Now I believe this pattern, expressing the results of a passionate search process and being confronted with the requirement to "teach" as a result while not having a body of knowledge to pass on, is actually quite common in highly creative people involved in similar processes. From some cursory reading it seems to me something similar was the case for Moshe Feldenkrais; in his case, the insight that he taught a "method of no-method" was somehow preserved, but is still at odds with the fact that there is now the "Feldenkrais method". It would be interesting to know whether there are other examples.
So, in a way, I arrive at a point similar to Grahams: how does a student best learn from such a person and process. One student of both Morihei Ueshiba and Feldenkrais told me about the later: "I did not study with Moshe to learn another profession. I wanted to understand how he thought" That seems a good strating point to me.
There are in fact methods of "no method" that are rather common in CMA, but rather than name them, I think it is fairer to state what makes some of them tick: Realize that any fixed form is an initial crutch just to give the student an idea of a larger context of application, that must be discarded as soon as the student gets the idea and can practice it in the larger context. The student must understand this even as he is being taught the crutch.
For instance, it may be easy to find "jin" standing in one spot, but to find "jin" in all movements is too difficult a starting point for basically anyone, so you first have the student find it in standing. If the student were to cling to standing, he would hinder his ability to express it eventually in all movement. So for any initial crutch, an equal amount of time must be spent learning how to get off the crutches. The longer you use the crutch, the worse off you are. Likewise that does not mean you won't go back and test your ability to use that crutch, as a test of your progress in the larger context, but that the crutch is never the vehicle of progress. This extends even to particular fixed movements or forms or waza. Progress is measured by widening scope.