Run through a web app that removes word junk.

docToHTML.sh

#!/bin/sh
input="$1"

# %%.* strips off everything after the first . in this string
filename="${input%%.*}"

textutil -convert html "$input" -output "$filename".html
curl -d html="$filename".html http://wordoff.org/api/clean
pandoc -f html -t markdown "$filename".html -o "$filename".markdown