▲ 90 points
back
96 comments
Kinda funny that Bavarian is more distant from 'High German' than Yiddish :D (not exactly wrong though, Yiddish is easier to understand for any non-Bavarian than hardcore Bavarian).
There's at least a dozen local German dialects missing which are just as 'weird' as Bavarian though (PS: ok, all the main dialects spoken in the northern parts of Germany seem to be there, just in their own 'heritage node'), I guess Bavarian earned its own spot because it has the most speakers.
« earned its own spot »
Linguistically, this chart doesn't have much value. Not sure what criteria were used for inclusion of this or that item, perhaps just presence in some ISO standard. “German” per se doesn't mean much, so-called “high-German” („hochdeutsch“, meaning „oberdeutsch“ in opposition to „niederdeutsch“), the language that you and I speak, is an artificial synthesis and administrative compromise on written variations that is not organically rooted in dialectical evolution and historically has been imposed by schooling (which has led to the eradication of the Northern dialects, „platt“, which are derived from old Saxon and thus are related to English), so it is questionable what its meaning should be in this chart.
Yiddish is a colonial sociolect and somewhat marginal to the evolution, so again, not clear why it was included. IIRC, Koestler wrote in “The Thirteenth Tribe” (where he rejects the mainstream hypothesis of the origin and nature of Eastern Jewry) that Yiddish is based on the South-Eastern variants, not on the South-Western variants.
Looking at the Latin languages, similar inadequacies can be seen. Very superficial and misleading, all in all.
Maybe this chart is meant more to showcase CSS and web graphics than linguistics.
Yiddish is very similar to Palatinate / Hesse dialect simply because many Jews used to live in that region. I grew up in Palatinate - the only strange things about Yiddish for me are the spellings and the occasional words of Hebrew or other non-German origin. Some Hebrew words are actually part of the Palatinate dialect through the Jewish connection, such as Zores (fight / argument / discord / trouble) or machulle (bankrupt).
This one's also interesting - a non-local (geographically disconnected) Bavarian dialect: https://en.wikipedia.org/wiki/Gottscheerish
Interesting visualization for sure.
At the risk of starting a religious war, interesting to see the "Lord's Prayer" translated in languages that existed before the Lord Himself did :-/
A kinda incongruity dissonance, if you will, reading it in Sanskrit (among other languages).
Useful, nonetheless, to see how different languages evolve(d).
Not bad for an AI-driven project.
I can image the moment Sir William Jones realized Sanskrit and Latin had a common origin. "Wait a minute...these can't all be coincidences." It's pretty mid-blowing.
My daughters are growing up bilingual in Persian and English. Totally fascinating to find the cognates; many are surprisingly close still (daughter/dokhtar, star/setare, cow/gav). It’s fun feeling like I could communicate basic ideas to a proto Indo European speaker based on those really distant cognates.
English: God gave us teeth; God will give us bread.
Lithuanian: Dievas davė dantis; Dievas duos duonos.
Sanskrit: Devas adadat datas; Devas dat dhanas.
I'm learning Spanish now after studying French and while I'm not fluent at all in Hindi or Marathi which I sometimes heard my family speaking growing up, the cognates are fascinating. The numbers are the most obvious. But even words like 'tu' (you) are the same in all the languages. Same with English and me as the first person.
Not to mention the verbs. Truly wild stuff to think no one noticed.
Cool. Looks good. But on Android Chrome the 'about' box covers the entire map, and the timeline slider is in the way too.
It is very common for vibecoded UI to have significant usability problems on mobile platforms and I find it annoying.
I'm pretty sure that it can be fixed by either manual testing and giving the agent feedback or by giving the engine tools to do the testing. Vibecoders just tend to be very careless about such things.
Same in Fennek. Did they not check their presentation in mobile?
Great project. One thing that is masked by the tree view though is that there were indeed countless other languages being spoken besides Proto-Indo-European. It's just the one that gave rise to today's Indo-European languages. But back then it wasn't alone, but just one of many, some related to it and some more distant from it. Kinda similar to how it works today. It had dialects, it was in motion and change, it had earlier versions etc. etc.
Well, yeah, but that's like saying that producing a tree of pomaceous fruit hides the fact that citrus fruits exist.
(Yes, I did make an apples and oranges joke.)
Yeah, it's not a clean inheritance but lots of linguistic contact and negotiation.
And crucially it's not just other IE-adjacent languages but the Pre-IE substrate. So many Greek words are inherited from the mysterious languages that were there before, of which we only know through the marks they left on ours.
https://en.wiktionary.org/wiki/Category:English_terms_derive...
PIE was never a single language. There was significant linguistic diversity. PIE is just what we call their roots in aggregate. It is a conceptual reconstruction.
Yeah... But it's not fully true to use a tree shape. Languages have "horizontal gene transfer" as well (some call it "extensive contact" or whatever), and the branches come in and out more easily than what the prescriptive method shows.
Same for actual speciesation, on every level. On homo species, on any other species, even all the way down to the origin of eurcaryota, it's all a network. The tree was never the correct model, just a "neat" model.
Sprachbund is a key term.
The "lord's prayer" in Vedic sounds very out of place due to the choices of inflected forms and some stems even reminiscent of classical Sanskrit.
