Repurposing Emoji for Personalised Communication:

Why 🍕 means “I love you”

Sarah Wiseman Sandy J. J. Gould
Goldsmiths, University of London University of Birmingham
London, UK Birmingham, UK
[email protected] [email protected]
ABSTRACT images, rather than text, has tempted some researchers to
The use of emoji in digital communication can convey a claim that emoji represent a type of ubiquitous language,
wealth of emotions and concepts that otherwise would take capable of being understood by people from varying
many words to express. Emoji have become a popular form languages and cultural backgrounds [11]. This claim has
of communication, with researchers claiming emoji been questioned by others, as evidence suggests that there
represent a type of “ubiquitous language” that can span are significant differences in the patterns of emoji use
different languages. In this paper however, we explore how between countries, languages, and cultures [18,19].
emoji are also used in highly personalised and purposefully Moreover, even if emoji usage and frequency were similar
secretive ways. We show that emoji are repurposed for between groups, a common alphabet does not mean that the
something other than their “intended” use between close emoji will be used in similar ways or retain their meaning
partners, family members and friends. We present the range between groups.
of reasons why certain emoji get chosen, including the
concept of “emoji affordance” and explore why repurposing Critics of the ubiquitous language theory highlight that
occurs. Normally used for speed, some emoji are instead cultural differences may lead interpretations of emoji to vary
used to convey intimate and personal sentiments that, for [13]. Emoji are images, and are presented without textual
many reasons, their users cannot express in words. We descriptions (although each emoji has a formal, Unicode-
discuss how this form of repurposing must be considered in assigned description). This ambiguity can cause emoji to
tasks such as emoji-based sentiment analysis. have multiple differing interpretations from culture to
culture.
Author Keywords
Emoji, Computer Mediated Communication; In November 2016, Apple device users were upset when the
Personalisation; Digital Intimacy company changed the rendering of the Peach emoji [10].
People were angry because they had assigned a meaning to
ACM Classification Keywords the emoji that that no longer made sense after the emoji had
H.5.m. Information interfaces and presentation (e.g., HCI): been redrawn. As the website Emojipedia reported, the Peach
Miscellaneous emoji (🍑), when used in messages or in social media, was
INTRODUCTION more likely to refer to buttocks than to the fruit [1]. By
Emoji are a popular way to communicate over digital media, altering the way the emoji was visually rendered, users felt
with Swiftkey (a mobile phone keyboard company) that Apple had removed their ability to assign a second
estimating that 6 billion messages containing emoji are sent meaning to the symbol. Emoji are inherently flexible in their
every day using their keyboard [via 24]. In 2015, the Oxford semantic meaning. The Unicode consortium, which
English Dictionary declared, for the first time ever, a determines which emoji will be brought into use,
“pictograph” as their word of the year – the Face With Tears acknowledges this fact and states on their website that emoji
of Joy emoji (😂) [15]. The website emojitracker.com, which are “encoded in Unicode based primarily on their general
updates the number of emoji being used in real time on appearance, not on an intended semantic” [22].
Twitter, also places the Face With Tears Of Joy smiley as the In this paper, we investigate the idea that varying
most popular emoji on the microblogging site. The use of interpretations can lead to emoji being assigned different
meaning. We focus specifically on usage at small scales,
looking at the “micro cultures” that exist between partners,
friends and family. We investigate how certain emoji are
repurposed between small communities, for example two
friends, and given additional or alternative meanings which
exist solely within those micro cultures.
RELATED LITERATURE The literature shows that emoji take on manifold meanings
Understandability of emoji depending on both the sender and receiver’s perception of
The variable meaning of emoji can cause potential problems the emoji, and is further complicated by divergent renderings
between users communicating via digital media. One key of emoji on different platforms. This flexibility of semantics
issue is that of understandability. Miller et al. [13] highlight is acknowledged by the Unicode Consortium who, in
a theory of psycholinguistics that states that issues of response to a question regarding the specific meaning of
misconstrual can occur between a speaker and addressee if emoji state “In fact, when used as emoji, many of these
the addressee does not successfully interpret the speaker’s characters acquire multiple meanings based on their
message. This problem is heightened when communicating appearance”[22].
using emoji, which have no set meaning – or whose meaning Cultural variability in emoji use and understanding
is presently developing. Having Unicode tags assigned by Prior to emoji, the emoticon was another visual method of
committee does not guarantee that the ‘intended’ semantic expressing an emotion or non-verbal cue [23]. Emoticons,
meaning of a given emoji will be shared by senders and unlike emoji, were created by the user from punctuation
recipients. After all, it is likely that most people never marks and symbols (an example being ‘:)’ to represent a
encounter the formal definitions the Unicode Consortium smiley face). A study of emoticon usage across cultures
assign to each emoji. highlighted a difference in emoticon style between Asian-
The problem of shared understanding is exacerbated by and European-countries, with Japan and South Korea
variations in how emoji are rendered. Although the Unicode preferring to use vertical emoji (e.g., ‘^_^’) and European
consortium name the emoji and assign them a code, it is up and English speaking countries preferring horizontally
to individual platforms to decide how emoji are rendered. aligned emoji (e.g., ‘:)’) [16].
