Twitch.TV and Parasocial Interaction: Understanding Twitch's Social Features Within a Parasocial Framework

(1)

Twitch.TV and Parasocial Interaction:

Understanding Twitch’s Social Features Within

a Parasocial Framework

Maren Zeinstra June 26th_{, 2017} Supervisor: Dhr. Dr. Markus Stauff Second reader: Dhr. Dr. Toni Pape Television and Cross-Media Culture

(2)

2

3 1. Introduction

1.1 Introduction

Twitchis an online platform founded in 2011 (https://twitch.tv) where users live-stream themselves while playing a video game, displaying gameplay andincluding webcam and microphoneto communicate with their viewers (image 1.1). According to Twitch, the community consists of more than 2 million live-streamers and close to 10 million visitors per day (About Twitch).Besides individual users, organizations broadcast their eSports tournaments where gamers and teams compete for prizes (Edge 34-5). Twitch additionally offers features for social contact between users by following, subscribing, adding friends and sending personal messages. Additionally, all channelshave a chat room right next to the live-stream where userscan communicate in real time (image 1.1). Twitch users can select streams from the list of channels they “follow” at the left side of the screen, or they can browse through all live channels, of both individual streamers and organized tournaments, filtered per game, community or popularity (image 1.2). Besides the heterogeneous gaming content, Twitch introduced a “Creative” section in October 2015, where users stream themselves while cooking, painting, designing, making music, et cetera. (Moorier)(image 1.2). Other than live-streaming content, users can watch previous broadcasts on demand. When doing sousers see a “chat replay”, showing the chat room as it was during the live broadcast, thereby effacing possibilities to participate in real time chat rooms.

Image 1.1 Screenshot, 20 June 2017, <http://twitch.tv/grimmmz>.

Image 1.2 Screenshot, 20 June 2017, <http://twitch.tv>.

Twitch is an especially interesting part of our current media environment, as it integrates multiple media forms into one hybrid platform. Just like traditional broadcasting on television, Twitch broadcasting is based on one-to-many communication, but combined with features that allow for

(4)

4

many-to-many communication, similar to those found on social media platforms. These social features are what distinguishes the platform from otherbroadcast media, as is also argued by Sjöblom and Hamari who state Twitch might be considered “yet another form of broadcast entertainment akin to online videos, but for many users it is a more manifold and holistic communication channel than mere video media content, particularly due to the high levels of interaction” (Sjöblom and Hamari 1). Many television broadcasters now also include interactive, online features, allowing for social television where users engage more actively with television content, but Twitch has successfully integrated direct real time communication in live broadcast settings (Wohn and Na). Twitch additionally shows similarities to YouTube, as both platforms allow individuals to film and broadcast themselves towards an audience that can comment on it. The difference with YouTube lies partially in the addressing of a niche by Twitch, focusing on the specific subculture of gamers who beforehand could only communicate within games or on text-based fora. YouTube has recently experienced a growth of game-related videos too, but Twitch is specifically designed for this subculture and its content. Additionally, while live-streaming is optional on YouTube, it is the norm on Twitch which increases immediacy.This emphasis on liveness connects back to television, as liveness is often considered the main mode of television too (White 75-6). A third media form appears as Twitch integrates video games into its platform, making them the content to be consumed. While video games are often considered an individual experience, Twitch uses exactly this solitary content to connect users in a social way (Scully-Blaker et al. 2026).

Twitch is thus a unique and hybrid part of our current media environment,showing a combination of live broadcast television, video games and social media platforms. Live, linear broadcasting is combined with the social features of social media and individual interactivity of video games.This combination of different media forms brings with it certain tensions as it affords a great variety of usages and allows for a strange combination of social and non-social behaviours. Twitch has not stopped developing and is constantly integrating new interfaces, applications, browsing options and social features (https://blog.twitch.tv/tagged/new-features). Although it is impossible for researchers to keep up with the continuous developments, these new features are all interesting as they display what elements Twitch considersto need improvement or enhancement. Here, it also becomes visible how Twitch emphasizes its own social features and attempts to enhance the social, communal experience of its users. Its continuous growth over the past six years prove the platform and its community are here to stay, showing Twitch should be taken seriously and is deserving of academic attention. Analyzing Twitch can help understand how there are multiple ways for being social online, even within one hybrid platform, and thereby is an interesting case-study for understanding our entire current media environment.

1.2 Discourse on Twitch

Over the past years, Twitch became object of research in a variety of academic fields. A majority of research emerges from computational, technological and mathematical fields. In 2014, Karine Pires and Gwendal Simon proposed implementing new streaming technologies in order to improve user experiences and Albert Haque designed algorithms to detect spam and trolls. These types of researches are focused on technology, designing specific codes or approaching the technical infrastructure of the platform. A second major trend of research focuses on motivations for using Twitch. These are especially interesting as there seems to be no clear consensus on the main motivation, and contradictions appear in the ways researchers approach the platform. Researches in this field are often based on uses and gratifications theories, assuming users are active agents who choose their media content according to their current needs. In 2016 Sjöblom and Hamari employ this theory to explore the motivations of Twitch users in their article “Why Do People Watch Others Play Video Games?”, where they relate escapism, acquiring knowledge, aesthetic pleasure, personal confidence and social motivationsto Twitch usage (6-8). Interestingly, they conclude subscription behaviour is positively correlated to only social motivations (Sjöblom and Hamari 8). This connects to their previous argument Twitch is distinctive especially due to its social features and possibilities for establishing social connections (1). Their results might answer a question raised in “Behind the Game: Exploring the Twitch Streaming Platform” by Deng et al. in

(5)

