[ExI] What's the fastest language to speak?

[Anders Sandberg]

A quick little check using Google Translate (so do not expect any 
scientific quality) of the story Eugene posted, where I translated it 
into different languages and checked the number of words and characters 
in the translation. In the case of Chinese, Japanese and Thai I used the 
translitterated versions since Word's wordcount didn't notice the word 
boundaries properly (and besides, ideographic characters are cheating :-)

Albanian: 958 words, 4,425 characters
Arabic: 799 words, 3,420 characters
Basque: 647 words, 4,221 characters
Chinese (traditional, translitterated): 805 words, 3,604 characters
Czech: 722 words, 3,782 characters
English: 845 words, 4083 characters
Finnish: 615 words, 3,969 characters
French: 905 words, 5,473 characters
German: 805 words, 4,636 characters
Icelandic: 787 words, 3,967 characters
Italian: 871 words, 4,410 characters.
Japanese (translitterated): 988 words, 4,542 characters
Latin: 644 words, 3593 characters
Hindi: 952 words, 3,597 characters
Hebrew: 674 words, 2,935 characters
Russian: 725 words, 4,158 characters
Swedish: 777 words, 3,930 characters
Swahili: 798 words, 4,193 characters
Tamil: 663 words, 4,481 characters
Thai (translitterated): 1,020 words, 4927 characters
Vietnamese: 1,146 words, 3,937 characters
Welsh: 938 words, 4,220 characters

This is of course not even close to the information measure used in the 
original paper it refers to. Still, we can see some fun stuff. 
Vietnamese used the most words but had a fairly average number of 
characters. French used the most characters but had a fairly average 
number of words. Swedish, Icelandic and English are close to the center 
of the distribution. Hebrew had fairly few words and short text - 
possibly because wovels were left out. Finnish used the least words, to 
nobodys surprise :-) Tamil used the longest average words, 6.7 
characters, and Vietnamese had the shortest, 3.4 characters.

So in terms of saving space, write in Hebrew.

Incidentally, the article referenced was this: François Pellegrino, 
Christophe Coupé, Egidio Marsico, /Across-Language Perspective on Speech 
Information Rate, /Language - Volume 87, Number 3, September 2011, pp. 
539-558
- I HATE when newspapers do not even give a hint where the studies they 
misinterpret come from. In this case they didn't even mention the 
researchers by name.

-- 
Anders Sandberg,
Future of Humanity Institute 
Oxford Martin School 
Faculty of Philosophy 
Oxford University