Currently the website Emojipedia lists nine different Cross-cultural studies have been conducted with the use of
versions of each emoji, each from a different operating emoji, too. Lu et al. [11] conducted a study of emoji usage
system or device manufacturer. This can result in serious across 212 countries, gathering data on how 3.88 million
miscommunication when two users are sending messages users use emoji in their messaging. This data highlighted
using different technology1. Studies have shown that the differences between countries in terms of emoji use. For
meaning of an emoji, as assessed by human raters, can example, whilst, on average, emoji were used in 7.1% of
change depending on the platform in which that emoji has messages sent during the study across all countries, in France
been displayed [13,21]. The valence of the same emoji can the average number of messages containing emoji was
vary from positive to negative depending on how it has been 19.8%. It was not just frequency that varied between
rendered. countries, but types of emoji – the top ten most frequent
However, it is not only the variation in rendering that can emoji were different for the top 10 countries by number of
participants. However, all of these countries used emoji
lead to differing interpretations of the same emoji. People
relating to face, hands and hearts most commonly.
ascribe widely varying sentiments to emoji that are
identically rendered [13,21]. This highlights that sometimes There does appear to be a link between the way in which
the assessment of sentiment is based solely on the viewer’s emoji are used and the language spoken in the country. Lu et
individual interpretation. Establishing this kind of shared al. [11] showed that there are great similarities in the patterns
understanding – and the confusion that arises when it does of emoji used by countries that speak the same language.
not exist – is a common challenge in interactional contexts, However, this also related to geographic proximity, as the
whether they be animations [8] or gestures [4]. Of course, emoji patterns between Brazil and Portugal were not as
these ambiguities often exist in written language too. Indeed, similar as those between Brazil and other South American
emoji are often used to make up for the difficulties in countries.
communicating emotion and intent in short text messages
[9]. This literature suggests that there is a high level of variability
between cultures in how emoji are used. Culture can also
The work we have considered so far assessed emoji on their affect how an emoji is understood [3]. When translating
own, with no textual context. Additional work has shown between two languages, translators need to be aware of how
that it is also possible to misunderstand the intention of an certain emoji are interpreted in different countries – for
emoji even when paired with text [12]. This makes using
emoji during conversations an interaction full of nuance that
often requires conversational partners develop a shared
understanding [6].

1
https://www.engadget.com/2014/04/30/you-may-be-
accidentally-sending-friends-a-hairy-heart-emoji/
example where Crying Face (😢) can mean sadness in one 3. With whom do you use the special emoji? Multiple
culture, in another might mean sleeping2. choice answer with ‘other’ free-text option
4. Which emoji is it? Short text response
Technology mediated personal communication
5. What does that emoji mean? Short text response
Much of the emoji literature has focused on cross-cultural
6. Why do you use this emoji? Short text response
and cross-language differences in meaning and
7. Why do you have an emoji shorthand for this thing?
understanding at national and continental scales. However, Short text response
technology mediated communication platforms are often
used to create very personal and private places for None of the questions were compulsory. If a participant
communication [14] meaning differences in usage can be answered “No” to the Question 2 then they were thanked for
observed in much smaller communities. Previous work has the time and the survey ended, they were not asked any
found that emoji can play and important role in the digital further questions.
creation and maintenance of personal relationships. Kelly
For Question 4, simple advice was given to participants on
and Watts explored how people “appropriate” emoji for
how to bring up the emoji keyboard on a Mac device
purposes other than expressing emotion, show that people
(ctrl+⌘+space). Entering emoji on Windows and Linux
use them to simply maintain connections with another
person, to add a playful element to communications, or to devices is more complicated and so participants using these
create a “shared uniqueness”, for example by telling stories platforms were advised to type a description of the emoji if
generated from randomly chosen emoji [9]. However, this they were unable to locate the emoji pictogram when
behaviour was reported from a small proportion of responding to the survey. Of the 72 valid responses, 17
participants and the usage was ephemeral; the meaning of participants reported the emoji in text description form.
