The key is to follow the prop response to torque from the motor (be it turbine or piston).
Constant speed propeller systems regulate a positive torque output from the motor and convert that torque into a regulated speed. Any increase in torque results in an increase in thrust by virtue or regulating a specified rpm limit with the prop control.
In your scenerio an engine is now attempting to regulate a negative torque (not a reduction in torque). In the case of a negative torque the engines prop control responds by attempting to increase prop rpm. HOWEVER the attempt to increase rpm is in relation to a positive torque applicaiton, since you now have a negative torque deviation the actual result is a reduction in rpm since there is now a negative torque being applied and the blades are commanded to a flatter pitch rather than a higher pitch which would be required to increase rpms...
Since your scenerio has resulted in a negative torque deviation, the prop (by virute of its design to regulate positive torque) will attempt to increase rpm by commanding a reduction in blade angle in relation to the relative wind. This reduction in blade angle will therefore reduce engine rpm with a significant increase in drag. In a negative torque situation the blade angle will always be the inverse of what is required to actually increase the rotational speed again because the system is simply designed to respond to torque from the engine.
In cruise, if you were to reduce the torque or in this case the manifold pressure prop rpms will rise despite the response to a reduction in torque, however it is a REDUCTION in positive torque being applied. As the MP is reduced the blade pitch will gradually return to a flatter pitch to compensate for the reduction in available torque. Eventually and as somebody has already mentioned the blades mechanical limits will be reached and the props ability to compensate for rotational rpm will result in a gradual reduction in engine rpm. If airspeed is slow enough a windmilling prop will cease to rotate, however the extreme rise in drag as a result of the flat blade pitch will more than likely cause a loss of directional control.
Because of the requirement to regulate positive torque, multi engine prop systems have to be fitted with a feather position that permits an expanded range of movement relative to blade angle past a flat pitch detent. This allows the prop to move from the flat blade pitch detent to a pitch that follows a rotational direction that reflects a negative torque angle until the blade is positioned in a low drag position relative to the airplanes relative wind.
A feathering mechanism requires a spring that forces the blades into the negative torque position, again there exists no ability for the hydraulic forces to act negatively upon the props internal mechanisms. The springs will force the prop beyond the limits of the hydraulic system thus feathering the prop and greatly reducing the resulting drag.