Cuban, I think you're thinking along the right lines. The only example of a long duct I can think of is on a NOTAR helicopter where the anti-torque is effected by ducting efflux from the engine to the end of the tail boom. I'm not sure of the exact physics behind the concept, but I reckon its along the lines of the harrier example.
As for 74's ideas on why the Kingair likes to roll over and play dead, its certainly not due to precession. In the stall case, its probably due to both props turning in the same direction (common on turboprops), causing a large torque (and also an asymmteric induced flow on the wing behind the engine) and therefore a large rolling moment. In the engine out case, you'll have less torque, but asymmetric thrust and induced flow over the wing on one side only, again, big rolling moment. This is one of the reasons why turboprops tend to have fairly sophisticated stall protection systems.
Regards,BGPM.
(PS, at any rate, if the engine did precess in response to a yawing or rolling torque, it would cause a pitch output)