Stylometry is not an exact science. As in traditional literary analysis, its aim is to help a scholar arrive at an informed perspective rather than a definitive answer. As when conducting purely qualitative research, it is difficult to know when a conclusion is the most informed one possible, or if a more illuminating result exists. We began our project knowing that these authors’ work were highly collaborative, and that due to limitations we will discuss below, we were working with an extremely restricted corpus. We knew that these two factors would make any analysis predicated upon such authors and such a corpus tenuous. Our aim in undertaking this research was thus primarily to determine whether we could bring any new evidence or insight to the existing uncertainty over the nature of the Stevensons’ collaboration on The Dynamiter, rather than resolve that question conclusively.
In the hopes of being precise and transparent about our limitations, below are the three most significant barriers we faced. Some we knew about before beginning our project, and some we discovered along the way. We hope this list will be useful to future scholars expanding on our work.
- Fanny and Robert Louis Stevenson are known to have had a fluid and collaborative writing relationship. They often read edited, and commented on each other’s work, and seemed sometimes to care more for their collective body of work than their individual legacy — for instance, when Robert Louis was too ill to write his own stories on a sea voyage to Auckland, he still “…corrected page proofs for Fanny in the afternoon.”1 The couple seemed to either relish in or be oblivious to this ambiguity — they left no (intentional) trace of their division of labor in authoring The Dynamiter.
- As discussed in the notes on our corpora, Fanny’s corpus had to be tailored to accommodate their differing sizes and content. Fanny’s published work was limited to ten short stories, as compared to the many novels, short stories, poems, plays, and travel essays from in compiling our corpus for her husband. Further complicating the comparative possibilities of the two corpuses is the fact that short story is the only genre in which Fanny wrote, whereas Robert Louis published text in almost every genre. We therefore limited ourselves to prose in an attempt to construct a roughly even comparative corpora.
- Our sample sizes of text were very small. It likely would have been easier to determine who wrote The Dynamiter than it was to attempt ascribing an author to each chapter. The more words there are to analyze and compare, the easier it is to locate interesting patterns. The Dynamiter is 75,000 words long, but each of its fourteen chapters is a fraction of that.