certain emoji did not permanently change. Procedure
The survey was advertised on Twitter and Facebook. We
In this paper we consider the specific cultures that develop
attempted to use snowball sampling by requesting “retweets”
in small groups: partners, families and friends. In particular,
or “shares” from both participants and non-participants. In
we focus on how these groups actively repurpose emoji to
total the survey was retweeted 37 times on Twitter and
serve new functions, their motivations for these
shared 7 times on Facebook. The survey was also advertised
repurposings, and how the affordances of different emoji
on the research ‘subreddit’ of Reddit, the content aggregating
influence how they get repurposed. We define “repurposing”
website. Survey completion was not reimbursed in any way,
as giving an emoji a specific and constant meaning beyond
so there was no particular incentive for participants to
the initial “intention” of the emoji designer; this meaning
mindlessly work through the whole survey. Given this lack
would be inaccessible to an outside observer without
of incentive, the survey was kept purposefully succinct and
explanation. What is specific to this paper, is the
did not require a large amount of demographic data from the
consideration of personalised repurposing. The act of
participants. By keeping the survey short, we aimed to
repurposing has been reported in large groups (e.g., [1]), but
reduce the drop-off rate. The survey could comfortably be
here we consider emoji repurposing in smaller micro
completed in two minutes.
communities. We contribute a three dimensional analysis of
emoji repurposing, exploring why users need to use an emoji Participants
at all, why that particular emoji was chosen, and what The survey was completed by 134 participants. Fifty-seven
sentiment they intend to convey. participants (43%) reported that they did not use a
repurposed emoji so did not provide any further data. One
Method
participant was removed for providing fabricated data
A brief web-based survey was created using the Google
intended to be amusing. Four participants’ data were
Forms web app. The survey was piloted before dissemination
removed for reporting emoji use considered to be widely
and the questions unchanged. However, Question 4 had
understood and therefore neither repurposed nor personal
further instructions appended to help participants enter an
emoji if they were on a desktop computer. The survey (for instance, the ☺ emoji to represent happiness and the 😉
contained the following questions (with response types are to be suggestive). In total 72 responses were used in the data
in italics): analysis. Of these participants, three were between 0-17
years old, 28 were between 18-29, 35 were between 30-39
1. How old are you? Age range response and six were over 40.
2. Is there an emoji you use that has a special meaning for
Data Processing
just you and the recipient? (The meaning would be
The survey was open for 14 days to collect responses. After
unclear to a third party seeing that emoji) Yes/No
this, the data was cleaned to remove null responses and
response
responses that were not considered to fit the request of being

2
https://www.acuitytranslations.com/blog/item/becky-
kinnersley
personal or repurposed. Emoji were all assigned their Friend, Family Member, Family group chat, friend group
Unicode name (e.g., 🛀 would be person taking a bath) and chat, other. Partner was the most commonly reported
were categorised according to the common categories used recipient for the repurposed emoji, with 47% of participants
on Emojipedia, Apple’s iOS keyboard, and Google’s reporting they used a special emoji with a partner. Figure 2
Android keyboard (Smileys & People, Animals & Nature, shows the distribution of people with whom participants used
Food & Drink, Activity, Travel & Places, Objects, Symbols, their repurposed emoji.
Flags). The free text data from questions 4-7 were then Five participants used the “Other” category to report their
openly coded for common themes. A Content Analysis emoji recipient. Two reported a sexual partner, two reported
procedure was followed, with open codes being generated to using the emoji in a public space (e.g., Twitter) and one
cover the question of how that emoji was chosen (relating to reported using it for themselves. Out of all 72 responses, 85%
Question 6 of the survey), why an emoji is needed (relating of respondents reported using the emoji with just a single
to Question 7 of the survey) and what theme the emoji other person, whereas only 15% reported using it in a group
covered (relating to Question 5 of the survey). These codes situation.
were generated by one researcher, and validated
independently by a second researcher. Discrepancies were
discussed and a final code assigned. For each of the three
components (e.g., “Why an emoji is needed”), only one code
was assigned for each response. No component could have
more than one code.

Figure 2 Percentage breakdown of the recipients of the

repurposed emoji
Emoji Sentiment
The sentiments of the emoji were determined from the
answers to Question 5 of the survey. The reported emoji were
Figure 1 Percentage of repurposed emoji originating from the coded based on the meaning the participant told us that they
emoji keyboard categories were aiming to express. These codes were then grouped to
RESULTS produce the following categories, reported here with their
explanations, percentage coverage and examples. Not all
Reported Emoji
emoji reported fit into these categories, 4% fell within an
Eighty-one emoji were reported in the survey. This set
“Unknown” category if the participant did not report the
comprised 69 unique emoji. Sixty-six emoji were used on
sentiment or it was not detectable from their other responses.
their own and 15 were used in sequences of two or three
emoji. The most commonly reported emoji was the octopus Affection – 21%
(🐙) which was reported by four separate participants. Figure This category covers emoji that relate to affectionately used
1 shows the number of emoji reported in each of the emoji emoji. This might mean emoji used to express romantic love
categories found on the standard emoji keyboard. The most between partners, or platonic love and affection between
common category was Animals & Nature, representing 41% friends and family. In 80% of cases this sentiment was
of the emoji reported in the survey. Note, there were no emoji reported between partners or sexual partners, however in
used from the Activity or Flags categories. 20% of cases it was used between friends. It was never
reported in group or public chats; it was only ever seen in
Emoji Recipients one-to-one conversations.
