Performative trolling: Szubanski, Gillard, Dawson and the nature of the utterance

Abstract

In 2012 the Australian public witnessed three important examples of trolling play out in the public sphere that are the focus of this paper: the trolling of Julia Gillard’s Facebook page when she attempted to discuss education policy, the anonymous trolling of Charlotte Dawson’s Twitter page, and the trolling of Magda Szubanski on YouTube after she came out on The Project. These attacks may seem similar in that a public persona has been ridiculed and denigrated in flamboyant onslaughts. However, we will argue that there are important differences in the effects of these attacks, and that underpinning these are differences relating to the individual persona, the social medium and the nature of the utterance. The attacks on Gillard and Szubanski are primarily descriptive attacks on a deliberate and somewhat stage-managed public performance of identity, not a call to action. On the other hand, the anonymous trolling of Charlotte Dawson, which led directly to her attempted suicide, is clearly a performative utterance from the start, meant to have consequences on the object of attack. In Dawson’s case, the separation between her public persona and her private self is far less distinct than in the case of Gillard or Szubanski. These instances demonstrate that trolling exists on a performative continuum, engaging in constant disruption, but also lending itself to the production of social action. The kind of impact trolling will have depends, thus, on the affordances of social media, the persona under attack, and on the very nature of the utterance itself.

Full text

Persona Studies 2016, 2 (1)
27
PERFORMATIVE TROLLING: SZUBANSKI,
GILLARD, DAWSON AND THE NATURE
OF THE UTTERANCE
BEL INDA MORRISS EY A ND SU SA N YELL
ABSTRACT
In 2012 the Australian public witnessed three important examples of
trolling play out in the public sphere that are the focus of this paper: the trolling of
Julia Gillard’s Facebook page when she attempted to discuss education policy, the
anonymous trolling of Charlotte Dawson’s Twitter page, and the trolling of Magda
Szubanski on YouTube after she came out on The Project. These attacks may seem
similar in that a public persona has been ridiculed and denigrated in flamboyant
onslaughts. However, we will argue that there are important differences in the
effects of these attacks, and that underpinning these are differences relating to the
individual persona, the social medium and the nature of the utterance. The attacks
on Gillard and Szubanski are primarily descriptive attacks on a deliberate and
somewhat stage-managed public performance of identity, not a call to action. On the
other hand, the anonymous trolling of Charlotte Dawson, which led directly to her
attempted suicide, is clearly a performative utterance from the start, meant to have
consequences on the object of attack. In Dawson’s case, the separation between her
public persona and her private self is far less distinct than in the case of Gillard or
Szubanski. These instances demonstrate that trolling exists on a performative
continuum, engaging in constant disruption, but also lending itself to the production
of social action. The kind of impact trolling will have depends, thus, on the
affordances of social media, the persona under attack, and on the very nature of the
utterance itself.
KEY WORDS
Troll; Public Debate; Celebrities
INTRODUCTION
In 2012, in Australia, one of our most successful comedians was denigrated as a ‘bush
pig’ merely for coming out as a lesbian; a host on a popular reality show was encouraged to ‘kill
herself’; and the then Prime Minister was ordered to cook dinner for one of her constituents and
asked about the colour of her pubic hair. Such is the nature of unpoliced public debate. In all
three cases, the social media sites of each of these women provided those with a sexist and
homophobic agenda an extraordinary lien to literally write whatever they wished. This was
accommodated through the phenomenon of trolling, which, while not especially new, certainly
took on a new vigour at this time. Although public figures have always been at the mercy of
debate and abuse, the public is generally well aware of the risk of legal action to punish specific

Morrissey & Yell
28
slander or vitriolic insult uttered in conventional media. Not so with the Internet, where trolls
seem to believe that they are able to write whatever they please, with few, if any, consequences
to themselves.
1
This paper will discuss three cases where the level of trolling was both unprecedented
and especially vicious. They include the trolling of Madga Szubanski on YouTube after she came
out on The Project; the trolling of Julia Gillard’s Facebook page when she attempted to hold a
forum on education policy; and finally, the trolling of the Twitter account of Australia’s Next Top
Model host, Charlotte Dawson, after she attempted to gain justice over a specific instance of
malice. In Szubanski’s case, the trolls’ comments, while evidently hurtful, seem almost generic to
the occasion: a woman comes out as a lesbian and is immediately denigrated for her looks. The
Gillard trolling contains many features of relatively generic attack as well: she is taunted with
insults which have been used against powerful women for decades in being told to ‘get back to
the kitchen’ and that her voice shrieks like ‘nails on a blackboard’. Far worse, in her case, were
comments relating to her recently deceased father, in clear reference to Alan Jones’ statement
that her father would have ‘died of shame’ at his daughter’s ‘lies’ (Hall). The Dawson example is,
by far, the most extreme. In this instance, this ‘mean’ reality TV judge was told to ‘please go hang
yourself’, to ‘stick her head in the toaster’ and sent images of bloodied corpses after she
defended one of her supporters from a nasty attack regarding the woman’s deceased husband
who had killed himself (Hornery and Hall, Baird). This assault had the most severe
consequences as Dawson was hospitalized for attempted suicide and sadly, she subsequently
succeeded in committing suicide in 2014.
At first glance, all three cases may seem similar in that a public persona was attacked
and denigrated in a flamboyant manner using the remarkable capacity of social media to do so.
In this paper we argue that there are important differences in the effects of these attacks,
relating to the performative and affective nature of the utterances in each case, the persona and
the social media platform.
J.L. Austin’s theory of the utterance posits a basic difference between illocutionary (or
descriptive) speech acts, and perlocutionary speech acts which function as a call to action. Using
this, admittedly controversial, distinction, it is relatively easy to argue that the trolling of
Szubanski and Gillard, while cruel, is nevertheless illocutionary. The trolling of Dawson, on the
other hand, is a clear example of the perlocutionary utterance, designed to have direct
consequences for the object of attack.
The instances that we discuss in this paper demonstrate that trolling exists on a
performative continuum, from disruption to the production of social action. Trolling can be
described as the Internet’s signature mode of discursive politics, at once disruptive for the sake
of an anarchistic vision of political life, while using the freedom of the Internet to insert well-
placed and necessary commentary in an idealistic vision of true democracy. Our examples show
that trolling can be both repetitively banal and tremendously effective, both utterly pointless
and utterly shameful. We assert that the kind of impact trolling will have depends on its
discursive context and on the very nature of the utterance itself. The force of the speech act is,
thus, bound up with its affective intensity.
Trolling has been defined in a variety of ways. For a number of writers, trolling is
considered to be a time-wasting activity, where a participant deceitfully posing as genuine
draws others into pointless and circular discussions (Donath, Dahlberg, Herring et al, Turner et
al). This relatively benign version of trolling takes its cue from a fishing term, where a baited
line is dragged behind a boat (Oxford English Dictionary), but could also be connected with the
Scandinavian hairy monster hiding beneath a bridge hoping to snare hapless strangers (Herring

