Attacking a sentence

[devilsadvocate]

The answer to your question might be better source material, i.e. books.

I just posted a related topic on this.

https://hashcat.net/forum/thread-6415.html

Thank you.

My current solution is using a combinator attack using n-grams (http://www.ngrams.info) on the left and google-top-10000 on the right. This seems to do a pretty good job of creating sentences with meaningfull content.

These could also be a starting point.

http://www.lyrics.com
http://www.abbreviations.com
http://www.biographies.net
http://www.convert.net
http://www.definitions.net
http://www.grammar.com
http://www.lyrics.com
http://www.phrases.net
http://www.poetry.net
http://www.quotes.net
http://www.references.net
http://www.rhymes.net
http://www.scripts.com
http://www.symbols.com
http://www.synonyms.net
http://www.uszip.com

When it comes to using wget, spider smartly and responsibly.  How you parse the data is entirely up to you.

Cheers.