Guess the Language
You get the first article of the Universal Declaration of Human Rights in one of 128 languages. Name a language and the row that comes back compares it with the answer: family, branch, writing system, and how far apart the two are on the map. You have 6 guesses.
- Same as the answer.
- On distance only: the answer comes from within 1,500 km of your guess.
- ↗The direction from your guess to the answer, as on a flat map.
Where the texts and the clues come from
Why is the text always the same article?
›
Why is the text always the same article?
›Article 1 of the Universal Declaration of Human Rights says the same thing in every language: people are born free and equal in dignity and rights, and should treat each other in a spirit of brotherhood. Because the meaning never changes, the only thing left to go on is the language itself. It is also the most translated document in the world, so it exists in languages that have very little else in print. The texts come from the UDHR in Unicode project, and only the ones it marks as complete and reviewed are used.
What do the colours mean?
›
What do the colours mean?
›Family, Branch and Script are green when your guess matches the answer and grey when it does not. Distance is green only on the right answer, yellow when the answer comes from within 1,500 km of your guess, and grey beyond that. The arrow under the distance points from your guess toward the answer.
Where is a language on the map?
›
Where is a language on the map?
›At the single point Glottolog gives for it, which marks where the language comes from, not where most of its speakers live today. English is placed in England, Spanish in Spain and Portuguese in Portugal, even though most speakers of all three live in the Americas. Afrikaans sits in South Africa, more than 9,600 km from English, although both are Germanic.
Which way does the arrow point?
›
Which way does the arrow point?
›The way you would move on an ordinary flat world map, going the short way round the Pacific. It rounds to one of eight directions. A great-circle bearing would be more exact but less useful: from Europe it points north to reach Hawaii.
Where do the families and branches come from?
›
Where do the families and branches come from?
›From Glottolog 5.3, the catalogue of the world's languages kept by the Max Planck Institute for Evolutionary Anthropology. The family is the top of Glottolog's tree, and the branch is a well-known group below it, such as Germanic, Slavic, Bantu or Semitic. Glottolog does not accept Niger-Congo as a proven family, so Swahili, Yoruba and Zulu appear under Atlantic-Congo. Basque is related to no other known language and is listed as an isolate.
Does the branch go green if the families differ?
›
Does the branch go green if the families differ?
›No. Branch only goes green when the family matches too. For the families with a single language in the pool, such as Korean, Quechua or Navajo, the branch repeats the family name.
How are the hints given?
›
How are the hints given?
›After three wrong guesses the answer's family is written above the search box, and after five its branch. They are there for texts in a script you cannot place, where the board alone gives you little to work with.
Why is Croatian in the game but not Bosnian?
›
Why is Croatian in the game but not Bosnian?
›The two texts are identical word for word, so nobody could tell the answer apart from the other one. Serbian stays in because its text is in Cyrillic and differs in a few words.
Which languages are left out?
›
Which languages are left out?
›Creoles such as Haitian Creole, because Glottolog files them under the European language they grew from and that is a call many players would dispute. Constructed languages such as Esperanto. Irish, because the only checked text available has spelling errors that could not be corrected against the original. And any language most people would not recognise by name, however many people speak it.
Can I type a language that is not in the list?
›
Can I type a language that is not in the list?
›No. The search box only offers languages from the pool, so a name it does not show cannot be the answer. It accepts common alternative names, so Farsi finds Persian, Filipino finds Tagalog and Bangla finds Bengali.
What gives a language away
Neighbours that are not related
Distance is the column that misleads most. Hungarian sits 190 km from Slovak on these points and is not related to it; its nearest relative in the pool is Estonian, 1,360 km away. A yellow distance next to a grey family is a sign you have found the right part of the map and the wrong family.
The Caucasus is the hardest place for this. Armenian, Georgian and Azerbaijani are all within 240 km of each other and belong to three different families, Indo-European, Kartvelian and Turkic, written in three different scripts. Basque, the one isolate here, is 320 km from Spanish.
Relatives on the other side of the world
Malagasy is spoken off the coast of Africa and is Austronesian, the family of Malay, Javanese and Maori. Its nearest relative in the pool is Malay, 6,540 km away, while Swahili is only 1,620 km from it. If a Latin-script text from Africa keeps coming back with a grey family against every African guess, try Asia.
Afrikaans is the same trap the other way round. It is Germanic, placed in South Africa, and 11,130 km from Icelandic, the farthest pair in the pool that share a branch.
Scripts that point the wrong way
81 of the 128 texts are in Latin letters, so a Latin script on its own says very little. The scripts that do narrow it down are the ones that break the pattern of their family. Maltese is Semitic, a relative of Arabic and Hebrew, and the only Semitic language in the pool written in Latin letters. Yiddish is Germanic and written in Hebrew letters. Tajik is a form of Persian and written in Cyrillic.
Hindi and Urdu are close enough in everyday speech to be understood across the border, and they come back with the same family and branch. The script column separates them: Devanagari for Hindi, Arabic for Urdu. Their points are 1,010 km apart, which is inside the yellow range.
20 of the 28 scripts in the game are used by a single language in the pool, among them Georgian, Armenian, Thaana in the Maldives, Cherokee and the syllabics used for Inuktitut. If you can name the script, those are a one-guess answer.
Pairs that are nearly the same text
Some answers come down to a few letters. Zulu and Xhosa open with the same fifty characters before they split. Danish and Norwegian share their first six words. Kinyarwanda and Kirundi begin the same way and their points are 160 km apart. Tibetan and Dzongkha use the same script and sit 120 km apart.
Serbian and Croatian say the same thing in almost the same words. The Serbian text is in Cyrillic, which settles it, and one word in the Croatian text matches the Croatian standard: it writes the vowel in the word for conscience as ije, where the Serbian text has e.
Long and short
The same article runs from 43 characters in Mandarin to 374 in Maori. Chinese characters each stand for a syllable with its meaning, so Mandarin and Cantonese need about a quarter of the length of most alphabetic texts. Japanese, at 85, mixes Chinese characters with its own syllable signs, and Korean, at 87, packs each syllable into one Hangul block.
Past that, length is a poor clue: Maori, the longest, is in Latin letters like most of the pool. A very short text is the one case where length helps, and it points to East Asia.
What the pool looks like
53 of the 128 languages are Indo-European, from Icelandic to Sinhala, because the family covers most of Europe and much of South Asia. The rest come from 19 other families, plus Basque, which has none. 9 of those families have a single language each in the pool, including Korean, Japanese, Georgian and Quechua, which is why guessing one of those either wins or rules out the whole family at once.