Persona Studies 2016, 2 (1)
29
et al 372). This latter example leads to far more malevolent conceptions of trolling behaviour.
Trolling has sometimes been contrasted with ‘flaming’, defined as insulting, provoking or
rebuking messages sent specifically to vulnerable or naïve participants (Herring et al 372).
However, the two frequently merge, as, in more recent years, trolling has become a catch-all
term for a suite of negative behaviours (Hardaker 224).
Rebecca Rafferty notes in her 2011 study that trolls who are deliberately insulting tend
to be motivated by three reasons: informal social control, dominance and entertainment.
Informal social control is usually carried out by people known to the victim and it is effectively a
form of ‘cyber sanctioning’ for perceived bad behaviour (27). Dominance, on the other hand, is
defined as ‘the attempt to hurt, humiliate, or influence the behaviour of another individual in
order to gain or regain access to some valued resource’ (Rafferty 33), and is generally carried
out by former partners in romantic relationships who are anxious to either regain control over
the partner, or to make them miserable. The most important of the three motivations for our
purposes in this paper is the entertainment motivation. In Rafferty’s study, entertainment was
given as the most popular reason for cyber bullying or online aggression (37). The desire is to
annoy or to provoke the victim into exhibiting anger or misery through persisting with
harassment long after being asked to stop (37). Anonymity is vitally important to this sort of
trolling as it allows perpetrators a sense of power over their victim, and a feeling of
invulnerability from any form of consequences to themselves (38). Such trolling is considered
‘successful’ when the victim finally responds in the way suggested, or when others applaud the
troll’s posts (Hardaker 233).
It would seem then that trolling, at least of the malicious sort, is conducted by people
who ‘take pleasure in disrupting the social order out of anger, perversity or contempt’ (Herring
et al 382). While trolls interviewed in Jenny Brockie’s Insight programme claimed their trolling
was motivated by the desire to create comic effect and to expose hypocrisy, illogical argument
or fallacious reasoning, or as a form of ideological resistance, one might argue that malicious
trolling, whether banal or incisive, would rarely be motivated by anything other than very
personal peeves. The relation between the public persona of the troll as manifested in their
online discourse, and their private selves is complex and apparently contradictory. As Elizabeth
Grice notes in relation to the case of troll Brenda Leyland, who committed suicide after being
exposed for her relentless online abuse of Kate and Gerry McCann, “the image of the incoherent,
half literate troller who spits poison as a reflex compensation for lack of self-esteem” is far from
the truth (22).
The notion of ‘persona’ is valuable here in distinguishing the effects (and affects)
experienced by public identities under attack in instances of trolling. The concept of persona as
we are using it here is based on the poststructuralist argument that identities are not
psychologically stable entities but are produced through acts of discourse. Identity is “not a
fixed property of individuals but … part of an ongoing process by which subjects are
constituted” (van Zoonen 123). The production of any persona is then an ongoing performance,
which is open to contestation. As Judith Butler has observed, such a production is the
“reiterative and citational practice by which discourse produces the effects that it names” (2).
Thus, persona is a fundamentally performative and denaturalised concept referring not to an
autonomous and stable identity, but to an array of constructions. Public figures such as Gillard,
Dawson and Szubanski operate with one or more public persona, in addition to their private
‘selves’, and use social media in a more or less ‘managed’ way to perform these personas.
Marshall and Barbour argue (following Jung) that persona is:

