Experiment with corpus of 1880s children's fiction
This is a quick experiment in poetry generation with one of the test corpora available in Alan Liu's DH Toychest. It was created for a meeting of the Digital Approaches Reading Group at WashU on 19 April 2016.
I make use of a few NLP-related Python tools including TextBlob and Allison Parrish's pronouncing. The tool creates syllabic verse rather than the more typical accented-syllabic style in English (think iambic pentameter).
Syllabic verse was used by 20th-century poets including Marianne Moore, Dylan Thomas, and Thomas Nashe. It is therefore quite a bit distant from the more verbose style of 1880s prose writing for children. But the different number of short sentences in each text (as reflected in the relative lengths of the poems) calls attention to the styles of the different novels, and often this small snapshot gives insight into the tone, characters, mood, and even plot of each text.