The fan curves that were posted while I was still writing this response are somewhat misleading. First off, they imply that there is maximum airflow with zero pressure. That is referring to the backpressure of the system, not the pressure differential across the blower. This is comparable to the resistance of a wire, where the lower the resistance,the higher the current flow. These graphs are not taking into account the pressure differential across the blower itself, which is what this discussion is about.
The point is that this curve might lead some to believe that maximum air flow will occur with zero pressure differential. Air cannot flow if there is no pressure differential. Just as electricity cannot flow without a voltage differential.
I'm not sure that's quite right - either the statement or the electrical analogy. The pressure shown in the curves is the static discharge pressure, which would be measured in the duct at the fan outlet. If it is reading zero, then the fan's operation is fully equivalent to a non-ducted fan. In other words, an open propeller in the room. Open propellers have no pressure differential at all, but certainly do work (in fact, requiring more power than if they had a pressure differential...). In fact, it is possible for a ducted fan to create a NEGATIVE duct pressure thanks to Bernoulli's principle - and yet it still does more work than if the discharge is blocked!
Air can flow in the absence of a pressure differential. Air has momentum, and once set in motion will continue to move, even without further pressure changes or against a pressure gradient. There is always a pressure differential across the aerodynamic elements (blades) of a fan, but not always between the inlet and outlet. It is the differential across the blades that causes air movement, but this does not imply that there must always be a pressure differential between the inlet and outlet. Clearly, there is none in an open propeller, nor is there any for a ducted fan with an expanding discharge duct. Consider the momentum of the air to be similar to inductance in a circuit. Current in a superconducting inductor will continue to flow in the absence of any voltage gradient. In fact, it will flow against a voltage gradient.
The other thing you will notice about the curves is that they never go down to zero airflow. The power required by the blower is dependent on both flow and pressure-differential. When the air flow is relatively large, the change in pressure differential is very small, and therefore, the less significant of the two components. But when the volume is very low, the pressure-differential component becomes more significant, because it changes more rapidly.
Within normal operating range, the blower works harder as the volume increases (due to a reduction in the backbpressure of the lines). However, below the normal operating range, you will see a slight increase in power required because the pressure increase. This is why there is a difference between blocking the inlet versus the outlet of a blower.
I would have attributed the lack of information at zero-flow to the problem of compressor stall, which, incidentally, reduces the load on the motor even more:
http://en.wikipedia.org/wiki/Compressor_stall