Where languages come from
The famous fan diagram says one thing very well and one thing not at all. It records descent: who split from whom, and roughly when. It says nothing about whether two languages resemble each other today.
Those come apart more often than the picture suggests. English sits in the Germanic branch and has borrowed so heavily from French and Latin that half its dictionary is Romance. Romanian is a close cousin of French and spent a thousand years surrounded by Slavic. Mandarin and Cantonese share a writing system and are not mutually intelligible at all.
So this page does both. The tree below is the diagram, made openable. The comparison above it measures the other thing — how many everyday words two languages still share — from the same vocabulary the rest of the site teaches. Where the two answers disagree, the disagreement is the interesting part.
Compare two
Any two languages on the tree. The left answer is history; the right answer is measured from this site's own word lists. They are not the same question and they do not always agree.
By descent
opposite sides of the family
Their most recent common ancestor is Proto-Indo-European — 7 branchings apart on the tree.
Proto-Indo-European → Germanic → West Germanic → English
Proto-Indo-European → Italic → Latino-Faliscan → Romance → French
By the words
11% scattered borrowings
Of 607 everyday concepts both languages have a word for, 65 are close enough that a learner would recognise one from the other — in spelling or in sound.
- bug · bug
- the cloud · le cloud
- zero · zéro
- six · six
- judo · judo
- karate · karaté
- taekwondo · taekwondo
- kickboxing · kick-boxing
Far apart on the tree, and closer in the word list than that would suggest — the mark of heavy borrowing between neighbours rather than a shared inheritance. English is the standing example: English kept Germanic words for the kitchen and the farm, and took Romance ones for everything abstract.
The tree
Open a branch to see inside. † means a language known only from writing. The ten Communicator teaches are marked, and five of them are not on this tree at all — they are in the four families underneath it.
- Not attested anywhere. Reconstructed backwards from its descendants, and spoken somewhere around 4500–2500 BC — the steppe north of the Black Sea is the usual guess.
All ten, measured
Percentage of the 284 everyday concepts whose words are close enough to recognise across the pair. Darker is closer.
| en | fr | es | pt | de | it | ka | ru | hi | zh | yue | vi | ko | ja | ms | tr | |
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| English | — | 11 | 8 | 7 | 14 | 9 | 3 | 5 | 9 | 0 | 1 | 2 | 5 | 3 | 6 | 4 |
| Français | 11 | — | 22 | 22 | 6 | 26 | 4 | 5 | 3 | 1 | 0 | 3 | 2 | 2 | 3 | 5 |
| Español | 8 | 22 | — | 56 | 5 | 37 | 3 | 4 | 3 | 0 | 0 | 2 | 2 | 2 | 4 | 4 |
| Português | 7 | 22 | 56 | — | 4 | 33 | 3 | 5 | 2 | 1 | 1 | 2 | 2 | 3 | 4 | 3 |
| Deutsch | 14 | 6 | 5 | 4 | — | 7 | 4 | 5 | 3 | 0 | 1 | 2 | 2 | 2 | 4 | 3 |
| Italiano | 9 | 26 | 37 | 33 | 7 | — | 5 | 5 | 3 | 0 | 0 | 2 | 3 | 3 | 3 | 4 |
| ქართული | 3 | 4 | 3 | 3 | 4 | 5 | — | 7 | 4 | 0 | 0 | 1 | 2 | 2 | 3 | 3 |
| Русский | 5 | 5 | 4 | 5 | 5 | 5 | 7 | — | 4 | 1 | 0 | 1 | 3 | 2 | 3 | 4 |
| हिन्दी | 9 | 3 | 3 | 2 | 3 | 3 | 4 | 4 | — | 1 | 1 | 2 | 4 | 4 | 5 | 4 |
| 普通话 | 0 | 1 | 0 | 1 | 0 | 0 | 0 | 1 | 1 | — | 18 | 3 | 1 | 1 | 0 | 0 |
| 廣東話 | 1 | 0 | 0 | 1 | 1 | 0 | 0 | 0 | 1 | 18 | — | 3 | 1 | 1 | 1 | 1 |
| Tiếng Việt | 2 | 3 | 2 | 2 | 2 | 2 | 1 | 1 | 2 | 3 | 3 | — | 1 | 1 | 2 | 2 |
| 한국어 | 5 | 2 | 2 | 2 | 2 | 3 | 2 | 3 | 4 | 1 | 1 | 1 | — | 8 | 3 | 2 |
| 日本語 | 3 | 2 | 2 | 3 | 2 | 3 | 2 | 2 | 4 | 1 | 1 | 1 | 8 | — | 3 | 1 |
| Bahasa Melayu | 6 | 3 | 4 | 4 | 4 | 3 | 3 | 3 | 5 | 0 | 1 | 2 | 3 | 3 | — | 3 |
| Türkçe | 4 | 5 | 4 | 3 | 3 | 4 | 3 | 4 | 4 | 0 | 1 | 2 | 2 | 1 | 3 | — |
What the percentage is, and is not
It is the share of 284 everyday concepts — kitchen, forest, toolbox, feelings, numbers — where the two languages' words are close enough that a learner would recognise one from the other, comparing spelling and pronunciation and taking whichever is closer.
- It is not a cognate count. A historical linguist asks whether two words descend from the same ancestor. This asks whether they still look or sound alike, which counts modern borrowings — computer, sofa, taxi — as matches. That is the right measure for a learner and the wrong one for a family tree.
- It undercounts real relatives. French eau and Spanish agua are the same Latin word and score nothing, because a thousand years of sound change left no resemblance. Every number here is a floor, not a total.
- Basic vocabulary is the hardest place to borrow. That is why English scores closer to German than to French on this list, even though the dictionary as a whole leans the other way. Knife, wolf and hammer stayed Germanic; the abstractions went Romance.
Want the same measurement the other way round — which words the whole world shares rather than which languages do? That is Shared.