In Question 3, participants were asked to report who they
used the emoji with. Set answers were provided: Partner,
These emoji were used to express intimacy, “Hug, cuddling. having flirtatious conversation but can be used to simply
Just generally being there for each other and reaching out” discuss sex.
– P46, to simply say “I love you” or as “A way of saying 'hi,
Mainly, however, these emoji would refer to specific sexual
I'm thinking of you'” – P49.
acts or be used to express arousal between partners.
Person – 19%
Funny – 7%
This category relates to emoji that are used in place of a
Some emoji were used for humour between the participant
specific person (or animal). This emoji could represent
and recipient. Many emoji in this category referenced an
someone external or internal to the conversation. Emoji
ongoing joke that began outside of the digital
expressing this sentiment were used for a range of recipients,
communication.
around a fifth of the time this sentiment was used, it was used
in group chats. Opinion – 6%
Four participants reported using emoji to express an opinion
The emoji were generated by using an emoji that is known to about either the recipient (“you/we are right” – P21) or
both to represent that person or animal, “We use it to about a situation (“naff/overdone/not-to-my-taste” – P42).
symbolise our Scottie dog” – P12 (a dog emoji was not used
in this instance). Many of these instances related to pet names Power – 4%
that the participant reported using “Fish is our nickname for The category of power relates to emoji used in a direct
each other” – P17. response to a perceived threat to show solidarity or power.
This category was used both with friends and also with a
Emotion – 15% wider audience on message boards and other forms of social
Some emoji were repurposed to allow the participant to media where multiple “recipients” can be expected.
express an emotion they were feeling. This form of
repurposing was found more commonly amongst friends and Examples of the use of emoji power repurposing include
family with fewer than a third (27%) of instances being used using the Triangular Flag or Red Circle (🚩 or 🔴 [both
with a partner. symbols are coloured red]) to show solidarity with the
“#REDInstead” movement (A movement against an Autism
Emotions expressed in this way ranged from negative “We
“curing” campaign). Another included the use of the Flexed
use it if one of us is feeling a bit down for no reason.” – P8
to positive “Expression of excitement” – P55. Some Biceps and Woman Dancing (💪💃) to represent feminism.
emotions expressed using repurposed emoji were neither This participant directly reported needing it due to “the
positive or negative, representing a feeling instead “We're patriarchy” – P34.
big and greedy” – P13. Reasoning behind chosen emoji
Conversational words – 8% Participants were asked to report why they chose a particular
Some participants reported using repurposed emoji for emoji for the sentiment they were trying to convey. The
simple conversational words to quickly express responses were coded and categorised. Some participants did
acknowledgement. not explain the reasoning behind the emoji choice and so 4%
of responses were not coded.
Conversational emoji included confirmation “Yes/OK/
thumbs-up/understood” – P23 (note, the emoji was not the Image of word – 19%
thumbs up symbol) or to confirm receipt of the message “I Often emoji were chosen because they represented a word
have read and understood your message” – P3. used in participants’ real world conversations. Often this
type of emoji was used as a direct translation of a “pet name”
Logistics – 8% (with penguin 🐧 being most popular, appearing twice).
The category of Logistics refers to emoji used when talking Another participant reported using the Pig Face emoji (🐷) to
about a place or plan. These emoji were used as direct
refer to the local pub, called The Three Piggies. The
substitutions. For example, 🗼 (Tokyo Tower) instead of repurposing that occurs in these scenarios requires that the
typing ‘Tokyo’. image represents something which is special to both parties.
An illustrative example would be that two different Although, for example, the penguin emoji is being used to
participants reported directly translating the name of their mean ‘penguin’, there is more meaning attached to that word
local pub into the emoji form. Another used a landmark from in these scenarios. Here ‘penguin’ does not refer to the bird,
the country that they were going to visit to refer specifically but to a particular person, meaning the emoji has been
to a holiday that was being planned. repurposed.
Sex – 7% Historical – 17%
These emoji were used when referring to sex. They were “Historical” refers to emoji that are used specifically due to
largely used with participants’ partners, however one a shared story between the users. The emoji would represent
participant reported using a sex-related repurposed emoji a particular aspect of that story, or could be the result of the
with a friend, suggesting the emoji is not just used when story. For example one participant reported using a particular
emoji because “I made a typo once and it stuck” – P23.
Historical was used to code any emoji where understanding avatar emoji are not spoken aloud, but are used just for
of the emoji relied on a specific shared story “It refers to written digital communication.
something that happened early on when we met” – P53 or
Play on Words – 8%
simply translating a shared joke in real life into a common Some emoji were chosen due to a play on words with the
shorthand “It's been a running joke for many years” – P71. emoji being used to spell out words phonetically (e.g., the
Evocative – 15% word “Hatchet” spelled out with 🎩💩 – P16).