5

2015. They explore how users are distributed across different streams, arguing no stream contains unique content as “the same games are streamed by many different people”, which raises questions on how viewers differentiate and choose between channels (Deng et al. 1). Sjöblom and Hamari’s results suggest that during these selection processes, users pick streamers that are socially attractive to them. Hamilton, Garretson and Kerne also emphasize the social aspects of Twitch in 2014 in “Streaming on Twitch: Fostering Participatory Communities of Play within Live Mixed Media” where they “describe processes through which stream communities form”(1315). They argue social interaction with likeminded individuals and the creation of “participatory online communities” is the main motivation for Twitch users (1319). They continue to argue “streams that draw more than 1.000 viewers” are too massive to gratify those social needs, as conversation breaks down and “meaningful social engagement” is lost (1318, 1320). The chat room becomes “an illegible waterfall of text” in these massive settings due to the great amount of messages, making meaningful interaction impossible (Hamilton, Garretson and Kerne 1321). These abovementioned researchers share an emphasis on Twitch’s social features. However, a second trend of researches shows a different approach. Katyoue et al. investigate Twitch usage in 2012, in “Watch me Playing, I am a Professional: a First Study on Video Game Live Streaming”, showing those massive streams account for 95% of the views (1183). Seemingly, viewers are drawn to these huge streams and do not necessarily switch to smaller streams, despite the alleged loss of meaningful individual interaction. Similarly, Deng et al. suggest viewer distribution on Twitch is extremely skewed, as many streams have no views, while popular streamers and professional events “dominate the views”, forming “predictable flash crowds” (6). This sustains the argument that massive streams form a majority of Twitch usage and are not as problematic for Twitch users as Hamilton, Garretson and Kerne make it seem. Additionally, Georgen, Duncan and Cook approach Twitch in a framework of “participatory spectatorship”, rather than the participatory community as described by Hamilton, Garretson and Kerne (581). Georgen, Duncan and Cook argue a “silent majority” constitutes the biggest part of viewers, referring to the great amount of passive users (581). They characterize participatory spectatorship as “the active observation of a sport or spectacle in the pursuit knowledge”, emphasizing the possibilities for “lurkers” to learn new skills through observation without necessarily actively participating in the chat room (581-2).An interesting tension surrounding the variety of Twitch usages and affordances appears, with emphasis on social behaviours versus emphasis on ‘anti’-social behaviours in passivity and massiveness. My analysis attempts to add a new framework for understanding Twitch, combining Twitch’s social features with these apparent anti-social behaviours, thereby not attempting to immediately solve tensions, but instead offer new tools or concepts for a broader understanding. 1.3 Research Question

A strand of researches emphasizes the importance of social motivations for using Twitch, whereas a second trend shows a more critical approach is needed as the majority of actual usage, both passive and massive,deviates from this social ideal. In order to explore this conflict, this research explores Twitch’s interfaces, content and behaviours using “parasocial interaction”. This concept, introduced by Horton and Wohl in 1956, constitutes the idea users can experience a “simulacrum of conversational give and take” (Horton and Wohl 374). This means, despite the lack of a genuine interactive, one-on-one reciprocity, users in some cases experience the illusion of partaking in social interaction, thereby gratifying social needs. Here I am inspired by Tsiotsou, who researches parasocial interaction on social media platforms and thereby shows parasocial and social can co-exist and be intertwined on these platforms (404). A second source of inspiration was offered by Hartmann and Goldhoorn who argue parasocial interaction and enjoyment are positively correlated, leading me to believe the parasocial, and perhaps its connection to the social, might help explain Twitch usages as social needs can be gratified through seemingly ‘anti’-social behaviours (1109-10).

The research question that will guide this analysis is: in what ways does the parasocial appear, develop and connect in relation to Twitch’s social features?And, additionally, can this parasocial framework help to explain or understand Twitch in a new light?To explore these questions, an analysis of the interfaces and textual features, and the ways users respond to them, will be

(6)

6

performed. This analysis will show whether Twitch can be considered to afford parasocial interaction with Twitch performers, “personae”, and whether this creates new understandings of both social and ‘anti’-social behaviours. I believe there is a constant mix of social and parasocial,true interaction and a simulacrum of interaction, where the balance is dynamic and constantly shifting due to the complicated hybridity of the platform. While many researches thus attempt to use the social and the parasocial as fixed descriptive features, I believe Twitch is a perfect platform for analyzing how these two can converge and become combined within one platform, where social features play a role in establishing a simulacrum of the social. This analysis does not wish to invalidate individual interactions and relationships that might occur on Twitch. However, this research focuses on the users that do not engage with these behaviours, as they form a majority of Twitch usage. I am not attempting empirical research and I do not wish to devaluate previous researches nor will I attempt to resolve the question of why people use Twitch. Rather, I hope my research adds new frameworks and tools for understanding Twitch, thereby adding to discourses on Twitch, parasocial interaction and contribute to the entire current media environment within Twitch takes up a notable position.

1.4 Theoretical Framework

In order to adequately structure my analysis, a more detailed exploration of the theoretical framework is necessary. In 1956 Horton and Wohl introduce parasocial interaction as the special bonds and relationships audiences develop with television performers, or “personae” (374). They argue via the illusion of a face-to-face connection, a sense of intimacy is created; viewers feel like they are addressed individually and become part of a “simulacrum of conversational give and take” (Horton and Wohl 374). This “intimacy at distance” is stimulated by informal, conversational styles and direct address (Horton and Wohl 374). Despite the fact this intimacy is one-sided, through continued watching viewers might feel like they truly know this person and start developing a relationship with similarities to interpersonal relationships (375). When the significance of the parasocial relationship grows, the media user will attempt to engage more often or more intensely with the persona which is how the relationship continues to develop.

Since its introduction, researchers have attempted to create a reliable and valid scale for measuring parasocial interaction. While Horton and Wohl delineate an interesting concept, their theory is not very precise and lacks an accurate approach. In order to productively employ their ideas, other researches that deduce and analyze more precise features of parasocial interaction need to be added. Alan M. Rubin is a notable researcher who has written multiple articles on the parasocial with a variety of colleagues and developed a 20-item scale for measuring parasocial interaction in 1985 (Rubin, Perse and Powell). In “Loneliness, Parasocial Interaction, and Local Television News Viewing”, Rubin, Perse and Powell attempt to find correlations between loneliness and parasocial interaction with television personae by developing a parasocial interaction scale (PSI-scale) based on a theoretical approach. The scale consists of 20 items on identification, affinity and attraction towards personae, like “I feel sorry for my favorite newscaster when [they make] a mistake”, “I look forward to watching my newscaster on tonight’s news”, “I like to compare my ideas with what my favorite newscaster says” (Rubin, Perse and Powell 167). This first PSI-scale is used and elaborated on by many subsequent researches.

