Comment by jvanderbot

So the question is one of modularity. Having one sound for each such phrase doesn't scale.