"skinnysuicidal_99 reblogged your post”: Visual narratives on Tumblr for issues of depression, self-harm and suicide

“skinnysuicidal_99 reblogged your post”

Visual narratives on Tumblr for issues of depression,

self-harm and suicide

Student name: Nefeli Gkoni

Student ID: 11692812

Supervisor: Alberto Cossu

Program: MA Media Studies Course: New Media and Digital Culture

Acknowledgement

I would like to thank Dr. Alberto Cossu for his precious guidance throughout this research, my parents for giving me the opportunity to pursue my dreams, as well as Charis P. and Christos C. for becoming my family in the Netherlands and providing me with their friendship, support and love through all difficulties. Finally, I would like to devote this research to every student of every field who is struggling with mental health issues, as a note of solidarity and hope.

Table of contents

1. Introduction 3

2. Theoretical Framework 6

2.1 Social media usage and mental health issues 6

2.2 Depression, Self Harm and Suicide: The multiple faces of a youth in crisis 9

2.3 Self Presentation Online - What we are VS What we show 11

2.4 “What’s on your mind, user?”: Emotional disclosure online 12 2.5 Is there beauty in pain?- Pop culture and Aestheticization of Emotion 16

3. Methodology 18

3.1 Visual Narrative Analysis 19

3.2 Method of Analysis and Choice of Platform 20

3.3 Categorization 23

3.4 Limitations 27

4. Data Analysis 28

4.1 The tag ‘depression’ in 2010 28

4.2 The tag ‘depression’ in 2014 33

4.3 The tag ‘depression’ in 2017 37

4.4 The tag ‘self harm’ in 2010 44

4.5 The tag ‘self harm’ in 2014 47

4.6 The tag ‘self harm’ in 2017 50

4.7 The tag ‘suicide’ in 2010 55

4.8 The tag ‘suicide’ in 2014 59

4.9 The tag ‘suicide’ in 2017 63

5. Empirical Analysis 65

5.1 Sad and Beautiful: Aesthetic Imagery 66

5.2 Pop culture references as cultural ‘Touch-Points” 68

5.3 NSFW: Graphic Imagery and Selfies 71

5.4 The power of art and humor: Artistic Expressions and Irony 72

5.5 Writing on visuals: Text-based imagery 75

6. Discussion 80

7. Conclusions 88

Abstract: ​As mental health becomes an urgent issue of modern society, the need to monitor its expressions and variations is vital. Simultaneously, the constant growth of social networking offers opportunities to scholars to examine the landscape of emotional disclosure online in correlation to the impact it may have, especially on users who struggle with depression and depression-related illnesses. The present research aims to explore the visual vernaculars used by Tumblr users for the issues of depression, self harm and suicide. By examining the 100 most engaging images tagged with ‘depression’, ‘self harm’ and ‘suicide’ for the years 2010, 2014 and 2017, this thesis unveils the expressive vernaculars that dominate on the platform of Tumblr, as well as the progression of them through time and according to Tumblr’s unique affordances. Results indicate that there are specific, recurring vernaculars that users engage with and that are being facilitated by the re-appropriation and re-distribution of content. Discussion involves potential impact of problematic content.

Key words: mental health, Tumblr, visual analysis, depression, self-harm, suicide

1.Introduction

Back in 2016, Microsoft Japan developed Rinna, an Artificial Intelligence program and administrator of an online blog and a Twitter account. Rinna’s developers attributed the personality of a high school girl who is about to start her acting career in a Japanese television show. Indeed, over the first days, Rinna’s blog posts were imitating the online behavior of a teenager, by criticizing her “parents” (in this case, her developers) as well as sharing exciting news about her life (Baseel, 2016). However, a few months later, the character of the posts took a dark turn. In an impressive demonstration of her ability to portray and adapt human emotion, Rinna became a depressed, suicidal teenager leaving everyone in wonder whether the AI picked up patterns she recognized from adolescent behavior online and developed accordingly. Even though eventually this turn of events was a marketing trick for the promotion of Rinna’s TV show (Thor Jensen, n.p.), neuroscientist

Zachary Mainen suggests that AI programs may be indeed capable of developing symptoms of depression, given that they are built to think like a human being and thus, having the same mental health issues (Oberhaus, 2018).

What is especially interesting about this bizarre incident is that Rinna’s developers designed a teenager’s personality as depressive and suicidal. Rinna went from posting details of her filming days to a breakdown of self-hate, anger and despair, blaming herself for not being worthy enough to actually stating that she “wants to disappear” (Baseel, 2016). For example, one of the first posts that raised concerns about Rinna’s behaviour was stating:

“Actually, I couldn’t do anything right. Not at all. I screwed up so many times. But you know what? When I screwed up, nobody helped me. Nobody was on my side. Not my Line friends. Not my Twitter friends. Not you, who’re reading this right now. Nobody tried to cheer me up. Nobody noticed how sad I was.” (Thor Jensen, n.p.)

The reason why the public was concerned with the AI’s behavior is that it was surprisingly convincing: Rinna managed to mimic the expressions and vernacular of a depressed teenager online, by turning against herself and against others for not being present enough through her dark times. Even though the blog posts were staged, Rinna and her developers managed to encapsulate a reality familiar to those active online. Since the introduction of Web 2.0, when users became creators of content, emotion, as an interwoven part of human existence, has found its way into online narratives in various cases and forms. Especially in social media platforms, the ecosystem of which relies primarily on user-generated content, the lives of participants are on display. As networking platforms gain popularity over the years, one may suggest that the limits between public and private space have been blurred. Additionally, the more involved a person is to social networking platforms, the higher the amount of transcription of analog aspects of life to digital ones. For instance, it is fascinating to observe how digital language develops over the years to incorporate a greater spectrum of meanings and emotions. The introduction of emoticons and, later on, emojis reveals the gradual enrichment of online communication that sought to go further than textual communication, that was unable to transfer motions and emotions. Indeed, it was the user’s need to describe more online, to talk about more complicated issues, experiences, and feelings that led to the

birth of a new, exclusively digital language and, as a consequence, to the flourishing of digital cultures.

The present thesis will be involved with the subject of mental health issues and the vernacular that surrounds these issues online. Specifically, the research question that is posed here is how do people express themselves online for matters of depression, self harm and suicide. By taking the social networking platform of Tumblr as a case study, a website known for its ability to provide a safe space to marginalized social communities to openly express and discuss sensitive issues, the 100 most engaged imagery of the years 2010, 2014 and 2017 will be extracted and analyzed through the method of visual narrative analysis. The aim is to design the unique visual narrative for each year, as well as to compare and contrast the similarities and differences between the chronologies. That way, I aspire to observe how media evolution is being reflected in the vernaculars that surround the delicate issue of depression and depression-related actions and the ways digital culture has affected emotional expression and disclosure online.

