This depends on the design of the sensor.

Basic sensors effectively have one analogue to digital convertor (ADC) per column of pixels. That would mean that if you wanted to reduce the horizontal resolution, you would be disabling some ADC's, which in turn would reduce how many pixels can be converted per second, giving you no benefit.

No. The camera image is read row-wise. One could perhaps stitch together the output of multiple cameras and RPis though differing perspectives may make this impractical under some circumstances.

Yes, you can do 320x240 with modified raspiraw parameters. But you will get the same framerate as for 640x240, for v1 as well as for v2 camera.

You can do the opposite and increase horizonatal resolution for line scanner type of application with still impressive framerates:

"v2 camera can do 3280x32@720fps and 3280x64@500fps with frame skip rate less than 1% (tools attached)."

https://www.raspberrypi.org/forums/viewtopic.php?f=43&t=1091...

You can do quite impressive gobal external shutter capturings (with v1 camera only!), eg. capturing a 109m/s airgun pellet in flight on a single multiple exposure frame 10x: https://github.com/Hermann-SW/Raspberry_v1_camera_global_ext...