Computational Linguistics 6, Exercise 3: two trees, two meanings

The chapter's output:

1. [S [NP i] [VP saw [NP [NP the man] [PP with [NP the telescope]]]]]
2. [S [NP i] [VP [VP saw [NP the man]] [PP with [NP the telescope]]]]

Meanings:
1. "with the telescope" is inside the NP "the man with the telescope": I saw a
   man who had a telescope.
2. "with the telescope" attaches to the VP "saw the man": I used a telescope to
   see the man.

What a program would need to prefer one: information outside this grammar.
Possible sources, as my suggestions:
- context (the previous sentences: was a telescope mentioned, who has one?);
- world knowledge (people use telescopes to see things);
- statistics from parsed corpora about which attachment is more common with
  these words (the Parsing article says most modern parsers are at least
  partly statistical and learn from hand-parsed corpora).

The Syntactic ambiguity article calls choosing the intended structure
"syntactic disambiguation". Which preference the statistics would give for
this sentence I have not tested, so I do not claim one.