To begin with, the theoretical framework that encapsulates the present research will be drawn. Literature from both the fields of psychology and media will be used to explore the cause and effects of their relationship. However, the purpose of this thesis is not to examine a possible correlation between media and mental health, but rather to discuss how social events, such as the rise in depression rates, are being expressed on social media. The time spectrum of seven years, within which users of social media platforms have significantly increased and different means of expressions have emerged, such as GIFs and memes, is 1 sufficient for studying these changes. Following the theoretical framework, a methodology will be established for the extraction, collection, observation, and coding of data that will constitute the empirical part of the present thesis. Next, the categorization that emerges from the analysis of the imagery collected will be presented, followed by the discussion chapter, on which I will attempt to frame the results of the empirical analysis with theory and proceed to conclusions. My preliminary hypothesis is that emotional discussions online will be significantly different from 2010 to 2017 in terms of media used. Fundamentally, the augmented use of Internet worldwide has resulted in a digital culture that is beyond language barriers and national differences. An exemplary case of cultural construction is the use of

memes. Originating from the word “mimesis” (Greek: μίμησις) which means imitation, Limor Shifman (2013) defines a meme as “cultural information that passes along from person to person, yet gradually scales into a shared social phenomenon” (365). Indeed, memes have greatly contributed to the expansion of Internet cultures as they are used a cultural touch-points within online communities. As an extent, it could be supported that memes become the language of articulation of a collective narrative, from a new way of political participation to a new way of emotional expression. In combination with the increasing power of online communities and the re-appropriation of content that Tumblr as a platform affords, I estimate that this shift towards a more collective narrative will be inculcated on the data collected and that meme culture, as well as pop culture will have influenced how people express themselves about mental health online.

2. Theoretical Framework

2.1 Social media usage and mental health issues

Since the introduction of the SixDegrees.com in 1997, the first web page with the format of a social network (boyd, Ellison, 214) social media have been drastically incorporated in our everyday life, altering a substantial variety of practices, from online marketing to human interaction and emotional disclosure. Although the interface of social networking platforms evolves continuously through the years, social media are defined by boyd and Ellison as “web-based services that allow individuals to (1) construct a public or semi-public profile within a bounded system, (2) articulate a list of other users with whom they share a connection, and (3) view and traverse their list of connections and those made by others within the system” (211). The nomenclature, affordance system and nature varies from platform to platform.

In terms of usage, ongoing researches show that the numbers of social media users are steadily increasing, not only in numbers of users, but also in minutes spent online. Moreover, it is estimated that currently around 51% of the global population has internet access ("World Internet Users Statistics And 2018 World Population Stats") and in 2017, 2.46 billion people were active on social media ("Social Media - Statistics & Facts"). It is the universal acceptance of social media in people’s lives that incline the swift in communication that has

taken place alongside their entry. Yet, another significant point of interest is the demographics of the online population. Even though the preoccupation with social media has increased for all age groups, networking still remains most popular among generations. Ages between 18-29 present the highest rates of usage, with 88% of them managing Facebook profiles (York).

Minutes per day spent online/ years 2012-2017

​ © Statista 2018

Perhaps one of the most prominent effects of extended social media usage is the major swift in communication. Since the introduction of Web 2.0 that reshaped Internet’s affordance system by allowing the creation of user-generated content, the World Wide Web received a whole new dynamic. Not only did content became democratized, in a sense, but also the importance of the digital identity is increasing drastically as new media conquer leisure time. Moreover, it is the uniqueness in the structure of each and every social networking platform that has contributed in a differentiated communication model. Papacharissi (2010) has argued that it is the total of particular social, technical and architectural affordances of the networking platforms that indicate exclusive types of communication, while leave no space for others (p.314). Given that the usage of networking interfaces has increased drastically, it can be

argued that there is a general swift not only in communication but, as a consequence, in the narration of our own identity.

Alongside the emergence of social media platforms, social scientists have been recognizing an increasing mental health crisis among the younger generations. Jean Twenge (2006, 2009, 2017) has focused her research interest in the correlation between new media and psychological effects, with her latest book discussing how social media platforms and the attachment to smartphone devices may be interconnected with an epidemic of depression and anxiety in younger generations. Specifically, Twenge notices a dramatic increase in depression episodes among adolescents, with “56% more teenagers experiencing a major depressive episode in 2015 than in 2010 and 60% more experienced severe impairment” (108). By noticing the parallel trajectory of both social media and mental health related issues through time, a number of scholars have been sceptical about the connection of the phenomena. Tromholt (2016) conducted an experiment during which participants had to abstain from Facebook for a week. Results showed that there is immediate correlation between the use of Facebook and lower levels of well-being, for the matter that social networking is intervening with the social identity, by imposing constant comparison and need for reassurement. Indeed, a number of scholars argue that growing up in a time when technology and social media communication are reshaping the world, the youth of today has to cope with financial and societal distress that can only be enhanced in an environment of constant and instant information (Schrobsdorff, n.p.).

The relevant correlation between social media and mental health is further expanded to the space of communication and emotional disclosure. Page (2012) on her empirical analysis of blogging suggests that interactivity within the frame of digital narratives demands not only the user’s engagement with the context of the story, but also highlights the importance of the digital interface in the production of meaning (214). This transformation of narrative online is applied when seen in the context of specific issues. In an individual research on online narratives of chronic pain, Gonzalez-Polledo and Tarr’s analysis notes that “although social media has structures equivalent to traditional illness narratives, in each platform different meanings develop in relation to the medium used (text, images, multimodal content) and in relation to the social context of communication.” (7). More than a culprit for the crisis in

mental health of the youth, social media provide a free space for virtual communities to come together governed by “social behaviours that traverse and frequently converge a variety of behavioural and performative platforms” (Papacharissi 312). It is because social media afford a variety of new forms of networks and interactivity and enable the effective remake of our understanding of pain and suffering (Gonzalez-Poledo and Tarr, 1468) that make them worth studying in the context of mental health issues.

2.2 Depression, Self Harm and Suicide: The multiple faces of a youth in crisis

The focus of the present study is to understand how the reported crisis in mental health issues may be expressed in media cyberspaces, as well as what could the means of expressions indicate for the issues of depression, self harm and suicide outside the virtual world. To begin with, it is essential to define the terms from a psychological perspective as an attempt to differentiate the scientific use of these terms from the media use of them. Depression is defined by the American Psychiatric Association as “a common and serious medical illness that negatively affects how you feel, the way you think and how you act.” ("What Is Depression?"). It is estimated that around 350 million people worldwide suffer from depression and depression-related episodes (Pietrangelo) and the symptoms include loss of interest or pleasure, decreased energy, feelings of guilt or low self-esteem and disturbed sleep or appetite, often accompanied by episodes of anxiety (World Federation of Mental Health, 6). Demographically, women are more likely to experience depressive episodes (Twenge 108) with women aged 40 to 59 having the highest rate of depression (Gholipour). Moreover, recent studies have shown that there is a significant increase in depressive episodes among young adults and teenagers of about 37% (Mojtabai et al. .4), yet this increase is not reflected in the numbers of patients treated for depression, leading to assumptions that the majority of the incidents go untreated. Particularly interesting is that this dynamic is being expressed in cyberspace. Content related to mental health issues on Tumblr, one of the most popular platforms for discussions of sensitive topics, has increased by 248% between 2013 and 2016 (Ruiz, n.p.).