Evocative emoji refer to those that represent a shared
Random – 8%
metaphor between two people. Unlike Historical emoji,
Some emoji were used because they were once typed
these do not refer to a particular incident or occurrence, but
randomly and later took on a particular meaning. Other times
a common understanding. For one participant, the use of the
a random emoji was chosen to convey a sentiment. For
Spouting Whale emoji (🐳) could express an emotion that
example, P59 uses the octopus emoji (🐙) to simply make
both participants understood “Definitely feels like it evokes
contact with their partner, stating “An octopus seemed as
that feeling!” – P8. Another participant also reported a
random as any other emoji.”
shared understanding but for an opinion, rather than a feeling
“Because it represents frivolities for us” – P68 when using Irony – 3%
the Nail Polish (💅) emoji. Emoji in this category were chosen because they represented
the complete opposite of the sentiment they were intended to
Two participants reported using a shared love of food to convey (for example picking a purposefully unattractive
evoke feelings towards their partner by using Pizza (🍕) emoji to represent love [P47] or sexual attraction [P1]).
“because we both love each other as much as we love pizza”
Reason for using an emoji
– P12 and Cheese Wedge (🧀) “Because it's cute and
Participants were asked to explain why they used an emoji at
succinct and we both love cheese.” – P28 to mean romantic
all – why was it necessary for them to include an icon rather
love. These do not relate to specific stories in their past, but
than just using words? These reasons were coded by two
a shared understanding of common feelings.
researchers, the following categories were found. As before,
Visual Affordance of emoji – 14% some participants did not report a reason and so could not be
This category introduces the concept of ‘visual affordance of included in this analysis (17%).
emoji’. These emoji are used for a particular situation only
Ease – 28%
because of the way that they are rendered, not because of
The most commonly reported reason for using an emoji was,
their semantic meaning. These often arose from participants
perhaps unsurprisingly, ease. Participants reported that an
seeking an emoji for a concept that they were unable to
emoji was “a lot faster to type” – P59. Another participant
express using the standard set of emoji. For example, one
expressed that using words felt foolish with emoji available
participant reports using the bath emoji (🛀) in place of a “It seemed silly to type the word when there was a little
coffin “my friend wanted a person in a coffin emoji, couldn't picture of it right there” – P27.
find one and used this instead, arguing it was ‘the closest
thing’” – P48. No Words – 13%
The code of no words was used to describe instances where
Another participant saw the possibility to read one of the participants used emoji because they either could not use
smileys as though it were communicating in sign language. words to express the sentiment or would at least find it hard
The participant used the ‘thinking face’ (🤔) to mean lesbian, to do so.
as the position of the thumb and forefinger on the chin
denotes “lesbian” in American Sign Language. Multiple participants reported needing to use a repurposed,
personalised emoji because the idea or sentiment they were
These emoji were all used not because of what they trying to express was too complex for words “Because it's a
represented, but specifically because of the way they have complicated concept, with lots of different applications.” –
been drawn. P15. Another participant simply found it difficult to be
Avatar – 11% earnest with their partner, choosing to tell them that they
Avatar-coded emoji are those that represent a particular loved them using an emoji “To avoid sincerity” – P47. One
person or place. The defining feature of the avatar is that this participant summed up the use of emoji used in this fashion:
is a purely digital way to refer to that person or place. Unlike “talking is hard” – P3.
the image of word code, the avatar has not been directly Intimacy – 13%
translated from conversation, there is no real-world verbal Intimate emoji use was recorded when the participant
equivalent for it. For example, two participants (P6 and P66) reported using an emoji purely for the reason of creating a
reported using two animal emoji (🐥🐱 and 😻🐭) to secret code that could only be understood by themselves and
represent themselves and their partners. Another participant the person they were conversing with. This code was used
(P24) used an avatar to refer to their partner on some for the purpose of creating an intimate bond over having a
websites specifically to avoid using their real name. These shared secret.
The intimacy was often found between friends (“It's best DISCUSSION
friend code.” – P20) to represent a special bond between two Our findings support the Unicode Consortium’s statement
people (“It's funny and a thing only we share so it's a bond that emoji often take on alternative meanings based upon the
between us.” – P70). It was also found between partners. way they have been drawn [22], however we show that emoji
repurposing goes beyond this. Although some emoji are
There was a sense of pride associated with having such a indeed repurposed due to their renderings (see e.g., the
secret bond, as though the participant and the recipient were Visual Affordance of Emoji subsection), many are
in a private club “It's a sibling in-joke that no one gets except repurposed for entirely different reasons, for example
us, even if we explain it” – P67. because they relate to something in the real world, or because
Cute – 11% they have been chosen randomly or ironically.
