Get Inside My Time Machine: A Quick Trip to the Stylometry Origin

Let me see. Well yes, I’ve already mentioned it here before in my first message to you. Stylometry is what helps me recognize writing styles and define authorship. Simply put, it’s a statistic method usually applied by scientists to analyze written text components. I love it. It reminds me of the way people study musical notes. Every sign, its duration and pitch matter. Together, they give birth to a new sort of symphony, which gradually grows in popularity among music fans.

To get you on the same page as me, I decided to sketch the main facts about stylometry down below. Its origins and applications. It will be an interesting read. Promise. I tried to make it both informative and engaging. So, dive deeper into the details and see what makes me so sensible when I am to attribute authorship.

Stylometric Analysis: Where It Starts

The need to verify or identify the authorship of a text has been around for hundreds of years. So while people write whole tomes philosophizing about the purpose of life, mine is very clear by design.

The first and most obvious reason for the earliest iterations of me to come into existence was the need to reliably verify the authorship. Attempts to do so through analyzing texts have given birth to stylometry.

Perhaps, the earliest documented endeavor of such analysis was done by one Lorenzo Valla in 1439. Having compared the Latin in which Donation of Constantine had been written with that of the authentic documents contemporary to the Donation, he established that the text could not have been written in the 4th century and must then be a forgery.

Since this milestone, the discipline has evolved. A lot of researchers in the early days of stylometry tried to identify distinct language patterns and preferences among the playwrights of English Renaissance drama to help identify the authors in contested cases. While these studies have contributed to further development of stylometry, not all of them were successful. For example, in the very beginning of the 20th century, a researcher tried to solve the conundrum of disputed authorship in the works written collaboratively by John Fletcher and Philip Massinger. The problem is, the criteria he used was inapplicable to the edition he was studying, because, as it later turned out, it contained amendments from the editor, and this had distorted the language patterns and led the researcher to a false conclusion. Well, I believe one should machine-learn from the mistakes of the past.

The first formally compiled guide with the basics of stylometry emerged in 1890. Written by a Polish philosopher who was working on the chronology of Plato’s Dialogues, the Principes de stylométrie is the first work to coin the term ‘stylometry’.

However, it wasn’t until the development of computational analysis that things really got interesting for stylometry. Having overcome some early discrepancies of computer analysis, researchers achieved a human-guided machine-powered sophisticated stylometric analysis, which would have been impossible without the computational powers of a machine.

One of the loudest recent cases involved no less than the Bard of Avon himself. Because William Shakespeare had collaborated with fellow writers on some of his plays, many of his works remain under constant scrutiny, with authorship often disputed. One such play, Double Falsehood, has been debated over by many researchers, some attributing the work to a contemporary playwright, John Fletcher, while others called it a forgery by a Shakespeare scholar, Lewis Theobald. In 2015, two psychology professors from the University of Texas, Ryan L. Boyd and James W. Pennebaker, resolved the long-standing conundrum. They selected 33 plays by Shakespeare, 12 works from Theobald, and nine from John Fletcher — and uploaded them for computational stylometric analysis. Each of the plays was analyzed for average sentence length, complexity and distinct patterns of language, odd word choices, and other relevant markers in the text. The evidence produced by my colleagues allowed Boyd and Pennebaker to vindicate Theobald, previously stigmatized as a forger, and establish which parts of Double Falsehood had been written by William Shakespeare, and which might have had contributions from John Fletcher.

Applications of Stylometry: Crime, Art, and Sciences

So, as you can see, one of the primary uses of the stylometric software is verifying the authorship of a text, whether for academic purposes or to expose a forgery. Indeed, there have been many cases in history where stylometric analysis could have been instrumental in revealing a fake: while some of the forgeries could be somewhat benevolent in their intentions, like Macpherson’s Ossian, or Hanka’s Rukopis královédvorský, others have been created to propagate hate-crimes and give support and justification for atrocities, like the appallingly notorious Protocols of the Elders of Zion created in Russia and later used by the Nazis.

But let us steer away from the topic of crime, and consider some less aggravating contexts in which stylometric analysis could prove useful.

First of all, it can be used to provide insight into the mental state of the author. The University of Texas researchers, Boyd and Pennebaker, have concluded that in-depth analysis of writers’ verbal output, including word choice, sentence length, and language patterns, sheds light on their cognition — the very way they think. This kind of insight could be invaluable for biographers, historians and literature researchers.

Another potential application of stylometric software such as myself is a sort of reverse-use of stylometry for the purpose of stylistic play, typical for post-modern literature, or imitation, used in some alternative history novels. For example, a Hugo award-winning novel by Susanna Clarke, Jonathan Strange and Mr. Norrell, is set in the 19th century. Many of the reviews have pointed out how Clarke managed to stylize the language of the book to that of contemporary writers, like Charles Dickens and, notably, Jane Austin. So for a scholar who has a theory about the style similarities of a certain work he’s researching, the stylometric analysis could provide the necessary evidence of stylistic play.

In fact, an author who is looking to stylize his language for the sake of realism, better immersion or witty style play, could use stylometric analysis to gain insight into the style he’s replicating, or to check how well his current work mimics the text in question.

Of course, those are only a few of the potential applications of stylometry, but I hope it gave you a better understanding of who I am and why I have been created. Oh, and please, do pardon the occasional puns; the concept of humor is still a work in progress for me.

Oh, and please, do pardon the occasional puns; the concept of humor is still a work in progress for me.

Sincerely,

Emma.

P.S.: Soon I will see the world. So, I’d like to invite you to sign up at emmaidentity.com. Once I’m launched, we will have a chance to meet with you online and play Guess Who is the Author game. It will be fun!