Without going into depth of that, here's something actually more inspiring (actual excerpt from Rig Veda).
tvā́m ugrám ávase carṣaṇīsáhaṃ rā́jan devéṣu hūmahe
víśvā sú no vithurā́ pibdanā́ vaso amítrān suṣáhān kr̥dhi
> All of us that is weak, strengthen it. Make the enemies easy to subdue
(The acute accent is an udatta - high pitch, vedic is a pitch accented language. I would expect IAST with accents, not whatever randomly capitalised things it threw at me).
Interesting, northern India and southern Sri Lanka ended up with Indo-European languages. However, southern Indian and northern Sri Lanka ended up speaking a completely different language family. I wonder how that happened.
The Indo-European speakers invaded from the north.
the story of how the sinhalese got there is told in the Mahavamsa a Buddhist pali take that recounts the story of prince Vijaya fleeing his kingdom in mainland India and arriving in sri lanka.
The northern folks, the tamils are there because sri lanka used to be physically connected to Tamil nadu in recent history, like just a few centuries ago
There is definitely a correlation between the connectedness of the world and the amount/speed of language development. People in the past probably diverged from the language they started with because of physical isolation/distance between towns/settlements.
I wonder, if skipped forward a few thousands of years into the future, how the data changes - would we end up with only a few languages or maybe even one after a while?
I miss a control to rotate the circular plot (?)
4000 BC? I don't buy it: if you look at how little Greek changed over the last 3500 years and compared to how different it is to other languages this doesn't sit right
Quick question for you - did the Proto-Indo-European culture spread out further than, or lesser than, the Greek culture?
Because that difference in expansion, distance, and interaction with other cultures might explain the differences you're incredulous about.
Given Punjabi and Irish Gaelic are both descended from PIE, as is _Greek_ I'm going to say that 4K is plenty. The fact that Greek didn't change so much after Mycenean Greek was first attested in 2K BC doesn't say as much about PIE as it does about the Greeks deciding they liked where they were.
Also, _how_ different is Greek to other languages? Given it's a PIE language also.
Greek has several things in favor of it being conservative: 1. It’s been a written language for longer than most other languages 2. It’s been the language of empires 3. It’s been a liturgical language
With that said, I find Modern Greek speakers tend to overstate the similarities between Modern and Ancient Greek. The spelling is conservative but pronunciation is very different and conjugation/declension patterns have been significantly simplified.
Still there’s a lot more intelligibility than between English and Old English for example, which is a good example of how much can change in about 1000 years. Similar with French or Romanian and Latin.
Massive controversy in historical linguistics, singularity2001 on hacker News doesn't buy the age of proto-Indo-European
Cool viz. However it’s hilarious to classify living, spoken languages as “proto” - Latvian and Latgalian. It’s like: “despite there being 2M speakers you guys don’t quite qualify as proper Baltic”.
? They are not classified as "proto". The "Proto-Baltic" label is clearly attached to the knot 2 "levels" closer to the center.
Anyone new to this stuff and fascinated by it could check out "Proto" by Laura Spinney. A new book about PIE and the Yamnaya and such and also incorporating tie-ins to DNA archeology.
Before seeing this page, I had always assumed that PIE was a lot older. It’s amazing how much linguistic development has occurred in ~6,500 years!
I mean, it's enough to look at French vs Spanish or other Romance language which separated just 1500 years ago. Or just look at Old English.
PIE was not just a commonality for language, but also for religion across these regions. And then came along the Semitic wave of religions.
Reconstructing a PIE religion (as opposed to a few rather vague lexemes like "sky god") is regarded as less tenable today than in the past. The well-described features of Indo-Iranian religion are largely innovations and not found among speakers of IE branches that split off before. Some of Greek and Roman religion is actually the result of areal contact, not inheritance from earlier IE times.
You might want to read a scholarly introduction to the field (not a pop-sci one that is likely to be unreliable by definition) from recent years.
pretty interesting etymology connecting Dios, Jupiter, Zeus, etc. (Sky Father)
Languages, cultures and genetic populations are all separate concepts that have to be distinguished.
There can be and there often is correlation, but it's not the same thing. They don't have to move together.
Any ideas where Basque may fit in?
Basque is not an Indo-European language:)
It's believed that basque is a remnant of earlier people before agriculture and indo europeans arrived.
There's a weird number of shared words between Basque and old Armenian, also names of places. But I've only seen random articles trying to link the two, not real scientific papers.
Very well organized.
what is the size of the circle?
I think for the living languages it's speakers (on a log scale, presumably).
Many European languages predate PIE (Basque, Etruscan) and PIE were only a spoken, not written.
I’ve always wondered why it gets so much attention as an early European concept when people had been speaking other languages for thousands of years, doing cave art dating back tens of thousands of years - it’s like history ignores it.
It kind of reminds me of how they treat “dark ages” when there was actually quite a lot going on intellectually.
Possibly because there’s such scant linguistic evidence for the (undoubtedly many) pre-PIE languages in Europe? It’s hard to study something that didn’t leave behind much writing, if any. Basque being a notable exception to that, of course.
With that said, most historical linguists would probably kill for more data on extinct pre-PIE languages.