##Alien Reading: Text Mining, Language Standardization, and the Humanities

The aspects that I’m most interested in this article is the relationship between topic modeling and figurative language. Rhody argues that topic models of poetry should be read differently from those based on scientific journals for two reasons. One reason is that poetry uses larger amount of figurative language and produces more meaning than “non-figurative” texts. I personally agree with her argument and would take it as a foundation for my opinion. For poetic language, topic modeling implies a necessarily reductive methodology of sorting it into regular and stable categories, which is somehow at odds with the laws of their creation. As we all know, the ambiguity of figurative language is appreciated in poetry and many textual genres. I’m not sure whether it is really necessary to use topic modeling on those textual genres that values ambiguity as kernel. Or should tools and techniques be developed from topic modeling to target specifically the figurative language?