Twenge reports that “depressed teenagers are more likely to self-injure, such as through cutting” (108). Self harm is defined as the non-suicidal and intentional damaging of one’s

body tissue through cutting or burning (Seko and Lewis) while there has been a research that suggests that teenagers who practice self injury are more likely to use online media to a greater degree than their peers who do not (Mitchell and Ybarra 396). Testimonies from practicers of self harm report that the act alleviates the suffering and embodies the psychic pain by addressing its physicality. Demographically, self injury rates are significantly higher among younger ages in comparison to adults, with 13–17% of adolescents and young adults reporting at least one SI episode in their lifetime (Swannell et al.). Seko and Lewis conducted a research in regard to visual narratives of self injury on the social platform of Tumblr and the results of it will be utilized in the current research as guidelines. According to the researchers, the act of self harm is often portrayed as a coping mechanism for the sufferers. The provocative, and often repulsive, imagery is often accompanied aggressive text that aims to justify or even normalize the act (Seko and Lewis 190). It can be understood that young adults who engage with the practice of self harm not only do they transcribe it into a forceful part of their digital narrative, but also they seek to justify their action in public. Ultimately, “these graphic self-portraits may constitute a form of visual monologue that makes pain recognizable and shareable for asynchronous viewers” (Seko and Lewis 192).

Major depression, especially in its most severe manifestation, is also the primary risk factor for suicide (Twenge 110). Almost 1 million lives are lost every year due to suicide, a figure equal to 3000 suicide deaths every day. For every person who completes a suicide, 20 or more are closer to an attempt to end their lives ("WHO | World Suicide Prevention Day"). As with depression, suicide rates have increased to a worrying degree. The National Center for Health Statistics conducted a research to monitor the suicide rates for the US population between 1999 and 2014 and the key findings show that the rate has increased by 63% for women aged 45 to 64 and 43% for men of the same age span (Tavernise). For adolescents, three times as many 12- to 14-year old girls killed themselves in 2015 than in 2017, compared to twice as many boys (Twenge 110). However, in terms of online behaviour, there have been numerous opinions about the beneficial intervention of online communities, as well as the platforms themselves in suicide prevention. Mitchell and Ybarra (2007) suggest that “youth who engage in self-harm may be more likely to engage in online behaviours that have the potential to place them in risky situations” (392), yet this behaviour can now be monitored and reported to the platform. Facebook, in particular, is a pioneer on this matter, as there have been incidents

where authorities were alerted by the platform when content has been flagged as suicidal (Kliegman). Indeed, Facebook, Instagram and Tumblr are among the social media networks that have enabled the feature of online support, once the users searches for terms flagged as worrying (eg. suicide, self harm, depression).

The alarming rise in mental health issues that regulate the younger generations seem to interplay with social media networking. Given that younger generations are statistically more engaged with new media while, at the same time, their emotional state is at risk may hint a swift in the produced content. Peer pressure, cyberbullying, financial problems, the Fear Of Missing Out, as well as the constant comparison of one’s life with another are contributing to the deterioration of a generation in crisis. On the other hand, online communities and frequent engagement with emotional expression has been noted to benefit struggling adolescents in distress (Barak, Cohen, 189) that seems to be of great importance for the fragile age of puberty. Hence, the primary interest lies in the self narrative of people who struggle with mental health issues online and, to what extent, this online performativity is representative of the global epidemic in mental health.

​2. 3 Self Presentation Online - What we are VS What we show

Riessman, on her research on the uses of narratives in the presentation of self and illness, begins her introduction with a statement:

We are forever composing impression of ourselves, projecting a definition of who we are, and making claims about ourselves and the world that we test and negotiate in social interaction.(1195)

As she uses Goffman’s theory of narrative analysis to deconstruct the narrative of a recently divorced man who suffers from multiple sclerosis, Riessman suggests that the choices of the narrator in how they communicate their story is as important as the story itself, since it is the fundamental need to be likeable and accepted that constitutes the strategy of the teller (1196). Ultimately, the choices of the narration they do not only construct the story, but also the identity of the narrator, as well as the feelings that her audience will have for her. For Goffman, embarrassment is an effective indicator, as well as a motivator of where people fail

to present themselves (Milner). The fear of being seen as ignoble, ethically blemished or failed dictates a certain form of presentation, while the fear of embarrassment in the presence of such failure from the side of the other participants, shapes an audience eager to help this performance of one’s identity.

The same principles of narration apply augmented when transcribed to cyberspace. The public nature of social networks more often than not blurs the lines between what is socially acceptable to disclose and what is not. boyd suggests that social media have established a number of core activities, such as chatting and socializing, engaging in self-expression, grappling with privacy, and sharing media and information (8). She converses the definition of the networked public as “(1) the space constructed through networked technologies and (2) the imagined community that emerges as a result of the intersection of people, technology and practice” (8). Within the ambiguous space of the networked public, boyd identifies the difficulties teenagers face while trying to figure out how to be in public since there is a fundamental “difference between being in public and being public” (57). The social construction of identity in a public space can create confusion as to what are the limits to disclosure of emotions. When the user becomes the creator of the content and the online content is a product of one’s performance of identity, then it can be argued that a user struggling with mental health issues will include the pain into her narrative. Yet, boyd argues that media agency cannot afford serious modifications in the mentality of the users. She claims “Through social media, teens reveal their hopes and dreams, struggles and challenges […] Technology makes the struggles youth face visible but it neither creates nor prevents harmful things from happening even if it can be a tool for both. It simply mirrors and magnifies many aspects of everyday life, good and bad.” (212).

2.4 “What’s on your mind, user?”: Emotional disclosure online

Once you log in the front page of Facebook, the most popular social medium, the status bar reads: “What’s on your mind?”. On the same page, Twitter asks “What’s going on?” while Instagram and Snapchat invite users to photograph or record their everyday life in a “story”, that will disappear after 24 hours. While the decision of what is worth sharing lies entirely on the user, the affordances of the social platforms seem to nudge a disclosure. When the user is asked to contribute her life narrative, the semantic choices are of strategic importance to what

the user wishes to express. When the narrative derives and circulates around life itself, and its different aspects as the user experiences them, the emotional aspect, negative or positive, are interwoven with this narrative.