Some participants used repurposed emoji because they were
As Kelly and Watts highlighted, emoji are often repurposed
cute. Every participant using an emoji for this reason used
to allow senders and receivers to share a secret form of
the word “cute” to explain its use. In these instances,
communication [9]. We also found this in our study.
participants found the idea of an emoji to be cuter than using
However, we also observed the practice occurring when
words to describe the sentiment or person “It's a cute way to
people just want to have faster interactions, when they want
send a reminder of him when the family is apart” – P11.
to say things that they cannot in words, and for many other
Funny – 10% reasons. What is key, is that the present study shows that this
These emoji were used because seeing a visual repurposing behaviour has permanence; emoji take on
representation of a concept, rather than seeing the words, was consistent alternative meanings that are used more than once.
considered amusing to the participant and their recipient.
One participant is unable to justify why the emoji is used, but Here we discuss more specific findings from the current
study.
reports that its continued use is a source of amusement “it
just happened one day and we stuck to it - it makes us laugh Emoji for maintaining relationships
for some reason” – P64. One common theme throughout the analysis is that of emoji
representing love and affection, and being used to maintain
Discretion – 6%
personal relationships. Similar usage was reported in Kelly
Discreet reasons for using emoji referred to either sex or
illegal activities. Using an emoji to refer to sex allowed some and Watts [9]. They report three differing uses for emoji to
participants to discuss sex without advertising it to others maintain such relationships, two of which are directly
who might be able to see the message. This social discretion relevant and observed within the current study: Maintaining
was particularly useful for one participant who was a Conversational Connection and Creating Shared and
communicating about sex with a partner who was not being Secret Uniqueness. Participants within the current study also
faithful to another “It's probably harder to be accused of engaged in emoji use solely with the purpose of maintaining
cheating when you're just messaging […] little illustrations” a conversational connection. However, unlike Kelly and
– P72. Watts’ participants [9], our participants chose a specific
emoji to use repeatedly to let another person know they were
Other participants wanted discretion due to concern that their thinking of them. Kelly and Watts’ participants reported
messages, which contained discussion of illegal activities, sending randomly chosen emoji to make a connection,
would be intercepted whilst being sent “So that you don't whereas participants in the current study had chosen a
have to type the names of drugs in non secure messaging” – particular emoji to mean “I am thinking about you”. In this
P56. study, it was the emoji and not the act of sending it that
Accident – 3%
helped maintain the conversational connection.
Two emoji were used purely through accident. One because Similar behaviour was reported in this study with regard to
it was a typo, and the other because a meaning had been creating a shared and secret uniqueness. In Kelly and Watts
assigned post hoc to an emoji that was sent in a random [9], one participant reported typing random emoji and then
string, stating that the emoji was used for “No special reason assigning meaning to them after their use. A participant in
apart from it just happened” – P44. the current study had engaged in a similar post hoc
Emphasis – 1% rationalization by attempting to make sense of an emoji after
In one case the participant used an emoji as a form of it had been randomly chosen. However, again there was a
emphasis. The emoji represented something like a subtle difference in the current study in that the randomly
punctuation mark to add extra power to what they were chosen emoji continued to have meaning after the amusing
saying when words could not enhance the sentence – activity of trying to invent meaning. That emoji went on to
“Because it's an important thing close to my heart and the have a place in the partners’ conversations afterwards.
emoji brings a better point to it” – P41. The idea of shared secrecy more generally has been
expanded in the current study – the category of Intimacy that
emerged when participants were asked why they used an
emoji and not text related directly to the idea of creating a repurposing of the Peach emoji in this way caused
secret uniqueness. Participants reported using emoji because complaints when Apple changed the rendering of the emoji
it represented a unique bond they had with their partner or which resulted in an image which did not have the same
friend – the emoji represented something only they visual affordances.
understood.
This is a potential issue for other users making use of emoji
Emoji as non-verbal communication affordance within micro cultures. For example, one user used
Emoji are often reported as being used to impart emotion to the Triangular Flag (🚩 [coloured red]) to represent the
text, or to provide information about non-verbal cues such as colour red in an online campaign. The use of this emoji by
intonation or intention [23] (for example using an emoji to this participant is based solely on the rendering of its colour.
express sarcasm [9]) However, this study highlighted that This colour information is not stored in the emoji’s
emoji are not only used to replace non-verbal actions that description at all, leaving it susceptible to a change of colour
would happen in face to face communication. Instead, the in the future, thus removing its visual affordance and leaving
repurposed emoji were sometimes used to express the emoji useless to this participant.
sentiments that could not be made verbal. For many users,
creating a shared emoji with special meaning was a way of Repurposing behaviour in other languages
conveying very complex feelings and thoughts that would Reusing “words” in a language for personal purposes is not
have not been difficult to reference and describe quickly in a novel phenomena, and is found within other languages.
