You got me there. Still not the reason why parser is better, though. The tokenizer converts strings literals to strings before the parser is involved, so the parser never even sees the escapes.