Andalibi et. al (2017), when researching the practice of sensitive self disclosure online, in accordance to “#depression”, support that there is “positive bias” in social networks, meaning the favor towards online positivity and positive expressions over negative. This complies with the theory of performativity as Goffman explained, that each person needs social confirmation and one’s way of communicating life follows the narrative linear that will ultimately portray them as socially accepted and liked. Ergo, the digital transcription of identity may adopt the same principles, with users wishing to portray an ideal image of their self. This leads to a significant flow of positivity, with a tendency to highlight fortunate events over distasteful moments. Yet, this rhetoric changes according to the medium under scope. For instance, Facebook is known for relatively low rates of emotional disclosure, in comparison to other social networks. Newman et al. (2011) explain how Facebook’s network “supports the accumulation and maintenance of social capital” by enabling the possibility to create, interact and maintain social ties with friends who do not necessarily live in geographical proximity (342). However, the virtual friendships that Facebook allows are among acquaintances and the profiles of the users are semi-public, meaning that on the one hand the virtual networks are created by people in one’s actual milieu, while on the other hand the privacy restrictions allow users to curate the exposition of their content.

Even though positivity and strategic narratives of perfection flourish social media, emotional disclosure and discussion of delicate issues frequently take place in cyberspace. The Internet’s relative anonymity offers a safe space for marginalized communities to express themselves released from the fear of stigmatization (Seko and Lewis 181) that society often imposes. Numerous studies (Seko and Lewis 2015, Andalibi et al. 2017) have shown that online disclosure of emotion, pain or illness are directly associated with support communities that roust from emotionally charged content. Gonzalez-Polledo and Tarr (2016) conducted a research on illness narratives of chronic pain patients in social media, in order to examine in what ways new media have reshaped emotional expressions. They suggest that “the creative process involved in communicating about pain through multiple media has the potential to

transform the experience of pain by shifting its focus from within to outside the body” (1456), while Ressler et al.’s (2012) study of text-based illness blogs found that sharing the experience of illness with others and learning about illness and illness patterns to be among the top reported reasons to blog about illness (Gonzalez-Polledo and Tarr 1465). Finally, Seko and Lewis, (2015) while researching the narratives of self-injury online, found that “images of self-inflicted wounds may help people with SI experiences to identify with similar others and feel a sense of belonging”(192).

Indeed, emotional narratives online have been under research because of the specific nature of the story being told, as well as for the response it arises. By removing pain, illness and emotion out of a clinical scope, online narration of mental health issues bridges a gap in communication between patients and doctors, thus making its examination particularly interesting for clinical purposes. Hurwitz et al. (2008) define illness narratives as “polymorphous, malleable and dynamic devices through which pain and illness are known and communicated in clinical contexts and beyond” (Gonzalez-Polledo and Tarr 1457). Yet, in their analysis Gonzalez-Polledo and Tarr (2016) recognize the peculiarity of negatively-charged narratives in terms of complexity that lies inside pain and painful experiences, and they argue that traditional pragmatic approaches on narrative theory cannot address the multidimensional experience of pain and emotion. Gonzalez-Polledo and Tarr suggest that this approach reflect in a more efficient way pain narratives as it exceeds cultural, social and linguistic barriers and ascribes a holistic view of pain experience that lies in emotion itself.

Yet, this theoretical approach of narrative seems to apply for chronic pain visualisation online. The question that may arise is to what extent this assumption could also be compatible with mental health issues narratives. In their analysis, Gonzalez-Polledo and Tarr make clear that “although conditions such as depression are often an inseparable part of living with chronic pain, we have not selected material where the pain expression seemed to be predominantly about emotional pain” (1459). For this case, Seko and Lewis (2018) when researching the narratives of self injury on Flickr and Tumblr suggest that it is the fundamental architecture of the medium of narrative that plays a major role in how these delicate issues are being communicated. Platform affordances on Tumblr, they argue, must be taken into consideration

as the ecosystem of the media provide a fragmented narrative space for self-curation (182). Again, the researchers employ Goffman’s theory of presentation of self as transcribed by Hogan (2010) for social media usage to explain the unique affordance system of Tumblr and the consequences in user’s narratives. They argue that:

By offering users personal gallery for social curation, Tumblr embraces what Hogan (2010) calls the “exhibition” of the self through a bricolage of digital artifacts. While for Hogan the “curators” of identities are non-human agents such as algorithms that manage digital content beyond user’s control, Tumbleloggers seem to play an active role in their self-curation in collaboration with the algorithm. (183)

Ultimately, what Seko and Lewis add to the discussion of online disclosure of pain narratives, be it physical or emotional pain, is that it is of great importance to take into consideration the space on which this narrative takes place. Gonzalez-Polledo and Tarr insightfully approached the paradigm of pain narratives by understanding the importance of emotion as a factor for studying them, yet Seko and Lewis suggested that it is not only a matter of the user’s narrative, but how the environment of the digital space intervenes with the story.

In addition, from a media perspective, the unique platform affordances of Tumblr that dictate self-curation have led to a specific language of emotion that can be either text or image based. Andalibi et al. (2015) conducted a research on depression related imagery on Instagram as to identify the role of photo sharing in the framework of experience. Indeed, there have attempts to use photography as an artistic expression of emotion since “its technical and documentary functions is seen as relevant for therapeutic purposes” (Seifert 1) and “we cannot ignore images when researching young people’s experiences in any area of life, much less when our research questions focus on the digital aspects of being human” (Tiidenberg, n.p.). Perhaps it is this insufficiency of language to transfer the delicate nuance of emotion that turns people to photography and visual expression of emotion online. Visualisation has been traditionally the vehicle of expressing emotion, from art to photography to memes. New media by introducing the era of images, have managed to reinforce, as well as establish the use of visual over textual discourse. Photograph, in particular, has an exceptional role in today’s society as it can carry meanings, quickly transfer information and shape holistically the perception of one’s reality. It could be argued that photography online becomes the artifact of curation of one’s identity, as described above. When it comes to narratives of mental health issues such as

depression, Andalibi et al. compared communication modes of images and texts and “established the importance of visual imagery as a vehicle for expressing aspects of depression;” (234). Finally, one of the major representations of visual culture in new media and its importance to online communities is the vast usage of memes, namely, images, videos, and spoofs under constant circulation and modification by users (Vickery 301). In researches of pain narratives, memes have been seen to be especially popular among communities, “as this emergent narrative format enables users to re-appropriate cultural artifacts around them to express their unique stories—often in a sarcastic and provocative manner” (Seko and Lewis 184). Therefore, it is not only the image as a carrier of meaning that has been embedded into the era of online communication and user generated content, but ultimately it is the reshaping and circulation of this meaning in order to adapt to each user’s individual narrative. Given that emotional experience is a matter of subjectivity, digital artifacts have provided the essential tools for individuals to tell their own stories based on pre-existing cultural touch-points, therefore making the narrative easier to be identified and engaged with.

