Simon J Greenhill Profile picture
I study how languages and cultures evolve. Primarily with phylogenies and other assorted computational methods. Based at @Biology_UoA. Also @simon@mastodon.nz

Apr 19, 2023, 12 tweets

Finally! #Grambank is out!

Grambank is a new database that documents patterns of grammatical variation in over 2400 of the world’s languages.

All available here: grambank.clld.org

Press Release: mpg.de/20186271/0418-…

Our first paper is out now in @ScienceAdvances where we investigate what shapes the structure of languages.

First, we quantify how similar each language is - we find massive differences globally. Only a few of these 2400 languages are identical in the ways they structure info.

We cluster languages by similarity which shows that most clustering is driven by related groups of language families:

Even when zooming into these clusters, we find interpretable patterns e.g. The Indo-European languages split into the contact languages (e.g. Haitian, Tok Pisin etc) and the non-contact languages (e.g French, Swedish etc).

To formally evaluate whether these patterns of similarity are driven by genealogical inheritance or language contact, we develop a new method to quantify the relative signal attributable to space and phylogeny: Almost always, phylogeny is more important than space.

... this means that there are some very stable grammatical traits over tens of thousands of years of language change.

Next we investigated which languages were the most *unusual*. We find that these are most often language isolates. Isolates represent 4% of Grambank’s languages in total, but they make up 19% of the most unusual languages.

...and this also patterns geographically with particular cultural and historical regions revealing consistent values of unusualness from low (Southeast Asia), mid (southern Africa), to high (Northern Africa and Europe):

Finally, we also ask what do we risk losing as we lose languages. We applied an ecological metric to quantify the loss of language diversity.

See the dark green bars in the below figure? that's what's left once we remove all the endangered languages in light green.

Frighteningly, some places like South America and Australia are expected to lose *all* of their indigenous linguistic diversity. Even other relatively 'safe' regions (Pacific, SE Asia or Europe) still show a dramatic loss of ~25%

This amazing project brought to you by >100 authors across many institutions who've all dedicated a lot of time and effort.

We're excited to make this resource available to the community and are looking forward to see what you come up with. Have fun!

Share this Scrolly Tale with your friends.

A Scrolly Tale is a new way to read Twitter threads with a more visually immersive experience.
Discover more beautiful Scrolly Tales like this.

Keep scrolling