New statistical methods shed light on medieval literary mystery
Dramatic battles, courtly intrigues and love stories - no, we are not referring to the latest fantasy novel, but to the chivalry romance Tirant lo Blanch.
Written in Catalonia in the 1460s, it is considered by many to be the world's first novel (150 years older than Don Quixote). The book is also famous for the debate around its authorship. Joanot Martorell and Martí Joan de Galba were both real-life knights and also both the authors of Tirant. Martorell died and his friend de Galba finished the book. The mystery concerns where the change-of-author took place.
The fall of Constantinople
Tirant lo Blanch translates to Tirant the White; he is an English knight and the novel's main character. The authors claim to tell a true story, but none of the characters are historical persons and the plot is not historically correct either. The story can rather be considered a sort of alternate history, where the authors wrote history as they wanted it to be. In reality, Constantinople was invaded by Ottoman Turks in 1453, and was never again held by Christian rulers. Contrarily, in the book, the city is saved by Tirant; and as in a classic fairytale, he is rewarded with the princess' hand in marriage. However, he, rather originally, dies of sickness before the wedding.
Measuring literary style
In order to analyze the book with statistical methods one has to measure something. We thus had to define some measurable quantities that could (hopefully) be used to differentiate between the two authors. Here we have chosen to look at the writing style, and more specifically at the word lengths in each chapter.
The book is long, with 487 chapters of varying length, and using the proportion of words of different length in each chapter proved to be fruitful. This means that we collected 10 numbers for each chapter: the proportion of one-letter words, the proportion of two-letter words, .. and so on, until the proportion of words with 10 or more letters.
Interested in other types of medieval culture? The death rate in Game of Thrones is realistic.
The most likely chapter
Equipped with relevant data, we could now start the analysis.
We wanted to obtain an estimate of the most likely change-of-author chapter, but also an idea of the uncertainty around this estimate. Working with certain statistical models and using one of the methods presented in our article, we obtained the following figure:
The figure, which summarizes the whole result as a so-called confidence curve, shows the change-of-author estimate and the uncertainty around it. On the horizontal axis we see a subset of the chapters of the book. We see that the figure "points" towards chapter 371, meaning that the second author, de Galba, most likely took over the writing around this chapter (according to our method).
Simultaneously, the figure conveys that other chapters are candidates for the change-of-author point, especially chapters directly before 371 and around chapter 345. This displays the uncertainty in the analysis, which actually is quite small (remember that there are 487 chapters in total).
Naturally, the conclusion above is dependent on how well word lengths actually reflect the authors' literary styles. We have done supplementary analysis, based on other measurable quantities which have given somewhat different results, for example based on sentence lengths (check the article for more information on this). Still, these other analyses point to roughly the same place in the chapter list.
Also note that de Galba is unlikely to have taken over the writing at exactly one point. It is more likely that he had to fill in some text at different places in the book.
We claim nonetheless that our analysis is convincing and should interest literary scholars that investigate Tirant lo Blanch. Finally, we hope this work may inspire more quantitative analyses in the humanities!
More from Titan in English:
Pregnant women taking chances with herbal remedies
Saving our Nordic languages in the computer age
FNs klimapanel peker på tre teknologier som kan bidra til å løse de globale klimaproblemene. – Her i Norge driver vi med «cherrypicking», sier kjernefysiker Sunniva Rose. – Det skal liksom ikke snakkes om kjernekraft. Hun gjør det likevel.