Auter and Palmgreen explain in 2000 in “Development and Validation of a Parasocial Interaction Measure: The Audience‐Persona Interaction Scale” why they have significant problems with the PSI-scale. Firstly, they argue the scale is merely theoretical as it does “not have [its] origins in open-ended qualitative viewer surveys” (80). Secondly, the scale is “univariate”, as it only assesses “the individual’s identification with their favorite character – disregarding related, yet important concepts” (80). They develop a new scale based on television personae, the audience-persona interaction (API) scale, adding the audience-persona’s problem-solving abilities and group interaction and identification to the equation (82-3). While Auter and Palmgreen focus on television personae, these two additions are relevant in video games too, as problem-solving abilities are reflected as skills within the video game, where users might wish they “could handle problems as well” or “would like to be more like [them]” (Auter and Palmgreen 82-3). Their items for group interaction concern “[their] interactions are similar to mine with friends”, “my friends are

(7)

7

like [them]”, “I felt included in the group”. These items become relevant for Twitch,as users can observe group interactions between various users, affect the bond they feel towards them (82-3). David Giles wishes to develop an even more extensive model for analyzing parasocial interaction in 2002, combining the PSI-scale and API-scale and adding more elements. Firstly, David Giles explicitly introduces a “continuum of social-parasocial encounters” (293). Rubin, Perse and Powell tend to view parasocial and social as opposites, where an individual engages in either social or parasocial behaviour, and parasocial might compensate for a lack of social contacts (159). Giles, on the other hand, argues singular interactions can be placed on a continuum ranging from social to parasocial via a four item scale consisting of “(a) number of persons involved”, “(b) physical distance between interactants”, “(c) social conventions”, “(d) potential relationship between the interactants” (Giles 293). A continuum is more useful than a dichotomy when analyzing the dynamics between social and parasocial, like this research attempts to do. A second useful addition by Giles is his analysis of relationship development. While there are many similarities to Rubin and his colleagues, focusing on developing attraction, affinity and identification, Giles is the first researcher to explore the affect of the context of viewing. This context of viewing consists of aspects such as co-viewing, discussing the persona with others, and other encounters with the persona (Giles 297). While Giles does not apply his approach to online settings, as he too merely addresses television personae, his additions are especially relevant for Twitch. Twitch personae are watched concurrently with other viewers, leading to co-viewing and discussion in the chat room, making this a relevant addition. Other than that, streamers often have social media accounts, leading to other encounters with them. This online contextual setting of the encounter might affect the nature of the relationship and will thus be taken into account.

In 2011 Hartmann and Goldhoorn argue that too many researchers overemphasize developmental features and conceptualize the parasocial as “an enduring relationship (i.e., friendship)” (1104). Hartmann and Goldhoorn develop their own scale, the Experience of Parasocial Interaction (EPSI) Scale, focusing on one singular media encounter and its textual features instead. They emphasize the experience of parasocial signs, such as the addressing style of the performer, the perceived attractiveness of the character, and the perspective-taking abilities of users (1107-9). According to them, parasocial interaction is rooted in an impression of mutual awareness, rather than an ongoing development over time (Hartmann and Goldhoorn 1107). While I respect their attempt to focus on essential textual features, they neglect developmental features too much. Horton and Wohl emphasize predictability of personae is crucial for sustaining parasocial interaction, as personae need to offer “a continuing relationship”, becoming “a regular and dependable event” (375). The continuity of personae can thus not be ignored. Additionally, Rubin and McHugh show attraction develops along with exposure, meaning attraction cannot be isolated to one encounter (283). Additionally, the online contexts cannot be neglected as contexts affect and change the relationship. I do wish to take into account their critique as a reminder that not all encounters lend themselves to parasocial communications, while additionally using their precise specification of direct address.

All of the abovementioned researches solely analyzed traditional television performers. There are few predecessors when it comes to applying parasocial interaction to online media or attempting to develop a valid scale for doing so. Most of the researchers that do, are primarily interested in the promotional and marketing values of online parasocial interaction, like the chapter written by Hoerner in 1999 in Advertising and the World Wide Web, and the article written by Thorson and Rodgers in 2006 on websites used for political campaigns. Both emphasize the value of generating a parasocial feel to websites in order to be more attractive or valuable to visitors. In “The Role of Social and Parasocial Relationships on Social Networking Sites Loyalty”, 2015, Rodoula Tsiotsou similarly argues employing parasocial strategies stimulates platform loyalty for social networking sites (410). In “Parasocial Interactions and Paracommunication with New Media Characters”, 2008, Tilo Hartmann applies parasocial interaction to a new media context in a very productive and useful way. Whereas traditional parasocial interaction only applied to non-interactive, non-reciprocal, asymmetrical situations with authentic characters, Hartmann explores how “paracommunication” can also occur in interactive, reciprocal settings (Hartmann 193). Just

(8)

8

like traditional mass media, new media environments can “create an illusion of mutual awareness and symbolic behavior[sic]” (Hartmann 181). Hartmann argues “perceived distance” is a more productive framework than reciprocity in new media environments, which Hartmann defines as “the degree of conviction the user has that [their] behavioral expressions do or do not have an impact on the symbolic behavior of the other[sic]” (Hartmann 184). By replacing the dichotomy of reciprocal versus non-reciprocal, parasocial elements appear within reciprocal, interactive settings (Hartmann 186). Hartmann exemplifies this with anonymous online interaction, where despite reciprocity, parasocial interaction appears due to anonymity and great physical distance(186). Hartmann’s addition of perceived distance is relevant when considering the inherent reciprocity in Twitch’s chat rooms, allowing me to consider these chat rooms within a parasocial framework. All these researches contribute their own elements to the discourse on the parasocial, most notably the recent application to online environments, allowing for a nuanced and elaborate model for analysis. These different texts provide useful concepts to help me apply the parasocial to Twitch. This research analyzes different engagements on Twitch and understands them on a social-parasocial continuum combining the textual, contextual and developmental features offered by this theoretical basis. As for textual elements, establishing direct address, mutual awareness, identification, attraction, authenticity and intimacy have returned as important elements. Repeated exposure is a developmental feature important for reinforcing these elements and developing a parasocial relationship. Contextually, co-viewers, discussants, and other encounters play important roles in these online settings. The concept of perceived distance is especially useful in describing the dynamics between social and parasocial in these online contexts, as it shows how they can intertwine within one encounter.

1.5 Structure

In order to analyze the relationship between social and parasocial on Twitch, this research is divided according to different social roles encountered on Twitch. In order to maintain analytical focus, I will not address Twitch’s “Creative section”, but rather focus on its video game content. According to Nathan Edge, this community consists of “professional and amateur gamers, teams, commentators, sponsors, spectators and fans” (34). The sponsors are not encountered personally, but via product placement of computers, screens, keyboards, computer mouses, headphones, chairs, logos on clothing and commercials during the broadcast. Therefore, sponsors will be left out of the equation in this analysis. Professional gamers, teams and commentators are encountered in professional eSports tournaments, while amateur gamers are encountered streaming on their own individual channels. The distinction is adequately described by Karhulahti, who distinguishes between personal and impersonal live-streaming. The impersonal live-streaming is “structured by a third party organization over the individual performers and their local audience”, while personal live-streaming consists of people who own, host and structure their own stream (Karhulahti 1). In the impersonal eSports live-streams, professional teams and gamers do not communicate with their audiences. Rather, the teams and gamers appear behind their computers focused on the games. The first chapter thus focuses on the commentators during impersonal live-streams, as they are actively interacting with Twitch audiences. The second chapter focuses on the personal live-streams and the third chapter explores the role of other viewers in both impersonal and personal live-streams. Within each chapter, I explore the parasocial and social features of the first encounter with the persona, the development of a parasocial relationship between user and persona and the tensions that arise when doing so. I will combine textual, developmental and contextual features that surfaced in the theoretical framework to show how Twitch’s social features can afford parasocial interaction and whether this approach helps shed new light on the variety of usages and behaviours on Twitch. In the first chapter, the parasocial relationships between users and commentators during competitive eSports events will be explored. The role commentators play within these events shows similarities to the persona as introduced by Horton and Wohl, which provides a great basis for the existence of parasocial interaction on Twitch. Despite these similarities, tensions emerge when exploring the online contexts surrounding these personae, where reciprocity is easily established.In the second chapter, this reciprocity is further explored by focusing on the