Morrissey & Yell
30
a strategic public identity, not necessarily in tension with an inner soul of self and
individual … but a way to manage the various dimensions of life and its public
formation of the self. Persona can be seen as something that needs to be managed
and … a personal practice that is performed in order to enter the social world in
some particular way (4).
For celebrities and other public figures who are required to negotiate the relation between the
public and private personae on a daily basis, there are significant risks if the boundary is
unstable or the distinction unclear. The consequences of a close versus distant relation between
their public and private persona can be profound, as these case studies will demonstrate.
2
We argue that we can learn a great deal about the nature and power of trolling and its
effects by using theories of the utterance – specifically speech act theory and implicature theory.
J.L Austin first introduced the distinction between utterances whose pragmatic function is to
describe a state of affairs, which he called constatives (3) and those that perform an action,
which he termed performatives (6). That is, to say or write something is not merely to produce a
sequence of words with certain meanings, but to perform a kind of act, which may then have
particular consequences. Austin (109) further distinguishes the act performed in saying
something (its illocutionary force) and the effect or result of the act (its perlocutionary force). In
relation to this paper, a troll may attempt to insult someone by calling them a ‘slut’ and the
person may indeed feel hurt or insulted (the perlocutionary act) or may merely laugh off the
insult as an obvious attempt to insult or provoke. Perlocutionary force is thus not predictable
from nor a direct consequence of the illocutionary act, and also cannot be judged based on the
grammatical form of the utterance; it is very much linked to the ripples of affect, as we’ll discuss
further below.
Numerous types of speech acts can be distinguished and these can be grouped into four
basic categories: statements, questions, offers and commands (Halliday 69). These categories
are based on the understanding that all utterances are performative in the sense that all
utterances, in social contexts of interaction, have illocutionary force. The purely constative act
which is a neutral statement of fact not designed to produce any corresponding response from
the hearer is a theoretical type which is rare in actual contexts of interaction, including online
discourse. Nevertheless, we argue that it is theoretically necessarily and analytically useful to
distinguish between speech acts that are primarily descriptive and those which are directed
towards inciting action. To understand the grammatical structure and performative nature of
trolling, we will consider the illocutionary and perlocutionary forces, not only of ‘obvious’
performatives (commands, offers, threats, and so on) but also various types of statements and
questions.
Trolls may make statements, ask questions, offer to perform services, or command
others to act. In many cases, these trolling utterances can be perceived as speech acts which
seem to flout, or perhaps ignore altogether, that which H. Paul Grice described as the
cooperative principle in conversation, and in discourse more generally.
3
Grice’s principle asserts
that participants in a dialogue will attempt to be as informative, truthful, relevant, clear and
unambiguous as they can in order to communicate effectively and to maintain relationships.
When these rules of communication are deliberately flouted, participants may deduce
alternative meanings since the literal meanings do not appear to comply with the cooperative
principle. These implicatures require participants to infer meanings other than those stated.
While it may seem that the language used in the three cases we are discussing is characterised
by a consistent violation of the cooperative principle, nevertheless these speech acts are clearly
meaningful, and so do take on the status of implicatures. For example, several of the utterances,
such as when Julia Gillard is told to ‘Get my dinner ready’, or Charlotte Dawson is instructed to

Persona Studies 2016, 2 (1)
31
‘put your head in the toaster’, are commands which are clearly not meant to be taken literally,
and therefore must be interpreted as insults.
Both are impossible acts: Julia Gillard is being addressed as down-trodden housewife in
relation to her interlocutor; Charlotte Dawson would find it impossible to place her head inside
a toaster. However, they form part of the conversational pact with the interlocutor that requires
an interpretation of their ‘relevance’. In both cases, they demand that the reader infer that these
women are not worthy of respect: the command to cook is used to denigrate Julia’s status as
Prime Minister; an act designed to deform her face, and thus her ability to carry on her career, is
prescribed for Charlotte.
To better understand the nature of such speech acts, we also need to consider the way
illocutionary force is bound up with affect. By affect, we mean the capacity of the body to affect
or be affected (Massumi, Parables 15). Following Spinoza, Brian Massumi further describes
affect as a ‘pre-personal intensity corresponding to the passage from one experiential state of
the body to another’ (Massumi, ‘Autonomy of Affect’ 222). We use it here particularly as it
relates to changes in emotional states, and include the body’s capacity to incite affects through
discourse.
While the concept of affect as it has been taken up by cultural studies is usually regarded
as outside signification and as non-discursive, Massumi does acknowledge the possibility that
affect can be discursively captured (see Gilbert 14). Atkinson and Yell argue that there is a
complex relation between affect and discourse – since utterances are enunciative acts which
emanate from bodies, even in cyberspace – and these bodies use discourse to react to bodily
experience and to transmit and evoke further affective states in themselves and others. Trolling
is a discursive act that provokes affects and incites further discourse and sometimes other
forms of embodied social action. We argue that the grammar of trolling speech acts cannot be
understood without consideration of its capacity to provoke affects. Indeed, this is the driving
force underpinning acts of trolling.
In the following we analyse both the types of affects and the types of speech acts
performed in these instances of trolling. The grammar of trolling utterances allows affect to be
linguistically ‘captured’ and transmitted or transmuted, with varied illocutionary and
sometimes perlocutionary force. Trolls use affective triggers including explicitly emotive terms
which present positive or negative evaluations, often of high intensity (‘fat’, ‘ugly’, ‘disgusting’),
as well as using expletives and abusive language or evoking scenarios which invite a strong
emotional response. These, when combined with directly performative utterances (e.g.
commands like ‘kill yourself’), produce powerful affective/perlocutionary force.
These affects are in turn further mitigated or enhanced by the affordances of different
social media platforms, such as YouTube, Facebook and Twitter. Twitter is a “microblogging
site” with a “directed friendship model” (Marwick and boyd 116) in which participants choose
to follow particular Twitter account holders. Facebook is also based on a directed friendship
model but its settings can be adapted to enable public or private access to content, and when
settings are public anyone can view or comment on the site’s content. Facebook sites can be
specific to an individual or used by a group or a public figure, enabling strategic presentation of
a persona which may be more or less detached from a private self. Both YouTube and Facebook
function as repositories or databases for content created by an individual or organization or
hosted on behalf of a third party (Paolillo 156). They have the additional feature of allowing
posts or comments to be added and to be visible long after the interaction has taken place. In
the case of Facebook, this content is usually directly related to the construction of a particular
persona, their tastes, interests, relationships and activities. Whereas YouTube, as a video

