Looking Behind the Text-To-Be-Seen: Analysing Twitter Bots as Electronic Literature

/ (1 of 3)


Master's Thesis in Visual Culture and Contemporary Art

Read the whole work


This thesis focuses on showing how Twitter bots can be analysed from the viewpoint of electronic literature (e-lit) and how the analysis differs from evaluating other works of e-lit. Although formal research on electronic literature goes back some decades, there is still not much research discussing bots in particular. By examining historical and contemporary textual generators, seminal theories on reading and writing e-lit and botmakers’ practical notes about their craft, this study attempts to build an understanding of the process of creating a bot and the essential characteristics related to different kinds of bots.

What makes the analysis of bots different from other textual generators is that the source code, which many theorists consider key in understanding works of e-lit, is rarely available for reading. This thesis proposes an alternative method for analysing bots, a framework for reverse-engineering the bot’s text generation procedures. By comparing the bot’s updates with one another, it is possible to notice the formulas and words repeated by the bot in order to better understand the authorial choices made in its design. The framework takes into account the special characteristics of different kinds of bots, focusing on grammar-based bots, which utilise fill-in-the-blank-type sentence structures to generate texts, and list-based bots, which methodically progress through large databases.

From a survey of contemporary bots and earlier works of electronic and procedural literature, it becomes evident that understanding programming code is not essential for either analysing or creating bots: it is more important to understand the mechanisms of combinatory text generation and the author’s role in writing and curating the materials used. Bots and text generators also often raise questions of authorship. However, a review of their creation process makes it clear that human creativity is essential for the production of computer-generated texts. With bots, the writing of texts turns into a second-order creation, the writing of word lists, templates and rules, to generate the text-to-be-seen, the output for the reader to encounter.