(9)

9

relationship between users and individual streamers which appears in inherent interactive settings with constant potential reciprocity, showing an inherently different persona. Social and parasocial elements are increasingly intertwined in this setting, further blurring distinctions between these two concepts. In the third chapter, I will explore the mutual relationships between viewers. Here, the traditional idea of a singular persona disappears, taking one more step away from traditional personae. The possibilities for parasocial relationships towards an entire community are analyzed, which I will refer to with the new concept ‘paracommunity’. Here, it appears that it is exactly Twitch’s social features that allow for such a strong parasocial connection to appear, inherently intertwining the social and parasocial and bringing its own, new values.

This analysis will add new interpretations and perspectives to Twitch usages. Even when actual interaction does not appear, social values might be fulfilled, and even in reciprocal situations the parasocial proves to be a productive framework. Through engaging with Twitch in a social, parasocial or hybrid way, with either commentators, streamers or viewers, different social needs might be gratified and a sense of community, or paracommunity, appears. The integration of the broadcast environment into an online setting provides notable similarities and distinctions to traditional television personae. As Twitch is constantly developing new features it might be argued that this analysis easily becomes outdated. However, within this constant development, a trend of constantly enhancing social features can be found, meaning this research into those social features remains relevant. I hope to show the complicated engagements on this unique hybrid platform make for an interesting research object, warranting more academic attention in the current media environment. While adding to discourses surrounding Twitch, this research is relevant for the entire media environment where all platforms are increasingly hybrid combinations of different social elements and features. Twitch takes up a notable position in this environment and can serve as an interesting case-study. Parasocial interaction might then prove to be a useful and valuable concept to apply to other social platforms as well, expanded by the concept “paracommunity”, hopefully stimulating others to apply this concept in future research. It is inspiring to see a relatively old concept like parasocial interaction still relevant in the 21st_century

(10)

10 Chapter 2. Commentators

2.1 Introduction

The variety of social roles encountered on Twitch structure this analysis, focussing on the commentators as personae in this first chapter. These commentators show interesting similarities to traditional television personae as described by Horton and Wohl, making them an interesting starting point for this analysis. Commentators are encountered during eSports events, such as tournaments, play-offs or championships, held in-person or online, organized and live-streamed by various professional organizations from the gaming industry (Edge 37; Karhulahti 1). The term eSports refers to electronic sports, “high-level games and spectating of digital games in a competitive atmosphere” (Edge 34). During tournaments, small groups of young, male commentators are responsible for guiding their audiences through the tournament, similar to traditional personae who seem to “exist for their audience only” (Horton and Wohl 374). Commentators welcome the viewers, summarize previous matches and introduce upcoming matches, the competing teams and their predictions for who will win. Image 2.1 shows a general example of commentators behind their desk during a major tournament 31 October 2016, similar to how talk shows hosts face their audiences. During the matches themselves, they are responsible for explaining the game and the strategies employed within them, similar to traditional sports casters. After the match, they might discuss and analyze the highlights of the match.The most popular eSports games are Counter-Strike: Global Offensive (CS:GO), Defense of the Ancients (DotA), StarCraft (SC) and League of Legends (LoL). These competitive multiplayer games consist of multiple matches between two professional teams, each lasting about an hour, leading towards the final match between the two remaining teams resolving who receives the trophy and prize money. This can last a week, a weekend or multiple consecutive weekends.In between matches, some channels will stop broadcasting, while others broadcast commercials or timers counting down until the next match. Some events take place in a stadium or event hall and can be attended by local audiences, though tournaments may also take place online or on smaller scales. TheseeSports broadcasts can be found on the great variety of channels owned by these organizations. The events often createpredictable and enormous “flash crowds”, making the channels easy to locate via Twitch’s browse features.The chat rooms allows users to comment upon the match, the gamers or the commentators, establishing reciprocity with other viewers, though the messages are often unreciprocated by commentators (Deng et al. 6).

Image 2.1 Screenshot, 13 April 2017, <https://twitch.tv/videos/98181519>.

This chapter explores the possible parasocial relationship between viewers and commentators, starting with an exploration of the first encounter and how this connects to traditional parasocial theories from my theoretical framework. A tension appears surrounding the role of the online

(11)

co-11

viewer within this relationship, which will be explored in the following paragraph. Then I will consider how parasocial relationships might develop, especially paying attention to repeated exposure. A new tension appears here, as repeated exposure is less predictable in these online environments. In the last paragraph, the appearance of cross-media encounters will be analyzed along with the role these encounters play in establishing a stable persona. This chapter shows, though Twitch’s environment shows many similarities to Horton and Wohl’s traditional concept, applying their ideas to Twitch’s online context along with its social features, means the concept alters significantly and new points of focus appear.

2.2 Encountering the Commentator

Users and commentatorsfirst meet each other at the start of an eSports broadcast, where commentators welcome the viewers, introduce themselves and introduce teams that will be playing in the upcoming match (image 2.1). In this first encounter direct address is employed, as the commentator “faces the spectator, uses the mode of direct address, talks as if he were conversing personally and privately”, creating the illusion they are aware of their viewers (Horton and Wohl 374). Direct address is divided into verbal and body address by Hartmann and Goldhoorn (1113-4). Bodily, commentators face each other and the camera, making eye contact with, and thereby acknowledging, the viewer. Verbally, they explicitly greet and welcome the viewer, often addressing them with a directed “you”. A typical example is the introduction to a DotA tournament, 11 December 2015, by Dakota “KotLGuy” Cox who faces the camera and says “hello everybody, welcome back to The Summit 4, we are getting into the nitty-gritty of day number three, the winner bracket final matchup of Evil Geniuses taking on Vici Gaming. We're getting ready to get underway into game number two” (https://youtu.be/mNfEip0GKxg). Through this explicit introductory welcome, KotLGuy creates an illusion of mutual awareness and attention while acknowledging his role as guide. The directed “hello everybody” engages viewers into a social process as this expressive behaviour is directed towards them (Hartmann 181). Simultaneously, KotLGuy uses “we” to include viewers and make them part of a communal happening. This usage of “we” is commonlyused by commentators to connect their viewers to them. The “welcome back” also emphasizes the seriality of the tournament and assumes viewers saw the previous matches. However, his introductions remain inclusive and complete, still providing sufficient information for those who missed previous games, making them feel equally included. It is these modes of direct address that immediately establish aparasocial interaction between viewers and commentators through a felt mutual awareness, inclusion and connection.

