Digital platforms have turned stance into a polarizing practice that goes beyond normal stancetaking.

By: Will Culkin
Course: Language and Digital Media (Ling 3800-800)
Advisor: Prof Kira Hall
LURA 2020

Sociolinguists use the term “stancetaking” when describing how speakers evaluate objects intellectually and aesthetically. One of the leading linguists in this area, John W. Du Bois (2007), has proposed that each instance of stancetaking involves a “stance triangle” in which two speakers create intersubjectivity through their evaluation of an object. As seen in the figure below, taken from Du Bois’s influential article, Subject 1 evaluates an Object and thereby positions themselves as taking a particular kind of stance. Subject 2 then aligns to that positioning by also evaluating the Object, positioning themselves towards the Object in relation to Subject 1’s positioning. However, Du Bois’s stance triangle was originally formulated for in-person, small-scale, single interactions. As digital factors change the interactional landscape, there is a need to reevaluate the stance triangle to account for digital communication. This was the topic of my investigation for the course Language and Digital Media.

As a case study, I worked with Facebook’s recommendation system - identified as the “controller” - to determine if it intensified stancetaking by creating a snowball effect to control the object being evaluated.  I argue that the effects of the recommendation system and its publicity of other people’s stances create a new type of stancetaking: superstance. To test my superstance hypothesis, I applied the theory to Facebook, allowing account holders to be users (or “subjects,” in Du Bois’s terms), content to be objects, and the algorithm to act as the controller. In my small-scale experiment, I created a new Facebook account in which I took stances on particular kinds of content by liking certain items and following certain accounts. I quickly saw the controller in action. The algorithm evaluated my actions and then positioned itself towards me by controlling what types of objects (news events covered by a conservative commentator, in this case) populated my feed. In other words, each mini-stance taken by the user toward social media content gets recorded by the controller. This lets the controller then show objects that it has decided a user will like, in order to encourage the user to spend more time on the site. However, this phenomenon creates an adverse effect as it pulls users farther apart, confirms biases, and doesn’t allow other voices to be heard. By compounding upon the user’s initial stance, the algorithm proliferates the occurrence of ministances and creates a loop of stancetaking which echoes and enhances upon itself. This echo chamber is the basis of the superstance. 

After establishing the existence of the algorithm’s superstance, I focused on the specific mechanisms at play. The narrow and simplistic design of digital media “reactions” are the essence of stancetaking in the modern age. The sheer access of objects and users is magnified to a global stage, and social norms are inhibited by disassociation from the interaction altogether. Furthermore, social media has created a new standard of communication so that prespecified reactions are simple and encouraged. In this way, digital media has predetermined a small repertoire of reactions for users to then select between. This further accelerates the speed of stancetaking as a simple click that “likes,” “dislikes,” “hearts,” or “cries” over an object, limiting the majority of stances to this specific menu. This systematic online stancetaking is a highly strategic and intentional element of digital media.  Each stance that is taken on social media sites like Facebook enables the site to position themselves towards the user in a manner they hope will get the user to engage more with the platform. The bottom-line goal of these sites is to draw users in, and user engagement is a pinnacle metric of success in the industry. While the foundational thinking in creating social media may not have been to support this superstance vortex, it is undoubtedly an effect. 

The polarization created by superstances, known to users as groups or curated feeds, would never have reached the divisive level that characterizes today’s climate if not for the algorithm acting as the controller. The mini-stances that one might recognize from Du Bois’s original stance triangle are empowered and augmented; in doing so, social dissent, division, and disorder is exacerbated as well. 

Bibliography:
Du Bois, John W. “The Stance Triangle.” Stancetaking in Discourse, 2007, pp. 139–182., doi:10.1075/pbns.164.07du. 
Thurlow, Crispin, and Kristine R. Mroczek. Digital Discourse Language in the New Media. Oxford University Press, 2011.