Beyond lexical frequencies: using R for text analysis in the digital humanities