During the match itself, the commentators are invisible and only present audibly. The mode of address shifts to a less direct mode, as both bodily and verbal direct address decrease. However, they do sustain their guiding role in commentating on the match for viewers, similar to how a sports caster is responsible for guiding viewers through a match, thereby sustaining the created parasocial connection of simulated mutual awareness to the viewers. Their vocal cues remain engaging, as during tense or exciting moments the commentators start talking louder and more enthusiastically. This phenomenon has turned so common that commentators are also called “shoutcasters”, referring to their enthusiastic, loud commentating style (Cheung and Huang 768).During a CS:GO match between NIP and F’natic, commentator Anders Blume was extremely surprised by an unexpected turn of events, at which he yelled “OH MY GOOD WHAT [incomprehensible screaming]. Oh my god, I am at a loss for words right now. These NIP retakes, they’re giving me a heart attack” (https://youtu.be/mtcLILJ3LIg). In another match between Cloud9 and Dignitas, Anders yells, without taking a breath, “HIKO ARE YOU KIDDING ME HE’S GONNA GO WIN THE ROUND HIKO MAKES IT A TRIPPLE KILL AND CLOUD9 WIN IT THIS IS UNREAL WHAT?!” (https://youtu.be/mtcLILJ3LIg). Anders’ lack for words and incomprehensible yelling, display honest surprise at that moment. The spontaneity of responses to live events, increases authenticity as they display honest excitement and enthusiasm to events unfolding in real time. This honesty and authenticity is argued by Hartmann to be essential for establishing and maintaining parasocial interaction, as viewers have to feel they are encountering a true, responsive and social human being (Hartmann 187). Additionally, witnessing the events simultaneously with the commentator in a shared liveness increases the sense of an intimate connection. This authentic and emotional shoutcasting is appreciated by viewers, not only in

(12)

12

CS:GO, but in other games too, proven by the compilation videos on YouTube where users collect the most memorable shoutcasting moments by their favourite commentators.

Additionally, this informal commentating style with laughter, yelling and cursing creates a relaxed, friendly atmosphere. There are often multiple commentators, allowing them to interact with each other like close friends might, making fun of each other via snide remarks or inside jokes. During a CS:GO tournament, 5 March 2017, commentator Lauren “Pansy” Scottasks co-commentator Jason “Moses” O’Toole “what are your thoughts on this?”, to which Moses answers “I have no idea”. Pansy replies with laughter and a sarcastic “great job, Moses, nice” (https://twitch.tv/videos/126623028). Displaying, or simulating, this friendship via recognizable group interaction helps the spectator feel “included in a group” and part of their intimate and spontaneous friendship, increasing the felt intimacy(Auter and Palmgreen 82-3). This relaxed style is maximized in a unique way by the channel BeyondTheSummit, where tournaments are broadcast from a regular living room (image 2.2)(https://twitch.tv/beyondthesummit). The commentators are on their couches in comfortable everyday clothes, without shoes, talking like friends watching television together (image 2.2). During these broadcasts, gamers and commentators walk into the kitchen behind them to grab drinks or snacks, creating a relaxed and intimate atmosphere where it feels like they are regular friends and the viewer is part of this informal friendship. While this example is an exception, BeyondTheSummit is known and liked for this informal, relaxed style. An interesting side note to the friendship between commentators is the names they address each other with, as they alternate between nicknames and real names. During one tournament, the commentator who was previously addressed as “Moses”, is now constantly addressed with his real name, “Jason”(https://twitch.tv/videos/98181519). Even within one event, commentators alternate in how they refer to each other.During a BeyondTheSummit broadcast, Joshua “Steel” Nissan is referred to as Josh, Joshuaor Steel, displaying there are apparently no general rules in distinguishing between real names and nicknames as both are used in similar friendly, informal contexts (https://twitch.tv/videos/137508335). At the start of an event, a bar on screen displays both real names and nicknames, sometimes along with Twitter usernames too, including viewers in all informal modes of address these commentators use for each other, further increasing the simulated friendship between viewer and commentator.

Image 2.2Screenshot, 25 April 2017, <https://twitch.tv/videos/137508335>.

The informality increases authenticity and intimacy as users can identify friendly group interaction, but also contributes to the attraction viewers feel towards commentators. Rubin and McHugh research different forms of attraction (“social, physical, and task attraction”) and its role in developing relationships, both mediated and unmediated (283-4). Social attraction appears when personae appear friendly, likeable and display “similar attitudes”, which is stimulated via a shared subculture (Rubin and McHugh 285). The way commentators address each other and their

(13)

13

audiences connects to that subculture, in humour, jargon, similar codes and values and enthusiastic responses to similar interests. A second form of attraction, task attraction, depends on the persona’s “ability to get the job done” and relates to the validity and precision of analyses by commentators (Rubin and McHugh 285). Commentators gain more respect and validity from personal gaming history, displaying great knowledge and precise observations. The most popular commentators are liked for a combination of precise analyses, funny remarks and honest enthusiasm, a combination of task and social attraction. These preferences are reflected in a YouTube-video, 14 March 2017, made by Duncan “Thorin” Shields, a commentator and eSports analyst himself, who argues “Sadokist” is currently the best commentator due to his precise analytical skills and his humour (https://youtu.be/bFxGEMGPHqU).The relevance of physical attraction seems to disappear, consistent with Rubin and McHugh who conclude physical attraction is insignificant for parasocial interaction (288). Interestingly, most commentators are male in a predominantly malesubculture, meaning a potential gender difference might disappear from sight (Horton and Wohl 381-2; Consalvo 86). This will be addressed in the next chapter.

Image 2.3Screenshot, 20 June 2017, <https://twitch.tv/videos/134436381>. Image 2.4 Screenshot, 20 June 2017, <https://twitch.tv/videos/137700083>.

