ON THE INTERPRETATION OF SARIN AND WAKKER'S "A SIMPLE AXIOMATIZATION OF NONADDITIVE EXPECTED UTILITY"