How Fast Do Languages Evolve? - Dyirbal glottochronology 1 of 2

Languages change over time, but do some evolve faster than others?
Let's say you fall into a coma hearing your language spoken all around you.
You snap awake in the same place but at a different time.
They're still speaking your language but you can't understand a word.
How long were you out?
We'll try to answer that question in the next two videos, starting this time with the speed
of language change.
Gerbil.
Ha-ha.
No, Dyirbal, the name of a language in Queensland, Australia, a very quirky language first brought
to our attention at the turn of the 1900s and full of so many unexpected features that
the grammar written about it earned the status of a linguistic epic.
Thanks to this book, even from my first linguistics class, I heard fantastic tales of Dyirbal's
fascinating features.
Features like... it has four genders: human males and animals, human females and fire
and water and harmful things, edible things, and other or inanimate objects.
It has incredibly free word order.
To the uninitiated, Dyirbal sentences might seem about as organized as word soup.
The people, the Dyirbalŋan, have a no vagueness policy.
They couldn't say vaguely: she made me laugh.
They'd have to be more specific: she told a joke and I laughed, she tripped and I laughed,
she tickled me and I laughed.
And in this language there's not even a general word for "language"!
If that's not enough to tickle you, consider my favorite of Dyirbal's eccentricities.
A woman had to speak very differently in earshot of her father-in-law, and
a man around his mother-in-law.
They called this Dyalŋuy.
It had the same grammar as normal speech, Guwal, but the content words were switched
out to avoid offense.
Dyalŋuy came in handy in other complex social situations, but it always involved speaking
with completely different vocabulary.
It sounds extreme, but it happened!
Dyirbal was so peculiar and so influential that linguists have even gone as far as to
say that, without Dixon's grammar, our understanding of the world's language diversity would be
much poorer.
Something odd happened when linguists returned to study the language just decades later.
Traditional Dyirbal was mostly a memory.
A new Dyirbal, Young Dyirbal, had taken over.
The class system had been reworked.
Word order was now fixed, with basic subject-verb-object sentences just like English.
The Mother-in-law language had been upended.
Some everyday words were being used in new ways and others were being forgotten for good.
Were we watching Dyirbal dismantle itself in front of us?
Can the core features of a language really evolve this fast?
What is the speed of language change, and how could we even measure it?
Well, ask Morris Swadesh.
He's the first person I know of who laid out a method for calculating
the rate of language change.
He developed a one-size-fits-all solution he swore would work for every language around
the world, from Norway all the way to the northern rainforest of Australia.
Take a list of words.
Basic words that kids learn at an early age.
The kind that, ideally, don't get replaced very much from generation to generation.
Compare it to the same list of words in a descendant language.
What you should find is that some words have been replaced, but many are still essentially
the same word.
Sure, some sounds will get tweaked, and it may take an expert's eye to recognize
their sameness.
So Dixon tells us that the Dyirbal word for dog is "guda".
But when he documented a neighboring language, Mbabaram, they told him that their word for
dog is... "dog".
Seriously.
Nothing to do with English though; it's a coincidence.
It turns out that this word is cognate with guda.
For Swadesh, that counts as a retention.
On the other hand, water is "bana" in all Dyirbal dialects except one where it's "gamu";
this counts as a change.
Once you know the proportion of retained words, next comes the fun part: glottochronology,
"timing tongues".
Swadesh thought that basic vocabulary works like radioactive decay, meaning that all languages
replace words at a constant rate over long periods.
And he felt vindicated when he checked his assumption against history, say, the time
it actually took for Latin to turn into the Romance languages.
The method looked solid.
After 1000 years a language keeps 86% of the 100 words on Swadesh's list.
It's an amazing thought.
Just looking at a list of words in a language, he could calculate how old it is and how far
it is from its relatives.
He could even construct a lexical chronometer!
That's the nerdiest thing I've said all day.
But it predicts how long it takes to hit major vocabulary milestones and end up with
a brand new language.
Remember our hypothetical coma patient?
Yeah, they'll be waiting a long time before Dyirbal changes into a new language.
Another 900 years, because languages change at a constant rate!
Except... they don't.
Come back next time to see why not and learn a bit more about Australia's languages as
we ask: how long can a language last before it becomes unrecognizable?
That's right, a two-parter!
Because this is too interesting and because my drawing hand needs a rest.
So stick around and subscribe for language.
What Latin Sounded Like - and how we know Euroversals - Are all European languages alike? Kanji Story - How Japan Overloaded Chinese Characters What Shakespeare's English Sounded Like - and how we know Basque - A Language of Mystery Does time work differently in different languages? - Hopi Time Hungarian explained - such long words, such an isolated language 10 Things Body Language Says About You Weird Phonemes - pronouncing the world's rarest sounds The Hardest Writing System! - an animated rant about learning Japanese

How Fast Do Languages Evolve? - Dyirbal glottochronology 1 of 2

Latest Posts