Morrissey & Yell
32
hosting site, is the least personal of the three platforms, often parasitic on other media (such as
movies, music videos and broadcast television) and not specific to an individual, unlike Twitter
and Facebook. All three social media afford public opportunities to participate in extended
dialogue (and thus opportunities for trolling) but in the case of Twitter, the short 140 character
tweets directly address the Twitter account holder and more closely approximate real-time
conversation whereas posts on Facebook and YouTube enable lengthier posts and can be
responded to over a period of hours, days or even months, and thus lack the rapid-fire
immediacy of Twitter.
CASE STUDY 1: MAGDA SZUBANSKI ON YOUTUBE
When Magda Szubanski publicly ‘came out’ on TV panel show The Project in February
2012 a clip was immediately posted on YouTube, where it rapidly accumulated a large number
of comments, and comments continued to be added over a period of some months. Amongst
apparently genuine attempts to debate Australian government policy and cultural attitudes to
same-sex marriage, and messages of personal support for Magda’s courage in expressing her
own sexuality and her support for gay marriage, were inflammatory statements denigrating
homosexuals in general (particularly lesbians), and homophobic insults directed at Magda.
The reaction of these YouTube trolls was perhaps unexpected, given Magda’s well-
known and well-loved public persona. She is possibly best known for her work on the television
series Kath and Kim, where she played the loveable, but gormless, Sharon, forever on the
receiving end of Kim’s jokes and aspersions. Magda previously had a well-known comic persona
in other shows as well, such as Fast Forward and Big Girl’s Blouse, where she once again played
likeable characters. Her acting career extended to playing Esme Hoggett in the Babe films, which
were internationally popular and the science fiction series Farscape. On top of her acting career,
Magda was also a spokesperson for Jenny Craig as she publicly battled her weight problems,
with the support of the Australian public. In short, Magda had developed a warm, funny,
charming public persona by the time she decided to reveal her sexuality. Even when she did,
this was not considered much of a ‘scoop’ in journalistic terms. As the Sydney Morning Herald
noted, it was ‘one of the entertainment world’s worst kept secrets’ (Quinn). Yet, the trolls were
vitriolic in their abuse, leading to the inescapable conclusion that it wasn’t Magda’s confession
they were concerned with in particular, but instead, the notion of lesbianism in general and the
idea of anyone ‘coming out’ at all.
The trolls were primarily concerned with the existence of any lesbians in their midst,
which is evident in their utterances that predominantly took the grammatical form of
descriptive third person statements – they were about Magda (or gays/lesbians in general)
rather than being addressed directly to her. The only time the second person (‘you’) was used in
a statement was to issue a generalised (future) threat to supporters of homosexuality (‘Magda is
going to hell and so are all you fag enablers’), albeit one not to be carried out by the speaker but
by an unnamed higher power. Some included the use of the first person to project the speaker’s
opinion (‘Wow I’m shocked… NOT’, and ‘I notice lots of fat cake eaters are carpet munchers’).
Such statements violate Grice’s principle for the speaker to be truthful (i.e. sincere) and
unambiguous, requiring other participants on the YouTube site to read the first example as
sarcasm and the second as an insult (and one in which the participant must be complicit in
recognising and decoding the metaphor of a ‘carpet muncher’).
Within the ongoing ‘conversation’ on this YouTube comments page, these evaluative
statements are speech acts which are designed to incite responses – the posters are not making
a statement as a disinterested contribution to knowledge, but an affectively loaded expression

Persona Studies 2016, 2 (1)
33
of opinion, designed to incite equally affectively loaded responses. The evaluative terms are
strongly negative homophobic terms of abuse (‘ugly dyke’, ‘carpet muncher’, ‘carpet licker’),
insults based on physical appearance (‘fat ugly unfuckable pig’, ‘a whale like her’, ‘disgusting to
look at’, ‘disgusting overweight bitches’) and include generic statements such as ‘no man in his
right mind would want to fuk [sic] a whale like her’.
These insults are perverse but also predictable and banal. The use of explicit
homophobic language is crude, unskillful and reductive, with its references to appearance, body
weight and physical attractiveness and the use of animal terms. Magda Szubanski was being
informed that her ugliness would preclude heterosexuality – effectively an implicature asking
readers to infer boundaries and makes specific that Magda is to be reviled as an outsider. The
language expresses primal affects of revulsion, disgust and contempt. While these statements
undoubtedly have the potential to hurt and contribute to a discursive universe in which
homophobia continues to exist, their reflexive rehearsal of primitive homophobia seems to
blunt the affective force of these utterances. One potential reading of these utterances is that the
trolls are taking a position just to get a reaction through such ‘ludicrous rants’ (Narraine 146).
However, the trolls gained little traction or support for their trolling. Some posters
responded to the trolls’ comments, but if anything, the trolls incited increased positive
responses in support of gay marriage and in praise of Magda Szubanski. Furthermore, there is
no evidence of any perlocutionary effects on the primary subject of these trolls, Magda herself,
who is not a participant in this conversation and apparently unaffected by the existence of these
comments. The effects (and affects) of the trolling are mitigated by the affordances of YouTube,
in that its primary function is as a video hosting site and it is not personal to an individual.
Furthermore, because both the original content (in this case the footage from Magda’s
appearance on The Project) and the sequence of comments on this, are archived on the site, the
debate does not necessarily occur in real time, nor does it necessarily involve the denigrated
individual as an interlocutor. Finally, the pace of the debate can be extended over days or
months, again diminishing its affective intensity as a concentrated onslaught of comments. In
terms of the definition of trolling, this is an example of ‘unsuccessful trolling’ (for the trolls), and
perhaps offers some support for the notion that trolling can indeed make a positive contribution
to internet discourse, since in this case it activated others to articulate and defend their views in
opposition to discursive positions that were intolerant, unreasonable or inflammatory.
CASE STUDY 2: JULIA GILLARD ON FACEBOOK
In October 2012 then Australian Prime Minister Julia Gillard engaged in what was
touted as a ‘first for a federal politician’ (Hall), conducting a live chat session on Facebook with
the aim of discussing education policy and the national curriculum with members of the public.
For approximately an hour the debate ran smoothly, with questions and answers on various
education-related topics, however the session then descended into abusive trolling, with the
most offensive remarks being posted after the Prime Minister had logged off. By ‘invading’ what
was intended as a civil and democratic dialogue on the Prime Minister’s Facebook page, these
trolls shifted the debate from public discourse to a highly personal attack. Unlike the trolls
targeting Magda Szubanski on YouTube, these trolls were on a site hosted by and for
communication with a specific person, albeit a public figure in her official capacity. The
institutional context in which these speech acts were performed is public but directly targeted,
thus sharpening their potential to enact not only illocutionary but perlocutionary force. Who is
meant to be affected and acted upon is quite explicit.

