Can Natural Language Processing Help Identify the Author(s) of the Book of Isaiah?

Many historians believe that the Biblical book of Isaiah was written by two authors approximately two hundred years apart, generally called First Isaiah and Second Isaiah. Some even believe that the second part was itself written by two or more authors. In this paper we use natural language processing techniques to study this hypothesis. We used the Stanford parser to parse the book of Isaiah. Using Student’s t and two measures of text complexity, average sentence length and average tree height, we were able to differentiate the second part of Second Isaiah, commonly called Third Isaiah, from the rest of the book. We then used MALLET’s implementation of LDA to identify ten topics in the book. Using ANOVA, we were able to find two topics that could differentiate selected parts of Isaiah. We then successfully used MALLET's implementation of the Naive Bayes algorithm to find differences between First Isaiah and Second Isaiah and also to differentiate the two parts of Second Isaiah. Finally, we showed that the same technique could be used to easily differentiate Isaiah from another prophetic book of the Bible, I Samuel.