Disembodied Pronunciation done well--by Rosetta Stone

Although there appears to be no readily accessible published research as to the efficacy of the widely promoted language teaching program, Rosetta Stone, if the testimonials on the website are to be believed,  it certainly works for at least some learners.  Having reviewed the Korean and ESL programs, I am struck by how well it does a little--a good business model.

Image: Rosetta stone.com

For the visual-auditory, literate learner (See previous posts on the myth of learning styles, however!)  for whom pronunciation will not be much of an issue, it offers reasonably good, low cost, individualized access to at least functional vocabulary and structure. In fact, that it does not do active pronunciation instruction (other than repetition with some feedback) may almost be a plus, as opposed to presenting it even more "disembodied," as is the case with many current, computer-based systems.

A haptic interface could certainly be developed to use with it. I suggested that when I talked to one of the designers a couple of years ago. Basically, he confidently informed me that according to Krashen and most experts in the field, comprehensible input and aural comprehension were generally sufficient for developing acceptable pronunciation--and selling the product. Well . . . duh. He was at least half right. And besides, recall that the pronunciation of the original Rosetta Stone took over three decades to figure out . . .