(14)

14

The chat room often responds to conversational topics initiated by commentators or their predictions for upcoming matches (https://twitch.tv/videos/134436381). This can be stimulated by commentators who ask their audiences to share their opinions or predictions. In image 2.3, users respond to predictions for a match between “FaZe” and “HellRaisers”(HR) by typing their favourite team’s names along with their anticipated final scores. Simultaneously, Spietres comments “THORING ON POINT”, refers to the preciseness of the predictions by commentator Thorin, showing appreciation for his skills. Some events initiate polls that register chat predictions, allowing the organization to display percentages on Twitch’s expectations for the match during the broadcast. This feature is not available during all tournaments, but the presence of this option might contribute to viewers’ motivations to respond and engage with the commentators, enhancing the parasocial feel of the encounter. Other than content, the chat room often responds to the way commentators behave or express themselves. During tense and exciting moments, commentators start speaking more enthusiastically and faster, not only shoutcasting, but also “speed”-casting. When this happens, the chat will be seen messaging “Rap God”, by which they jokingly refer to the similarities to the high speed of lyrics in rap songs, additionally displaying appreciation by referring to them as “God” (image 2.4). These types of responses are common, proving viewers socially engage with both content and performances of commentators. Participation in these responses to commentators might contribute to theconnection viewers feel towards them, despite the fact that the relationship remains one-sided as commentators do not reciprocate. An interesting connection appears between social interaction and parasocial interaction, as the social contributes and stimulates the parasocial. However, these social conversations might also be considered part of the social relationship towards viewers, rather thancommentators. Especially in massive streams with busy chat rooms, participation in the chat means spectators need to divide their attention between the match, the commentator and the chat. The chat might here be argued to distract viewers from the intimate relationship towards the persona,contradictory to television spectators who are less aware of co-viewers and thereby more individually connected to the persona.

This paragraph shows a simulated connection between viewer and commentator via direct address, creating the illusion of mutual awareness, intimacy, attraction and motivation for developing a relationship. The features of the commentator and content of the broadcast thus allow for a traditional parasocial interaction to appear. However, the added features in Twitch’s chat room and the constant visibility of co-viewers do affect this parasocial experience. On the one hand, engaging with Twitch’s chat rooms canincrease the parasocial bond users feel, while on the other hand it might decrease intimacy and distract users from their individual bond to the persona. As the exact nature of their role is not yet clear, the next paragraph will explore the contextual co-viewer and Twitch’s chat room and the ways they affect parasocial interaction.

2.3 Co-Viewers

This paragraph explores the way co-viewersin Twitch’s chat rooms affect the bond between viewers and commentators as it appears they decrease the degree of intimacy and attention towards commentators, but they might also be argued to play a sustaining role. Horton and Wohl describe “discussions which the spectator may undertake with other people” about the persona might influence the relationship (379).Horton and Wohl also argue visible studio audiences are strategically “coaching the audience” into appropriate roles by showing responses towards the host through applauding, cheering or booing (Horton and Wohl 377). Seeing these responses towards hosts makes it easier to imagine one’s own interaction with them. Giles more explicitly divides these contributions of “other users” up into “co-viewers” as coaches or guides and “discussants” as people to discuss the persona with (299). These two functions of other viewers are now compared to both Twitch’s local studio audiences and Twitch’s online co-viewers.

When comparing local eSports audiences to traditional studio audiences, similarities emerge in their position, both directly facing the broadcast content live. However, eSports audiences are not necessarily aware and responding to commentators, nor do commentators explicitly address these local audiences as they seem to merely exist for the audiences at home. A relationship as established between traditional studio audiences and television hosts does not appear, as they show no interaction to each other. These local audiences can sustain emotional involvement to the

(15)

15

match itself, as they do respond to events within the game. However, they do not position themselves in a role towards the commentators, meaning the traditional coaching or guiding role of studio audiences disappears in these eSports events. Rather, this traditional guiding role of studio audiences is taken up by online co-viewers in the live chat room offered and displayed by Twitch. Here, active, real timeresponses to the commentators appear, cheering, booing, responding to content or behaviours, displaying possible roles for other audience members. Conversational topics are taken up in the chat room, stimulating others to consider or share their own predictions for the upcoming match or to also engage with the conversation. A fun example of this is when, during Smash Summit Spring 2017, 3 March 2017, the commentators are informally discussing their breakfasts (http://twitch.tv/videos/125880494). Chat users message “I had an apple this

morning”, “eat fruits last”, “fruits are not good for you”, “fruit for breakfast is amazing”, “fruit is dessert”. The chat room is mingling in the conversation, replying to statements made by the commentators as if part of a reciprocal conversation, while in reality commentators do not read or respond to these messages. As certain viewers respond, others are stimulated to engage with thisvisible conversation too while simultaneously faced with the active possibility to do so in the chat room. In this sense, spectators are thus guided by their online co-viewers. It might then be argued participation in the chat room is guided or coached by organizations, who might initiate polls, or by Twitch, with an interface where viewers are faced with co-viewers and actively invited into the conversations in the chat room.

Besides guiding and stimulating viewers, chat participants also function as discussants. By reiterating beliefs, displaying appreciation or invalidation of commentators, attraction is affected. During Smash Summit Spring 2017, chat users address the absence of commentator Aziz “Hax$” Al-Yami (http://twitch.tv/videos/1258804940). Comments like “where the fuck is Hax?”, “WHERE IS

HAX$”, “WE NEED HAX”, “GIVE ME HAX”, or simply “HAX”, express an appreciation of Hax$, increasing his overall validity and attraction.Simultaneously, appreciation and validation of commentators that were present at that moment is affected negatively, as they apparently do not measure up to Hax$. Twitch’s online co-viewers thus take up a double sustaining function, where they are both comparable to studio audiences in guiding or stimulating behaviours and additionally function as discussants in displaying, reiterating or negotiating beliefs about the commentator. In these eSports events, the amount of viewers is often very high (Deng et al. 6). Giles suggests not only the content of expressions, but the “number of persons involved” also affects the parasocial (293). Here, a tension between massiveness and intimacy appears. As there are many other viewers involved, the sense of individual relevance and intimacy is reduced. Viewers are constantly aware of the amount of other viewers engaging with the same persona, which stimulates a more critical look upon the persona and decreases intimacy. Increased audience sizes thereby might distract the viewer from theirconnection with the commentator, shifting attention more towards these co-viewers. Additionally, as audience sizes grow it is often argued meaningful conversations are harder to establish, which decreases their sustaining function as guide or substantive discussant. While co-viewers can thus sustain parasocial relationships, an increased amount of co-viewers might actually reduce their stimulating function.

