I'm not going to attempt to wade into what you've proposed as the criteria for different belt levels. However, I think that some of your basic premises, those driving your criteria, aren't valid.
Appledog wrote: When I attempt to introduce shenfa such as chan si gong in the form of basic exercises, these exercises have no real function (and actually no real form) yet I am introducing various forms to point out the internal body movement. The student is left, presumably, with the shenfa but without any knowledge of how to apply it in a martial arts situation. In Chansigong exercises, for example, once the student "gets" the desired quality of movement, he or she is expected to use the turning points to release energy at any point along the circle. Therefore, showing the student various martial arts applications in isolation does not help them to learn this shenfa, but, showing them the shenfa alone does not necessarily show them how to execute the movements. [b](edit: remember, as a newbie, said student may not have any idea what any particular 'martial arts move' could be and may not be able to intuit any applications on their own. so shenfa isn't enough).
There are three components to teaching silk reeling exercises. These are, the body mechanics, the "energetics" (qi stuff) and applications. They should go hand in hand. To start with, silk reeling exercises involve moving qi from the dan tian out to the extremities and then returning qi from the extremities back to the dan tian. It is a circuit. Part of teaching silk reeling should be gross tracing of the route taken by the qi in the process of the choreography of the "circles". Qi starts in the dan tian, goes to the ming men, up the back to the back of the shoulder, to the elbow, to the fingers. On the way back, it goes from the fingers to the elbow to the rib cage to the dan tian. Qi to the shoulder is the basis of a shoulder kao. Qi to the elbow is the basis of an elbow strike (zhou kao, ji zhou, for example). Qi to the fingers is often a push or strike. And so on. Knowledge of basic uses of these actions can be integrated very, very early on in the teaching of basic silk reeling, if the teacher knows the material well and is skilled at teaching it. Skill in those applications is a different issue, and comes with time and practice.
3. Since Taijiquan is intended to be learned by rote from childhood, martial development in the initial 5 to 10 years is expected to be of little concern to the student.
While it is common that members of a martial family start learning martial arts at an early age, I don't think there is a compelling amount of evidence that proves that Taijiquan was
intended to be learned by rote from childhood. There are skilled Taijiquan practitioners who did not learn Taijiquan starting when they were children.
Overall the problem appears that it is very hard to teach the civil and martial sides of tai chi together, due to the massive amount of knowledge wrapped up in the circling alone (let alone other kinds of movement) and that due to the large amount of information packed into the art that way it is not possible to teach the shenfa if you pick out only the most effective moves from the form.
I think the problem is that being good at teaching is a skill unto itself: most teachers are not very good at teaching, regardless of how skilled and/or knowledgeable they are at their respective arts.