Turbo, I follow your argument which in some ways is similar to the views in –
Why "Human Error" Is A Meaningless Concept .
However, in reality, the industry is ‘stuck’ with TEM and thus there is a need to explain what and/or how this activity (TEM) is to be implemented.
The problem is akin to using the DECIDE model for decision making (a good representation of the process), but apparently not at all how humans make decisions. Here, one solution is to generate the knowledge and experience required for situation based (naturalistic) decision making.
Similarly, we have a TEM model which depicts a process, but does little to help manage whatever our notion of threat or error is.
I like your ideas where after an action, if the outcome is not as required (towards the goal), then the assessed situation is used as a new starting point. Is this management or mitigation?
In some ways, this turns the TEM process into an iteration of situation assessment and decision making, which appears to be the basis of most activities.
But if you wish to stretch you mind then back to Woods & Cook in #3 or try
PRIME (about page 136).