Tip:
Highlight text to annotate it
X
Here we see the answers, and although neither of these sentences or the parses of these sentences are very good--
"raises raises interest rates" is an unusual sentence--it does seem clear that the parse tree on the left is better than the one on the right.
"Raises interest rates" is a very good verb phrase, and it's just "raises" is an unusual noun phrase for that.
And so it seems proper that the probability here of .012% is 16 times higher than the probability
of this more unusual tree. Thus, it does seem like this idea of a probabilistic, context-free grammar is capturing something
about what makes one parse tree better than another parse tree.