Morrissey & Yell
34
Julia Gillard, Australia’s first female Prime Minister, was a divisive figure in Australian
public life. Holding high public office makes very particular requirements of women in terms of
persona management and gender identity. As Anne Cranny-Francis notes, “female politicians
constantly deal with attacks that challenge their femininity rather than their policies and
values” (26). What was publicly known of Gillard’s private life presented a persona lacking in
stereotypically feminine qualities – she was in a de facto relationship with a (male) hairdresser,
childless, and apparently not skilled in the domestic arts (much was made of the empty fruit
bowl photographed in her kitchen). In public life, she was a strong feminist politician, heading
up a minority government where she was required to negotiate the balance of power with
various independents. She was an object of hatred and vilification by right-wing groups, as
expressed in the infamous “Ditch the witch” and “Juliar” signs held by protesters at a March
2011 rally against her government’s planned introduction of a carbon tax. Yet Gillard was also
much admired, particularly by feminist groups, for her strong and explicit feminist principles.
The very same day the trolling incident took place, Gillard made her famous ‘misogynism’
speech in the Australian parliament (October 8, 2012), in which she defended herself and
attacked the Opposition Leader, Tony Abbott, for his gendered double standards (“Gillard labels
Abbott a misogynist”). Subsequently anti-feminist discourse against her has lost considerable
ground as a legitimate form of attack, most particularly in the light of the immense public
support not only within Australia but worldwide, ironically garnered substantially via another
social medium, YouTube, where her speech had 2,764,308 views by February 2016 (with
32,243 likes and 2,029 dislikes).
Gillard’s public persona was, in October 2012, one that had already been the target of
abuse and personal attacks. In the trolling of her Facebook forum on education, the context was
specifically a dialogue with Gillard’s public persona, however this didn’t prevent the trolls from
attempting to target her in quite personal ways. The grammatical form and types of speech act
employed demonstrate an escalated kind of trolling in comparison to our first case study. In this
instance, the trolling utterances included not only constative but performative utterances, using
the second person (‘you’) and directly addressed to Julia Gillard. In what constitutes one of the
milder insults, one poster stated ‘Hopefully when you say “I have to go now” its [sic] back to
your home planet’. The attacks drew on anti-feminist and misogynist rhetoric, some of it
amounting to nothing more than primitive name-calling, such as the drawn out hissing insult
‘SSSSSSSLLLLLLLLLLLLLLLLUUUUUUUUUUUUUUTTTTTTTTTTTTTTT!!!!!!!!!’). This utterance,
while expressing a strong affect of sexual contempt, does not require any action from the Prime
Minister.
Likewise, the trolls also asked rhetorical questions, which expect no response but have
the illocutionary force of insults because of the implicatures they generate. Asking Julia Gillard
‘How’s your dad?’ was deliberately cruel and deeply offensive given the wide public knowledge
that the Prime Minister had recently lost her father – and that talkback host Alan Jones had
made an after dinner speech castigating Julia Gillard by saying her father ‘died of shame’ at his
daughter’s lies. It fails the Gricean principle of relevance at a literal level (the question is quite
unrelated to the stated topic of education), but has strong relevance and affective resonance in
relation to these recent events, which would have been well-known to all accessing the
Facebook site. Another poster asked ‘Are your pubes as radiant, shiny and glorious as mine?’ –
again, not a sincere request for information, but a boundary-violating and lubricious speech act,
with the illocutionary force of an insult.
These statements and questions, while sexist, derogatory and insulting, have no
particular perlocutionary force, although they do have strong affective potential, and contribute
to the overheated environment of the Facebook page. Some trolls did address direct commands

Persona Studies 2016, 2 (1)
35
to Gillard, including telling her to ‘Lube up’, ‘Get my dinner ready’ and ‘McPiss off’ (Hall, Whyte).
‘Lube up’ and ‘Get my dinner ready’, using sexist discourse to put down the Prime Minister and
attempt to assert the primacy of phallocentric power over political power.
4
Here again the trolls are violating the Gricean principle of relevance, in order to
generate the implicature that it is Gillard’s gender which is of overriding relevance, not her
political performance or her education policies. All of these commands, issued apparently by
predominantly male trolls, assert the speaker’s power over the addressee and their supposed
right to tell this upstart woman what to do. But in reality Julia Gillard is never going to cook one
troll’s dinner for him, nor perform sexual acts for another; these utterances confuse the public
persona with the private individual. The commands are ultimately hollow illocutionary acts
with little perlocutionary force.
The trolls on Julia Gillard’s Facebook page were enacting a specific form of anti-
feminism, and in the ensuing media coverage and online debates men, as well as women, joined
in condemning the trolls, while others attacked Gillard for having a ‘glass jaw’ and suggested
‘playing the victim is working wonders for her poll results’ (comments posted to Crikey.com,
Whyte).
The Gillard Facebook debate constitutes an example of partially successful trolling.
While Facebook as a social media site enables dialogue directly with the site’s ‘owner’ as well as
hosting content, it was being used in this case for a managed debate which unfolded over a
specific and limited time-frame as an event in a politician’s public engagement schedule. The
affordances of Facebook enabled comments to be archived and therefore on the record for
intended target (Gillard) to read later but also allowed for the most offensive of these posts to
be removed by her staff, meaning she may not have seen them. The trolls’ attacks did succeed to
some extent in derailing the Facebook Education debate, but the strong affects of contempt
articulated by the trolls’ speech acts did not appear to have a direct effect on their target, who
was in any case constituted through a public performance in her persona as Prime Minister,
separate from the private individual. Their ineffectiveness was compounded by the fact that
most were made after she had logged off the site, perhaps an indication of the false bravado of
many trolls, who use the online environment to act out aggressive behaviour they would never
display face-to-face. As in the Magda Szubanski case, the trolls’ offensive and sexist speech acts
did not succeed in the perlocutionary act of creating a ‘victim’, but rather incited a number of
positive responses and expressions of support.
CASE STUDY 3: CHARLOTTE DAWSON ON TWITTER
On August 30, 2012, Charlotte Dawson was admitted to hospital, following an overdose
of prescription medication. This desperate act was the culmination of weeks of online
aggression on her Twitter account, during which her battles with depression were made clear.
Even more pressing, in terms of her suicide attempt, were the several hours of severe abuse
prior to the act, during which Dawson was repeatedly told to ‘go hang yourself’, ‘kill yourself’,
‘put your head in the toaster’; described as a whore and as a ‘sad, ugly moll’; and sent images of
bloodied corpses in an effort to make her understand her persecutors meant business (Hornery
and Hall, Baird, Tarasov). At last, Dawson gave in to the will of the mob, posting a picture of a
hand holding pills and two messages: ‘You win x’ and ‘Hope this ends the misery.’
Dawson’s final tweet on that day demonstrates the extraordinary impact anonymous
malicious trolling can have. For, although the vicious attack began only after Dawson managed
to name and shame one of her abusers, Tanya Heti, the rest of the trolls remain unaccounted for.
Heti infuriated Dawson days before when she tweeted that she could understand why the

