It could've been interesting to create a language learning program that would determine one's active vocabulary (e.g. based on one's contributions to the social networks and/or public chat transcripts - suppose for simplicity that privacy issues could be solved here somehow), compile more or less close analog in the target language and craft the study path based on that. Of course, such approach might have missed the point of expanding one's horizons with learning a new language. But it might also encourage the learners to use the new language sooner and more efficiently if they could do so without deviating too much from their original identity.
On the other hand, getting the proper feel for the language without living in the country where that language is spoken might be too challenging for a human being. How many languages would it be even possible to learn properly and what is the price for that? I don't know about any strictly scientific studies in this area.
Personally, I often feel that my English is either dry as winter leaves or rough like a cartoon drawing (or both), and my Dutch dwells on the pre-teen level, which results in a personality switch every time I switch the language.
I wonder how many other seasoned ESL speakers have similar experiences? It might only be the thing for those who, like me, started actively using another language relatively late.
It is great that the modern tools (like online translation) help to reduce language barriers, but could these barriers one day disappear completely? So many misunderstandings, from personal to country level, might go away then. (One hopes...) Would it be possible for a human both preserve their own identity and easily "map" it into any other language / culture?
Of course, there is more to the game than just language (Le Ton Beau de Marot by Hofstadter is explaining that much better than I ever could), but one has to start somewhere.