text. In this way, emoji need not be seen as something that This occurs, for instance, in home signing, a type of sign
can add additional non-verbal information, but as a way of language that is only understood by a small group of people
communicating important ideas in entirely non-verbal ways. in the home, and which would not necessarily be understood
As P49 explains “It's been a complicated friendship. Sending by outsiders [17]. Within the current study we see
[emoji] says a lot without having to find the words.” participants creating similar personalised understandings of
emoji within family units. Private languages have also been
Real life and the digital world noted between twins in the early stages of language
This study revealed how emoji repurposing can be split into development [20], highlighting that the process of creating a
two categories – digital-based and real world-based origin. shared secret languages within family groups is a phenomena
For example, the Avatar and Random categories that in other media.
emerged when discussing the reasoning behind the emoji
were both used exclusively in the computer mediated One reported use of repurposed emoji in the current study
communication. The avatar was not used in the real world to was to allow users to talk about illegal activity safely and in
refer to a person or pet, but was exclusive to the emoji a hidden manner. Secret languages have similarly been used
context. In comparison, the Image of word and Historical by criminals in order to disguise illegal activity from
categories both refer to emoji generated as a result of an authorities [7] and historically by people within the gay
existing real-world event or reference, both are used to community who also needed to remain hidden in order to
translate an occurrence in the real world into a digital protect themselves in a society which considered their
equivalent. actions criminal [2].

This shows that repurposed emoji can be both the cause of a The use of repurposed emoji in this study also reflects the
new understanding within a microculture (e.g., accidentally way that “nicknames” are generated and used in spoken
making a typo and attributing it meaning post hoc) or the languages: names are shortened for efficiency (Ease in this
symbolic representation of an existing understanding (e.g., study), nicknames can also arise because they are ironically
choosing a symbol to represent an in joke that has already opposite to the person they describe (Irony) or can use, for
been established in face to face communication). example, an animal or other word entirely to refer to
someone (Image of Word, Historical).
Visual affordance of emoji
Some users reported using the visual representation of an In these ways, the findings in the current study do not detail
emoji and repurposing it to mean something visually similar. novel or new human behaviour, but instead highlight how
Wijeratne, Balasuriya, Sheth, and Doran [25] explain that such common behaviours can manifest in new forms of
similarity of emoji should be based upon their semantic technology mediated communication. In particular , we show
meaning, not their visual similarities. However, here we have how the affordances and constraints of the specific context
observed that people do indeed make use of that visual of emoji influence the expression of this common human
feature of emoji when ascribing meaning to it. behaviour.
Using emoji in this way is a phenomena that has been noted Limitations
on a larger scale in the study conducted by the Emojipedia There are a number of limitations in the current study, which
website, who discovered that the Peach emoji (🍑) is most we believe represents an initial exploration of the
commonly used to refer to buttocks, rather than the fruit [1]. repurposing of emoji.
This is due to its visual similarity. The wide spread
We would firstly like to address the demographic data Additionally, the redrawing of emoji must be carefully
collected. A future extension of this work could explore how considered. Just as users reacted badly to Apple redrawing
repurposing varies across culture. Just as the interpretation the Peach emoji, smaller communities of other users may be
and use of emoji changes depending on country of use, the affected by redrawing of any number of emoji currently in
act of repurposing may also vary. Understanding the device use because of how they look, and not what they represent.
upon which the emoji is most commonly used, and the
ACKNOWLEDGEMENTS
medium would also help in understanding whether
We would like to thank all the survey respondents for sharing
repurposing is limited to certain aspects of personal digital
the ways in which they repurpose emoji. We appreciate the
communication.
candidness of replies and have striven to ensure anonymity
This survey allowed users to enter free text to describe their of all participants. We would also like to thank the reviewers
responses. In doing so, we were able to generate a set of and committee members who reviewed this work for their
codes to explain and explore the use of repurposed emoji. valuable insights and helpful suggestions for improvements.
Future work could expand upon these codes by additionally REFERENCES
collecting interview data to further explore the origins and 1. Hamdan Azhar. 2016. How We Really Use The
usage of repurposed emoji. Peach. Emojipedia. Retrieved from
Future Work https://blog.emojipedia.org/how-we-really-use-the-
The current study represents an initial exploration of the peach/
phenomenon of personalised emoji repurposing. Building 2. Paul Baker. 2003. Polari-the lost language of gay
upon this work, we can see that future work might consider men. Routledge.
how repurposing varies across geographic location [3] or by
age or other demographic characteristics [5]. 3. Francesco Barbieri, German Kruszewski, Francesco
Ronzano, and Horacio Saggion. 2016. How
Further work can be conducted to investigate the wider Cosmopolitan Are Emojis?: Exploring Emojis
applicability of the codes generated in this paper. Is the list Usage and Meaning over Different Languages with
of codes generated comprehensive, for instance? Distributional Semantics. Proceedings of the 2016
CONCLUSION ACM on Multimedia Conference, 531–535.
