Testing the "activation reflects encoding transparency" principle in writing systems using repetition blindness.