Morrissey & Yell
36
husband of a follower of Dawson would have committed suicide (Baird). She also responded
with the message: ‘On behalf of New Zealand we would like you to please go hang yourself’, after
Dawson had described her native country as ‘small, nasty and vindictive’ (AAP, ‘Posts put in
perspective’). Dawson discovered the university Heti worked for, and had her suspended from
employment. At this point, the real barrage of anonymous threats, insults and commands began.
Dawson is an interesting figure in any discussion of malicious trolling. No stranger to
conflict herself, as the well-publicised ‘mean’ judge on Australia’s Next Top Model, she insulted
would-be models for a living. For some, her performance of this role was enough to deny her
sympathy, or to at least temper it (Hornery). However, to engage in this form of debate is to
misrecognise the fundamental differences between a public and a private persona. Dawson was
a public figure, playing a part; she cannot be simply and unproblematically collapsed into a
cardboard cut-out caricature of herself.
Many trolls argue that their online abuse is playful (Insight, Johnston). If we look at the
structure of the utterances in Dawson’s case, it becomes very hard to believe that these attacks
could have been anything other than serious. The vast majority of the statements made to
Dawson were clearly perlocutionary performative utterances; they were commands to take
action. As the crescendo built, and the mantra ‘hang yourself’ was repeated, the commands
gathered an urgency demanding immediate response. This command is hardly subtle, and
requires no inference from the reader. It cannot be argued that this injunction is merely acting
as an insult, desiring nothing more than that the reader leaves the writer alone. The addition of
images of corpses, making visually evident the expected and required outcome, put paid to that.
Dawson is told what to do, and preferably, how to do it. There is no flouting of the
conversational principle here; the message is self-evident.
Other speech acts directed at Dawson are more problematic. Dawson was told to ‘put
your head in the toaster’, and described as ‘fat’, ‘ugly’ and a ‘sad moll’. The first command
demands an impossible act that, were it indeed possible, would destroy Dawson’s face. Because
it is literally impossible to carry out, it therefore generates an implicature. Flouting the
conversational principle of clear, unambiguous speech, this strange command demands
inference on the part of the reader that Dawson should not only die but be disfigured in her
death. There is clear reference here to her career as a judge on a show evaluating the
performance of models. The constatives on Dawson’s appearance disparage her looks and thus
challenge her right to speak at all regarding the looks and the style of others. The private
individual Charlotte Dawson has clearly been confused and conflated with her job here; her
public persona becomes her only persona and her performance of the performative ‘mean
judge’ is equated with her personality.
This kind of attack, especially on women, is hardly new. Dawson herself had been the
target of such vitriol many times before. On this occasion, however, she allowed it to affect her
to such a degree that she took her pain out on her own body. Suddenly, the rude constatives and
the appallingly violent performatives produced affect in the off-line world. In discussing why
this should be the case, it is necessary to consider the discursive context of the statements, and
the medium in which they were made, to determine why they had such affective intensity.
Dawson was a self-proclaimed avid Twitter devotee: “I like Twitter. I like the sense of
community it creates, the ideas it gives us access to, its ability to make people feel closer
intellectually and personally than they might otherwise be geographically or demographically.”
(Byrnes). Even after her attack, she had no intention of stopping her use of the medium and
indeed did not until her suicide in February 2014.

