My assumptions about each function were: 
1) tokenize: that tokenize would only split up words by
whitespace (tabs, spaces, and newlines), and not do anything
with punctuation. 
2) bestMatch: that if bestMatch were called with a correctly
spelled word, it would return a list containing that
correctly spelled word. Moreover that it would be
case-sensitive. 
3) addWord: that it would be case sensitive.