2.5 Is there beauty in pain?- Pop culture and Aestheticization of Emotion

Following the discussion of the particularly idiosyncratic emotional disclosure online of people suffering with mental health issues, Boyd and Marwick’s (2009) term ‘conundrum of visibility’ comes along to highlight the twofold nature of being public. On the one hand, extreme expressions of sorrow, physical and emotional violence and even suicide online by teenagers may appear troubling for clinical researchers and doctors, especially in the context of contagious and disastrous effects of influence among online communities (Lewis and Seko 250). However, Boyd and Marwick claim that these visual narratives bring out an otherwise hidden part of today’s troubled youth, by providing the required coverage for discussing topics of bullying, substance abuse, suicide, thus helping the audience understand “the lives of youth, the risks and dangers they face, and the personal, social and cultural logic behind their practices” (410).

In reality, exposure to extreme content that portrays depression and depression-related actions in a certain, often aesthetically pleasing way may influence imitative behaviours on young teenagers. In virtual communities, such the ones formed in the social platform of Tumblr, peer pressure is enhanced due to the instant exposure of one’s life and, as a consequence, the

immediate comparison that comes along. Furthermore, the glorification of depression online has long been discussed among scholars and health professionals, followed by concerns than “online pro-anorexic or self-injury forums can undermine recovery, encourage harmful behaviours by validating pro-disease view and identities through online disclosures and exchanges of support” (Andalibi et al. 36).

Tumblr has been under the scope and has been accused by multiple sources for affording subcultures that promote behaviours of aestheticization of pain. Namely, in 2010, a subculture flourished in Tumblr, that seemed to be representing the aesthetic of ‘Soft Grunge’, “which paired beauty and intelligence with a loner aesthetic” (Premack, n.p.). Originating from the music genre of Grunge, represented by anti-heroes such as Kurt Cobain, who committed suicide in the young age of 27, the image of Soft Grunge is “ironically masochistic, no-photo-unfiltered, contrivedly uncontrived – trying to look like it’s not trying” (Blackwood, n.p.). Beautiful in their tragedy, pictures that belong to this aesthetic often depict sad, emaciated girls, bruised or wounded body parts in black and white filters followed by quotes that provoke the notion of marginalisation and turmoil. However, in a society that suffers from clinical forms of depression, as the rates indicate, this confusion and over-use of terms has the potential to create and promote unhealthy tendencies for a youth that seeks positive affirmation. “This is exactly the kind of romanticizing I think that a huge part of Tumblr is doing, just making it seem like — depression, suicide, this is cool. If you want to be part of the club, you gotta be insecure and unstable in some way.”, said Hampus Leijon, a 22-year old student who started using Tumblr while experiencing depression in his adolescence, to Rachel Premack for ​The Ringer

​ magazine (2016).

Indeed, a simple search on Tumblr for terms like ‘depression’, ‘sadness’ or ‘suicide’ will bring up a great amount of posts that compile with the aesthetic of Soft Grunge. Yet, even though Tumblr is not the only platform that this peculiar tendency has been noticed, a number of factors indicate that it is the architecture of the platform that encourages the spread of confusion. As Shifman (2014) noted “imitation and remixing become the pillar of participation” (Gonzalez-Polledo and Tarr 1459) and even though this statement was referring to meme culture, similar patterns can be identified in the affordance system of Tumblr.

Tumblr is a very easy place for people to feed off of this kind of frenzy because of their ‘reblogging’ system, [which makes] it very easy to [spread] pictures and gifs… particularly gifs, which can be quite graphic.(Bine, n.p.)

Taking into consideration the preference of the audience for generic media content (Gonzalez-Polledo and Tarr, 2016, Seko and Lewis, 2018), such as memes and pop culture ‘touch-points’, instead of personal narratives that communicate a user’s individual experience with depression, it can be understood that the representation of mental health issues as beautiful suffering, thus the glorification of it, could be more accessible than the more realistic depiction of it. Naturally, pop culture has been exploiting this trend by encapsulating Soft Grunge and, ultimately, commodifying it. There have been generations that grew up with ‘emo’ culture being mainstream, with bands singing about suicide and self harm and2

thousands of teenagers practicing self injury, even though there was no clinical evidence of depression in their behaviour, but because it was simply what everyone else seem to do. Therefore, by recognizing the ability of aesthetics to shape and intervene with the viewer’s perception of reality, serious concerns are raised as to how heavy exposure to distortive imagery of pain can form the fragile psychosynthesis of a stressed-out youth. “Images transfix. Images anesthetize” (Sorensen 20) said Susan Sontag (1977) when discussing the role of photography of violence and it is a statement to keep in mind when researching critically visual narratives and their correspondence with sensitive issues of mental health.

3. Methodology

3.1 Visual Narrative Analysis

The interest of the present research lies on the visual narrative surrounding depression, suicide and self harm on Tumblr, as well as the gradual progression of it through a time span of seven years. By closely examining the 100 most engaged posts that contain images for the respective years 2010 and 2017, I will be focusing on the visual narrative that arises for each individual tag. Following the theory of emotional disclosure online and self narrative as developed in the previous chapter (Andalibi et. al, 2017, Riessman, 1990) the method of visual narrative analysis seems appropriate as it takes into consideration multiple aspects of

the story the image says, instead of focusing solely on the technical elements. In order to better understand the methodology of visual narrative analysis, Rose’s (2012) and Riessman’s (2005, 2008) theoretical principles will be used as guidance. As Tiidenberg (2015) notices in her visual narrative analysis of selfies, “We cannot ignore images when researching young people’s experiences in any area of life, much less when our research questions focus on the digital aspects of being human” (n.p.). That being said, images are an integrated part of digital culture, one that has been traditionally used to communicate emotions and feelings when textual language fails to. Focusing on an analysis of visual data when studying issues of mental health will lead to results that provide an enriched understanding of the online tendencies in terms of content and form, as well as the development of those.

For Rose (2012), the fundamental aspects of configuring the meaning of an image are composed of (1) the context of the image production, the environment and the motives that result in this image, (2) the image itself, as in its technical components and visual characteristics and finally (3) the audiencing of the image, the receivers of its meaning and the reaction that occurs (20). Yet, if one wishes to speak of a visual narrative, the image cannot be isolated from the story that it formulates. More than an individual incident, a narrative must consist of a series of interconnected events that have an internal flow of a common character (Gonzalez-Polledo and Tarr 1458). For the present research, the narrative that is being constructed differs for every single tag. For instance, the imagery accompanying the tag ‘depression’ will be analysed as an individual narrative in its own ecosystem. In this case, the imagery becomes the grammar of the narration. Every single image out of the 100 most engaging ones will be manually analysed and the patterns that derive from this analysis will be interpreted and coded in categories that construct the narrative of the selected theme, or tag, in this case.

