http://digital.georgetown.edu/gurt/1999/...999_07.pdf

Seeing as it's a decade old I'm sure many people here are familiar with this whitepaper, but I just encountered it today for the first time. It's a summary of the lessons learned by the teaching staff at the Foreign Service Institute, which trains U.S. government employees in foreign languages.

Among self-study learners of other languages, FSI has a particular reputation for being a drill-heavy course, almost antithetical to the AJATT-inspired methods advocated on this forum. This is due to the publicly available training methods they developed in the 50's and 60's which follow that approach. It turns out they've advanced quite far since then, however. To quote a footnote of the above-linked document:

FSI Wrote:It is sometimes said at FSI that we began forty to fifty years ago with a metaphor of “teaching the course,” but that, as the years have passed and we have understood more, we have moved from that concept to “teaching the class,” to “teaching the students,” to “teaching each student,” to the present metaphor of “helping each student find ways to learn.”

It makes for a very interesting read.

It's too bad their modern approach is not so accessible to the outside world.

Some more gems:

FSI Wrote:Despite what some published research has indicated, for example Brecht, Davidson, and Ginsberg (1993), our experience is that in-country immersion is most effective where the learner is at higher levels of proficiency.

FSI Wrote:For knowledge of one language to be a real advantage in learning another, however, it needs to be at a significant level. Thain and Jackson (n.d.) and an interagency group determined recently that this kind of advantage takes effect at a three-level proficiency [ED- S/R-3: General professional proficiency: Able to speak accurately and with enough vocabulary to handle social representation and professional discussions within special fields of knowledge; able to read most materials found in daily newspapers.] or better. Below that level, knowledge of a second language does not appear to make any useful difference in acquisition of a related third language.

In fact, our experience at FSI—based on work with such related languages as Thai and Lao,German and Dutch, Russian and Ukrainian, French and Italian, and Spanish and Portuguese—is that a relatively weak knowledge of one language may be an actual hindrance in trying to learn a related third language.

FSI Wrote:Segalowitz and his collaborators have shown us that iteration of relatively easy processing tasks is crucial to developing reading skill.

FSI Wrote:the whole point of language pedagogy is that it is a way of short-circuiting the slow process of natural discovery and can make arrangements for learning to happen more easily and more efficientlythan it does in ‘natural surroundings’