linguistics gets a particularly rough time of it, in part because the public in general knows next to nothing about how linguistics works as a field, but also there seems to be a lot of anti-linguist bias in this kind of reporting (which might be the result of anti-Chomsky bias?). But you see it here, and in other similar articles, where “scientists” have some new approach to global-scale language studies, and in contrast, linguists (who apparently don’t count as scientists) are painted as these backwards, anti-technology luddites stuck in our ways and unwilling to accept new research like this. And that annoys me because there are very good reasons to be against new research like this *and* still be interested in technology and new approaches to linguistics.
A big part of the problem is that these “exciting new discoveries about language” come out of studies led by biologists, computer scientists, and mathematicians/statisticians without any input from linguists, who actually study language and are familiar with the field, its methods, and what works versus what doesn’t. So these studies don’t take into account really basic facts about how language differs from genetic material. You don’t just inherit a language from your parents (or whoever raised you) and that’s it. Pronunciation shifts over time in individual speakers, kids are more sensitive to the pronunciation of their peers than their parents, and it’s generally environmentally-sensitive in ways that make it differ from biology. But over the past couple hundred years of research, historical linguists have come up with ways to work with these facts in methodologically-sound ways.
But then “scientists” can’t be bothered to learn any of this before diving right in to work on language. And apparently they can’t be bothered to use good scientific principles either. The globalpost article says:
"Atkinson also hypothesized that languages with the most sounds would be the oldest, while those spoken by smaller breakaway groups would utilize fewer sounds as variation and complexity diminished."
But you can’t have a hypothesis as one of the basic assumptions of your study. A hypothesis is something you *test*. So, if you wanted to hypothesize that languages with the most sounds are the oldest, you would take data from language families that are pretty well worked-out, like Indo-European or Austronesian or Algonquian, and see if the languages that branched off first (the so-called ‘oldest’) really do have the most phonemes. And then if that works and gets you results close to the results that philologists and linguists have worked out using tried methods, *then* you apply your method to the entire world.
But this is all apparently based on an earlier study that found that “the number of phonemes in a language increases with the number of people who speak it.” And that’s kind of interesting, but it’s important to think about *why* that might be the case. I suspect it’s because in today’s world, if a language has a lot of speakers, it’s because it’s a trade language or a language of colonialization. And both of those mean that there will be many speakers for whom it’s a 2nd (or greater) language, and they’ll carry over a lot of pronunciation habits from their first language. For example, there are cool things going on with retroflexion in Indian Englishes, so you could say that retroflex consonants are phonemes of (varieties of) English. But this study seems to assume that it’s like, once you reach a certain number of speakers, a voice from the heavens comes down and says “Congratulations, you have unlocked aspiration! Enjoy your /tʰ/.” And then when smaller groups branch off and their speech evolves to the point that it can be considered a new language, we’re supposed to assume they suddenly lose a lot of phonemes because they don’t have the speakers to support them, which is utter nonsense. Languages lose and gain phonemes for a lot of reasons over time.
On top of these methodological concerns, it’s not even that straightforward how to count phonemes, especially in languages which seem to have a lot of them. Like, the “African click languages” that the articles mention as having the most phonemes, some linguists think that some of the clicks should be counted as made up of clusters of phonemes, and if you count them that way, these languages have about the same number of phonemes as the Caucasian languages (like Ubykh). I would also be interested to know how and if they counted suprasegmental phonemes like contrastive stress, pitch, or tones.
The sounds of a language are also *highly* subject to change based on the sounds of neighboring languages. Like, these Khoisan languages with the clicks, they aren’t even thought to be related to each other, which I think is pretty damaging for a theory of a single south African origin of language. And the clicks spread to neighboring Bantu languages as well as some other unrelated languages, which gives them a higher phoneme count without being closely related or the same “age”.
These problems with phonemes are why they’re never used as basis for showing which languages are related. They’re highly unstable. Like, if you can show regular sound correspondence in related words, that’s helpful, but you can still get thrown off if two languages have a lot of borrowings. I think shared morphology is used as the gold standard these days. I’m pretty disappointed that Don Ringe and Brian Joseph we’re quoted saying positive things about this study in the NYT article. But neither of them is a phonologist, so I guess that’s the problem. And I think they’re both a little too excited about statistical/mathematical models of language, and don’t remember to be critical too.
So, apparently this study is pretty old, and some other linguists officially debunked it a couple of years ago. They re-ran the experiment with the original data, and found that it points to a number of different origins for language. You can read their full article in all its glory.
Also, I’ll add that it was super classy of globalpost to use a picture of a random South Sudanese boy for their article about the origin of language being in a completely different part of Africa. (Two whole bars away, by the NYT map graphic.)
And I kind of wonder if publishing the original article (& follow-ups) was just Science going after click-bait. Surely someone in the article review process must’ve noticed how messed up it was.