This paragraph shows online co-viewers can indeed sustain the parasocial, functioning as both discussant and studio audience. The visibilityof co-viewers stimulates parasocial interaction by displaying possible attitudes and responses, while it simultaneously allows for live, synchronous discussions concerning the persona, affecting attraction and validation. Contrariwise, in massive settings with great audience sizes, thissustaining role is reduced due to decreased intimacy, decreased meaningful conversation and dividedattention. This dual role will be explored further in later chapters. This paragraph shows tensions between massiveness and intimacy, as well as connections between social features and parasocial experiences. Participation in the chat room might be considered a social activity, while simultaneously serving parasocial functions, blurring the definition of what it means to behave socially or parasocially.

(16)

16

2.4 Developing a Relationship

The next step within this analysis will be exploring whether and how parasocial interactions between viewer and commentator grow more intimate, as one parasocial experience is not equal to the development of a relationship. This intimacy is developed over time with repeated exposure, as commentators often return for multiple tournaments of specific games. Within one single tournament repeated encounters also occur, as there is often a small crew of commentators who alternate in guiding the viewer during the entire event. Rubin and McHugh analyze the development of relationships, arguing repeated exposure is a strategy actively employed by viewers to decrease uncertainty and increase intimacy with a persona (281-2). According to them, repeated exposure is necessary for “uncertainty reduction” and increasing attraction, intimacy and importance of the relationship (282-3). Repeated exposure thus helps understand a persona better, increasing the sense that the viewer “knows” the commentator “more profoundly than others do; that he ‘understands’ his character and appreciates his values and motives” (Horton and Wohl 375). Additionally, repeated exposure is essential for understanding certain jokes and references commentators make. The previously mentioned quote from Anders Blume who enthusiastically yells “are you kidding me?!”, has become a well-known quote within the CS:GO community and is often jokingly used and referred to by other commentators. Via repeated exposure a shared history and shared knowledge is established, necessary for understanding these types of references and increasing felt intimacy.

Commentators often emphasize their own seriality and returning nature. In January 2017, RiotGames started a LoL-competition between American universities, the “uLoL Campus Series 2017”. Each Tuesday, two university teams play a match, hosted by Tom Searfoss and Mark Zimmerman. In the first broadcast, 23 January 2017, they welcome viewers to the series, promising “to make this season the best one yet”, thereby already emphasizing their returning character and inviting viewers to join the entire season (https://twitch.tv/videos/116861380). After the first match Tom employs direct address while inviting their viewers to return, “that’s our show for today but thank you for watching and we’ll hope to see you again next week”. The second match of the series, 31 January 2017, Tom starts with: “Hello and welcome back to the 2017 uLoL campus series, I’m Tom Searfoss and joining me once again this week at the desk is Mark Zimmerman,” to which Mark replies “yeah, I’m happy to be back” (https://twitch.tv/videos/118593785). Their seriality is thus already normalized and emphasized in this second week. In the tenth week of the series, 28 March 2017, Tom introduces Mark as his co-commentator, “basically as always”, emphasizing their own regularity and predictability (https://twitch.tv/videos/131682106).

While repeated encounters are possible, most events and channels are rather unpredictable. The fixed weekly broadcast of the previous mentioned uLoL-series is exceptional, which becomes clear when further examining their channel, where multiple tournaments are happening simultaneously and broadcasts alternate each other irregularly (https://twitch.tv/riotgames/videos). Some weeks have daily broadcasts, or even multiple broadcasts a day, while other weeks have no live broadcasts at all or only “reruns” of tournaments that have already been broadcast. Additionally, while the uLoL campus series can tell their viewers to return next week, most eSports events end their broadcast with vague statements on the next encounter, like “we hope to see you next time for more”, without any specifics on what and when (https://twitch.tv/videos/130180679). This is problematic, as Horton and Wohl argue it is important to making a persona’s “appearance a regular and dependable event, to be counted on, planned for, and integrated into the routines of daily life” (Horton and Wohl 375). While television broadcasting is highly institutionalized and centrally organized, providing clear information on all upcoming broadcasts in advance, Twitch does not provide any information, announcements or guidance for its viewers and there seem to be no fixed days or times for broadcasts. This is amplified by the fact eSports crossvarious time zones, meaning certain events are broadcast in the middle of the night in certain countries.Viewers have to know their way around multiple websites and social media platforms to find information on the times and channels of future events as it is often the websites of the organizations themselves that provide the necessary information (https://hltv.org,https://lolesports.com,

(17)

17

viewers cannot depend upon a singular platform for reliable information on future encounters, but have to spread their attention across multiple platforms.

Not only channels and times of events are irregular and unpredictable, it is often also unknown which commentators will be present. In some cases, announcements for events contain information on the entire crew, but even then it will often be unpredictable which exact commentator appears at what time and what match ( https://hltv.org/news/19241/dreamhack-winter-talent-announced). Generally speaking it is but at the start of the match that viewers know which commentators will be casting. This balance between recurring, yet unpredictable, commentators is exemplified by Bobby “Scar” Scarnewman’s introduction during Smash Summit Spring 2017, “we are here as we were time and time again, except we weren’t here last time, which sucked, but we’re here now” (http://twitch.tv/videos/125880494). He displays his recurrence and seriality, while simultaneously addressing his unpredictability. It is important to remember eSports broadcasts are archived on Twitch, allowing viewers to watch them on demand, creating stable other encounters that are always accessible. However, when watching archived broadcasts, the sense of sharing a live experience is lost, decreasing intimacy. Additionally, Florian Hoof argues liveness is extra valuable for sports events, as the value of sports footage has “diminished greatly after the final score is known” (88). Lastly, even in this stability, it can be a puzzling search to find a broadcast with specific commentators. While watching certain television shows ensures viewers of certain personae, the constant unpredictable alternating of commentators between events, channels and organizations makes them hard to locate on demand.

There is thus a crucial difference between the regularity and predictability of traditional television personae and online commentator personae. The same commentators do return within the community of one game leading to repeated exposure. However the events and the presence of specific commentators are irregular and unpredictable. Whereas television personae appear according to an organized schedule, eSports lack such a fixed, regular schedule. The development of a parasocial relationship is more fragile than in traditional organized televisual contexts. Viewers can no longer depend upon a single source for regular exposure to the persona, but have to spread their attention across multiple websites to ensure predictability and stability. However, other websites also become spaces wherepersonae can create a social presence outside of their mere Twitch appearances. These are described by Giles as “other appearances”, like interviews, magazines or other television programs, which he argues are relevant contextual elements contributing to parasocial relationships (297). These appearances, according to Giles, sustain an ongoing relationship, reinforcing authenticity and attraction (297). These encounters outside of Twitch, especially on social media platforms, and their potential role in creating stability, regularity and authenticity, are explored in the upcoming paragraph.

