I build controllers -- as in, I buy blank circuit board, ICs, resistors, capacitors, diodes, and mosfets, and solder them into submission.
I can't necessarily speak for every controller design out there, but for mine ( http://www.glytch.net/~emolatur/electro ... esc24v.pdf
) you won't do any damage by attaching the outputs in parallel.
However, a PWM motor controller is a switching device -- it works by turning the motor on/off at a high speed (440Hz in the one I linked, 2.5kHz in the latest, higher-power version I'm working on) -- the problem with getting them to switch simultaneously. Just tying the throttle signals ("CONTROL_IN" in the diagram) together won't do it. If controller A and controller B are both at 50%, that does mean that each controller will turn the output "on" for 50% of the time... however, there's nothing enforcing that it has to be the same 50%. Controller A might turn the power on for half of the time and then controller B might turn it on for the other half, resulting in an output that is on all of the time -- 100% output. 0% is relatively guaranteed to be 0% (if nothing is broken) and 100% is relatively guaranteed to be 100% (same conditions) but all positions in between could very well be more than what you expect them to be.
If you're going to try paralleling controllers, and you expect them to behave predictably, keep the outputs seperate.
The other potential quirk is related to the current limit function. Everything has some tolerance - very few things are exact. Each controller I build ends up with a slightly different value shunt resistor. One adjusts the ratio of R12 and R13 to set what current the output is limited to. At takeoff, when the current draw is highest (or if you're going uphill), one controller might output slightly higher current than the other... so your vehicle might tend to steer left or right at that point. It is likely that any difference would be very tiny, but it's still food for thought - especially if you're ever going to try two different model controllers.