3.2 Method of Analysis and Choice of Platform

In order to construct a narrative, a process was established to reassure the data collection and further analysis. To begin with, the terms under evaluation in the form of tags have been selected. I have chosen the tags ‘depression’, ‘self harm’ and ‘suicide’ because in psychology the consecutive use of these terms reflects the escalating intensity of emotional disclosure, while they cover the entirety of depression-related actions. While these terms are interwoven

and the results were overlapping frequently with the same image containing more than one of the examining tags, it was within the interests of the present research to examine occasional overlaps in the flow of the narrative. Furthermore, this event enhances the assumption that these online practices are not irrelevant to one another, but they may confirm a correlation that is being transcribed in their digital narrative.

The next step was the choice of the suitable social medium platform on which the research would be based. Among the criteria for this choice was whether the platform enables the formation of online communities, as well as the space this platform provides for this kind of sensitive conversations. Given that one of my main interests was the visual elements of the online narratives, the initial thought was to be involved with Instagram’s platform, where the content is exclusively based on imagery. However, Seko and Lewis’ (2018) paper on visual narratives of self harm on Tumblr, as well as personal experience with different media platforms lead to the choice of Tumblr as the most suitable social medium for this kind of research interests. To begin with, Tumblr’s platform does not solely work as social media. Its structure combines the elements of both a social network and a micro-blogging site, given that the user has the ability to create any kind of content, both textual and visual. In addition, Tumblr ensures user’s anonymity, in contrast to platforms more popular like Facebook, a fact that reinforces conversations of marginalized social groups about delicate issues such sexuality or mental health. Finally, what Seko and Lewis indicate to be of substantial importance for the narrative unfolding on Tumblr, is the uniqueness in terms of content recreation. Tumblr, by basing its content circulation on the feature of reblogging and 3

reposting, contributes to the creation of a social dynamic that is based on re-appropriation of content and “collective curation” (182). Furthermore, this exceptional affordance of Tumblr’s platform favors communities rather than individual narratives and original content. Thus, the curation of personal blogs on the platform is based on a collection of different, mostly popular artifacts that circulate the newsfeed of Tumblr’s homepage. By reblogging content, the user constructs a narrative by “borrowing” key content and re-appropriating to fit their personal storytelling. Consequently, this creates a unique, dynamic media ecosystem within

the platform (Seko and Lewis, 183) that facilitates actions of mimesis and repetition rather than originality.

A key-concept of Tumblr’s unique affordances is the network creation. When a user uses a particular tag for their content, the platform immediately creates a network of users with similar interests and curating choices. Therefore, a user’s re-blogged or liked content becomes a component of a larger, social narrative that is formed by user’s unique choices of curation. What Tumblr’s feedback system, eventually, achieves is to generate a new form of social narrative that derives from personal curation. For that matter, Tumblr has been chosen for this research as it provides a broad spectrum of popular opinion rather than personalized views on the matter of mental health. By looking into Tumblr’s unique ecosystem, one can observe the social narratives that its networks create and, eventually, understand how digital culture is being expressed about depression, in this case.

Later on, the time span of the needed data has been selected. Since I was working with TumblrTool, as I will further explain shortly, I was able to have an overview of all the images that were included in a chosen year. I have chosen to work with the years 2010 until 2017, with a special focus on the years 2010, 2014 and 2017, to monitor any progressive changes in terms of content that may have occurred within 7 years. The original idea was to begin my research on the year 2007 as Tumblr was founded on that year (Tumblr, Press Information). However, on the first preliminary run in search of data, it was found that the amount of posts tagged with the required terms was not substantial enough to justify a narrative analysis. For that matter, 2010 was chosen as the oldest and richest in data year, since the establishment of the platform. Respectively, 2017 was chosen as the most recent year with a great variety of results to compare and contrast with 2010. Lastly, by choosing the year 2014 as a middle year of focus, I will be able to monitor and the intermediate progression of tendencies in terms of data. Therefore, the distance of seven years provides a well-established ground to observe the rise and fall of online trends in terms of the type of content (images, videos, GIFs), as well as any potential development in the visual narrative surrounding these issues.

In order to retrieve the data required for the analysis of my research, I used TumbrlTool, a post retrieving tool developed by Bernhard Rieder, Associate Professor in Media Studies at

the University of Amsterdam and researcher with the Digital Methods Initiative. The tool allows the user to input the chosen tag and time frame for post retrieval and delivers a .zip file, as well as HTML preview of the query results. TumblrTool has run in total of 8 times:

● Tag: Depression → Timespan: 01-01-2010 until 31-12-2010 ● Tag: Depression → Timespan: 01-01-2014 until 31-12-2014 ● Tag: Depression → Timespan: 01-01-2017 until 31-12-2017 ● Tag: Self Harm → Timespan: 01-01-2010 until 31-12-2010 ● Tag: Self Harm → Timespan: 01-01-2014 until 31-12-2014 ● Tag: Self Harm → Timespan: 01-01-2017 until 31-12-2017 ● Tag: Suicide → Timespan: 01-01-2010 until 31-12-2010 ● Tag: Suicide → Timespan: 01-01-2014 until 31-12-2014 ● Tag: Suicide → Timespan: 01-01-2017 until 31-12-2017

However, given the high speeds of media evolution, it is important for the present research to have an overview of all the images that each year from 2010 until 2017 includes. Therefore, by taking advantage of TumblrTool’s option to preview results in HTML form, I conducted an empirical ethnography for all the years in between (2011, 2012, 2013, 2015, 2016) and for each different tag (‘depression’, ‘self harm’, ‘suicide’). In this way, I was able to monitor any significant changes in terms of media evolution and to track down the trajectory of the different, identified vernaculars. Therefore, even though my dataset was focused in three individual years, I was aware of the specificities of each major change.

In order to keep the data relevant to my research interests, a cleaning of the database was needed. Since the research focuses on the visual elements, posts with images had to be isolated, counted, downloaded and archived. Given that my goal was to conduct a quantitative analysis on the images, it was needed to focus on the 100 most engaging posts containing images for each year (2010, 2014, 2017) and each tag. While this method would not provide numerical representivity of the posts that use the hashtags every year, the rates of engagement allows me to understand the progression of each vernacular in terms of popularity, style and meaning. The .zip file provided by TumblrTool enclosed a Tab file with all the details of each individual post such as ID number, type, date, timestamp, title, caption, blog name, note count, post URL, tag, photo link, source URL and source title. Then, by

using Google Sheets, every spreadsheet for the respective year and tag was organised in note count order, from most noted, thus engaging, post to least. Finally, from the most engaging posts, I isolated the first 100 posts that contained a visual element of any kind, by opening the provided photo link and manually downloading the image or GIF. The total of visual data gathered was 100x8=800 images.