Persona Studies 2016, 2 (1)
37
This statement, and her devotion to Twitter, lead us some way to discovering the
reasons for her response to her abusers. Twitter is a conversational medium; fast moving,
information can be shared back and forth with the speed of an off-line conversation. It doesn’t
have all the baggage of Facebook, which asks for an enormous amount of information on its
users, and encourages them to keep up their profile by adding posts and attachments that are
kept in a relatively static environment, so that it is easy to follow a thread from weeks prior.
Twitter, on the other hand, is all about the present, demanding its followers keep up with
current developments at all times, insisting on responses constantly. Twitter is designed to be a
relatively simple form of social media, allowing its participants the feeling of being in the same
time as those they follow, regardless of their geographical distance. It is the very nature of the
medium, and the spare nature of the conversation it encourages, which creates its extraordinary
affective intensity. No-one in the off-line world would ever have conversations such as those
held on Twitter, where simple declarations or commands take the place of extended discourse.
On Twitter, phatic communication is cut to a minimum, resulting in a world where social cues
are frequently lost.
On August 30, 2012, Charlotte Dawson received a string of these simply worded
messages, exhorting her to commit suicide in no uncertain terms. The phrase ‘hang yourself’
was repeated so often it became an act in and of itself: to write ‘hang yourself’ was to commit an
act, to ‘hang’ Dawson both metaphorically in stopping her voice for a time, and literally in its
clear demand which was ultimately followed through, although not to the letter. The closeness
of the Twitter world was shattered for Dawson by those words. The community she had enjoyed
engaging with turned against her, the intimacy she had experienced evaporated. The discursive
context of Twitter, then, with its swift, unambiguous messages created such intensity of affect in
Dawson that she overdosed. Yet, even in extremis, she reached out, yet again, with her tiny
suicide note: ‘You win xx’.
CONCLUSION
In the three cases examined in this paper, the trolls conformed to certain stereotypical
behaviour. Firstly, they chose to pick on a woman and a celebrity; two of the prime sufferers of
vitriolic abuse online, according to Jason Wilson (cited in Hornery and Hall). Secondly, most of
them prized their assumed anonymity, believing that this privilege guaranteed them free reign
to say whatever they pleased. For, under the blanket of anonymity, trolls are able to ignore
social constraints and write things they would never normally be able to say in the off-line
world (Johnston).
While the cases share some similarities, we have demonstrated that the effectiveness of
trolling differs with respect to the performative and affective nature of the utterances in each
case. Using our three case studies, we have demonstrated a continuum of speech acts, from the
constative and primarily illocutionary, to the performative and perlocutionary. In doing so, we
have shown how the impact of the perlocutionary, combined with specific discursive contexts,
can lend the utterance greater and greater power.
In the Szubanski case, the trolling took place on YouTube in response to a clip put up by
producers of The Project. The trolls did make rude comments about Szubanski, but they were
merely descriptive and generic. There is also no evidence that Szubanski even saw these
comments as she did not maintain the site. Here we have an example of highly ineffective
trolling. The utterances are generally constative, and thus have only illocutionary force. There is
no demand for Szubanski to act in a particular way. The trolling of Julia Gillard’s education
forum on her Facebook site is only partially effective. The trolling took place after she had

Morrissey & Yell
38
logged off, so was never seen by her in real time. Consisting of mostly garden-variety, constative
insults and illocutionary rhetorical questions and commands, the comments lacked any
relevance to the topic of conversation, and thus took on no more significance than the posturing
adopted by playground bullies.
Nevertheless, action was taken: the posts were removed, and the Prime Minister began
targeting trolls in an attempt to work out ways to catch them (Matheson). Yet, as with
Szubanski, Gillard herself did not appear outwardly perturbed by the ludicrous comments
aimed at her public persona. This is due to the relative lack of intimacy possible on a Facebook
page dedicated to politics, and to the inescapable fact that she elicited such messages because of
her high status and public role, which undercuts the effectiveness of the trolls’ attempts to
humiliate her. In the Charlotte Dawson incident, we see perlocutionary demands made
repetitively and at fever pitch, insisting on a response. The discursive context of Twitter,
perhaps the most intimate of the social media, contributed to increase the effectiveness of the
statements and commands. Dawson maintained her own page and she saw every post as it came
through, in real time. Eventually, the trolls succeeded, and Dawson attempted suicide in the off-
line world.
It is evident, then, that discursive context is essential in determining affective intensity
and perlocutionary consequences: the more intimate a medium, the more results trolls can
expect from malicious messages. The form of the medium is important as well; slower moving
social media sites, such as YouTube and Facebook, are often read asynchronously which, in
these instances at least, tended to distract from their power. Twitter, on the other hand, with its
extraordinarily swift turnaround, demands constant attention, and thus an online campaign can
be devastatingly destructive. The affordances of social media also enable greater or lesser
degrees of separation between a public figure and their audience; from a more public and
managed persona to a relatively intimate and private ‘self’. Finally, the structure of the
utterance differs on each medium, from the relatively lengthy messages possible on YouTube
and Facebook, to the short, sharp commands and comments available on Twitter. The sense of
intimacy on Twitter is also problematic for those subjected to trolling, as it encourages a sense
of community through its real-time messaging which is as close to face-time as it is possible to
get. The loss of this community can have dire consequences, as Dawson observed after her 2012
suicide attempt.
To conclude, malicious trolling is both possible and effective on all three of the social
media platforms we looked at in this paper. However, the posts which have the greatest
effectiveness, and especially those that create trouble in the off-line world, combine intensely
personal and specific insults and commands with a page on a site that the victim considers
particular to them, and which is experienced by the victim in real time. The more communal a
page becomes, the less likely that trolling will be effective in a personal sense to an individual
victim. The more perlocutionary and performative an utterance is, the more likely it will create
long-lasting effects on its target. In the final degree, as we have demonstrated in this paper,
malicious trolling can potentially lead to the death of the victim. Never have mere words been
more powerful than in the virtual societies we now inhabit.
1
Australia’s first conviction for internet trolling was in 2011, and involved the defacing
of Facebook tribute pages with child pornography (Keim). The Australian Criminal Code Part
END NOTES

