Researchers and professionals in any number of fields that work with motor skill development know better. To the extent that L2 pronunciation is a motor skill, two intriguing new studies suggest something of what is involved in effective drill and practice.
The first, by Wymbs, Bastian, and Celnik, "Motor skills are strengthened through (memory) reconsolidation", (Summarized by ScienceDaily) suggests that if you practice a slightly modified version of a skill that you want to master, you actually learn more and faster than if you just keep practicing the exact same thing multiple times in a row. The second, by Castellanos and colleagues at Drexel University, "Surgical trainees retain information, master skills better when honed beyond proficiency" (also summarized by ScienceDaily) looked at how "overlearning" figures in to skill development.
Here's the fascinating parallel to pronunciation work. It is almost a given today that the goal of pronunciation instruction should be intelligibility--not accuracy. If students go further personally, good deal, but it is not our responsibility to help them or encourage them in that. Furthermore, "physical" work with pronunciation is generally not seen as being as essential as is metacognitive engagement. In other words, our first priority must be providing as much explanation and insight as possible. With that, a good percentage of learners will pretty much on their own get close to intelligibility. And I think is probably true, in fact.
But the research studies seem to tell us that to even get to "intelligibility" most need to go beyond it--and have to do that with at least varied physical practice. This is a great example of the "Product equals process" fallacy. In other words, how you get to, let's say an acceptable, functional spoken use of the consonant, th, for example, may not be theoretically consistent with the final goal, itself. In other words, "enlightened" drill, as disconnected from authentic interpersonal communication as it can be, is simply essential for most learners, if they are going to finally land on the high side of intelligibility.
One of the principles of the Lessac system is that pronunciation must be "drilled" extensively as homework--but not consciously integrated into spontaneous speaking. When done right, as I have seen literally hundreds of times over the years, the new, improved sounds just begin showing up in conversation. Learners are trained to recognize miscues and changes but not to emotionally react to them; just note them.
I am working on a presentation for the annual BCTEAL Island Regional Conference next week: Do your homework, which I will report on later. Those two studies, along with Lessac, suggest a great deal as to how we should structure and manage effective pronunciation homework.
Johns Hopkins Medicine. (2016, January 28). Want to learn a new skill? Faster? Change up your practice sessions. ScienceDaily. Retrieved February 5, 2016 from www.sciencedaily.com/releases/2016/01/160128130955.htm
American College of Surgeons. (2015, October 6). Surgical trainees retain information, master skills better when honed beyond proficiency: New study finds that overlearning can be a highly effective surgical training approach that shortens the learning curve. ScienceDaily. Retrieved February 9, 2016 from www.sciencedaily.com/releases/2015/10/151006144518.htm