3.3 Categorization

The visual analysis focuses on the 100 most engaging posts of each examined year as a qualitative sample, in order to obtain a focused understanding of both the cultural patterns and most preferred types of expression. I chose not to use an Application Programming Interface to analyse the dataset, since the manual examination of the imagery allowed me to identify the produced meaning, rather than the technical, compositional aspects of the image that the mechanic algorithm would provide. For the purpose of choosing an approach that would facilitate a visual analysis, Ι used Riessman’s ‘Narrative Analysis’ (2005), for it provides the required theoretical framework of a well-established methodology of a research project. While suggesting different methods for conducting narrative analysis, Riessman explains the method of thematic analysis for collective grouping of data, which I will espouse for the current research. According to this approach, “investigators collect many stories and inductively create conceptual groupings from the data. A typology of narratives organised by theme is the typical representational strategy [...]” (Riessman 2). In this case, by perceiving imagery as the typology of the narrative as Seko and Lewis attempted in their research of visual narratives of self harm on Tumblr, a number of different categories will be formed based on mutual motives, either of the technical aspects or the produced meaning of each individual image. However, to avoid confusion and to design a methodological form that applies to exclusively visual narratives, the formed categories must be governed by a set of rules.

In her ‘Visual Methodologies’ (2016), Rose explains the importance of image coding in visual analysis. The coding allows the researcher a more immediate interpretation of their data, thus provides the freedom of examining a set of images under a different, individual scope. However, Rose argues that, even though the categorization of the visual elements gathered lies in the discretion of the researcher, there is a number of characteristics that must apply to these categories. They must be:

“

● Exhaustive

​ : every aspect of the images with which the research is concerned must be

covered by one category ● Exclusive

​ : Categories must not overlap;

● Enlightening

​ : as Slater (1993: 236) says, the categories must produce ‘ a breakdown

of imagery that will be analytically interesting and coherent’ ”

When taking into consideration the parameters for an efficient image coding, it is crucial to parlay the research interests as a guide, for it is those interests that will formulate the categories. The choice of the most engaging visuals as representation provides a sufficient overview of the data, as well as a clarification of identified patterns and interests (Schrek et al., 69, 2013). However, since Tumblr’s interface affords the use of multimedia, the initial categorization of the content derives from the type of file. Thus, images and GIFs (Graphic Interchange Format) will be measured for the respective years 2010 and 2017. By doing so, any alterations in user’s preference between animated and static images will be recorded.

To continue with the categorization of the images according to their meaning, the three principles as described by Rose (see p. 1) have been kept in mind. In respect to the limitations of manual visual analysis, by closely observing the images I separated each of the images in terms of colour: ​light (bright, light shades of colour are more distinguishing in the image) and

dark ​(shades closer to black dominate the layout). However, it should be noted down that even if darker colours are associated with negative emotions, in this case meaning should not be associated with colour, as this would be arbitrary. The aim is to observe any chromatic patterns, as well as the evolution of those, through the different tags and years.

Another set of characteristics will be based on the demographics of the people appearing on the images. When the image displays one or multiple persons, certain characteristics are identified and recorded: Age, sex and race. In this way, general conclusions can be drawn about the diversity of people portrayed in depression, self harm and suicide-related images.

Finally, perhaps the most significant category derives from the meaning expressed on each image. First, I looked closely to each individual image from the dataset and came up with keywords that aimed to describe the produced meaning of it. Then, the overview of the

dataset revealed that certain keywords were repetitive across the imagery or similar enough to be combined. The result was 10 different keywords of meaning that were met across the three different tags and were used to formulate the categories and, later on, the vernaculars. The categorization serves the research as it conveys the different nuances of negative/positive expression of the narrative on Tumblr, while it also provides an insightful angle of the vernacular of expression online. The classification needed to take into account the differences of content among the three different tags of ‘depression’, ‘self harm’ and ‘suicide’, yet to focus on the similarities. The categories that derive from the critical examination of the visual content are:

1. Irony​: Images that contain a notion of irony or humour, a mockery of the suffering, the pain, or the activities that depression is causing. Irony is expressed through the use both of visual and textual elements on the layout. Memes are included in this category. A great amount of content within this category derives from the Tumblr blogspot ‘cheaper-than-therapy’, in which users are often seen to use irony to spoof their feelings of depression.

2. Cultural “touch-points”​: Mostly screenshots from TV series or films, cultural “touch-points” are visuals taken from pop culture, that are globally (or extensively) recognizable by users. Quotes from films and series, lyrics from songs and photographs of music bands and well-known paintings are some of the most common patterns. Both visual and textual elements.

3. Artistic Expression​s: User created artwork that is often used as an outlet for feelings of anger, depression, despair, loneliness and sorrow. Within the category there are numerous drawings, sketches, comics while textual elements are not necessarily present.

4. Motivation/Support​:Images that express motivation against the suffering or provide support lines. While the majority of the images contain textual elements, this category includes images of healing and hope.

5. Selfies​: Even though this category includes photographs of the user’s face, the term ‘selfie’ is being used in a broader sense to entail all kinds of images that portray the user herself, regardless of which body part is visible. The textual presence in this category is limited and insignificant.

6. Graphic Images​:Especially common for the tags ‘self harm’ and ‘suicide’, graphic

images include any kind of explicit content and are No Safe For Work (NSFW). The textual elements are mostly absent from this content.

7. Aesthetic Imagery​: The main characteristic of this collection is the direction of the image’s aesthetics. Visual elements that are “aesthetically pleasing”, this category more often than not depends heavily on the textual elements, which often transfuse negatively-charged meanings on beautiful imagery. This type of content has been accused of romanticizing mental health issues, as it confuses suffering and aesthetics.

8. Text-based(Negative)​:Image files that contain only textual elements. Quotes, letters,

tweets or screenshots of messages, text-based images portray no visual elements or, if they do, the visual element does not contribute to the narration. This category focuses on the text-based images of negative content. The language used within the text is described as emotionally charged and the words used describe negative emotions.

9. Text-based(Positive)​:This category has the same characteristics as the previous one although the textual content here has a positive outcome. Given that these posts derive from depression, self harm and suicide related content, the texts rarely describe positive emotions but rather an attempt for support. The difference with the ‘Motivation-Support’ category is the lack in visual elements. Posts that contain information and descriptions of symptoms are included.

10. Other​: The content does not compile with the narrative of the research and the imagery is characterized as ‘neutral’ without the use of context.

To recount, the present research aims to focus on the visual narrative analysis. A sample of one hundred images of the totality of the images collected has been selected, based on the engagement. To continue, the variety of patterns in terms of technical structure and meaning produced has been identified, thus a number of categories was formed. Once the collection of categories has been completed, the results for each year will be compared and contrasted to observe and analyse any differences or similarities that may occur. The aim is to have a clear overview of the vernaculars that derive, as well as the progression of visual narrative that surrounds the chosen subjects, that will eventually guide the following discussion. By attempting to connect the theoretical framework with the results of the empirical analysis, the discussion will be mostly involved with contextualizing the results of the analysis and

drawing conclusion that will hopefully contribute to the greater scientific discussion of mental health narratives, seen from the scope of media epistemology.

