On the language of physical science