Well, if conjecture is up on the controllers, I would go with radio difficulty, rather than "letting them go". A break in procedure is unacceptable, or certainly should be. Report TASIL hadn't come up, 447 was, what? Upset? trying to navigate out of the cell? No attempt to follow procedure is too lax to entertain, No? The radio inop/switching system may have played a role in lack of comm? Occam would say that the a/c was certainly in trouble on or about the comm failure, so the onus is on the ground procedures, which is handy, we have that data, or should. BEA does, right?
bear