As Sumo fan, haptician (practitioner of haptic pronunciation teaching) and veteran, one of my favorite metaphors for ongoing interaction in the (pronunciation) classroom has always been "H2H" (hand2hand combat.) Research by Mojtahedi, Fu and Santello, of Arizona State University - Tempe highlights an important variable in such engagement, evident in the title: On the role of physical interaction on performance of object manipulation by dyads.
Granted, there is a difference between two people holding on to a piece of metal and guiding it around together, cooperatively--and an instructor being mirrored in gesturing by students across the room, synchronized with speaking words and phrases. Research in mirror neurons in the brain, however, would suggest that the difference is far less than one might think. In a very real sense, if you are paying close attention, watching something being done is experienced and managed in the brain very much like doing it yourself.
Now hold that thought for a minute while we go on to the next, related study, How spatial navigation correlates with language by Vukovich and Shtyrov at the HSE Centre for Cognition and Decision Making. In this study, subjects were first identified as to whether they were more "egocentric" or "allocentric" in their ability to grasp the perspective of another person, somewhat independent of their own position in space or time. (A concept somewhat analogous to field dependence/independence.)
What they discovered was that subjects who were (spatially) allocentric were also better at understanding oral instructions that required differing responses, depending on whether the subject pronoun of the description was 1st person singular or 3rd person. And more importantly the same areas of the brain were "lighting up", meaning processing the problem, for both language and spatial navigation.
Now juxtapose that with the finding of the other research which demonstrated that side-by-side (SxS) rather than face-to-face (F2F) "help" on the H2H task was more effective. F2F assistive engagement requires, in part the transposing of the movement of the person facing you to the opposite side of your body, an operation that we discovered a decade ago in haptic pronunciation teaching was exceedingly difficult for some instructors and students.
So what we have is a complex of the factors affecting success in gesture work: (probably) inherited ego or allo-centric tendencies which will impact how well one can accommodate a model moving in front of you, taking on the same handedness, as opposed to mirror image, and fact that some, less skillful learners are assisted more effectively by a partner SxS instead of standing F2F.
In other words, both studies seem to be getting at the same underlying variable or issue for us: why some gestural work works and some doesn't. This is potentially an important finding for haptic pronunciation teaching or just use of gesture in teaching in general, one that should impact our "standing" in the classroom, where we locate ourselves relative to learners when we manage or conduct gesture.
Sometimes facing your problem is not the answer!
Mojtahedi K, Fu Q and Santello M (2017) On the Role of Physical Interaction on Performance of Object Manipulation by Dyads. Front. Hum. Neurosci. 11:533. doi: 10.3389/fnhum.2017.00533
Nikola Vukovic et al, Cortical networks for reference-frame processing are shared by language and spatial navigation systems, NeuroImage (2017). DOI: 10.1016/j.neuroimage.2017.08.041