During my time at the London Institute of Education (2005-7), I was deployed in the TESOL section and worked with an outspokenly international group of students. These students were recruited after a rigorous selection in which superior IELTS scores were mandatory. All were, consequently, "fluent" in "English" when they arrived in London. The scare quotes around both terms above will become clear as we go on. For the thing is: all of these young people were highly skilled globalized junior academics, but many of them were unhappy in London.

I talked to a great many of them and started making observation notes on their English conversational proficiency. I also asked them how they felt about their English proficiency, and when one of them replied "I can't understand their [i.e. UK English] jokes and that frustrates me", my curiosity was triggered. I started talking to them on the specific bit of English they felt they lacked in London. The answers were highly diverse, but some stood out. One recurrent answer was: I don't have the English that can help me find a boyfriend/girlfriend - the English one needs to flirt and enter into a love relationship with someone. Another was: I don't have the English I need to understand entertainment shows on British TV. And yet another: when I go out for drinks with British friends, I just can't understand a lot of what they're saying in the pub. Many articulated frustrations about the fact that their limited English proficiency made it very difficult for them to come across as an interesting, witty, creative and nice person. Many felt socially awkward and lonely, and had the impression that making real friends was terribly hard, given the constraints they experienced in informal social interaction with others.

Their responses reminded me of my own experience teaching and living in Chicago in the Winter of 2003. From 9 to 5, I would be talking shop there, and interlocutors would have perceived me as highly articulate and confident, perhaps even eloquent in English. As soon as I left the UofC campus and went shopping, however, I felt I was lacking almost all of the English I needed to identify the right meat cuts, vegetables or cleaning products. And one of my most catastrophic communicative experiences was when I had to call a plumber about a drainage problem in my bathroom: I lacked literally every bit of English required to adequately explain the problem and was reduced to begging the plumber to come over and see for himself. On campus, I was a "near-native" user of English, while in the supermarket or with the plumber I must have sounded like just another immigrant struggling with basic English vocabulary.

Such anecdotes are relevant for at least three reasons.

• They show us that "language learning" is effectively register learning. My students and myself had acquired the academic register characterizing contemporary globalized academic practices and culture; we had not, au contraire, acquired the registers that controlled specific informal social and cultural communication modes, and could consequently not perform the roles we were supposed to play in and through them.
• In the case of my students, they also show that "language" testing is in actual fact register testing: high IELTS scores indicate a high level of active and passive proficiency in a limited set of registers and genres qualified with a  (rather unhelpful) umbrella term as Academic English. They do not indicate a general socioculturally adequate competence in English, and do not as such announce a generative or cumulative competence. That is: having achieved high levels of academic register-genre proficiency does not automatically generate (or even facilitate) competences outside the domains covered by such registers and genres; such specific register-genre competences must be learned separately.
• And most importantly, they show us a thing or two about integration. Let me elaborate that latter point.

There is, in the context of migration and superdiversity, a policy response which is widespread across Europe (and further afield) in which language learning is proposed as the key to "integration". The latter is a word in search of a clear definition (and has been for decades), but in actual practice, it is usually paraphrased as "participation in social life", with some emphasis on facilitating entrance into the labor market. Observe that "integration" is usually presented as one single process in which someone presently "not part of society" will become part of that society by a unilateral effort of adaptation, in which language learning is crucial since - one frequently reads - one cannot participate in the life of a community without communicating with other members.

What we now know is that

• Integration is not a single process but a multiple one, in which several very different forms of "integration" need to be achieved, into numerous specific social milieux and niches, each organized and characterized by their own sociocultural normative codes, in order to be, let us say, happy as a social and cultural being.
• Integration into the "most important" social milieu - academic work in the case of my students, the labor market in the eyes of many policy makers - does not guarantee integration into the different milieux and niches that make up social life outside the "most important" segment of it. As my own experience showed, one can be highly integrated in the segment of labor and the sociocultural milieu that sustains it, and poorly integrated (even highly marginal) in several other social milieux. In fact, this assemblage of different degrees of "integration" in which one is simultaneously very well integrated in some segments of sociocultural life, less integrated in some others and not integrated at all in another set of them - is perhaps the default mode of "integration" any person would have in social life in general, at any point of time.
• Consequently, teaching competences and skills deemed useful for "integration" would seem to require a very precise diagnostic stage in which the specific register-genre needs valid for targeted social milieux (and thus defining a range of very different integration processes) can be identified and followed up by more precise and specific knowledge transfer.

Being "fully integrated" as a person, when one investigates it in some detail, actually refers to a set of experiences of satisfaction - happiness, let us say - derived from a perceived smoothness in social contact beyond the borders of narrowly conceived and functionally defined social milieux such as that of labor. It actually means that one is integrated into the full set of social milieux experienced as crucial for a satisfying social life. When we teach people the language they need for this purpose, we have to teach them the specific bits of language that make them happy. The term "happy" sounds funny, perhaps, and there is no tradition in language teaching where it has ever been central. I suggest we take it very seriously.