Persona Studies 2016, 2 (1)
39
10.6 covers offences committed using the internet, including using a carriage service to menace,
harass or cause offence (The Findlaw Team).
2
In 2015, Lena Durham, creator and writer of the HBO series Girls, chose to delegate the
management of her Twitter account because of abusive tweets directed at her (Milsom).
3
Grice entitled his original paper “Logic and conversation”, however his arguments
apply to all dialogic communication, whether synchronous or asynchronous, as it accounts for
the inferences made by the addressee or audience (in the case of YouTube and Facebook this
includes those reading posts “after the event”).
4
The futility of this gesture was duly noted and commented on by others: “self-
emasculating morons looking for a group hug” and “Some of you blokes need to have your right
to vote revoked”, comments posted on Crikey.com (see Whyte).
WORKS CITED
AAP. “Posts put in perspective.” Newcastle Herald 31 Aug. 2012: 3. Print.
Atkinson, Paul & Yell, Susan. “Affect, time and the enunciative body.” Southern Review 38.2
(2006): 40-57. Print.
Austin J. L. How To Do Things With Words. Ed. J. O. Urmson. Oxford: Clarendon, 1962. Print.
Baird, Julia. “Inhuman, abusive trolls deserve exposure.” Sydney Morning Herald 1 Sept. 2012.
Web. 8 Feb. 2016.
Butler, Judith. Bodies That Matter. New York: Routledge, 1993. Print.
Byrnes, Holly. “Charlotte Dawson won’t be silenced by Twitter trolls.” The Daily Telegraph
[Melbourne] 11 Sept. 2012. Web. 8 Feb. 2016.
Cranny-Francis, Anne. The Body in the Text. Carlton South: Melbourne University Press, 1995.
Print.
Dahlberg, Lincoln. “Computer-mediated communications and the public sphere: A critical
analysis.” Journal of Computer-Mediated Communication 7.1 (2001):1-26. Web. 5 Feb.
2016.
Donath, Judith S. “Identity and deception in the virtual community.” Communities in Cyberspace.
Ed. Marc A. Smith and Peter Kollock. London: Routledge, 1999. 29-59. Print.
“Gillard labels Abbott a misogynist.” ABC News (Australia). 8 Oct. 2012. Web. 9 Feb. 2016.
Grice, Elizabeth. “The sad, secret viciousness of the village troll.” Sunday Age [Melbourne] 12
Oct. 2014, Extra: 22. Print.
Grice, H. Paul. “Logic and conversation.” Syntax and Semantics. Ed. P. Cole and J.L. Morgan. New
York: Academic Press, 1975. 41-58. Print.
Hall, Bianca. “A first for the PM but abusers take advantage.” Sydney Morning Herald 9 Oct. 2012.
Web. 5 Feb. 2016.
Halliday, M.A.K. Introduction to Functional Grammar. London: Edward Arnold, 1985. Print.
Hardaker, Claire. “Trolling in asynchronous computer-mediated communication: From user
discussions to academic definitions.” Journal of Politeness Research 6.2 (2010): 25-242.
Print.
Herring, Susan, Kirk Job-Sluder, Rebecca Scheckler, & Sasha Barab. “Searching for safety online:
Managing “trolling” in a feminist forum.” The Information Society: An International
Journal 18.5 (2002): 371-384. Print.
Hornery, Andrew. “Dawson fights demons, trolls.” Sydney Morning Herald 1 Sept. 2012. Web. 8
Feb. 2016.

Morrissey & Yell
40
Hornery, Andrew and Bianca Hall. “Twitter torment: TV personality taken to hospital after
abusive online attacks.” Sydney Morning Herald 31 Aug. 2012. Web. 8 Feb. 2016.
Huffaker, David A. & Sandra L. Calvert. “Gender, identity, and language use in teenage blogs.”
Journal of Computer-Mediated Communication 10.2 (2005): np. Web. 5 Feb. 2016.
Johnston, Megan. “The devil inside.” Sydney Morning Herald 22 Sept. 2012. Web. 8 Feb. 2016.
Julia Gillard – Politician. Facebook. 9 Oct. 2012. Web. 16 Jan. 2013.
Keim, Tony. “Internet ‘troll’ Bradley Paul Hampson refused bail pending his appeal.” Courier
Mail [Brisbane]. 3 May 2011. Web. 4 Feb. 2016.
“Magda Szubanski on Gay Marriage – The Project.” 14 Feb 2012. Online video clip and
associated posts. YouTube. Web. 10 Dec. 2012.
Marshall, P. David and Kim Barbour. “Making intellectual room for Persona Studies: A new
consciousness and a shifted perspective.” Persona Studies 1.1 (2015): 1-12. Web. 20 Apr.
2016.
Marwick, Alice E. and danah boyd. “I tweet honestly, I tweet passionately: Twitter users, context
collapse and the imagined audience.” New Media Society 13 (2011): 114-133. Web. 20
Apr. 2016.
Massumi, Brian. “The Autonomy of Affect.” Deleuze: A Critical Reader. Ed. Paul Patton. Oxford:
Blackwell, 1996: 217-239. Print.
Massumi, Brian. Parables for the Virtual: Movement, Affect, Sensation. Durham: Duke University
Press, 2002. Print.
Matheson, Melissa. “Julia Gillard closing cyber troll net.” The Daily Telegraph [Sydney], 17 Jan.
2013. Web. 8 Feb. 2016.
Narraine, Ryan. “The 10 biggest web annoyances.” PC World (Dec. 2007):141-148. Print.
Paolillo, John C. “Structure and network in the YouTube core.” Proceedings of the 21st Hawaii
International Conference on Systems Sciences. 7-10 Jan. 2008: 156. Web. 21 Apr. 2016
Quinn, Carl. “Gay-gay-gay-gay – Magda comes out” Sydney Morning Herald 15 Feb. 2012. Web. 4
Feb. 2016.
Rafferty, Rebecca S. “Motivations behind cyber bullying and online aggression: Cyber sanctions,
dominance, and trolling online.” Master of Arts thesis, College of Arts and Sciences, Ohio
University, 2011. Web. 5 Feb. 2016.
Starke, Petra. “Julia Gillard the target of abuse on Facebook after trolls hijack live chat.” Herald
Sun, 9 Oct. 2012. Web. 14 Jan. 2013.
Tarasov, Anne. “Dawson: I will recover.” Sydney Morning Herald 2 Sept. 2012. Web. 8 Feb. 2016.
The Findlaw Team. “Internet Trolls can be prosecuted under Australian law.” Thomson Reuters
(Professional) Australia. 2016. Web. 5 Feb. 2016.
“Trolls.” Insight. SBS. Host Jenny Brockie. 16 Oct. 2012. Television broadcast.
Turner, Tammara Combs, Marc A. Smith, Danyel Fisher and Howard T. Welser. “Picturing
Usenet: Mapping computer-mediated collective action.” Journal of Computer-Mediated
Communication 10.4 (2005): np. Web. 5 Feb. 2016.
Van Zoonen, Liesbet. Feminist Media Studies. London: Sage, 1994. Print.
Whyte, Sally. “Gillard cops sexist spray in Facebook live chat.” Crikey 8 Oct. 2012. Web. 12 Jan.
2013.