This is a physics question, popularised by Richard Feynman that still has no agreed theoretical solution, but experimentally it is known what happens.
The problem is to take a sprinkler, put it underwater. If you blow water through it, it will turn around in the obvious direction. The problem is simple: if you suck the water through it which way does it turn?
One theory says that the water follows the same curve (albeit in the opposite direction) whether you suck or blow, so the sprinkler should turn the same way, as the momentum gives the same turning moment.
The next theory says that the water gets sucked in, so the sprinkler will be pulled along by it.
Another theory says that everything cancels, so it doesn't move.
Feynman's answer was that it does not move, and his impromptu experimental demonstration of that ended by getting everyone very wet; however it supported his conclusion before the glass broke.
More recent, sensitive experiments have shown that the sprinkler does in fact move as if it were sucked along by the nozzle. The effect is small though. No generally accepted explanation for this exists, although it doesn't seem to violate any known physics.