Going back to the Ark Royal example, if the ship could decelerate instantly to a standstill, the plane wouldn't loose any airspeed would it?
Try walking along a train carriage, against it's direction of travel (i.e. 'upwind'), at say 5mph ('airspeed'). Now ask the train driver to slam its brakes on. What will your trajectory be? Still think the plane won't lose any airspeed?
Islander 2's comment that pitch and bank indications also cycle during the turn is way off the mark.
I suppose the standard instrument texts (try Pallett:
Aircraft Instruments) devote several pages to the turning errors in gyroscopic attitude indicators for the hell of it, then!
