Do we know how people identify spoken words