1. Forum
  2. >
  3. Topic: Duolingo
  4. >
  5. L2/L1 Ratio as a Measure of L…


L2/L1 Ratio as a Measure of Language “Importance”


We saw in the forums yesterday a discussion on the relative “usefulness” of French versus German, and the point was raised that French is one of the very few major world languages where the number of non-native speakers (“L2” speakers) significantly outnumber the number of native speakers (“L1” speakers).

This made me wonder how many other languages have this distinction and whether the ratio of L2 to L1 speakers (the “L2/L1 Ratio”) can help assess the “usefulness” of any particular language.

Anyway, based on Ethnologue data reproduced on Wikipedia (https://en.m.wikipedia.org/wiki/List_of_languages_by_total_number_of_speakers), coupled with a few extra searches, it seems that there are four major languages with significantly more L2 than L1 speakers, namely:

Swahili (L2/L1 Ratio of 5.7)

Malay-Indonesian (2.6)

French (2.0)

English (1.6)

Persian also merits mention, apparently having around an equal number of L1 and L2 speakers. I confess that this comes as somewhat of a surprise to me and I find myself wondering why there are so many L2 Persian speakers.

Besides these languages, there is a second group of languages where the L2/L1 ratio is over 0.5, including Arabic (0.5), Hindi-Urdu (0.7), German (0.7), Russian (0.7) and Hausa (0.8). Interestingly, neither Spanish nor Portuguese form part of this group.

I would hypothesise that the L2/L1 ratio offers a measure of the degree to which a language is used as a means of communication between people of different linguistic backgrounds. English and French have long served this role internationally. Swahili and Malay-Indonesian — while not typically considered “major world languages” are both linguae francae of large regions characterised by great linguistic diversity. I might also add—reiterating a point I have made previously—that Indonesian really does not often get the attention it deserves.

In any event, I open the Duolingo floor to further discussion, and wish everyone a happy new year!

December 31, 2017



Thanks for compiling these numbers. Basically you have shown that any argument considering the L2/L1-ratio is not really worthwhile. Clearly anybody who does not know any languages should be advised to learn English first.

Maybe L2/L1*(L2+L1) would be a better metric, i.e. the total amount of speakers normalized by the ratio of language learners. If you already have all the numbers in a spreadsheet maybe you can calculate this and share the result?

Btw, what is the L2/L1 of languages like Esperanto, Klingon, Elvish etc.? There are a few children raised in Esperanto, but for Klingon and Elvish it should be infinite!


Thanks for this.

To deal with the case of constructed or dead languages (which skew the L2/L1 ratio), I limited my calculations to “major” world languages (ie, those included in the list of top languages as compiled by Ethnologue). Certainly, if Esperanto had tens of millions of speakers, its L2/L1 ratio would be off the charts. This would actually be a compelling argument for anyone to learn Esperanto, even as a “first” L2, since it would show that Esperanto had reached a critical mass of speakers and was fulfilling its mission of becoming a world auxiliary language.

I have run the numbers on what I will call the “Territrades Metric” for the languges on my spreadsheet. The “most important” languages as measured by the Territrades Metric:

  1. English (1617)

  2. Malay-Indonesian (744)

  3. Swahili (609)

  4. French (461)

  5. Hindi-Urdu (356)

  6. Mandarin (235)

  7. Russian (196)

  8. Arabic (192)

  9. Persian (123)

  10. Hausa (115)

  11. Spanish (110)

As may be expected, English dominates, followed by the same languages that topped the L2/L1 list: namely, Malay-Indonesian, Swahili and French.


I have a number of Indonesian colleagues, and they have told me that few people among the millions on its 13,000 islands speak the same language. Indonesians speak their local language first and Indonesian second - some not very fluently at all, as they may indeed only use it in school. One friend told me he and his wife, although coming from locations not too far apart spoke different languages and could not understand each other...so they both use standard Indonesian at home. In an environment like this you can see why L2 would be dominant.

My SIL is Malaysian, and there is a huge minority Indian and Chinese population in Malaysia that may speak other languages in preference to Malaysian, thus creating a L2 shift.


According to the Wikipedia link you provided, English ranks #1 with 611 million L2 speakers, far outpacing Hindustani with 215 million speakers. That sort of jibes with common sense, English being the lingua franca of the world for both business and politics. You have the British Commonwealth which spans the globe, and the American Empire which replaced Spanish in her colonies. English is also the language of science. The most prestigious journals are Science (American) and Nature (British). And of course, computer languages are based on English, so it dominates in IT. English in today's world is like Greek in the ancient world. If you want to study the past, you can't get around it -- you have know Greek. Even the Hebrew bible was translated into Greek so that the Jewish people could read it, not the Greeks. Romans conquered the world, but studied and copied Greek language and customs. China is rising as a world power, but their language will not replace English, as it is too entrenched and Chinese is not developed in ways that English is. Chinese intellectuals, like everyone else, learn English to practice their craft.


> her colonies

Who's she?


i can't speak for other languages but for malay in context of malaysian society, it is not very useful these days. i've come across lots of my fellow malaysians, particularly the non-malays, not being fluent at all in the language. usually we just talk in english. or manglish.


The L2/L1 ratio for Latin is infinity, or at least a very large number, since it is the mother tongue of very few people. Same with many constructed languages and dead languages.


It sounds better than it actually is. Not that many people actually are fluent in Latin to the point that they can have a conversation, listen to a somewhat long audio segment or even read classical authors without resorting to translation methods one has been taught in high school and at uni. So calling someone L2 in Latin isn't really comparable to L2 in English, or any spoken living language for that matter.


I happen to be friends with one such latinist, who has given speeches and translated songs and poetry to Latin, for example, as a hobby. They do exist. (If even one such person exists, and if nobody speaks Latin as their mother tongue, then the L2/L1 ratio is infinite.)


I still think it's different to write a speech or translate a poem, in which case you are able to check every little grammatical detail before presenting your work - and even if you make a small mistake, few people will notice - then to be able to hold a conversation in a modern language as a L2 speaker.

I know some people have learned to actually be able to speak Latin and hold basic to medium level conversations in it. I'm just pointing out that even those people are not nearly at the same fluency level as L2 speakers of any modern language would be.

I agree that even if one person exists whom we could truly call a L2 speaker of the language, then yes, the ratio would be infinite. But that's more a maths fact than really saying anything about the status of Latin in our modern world.


The original founder of the Weekly Incubator Update approves of this message!

Learn a language in just 5 minutes a day. For free.