Your question details diverge into several separate, but related questions. I'd recommend breaking them up for more detailed discussions of each individually, but moving onward.

To the original question (as flavored by your details) I would say the analogy of an infant holds well. Therefore, think about the mix of self-guided and other-guided "tutorials" life gives us examples like:

• Manipulating and identifying objects.
• Basic ambulation, plus climbing.
• Testing one's physical limits; balance, grip, etc.
• Conducting dialog.
• Giving and taking an object.
• Playing with objects alone or with others.
• Falling, striking or being struck, and falling ill.
• Improving skills for precision and efficiency.

Obviously the list goes on ad infinitum, but you can see the relationship between early child development, and your question. My first two examples are too obvious; the third entails any game-style "platforming" elements, i.e., finesse in the controls.

Rich virtual worlds are (IMO) best taught with the utmost fundamentals at heart. I wouldn't worry about how relateable the virtuality experience is for non-techies, beyond tracking improvements in devices that reduce the learning curve for basic controls, which is research mostly driven by the video game industry (i.e. Xbox Kinect, Nintendo Wii-U) and, much more recently, the phone peripherals marketplace (check out Google Glass).

The question of rising from child to master in a virtual world is an expansive topic that barely exists outside the realm of science fiction, for now; simply because only few systems (e.g. Second Life, LambdaMOO) have enabled both gameplay, and hacker-esque control beyond the surface elements, in the same interface. It touches on education models (see LambdaMOO's checkered history for more on this) and revolves wholly around the model of the virtual world both internally, and as part of it's creator's goals, and their resources/willingness for enabling those goals (and discouraging others).