2.5 Other Encounters

Many commentators own profiles on Twitch, Facebook, Twitter and Instagram, where they post updates on both their professional and personal lives. Announcements of broadcasts and URLs to the right channels are posted, facilitating easier access to information on eSports events and repeated encounters with the commentators. Additionally, social media profiles give insight into their personality, confirming existing perceptions and offering new information on their “backstage behavior[sic]” (Auter 176). This backstage behaviour is revealed as sides of their personality, invisible in the regular Twitch encounter, appear, establishing the persona as a fully rounded, authentic human being, necessary for an intimate connection(Auter 176). Commentator Henry Greer often posts personal updates on Twitter to which other commentators, Alex “Machine” Richardson and Jason “Moses” O'Toole, reply (image 2.5). This confirms the friendship they also portray during tournaments. Through jokes and personal anecdotes, intimacy and authenticity increase, as viewers may feel they are growing closer to the persona and getting to know them better. It seems the social media profile itself becomes a stable and predictable encounter, where intimate relationships appear as users gain access to the “veridical self”, “the real person behind the projected self that is offered for public consumption” (Barry 254). However, when maintaining an online profile, people construct their identity, consciously differentiating between public and private information. Though spectators may feel like they truly know their commentators, in reality

(18)

18

they are faced with a constructed balance between public and private, professional and personal (Marwick and boyd 128). It might then be argued it is not “backstage” behaviour that appears on these platforms, as awareness of their audience and active construction of an online personality turns these expressions into a part of their performance as persona too.

Image 2.5 Screenshot, 20 June 2017, <https://twitter.com/henrygcsgo>.

Social media accounts prove there is a real person behind the persona, thereby not only establishing authenticity, but also the possibility for future contact (Hartmann 185). Hartmann argues the “general possibility for contact with a character enhances social involvement” (Hartmann 191). Similarly, Giles emphasizes the “potential relationship between interactants” influences the parasocial relationship (293). Some conceptualizations of the parasocial assume all viewers wish to develop their parasocial relationship towards the social, actively pursuing reciprocal interaction with a persona. This is not necessarily true, as the mere potential for contact with an authentic persona might be sufficient for a great amount of users. This has also been shown by Thorson and Rodgers, who research how users value and perceive the interactivity of political weblogs and conclude “perceptions of the opportunity to interact may be sufficient”, “regardless of whether they make use of the interactive features” (41). It is thus not necessary for users to engage with these interactive, social options, as the mere possibility to do so might be a valuable addition to the parasocial relationship.

However, it cannot be denied social media have considerably lowered barriers for attempting to establishsocial contact (Tsiotsou 402). Previously, a small number of people sent letters to their favourite personae, but showing appreciation or opinions and hoping for replies is significantly easier now (Horton and Wohl 383). While the potential for contact is rarely fulfilled with popular commentators who have massive lists of followers, commentators do respond to messages in certain cases, thereby creating a reciprocal situation. Reciprocity traditionally does not fit within a parasocial relationship, as it is often considered inherently social rather than parasocial. However, Hartmann rightfully argues reciprocity is not the essence of the distinction between social versus parasocial, as reciprocal contact might still contain parasocial elements (182-3). Perceived distance between viewer and commentator, while significantly decreased, remains rather great due to great anonymity and minimal potential for further developing a mutual social relationship (Hartmann 184). By emphasizing this perceived distance rather than reciprocity, it becomes clear

(19)

19

that these reciprocal, social settings can still be considered within a parasocial framework, proving reciprocal and parasocial do not exclude each other, but can become intertwined (Hartmann 184). This paragraph shows social media profiles are responsible for ensuring predictable, stable and regular exposure to the persona on Twitch, as they are used for information on upcoming Twitch encounters. Additionally, they become stable, constantly accessible encounters of themselves as well where intimacy and authenticity are established. Whereas “other encounters” are often considered secondary and subordinate in traditional accounts of parasocial interaction, social media profiles of commentators take on a prominent role. Commentators cannot be isolated to a singular platform, but have to be understood in their online, cross-media context. This also means it is difficult to separate between backstage behaviour and primary performances, as social media profiles and expressions become part of the persona’s performance. Simultaneously, this online context increases the potential for social interactivity and reciprocal contact with commentators, which makes it harder to distinguish betweensocial and parasocial interaction. Strategies for social contact are employed within parasocial relationships, while commentators employ parasocial strategies on social websites. The line between the two categories blurs and the relationship between viewer and commentatordoes not fit within one definitive category.

2.6 Conclusions

In this chapter, a lot of similarities between Twitch’s broadcasting of live events and traditional television broadcasting appeared, as direct address, attraction and intimacy return as important factors in simulating conversation between viewer and persona, allowing viewers to feel like they become part of an intimate interaction and relationship with the commentator. This chapter has thus established a parasocial interaction between viewers and commentators, providing a good basis for my further analysis into the parasocial on Twitch. Some tensions appeared surrounding the role of certain social, contextual features that will be taken up further in upcoming chapters. A first concern is the online co-viewer, who balances between sustaining and disturbing parasocial interaction. While co-viewers can stimulate parasocial engagement and bonds, massive audience sizes decrease intimacy and felt mutual attention and awareness.

The parasocial relationship develops with repeated exposure, increasing authenticity, intimacy, affinity and attraction, similar to traditional theories. However, a second tension appears surrounding the fixed schedule according to which repeated exposure takes place in traditional mass media settings, which is largely absent on Twitch. The commentator is less predictable and fixed, increasing uncertainty and instability, meaningviewers have to access other websites for information and announcements. Social media accounts of commentators can facilitate and inform viewers on these future encounters, while simultaneously becoming stable encounters of themselves. Thus, even in a setting similar to television broadcast settings, online contextual elements are important to take into account, as both the contextual “other user” and the contextual “other encounter” become a prominent part of the relationship.

Another tension surfaces as the existence of social features contributes to the parasocial experience. The existence of the socially interactive chat room within Twitch stimulates connecting and responding to the commentator, while social media accounts outside of Twitch facilitate potential reciprocal contact. This chapter shows how reciprocity can still be understood within a parasocial framework, as perceived distance is rather great. However, these added possibilities for actual contact do affect the parasocial relationship between viewer and commentator and significantly blur the lines between what is social and what is parasocial. The balance between the social and the parasocial is inherently different from the parasocial relationships in televisual environments due to the added interactive, reciprocal features. While this chapter thus establishes the social features of Twitch indeed affect the parasocial relationships that appear, the tensions that surfaced warrant further exploration. In order to do so, the next chapter will analyze the personal live-streamer, in contrast to the eSports as impersonal live-streams, and its interactive and inherently social, reciprocal chat room (Karhulahti 1).

Twitch.TV and Parasocial Interaction: Understanding Twitch's Social Features Within a Parasocial Framework