Multilingualism and statistics

I am a big fan of statistics, which my father -a former university mathematics teacher- always considered “idle mathematics”. We used to watch almost one basketball game on TV per week and he would always point out how shooting stats are faulty in the following way: if a player has a ratio (I will use the typical Spanish TV stats representation) of 33 % (1/3 or 1 basket out of 3 attempts) and another player has a ratio of 33 % (3/9 or 3 baskets out of 9 attempts) you are not considering how, when and where the shots were made when seeing those stats out of the blue and they both seem equally bad. Mind that I am talking about the 90s, well before big data was introduced and savvy data gurus would be part of any decent sports team. He had a very good point which is worth extrapolating (or at least taking the risk to do so).

I came across this article in Forbes which is worth mentioning because of this map, originally from Jakub Marian and his web-site, always full of interesting posts:

The numbers seem clear cut, as Mr Marian himself explains:

Whole numbers in the following picture agree with the classical median, e.g. the 2 for Germany indicates that if you ordered all Germans according to the number of languages spoken and asked the one in the middle, he or she would speak exactly two languages (so it is reasonable to say that a typical German speaks two languages).

The data is from a EU survey from 2012. Now, I do not have the survey to check it up, but by just seein it, there are a few points that I would like to rise anyway.

The first one would be what are you asking for. Foreign languages? Are the regions / countries / state with two or more official languages and its population considered differently in the statistics? For example, let’s take the Baltic republics with 2,7 languages per person, where I assume the local official language and Russian, together with English (or Polish in Lithuania or Finnish in Estonia), are to be considered to understand the statistics. If we would take Russian from the equation (and I do not see a reason to do so except the point I want to make), the result is 1,7, very similar to Poland and Southern Europe. Galicia, Catalonia or the Basque country would easily yield that very same result (with English, or Portuguese in Galician and French in Catalonia and the Basque Country).

The second point is whom do you ask. Everybody? Nationals and permant residents only? I have already pointed out the role that a minority or (former) co-official language plays, and emigration acts in the very same way. As someone living in Germany for quite some time, I find 2.0 a surprisingly good result. I would assume that a good deal of that result comes from migrants’ first or second language skills and not necessarily from a flock of highly educated Germans who master English. The same applies to Austria: German and the emigrant’s languages (or Hungarian, Croatian and Slovene) and English make up the 2.2. That will not go well with some ultraconservatives and “traditionalists” out there.

The third point is that I am squeezing my head to understand the Benelux results. I take Slovenia as an example (another country I know very well): Slovene and “Serbocroatian” would make the 2.0 and then English and German or Italian (sometimes any of these last two or both instead of “Serbocroatian”) would yield the final 3.0. Now, I have my doubts of the active competency in “Serbocroatian” of any Slovene younger than 35 but that is where Italian and German step in.

Luxemburg is easy to explain: Luxemburgish, French and German make up that nice 3, English along with probably Portuguese or Turkish add the 0.6 to get by far the best result in Europe. No doubt about it.

In Belgium I would play the stats like this: French or Dutch on the one hand plus English and then the remaining 0.6 comes from any of the other 2 official languages (Dutch, French and German).

The Netherlands is surprising. I assume the following: Dutch and English make up that 2.0. Some degree of German (as in Slovenia with “Serbocroatian”) with questionable degree of active output (the same, by the way, as in Galicia with Portuguese) and English would make up that nice 3.0. The remaining 0.2 would be either Frisian or the emigrants’ language, yielding a questionable 3.2

What about Italy? I would be brave enough to give an extra 1.0 as I believe that the vast majority of the population is bilingual / diglossic in standard Italian and dialetto. On the other hand, I wonder whether the Swedes and Danes consider that they speak their neighbours’ language (as the Dutch with German).

Nice statistics do provide you nice data but interpreting it casts some shadow of understanding of what is behind. I assume that, in average, most of Europeans are equally regarding language skills: they speak their mother tongue and have a decent degree of communicating in a different language. What matter here is the self-evaluation of what speaking a language is and what communication skills are. I assume that the poll would consider that piece of information but point it out here does not harm… and enables me to play with stats.