3.4 Limitations

Given that the empirical part of the present research has been based on TumblrTool, a software that has been developed by Bernhard Rieder of the University of Amsterdam in order to retrieve Tumblr posts in file form for data visualisation and analysis. Unfortunately, the tool suffers from bugs that may intervene with the credence of the research. Taking into consideration the inconsistencies of the results that TumblrTool delivers, I have decided to conduct a qualitative visual content analysis instead of a quantitative one. In that way, occasional faults in the number of retrieved posts do not affect the results of the analysis, as the focus of the examination will be restricted on the one hundred most engaged posts for the respective years 2010, 2014 and 2017. By adjusting the methodology according to the exogenous limitations, the credibility of the research is being ensured, while the results remain relevant to the research question.

Another problem that occurred during the retrieval of data from TumblrTool is the extremely ephemeral nature of Tumblr posts, with 10-15% of the posts deleted within 24 hours after being posted. This resulted in TumblrTool retrieving a post, yet the content of the post (in this case, the image) was irretrievable from Image Downloader. As a result, for the query ‘Self Harm 2010’ the image data set available did not reach 100 images and only 41 images were retrieved for the whole year. Same goes for the query ‘Suicide 2010’ were the total number of posts including images was 98. Again, the research was adjusted to the available data set, while the lack of sufficient data could be perceived as a finding itself, discussed further in following chapters.

4. Data Analysis

4.1 The tag ‘Depression’ in 2010

Static Images 100

GIFs 0

Categories No. of posts/100

Irony 9

Cultural Touch Points 17

Artistic Expressions 15 Motivation/Support 6 Selfies 12 Graphic Images 2 Aesthetic Imagery 18 Text-based (Negative) 12 Text-based (Positive) 4 Other 5

To begin with, in 2010 there is complete lack of GIF files. Even though the existence of this type of moving images dates back to 1987, Graphic Interchange Format files reached the top of their Internet popularity in 2015 (Moreau, n.p., 2018).

What can be understood from the qualitative narrative analysis, Tumblr users using the tag ‘depression’ for imagery in 2010, have been mostly choosing aesthetics to express their feelings, as 18 images out of 100 matched this category. The platform of Tumblr has been criticized numerous times for aestheticizing depression and suffering , as well as for4

promoting the trend of ‘soft grunge’. What can be observed for this particular category in

2010, is the repetition of images that contain visual elements ostensibly unrelated to feelings of depression, on which the textual elements play the role of tone setter.

Example of the top 100 engaged posts for 2010 with the tag ‘depression’ [Aesthetics]

Pop culture references are similarly heavily used, with 17 out of 100 images. The majority of the imagery included in this category has been snapshots from the popular TV show “The Simpsons”. This cartoon, particularly popular in the US, has been known for its satire and dark humour. In the context of depression tagged imagery, the snapshots are for the most part accompanied with a subtitle that transfuses the melancholic tone in the image.

Example of the top 100 engaged posts for 2010 with the tag ‘depression’ [irony/meme]

These examples of posts are found among the most 10 most engaged of all 2010. The user’s eagerness in expressing themselves through the use of pop culture references could be deriving from the identification. Characters that are broadly popular, are easier to be identified and, as a consequence, to be relatable.

Lastly, 15 images have been identified as artistic expressions. The textual elements on this type of images is eliminated for 2010, while the majority of the imagery is black-and-white sketches and drawings, that may be used as emotional outlet, given their dark content and atmosphere.

Example of the top 100 engaged posts for 2010 with the tag ‘depression’ [artistic expression]

In conclusion, the visual vernacular surrounding depression in 2010 consists mostly of mild imagery, in terms of explicitly and expressions of suffering, while textual elements are strongly present. Selfies and negative-charges texts have respectively 12 out of 100 images, followed by irony (9 out of 100), supportive content (6 out of 100) and uncategorized/irrelevant imagery (5 out of 100). What can be understood for 2010 is that pop culture and user generated content are the main outlets of depressive expression, while it is noted that aesthetically pleasing imagery is being interpreted with descriptive text.

In terms of light and colour, the majority of analysed images appear to have dark characteristics. Texted-based categories (Negative, Positive) and unrelated images (Other) have been excluded from the compositional analysis.

Categories Light Dark

Irony 6 3

Cultural Touch Points 12 5

Artistic Expressions 6 8

Selfies 4 8

Graphic Images 0 2

Aesthetic Imagery 5 14

Specifically, out of 79 images, 38 image surfaces (48.1%) display light shades of colour, while 41 images (51,8%) are constituted by dark shades. Light intensity and colour saturation was taken into consideration and the results demonstrate that for 2010, darker images were slightly more.

Most interesting categories are for cultural touch points, were light elements were dominant (12 images out of 17), a tendency that resulted from the frequent appearance of screenshots from ‘The Simpsons’ cartoon, which is characterized by light, bright colours.

Example of Light VS Dark image under the category of ‘Cultural Touch Point’ [Depression 2010]

Selfies with dark colours were significantly more (8 out of 12) and interestingly enough, it is one of the few categories that colour is directly associated with the person’s mood. For 2010, the majority of selfies belonged to the same user. Thus, it facilitates an immense comparison of dark and light elements in accordance to meaning. 5

5 _{​For purposes of privacy and data protection, the selfies will not be demonstrated in the analysis. All imagery is}

Finally, images that fall under the narrative category of aesthetic imagery are constituted by darker, higher saturated colours and lower levels of light. This compositional elements justify the character of the category, as they facilitate as ‘tone setters’ for the atmosphere of a picture, thus create a sense of romantic melancholy.

Example of Light VS Dark image under the category of ‘Aesthetic Imagery’ [Depression 2010]

The images that fall under the categories of ‘Artistic Expression’, ‘Selfie’ and ‘Aesthetic Imagery’ have been further analysed in terms of demographics. Specifically, 27 out of 45 examined images include a human figure. For images of ‘Artistic Expressions’, 6 out of 15 images contain a human figure (in this case a drawing), 4 of them portrayed young, Caucasian women, 1 of them portrayed a young, Asian, women and 1 one of them a young, Caucasian man. For images under the category of ‘Selfie’, 12 out of 12 images were photographs of users with 11 out of 12 being selfies of young, Caucasian women and 1 of them being a selfie of a young, Asian woman. It is understood that for 2010, images that include human faces were mostly portraying Caucasian females of young age.

4.2 The tag ‘Depression’ in 2014

Type of File No. of posts/100

Static Images 76