In this paper we have highlighted how people using http://doi.org/10.1145/2964284.2967278
computer and mobile mediated communication have come to 4. Edwin Chan, Teddy Seyed, Wolfgang Stuerzlinger,
repurpose certain emoji. By using the ambiguity in meaning Xing-Dong Yang, and Frank Maurer. 2016. User
that emoji naturally have, users can create a shared personal Elicitation on Single-hand Microgestures.
meaning between themselves and another person, or small Proceedings of the 2016 CHI Conference on
group, creating a new cultural understanding of particular Human Factors in Computing Systems, 3403–3414.
emoji. This in turn can help people feel closer to one another. http://doi.org/10.1145/2858036.2858589
Previous work has tended to explore emoji understanding on 5. Zhenpeng Chen, Xuan Lu, Sheng Shen, Wei Ai,
a large cultural level, whereas this work looks at emoji Xuanzhe Liu, and Qiaozhu Mei. 2017. Through a
understanding at a micro scale. Gender Lens: An Empirical Study of Emoji Usage
Results of this paper will contribute to future work into emoji over Large-Scale Android Users. 1–20. Retrieved
sentiment analysis, as the work highlights that emoji do not from http://arxiv.org/abs/1705.05546
always correspond to their intended, nor culturally accepted 6. Henriette Cramer, Paloma de Juan, and Joel
meanings. At times emoji are chosen at random to mean a Tetreault. 2016. Sender-intended Functions of
specific concept, and equally some emoji are chosen Emojis in US Messaging. Proceedings of the 18th
purposefully because they convey the complete opposite of International Conference on Human-Computer
the intended sentiment. This implies caution is required Interaction with Mobile Devices and Services, 504–
when using machine learning techniques to understand the 509. http://doi.org/10.1145/2935334.2935370
meaning and use of emoji.
7. Maurizio Gotti. 1999. The language of thieves and
As with other papers in this area [13,21], we also conclude vagabonds: 17th and 18th century canting
that a universal rendering of emoji may be required to lexicography in England. Walter de Gruyter.
standardise how they appear across platforms. It is well 8. Jialun “Aaron” Jiang, Jed R. Brubaker, and Casey
understood that emoji facial expressions can be Fiesler. 2017. Understanding Diverse
misconstrued, for example, but this paper also highlights that Interpretations of Animated GIFs. CHI’17 Extended
discrepancies in rendering can also affect a whole range of Abstracts, 1726–1732.
emoji that people choose to use due to the way they are http://doi.org/10.1145/3027063.3053139
drawn. The current method of rendering emoji in a range of
different ways may be preventing users from using the 9. Ryan Kelly and Leon Watts. 2015. Characterising
implicit affordance of emoji. the Inventive Appropriation of Emoji as
 Relationally Meaningful in Mediated Close Conference Companion Publication on Designing
Personal Relationships. Experiences of Technology Interactive Systems - DIS ’17 Companion, ACM
Appropriation: Unanticipated Users, Usage, Press, 7–12.
Circumstances, and Design. http://doi.org/10.1145/3064857.3079109
10. Brian Koerber. 2016. People are mad as hell after 19. Swiftkey. 2015. Swiftkey Emoji Report. Retrieved
Apple ruins the peach emoji. Mashable. Retrieved from
from http://mashable.com/2016/11/01/peach-emoji- http://www.scribd.com/doc/262594751/SwiftKey-
butt-no-more/#y0IZqe3q7iqd Emoji-Report
11. Xuan Lu, Wei Ai, Xuanzhe Liu, et al. 2016. 20. K Thorpe, R Greenwood, a Eivers, and M Rutter.
Learning from the ubiquitous language. 2001. Prevalence and developmental course of
Proceedings of the 2016 ACM International Joint “secret language”. International journal of
Conference on Pervasive and Ubiquitous language & communication disorders / Royal
Computing - UbiComp ’16, ACM Press, 770–780. College of Speech & Language Therapists 36, 1,
http://doi.org/10.1145/2971648.2971724 43–62. http://doi.org/10.1080/13682820120395
12. Hannah Miller, Daniel Kluver, Jacob Thebault- 21. Garreth W Tigwell and David R. Flatla. 2016. “ Oh
Spieker, Loren Terveen, and Brent Hecht. 2017. that’s what you meant !”: Reducing Emoji
Understanding Emoji Ambiguity in Context: The Misunderstanding. Proceedings of the 18th
Role of Text in Emoji-Related Miscommunication. International Conference on Human-Computer
Proceedings of the Eleventh International AAAI Interaction with Mobile Devices and Services
Conference on Web and Social Media, 152–161. Adjunct (MobileHCI ’16), 859–866.
Retrieved from http://doi.org/2957265.2961844
https://www.aaai.org/ocs/index.php/ICWSM/ICWS 22. Unicode Inc. 2017. Emoji and Pictographs.
M17/paper/view/15703 Retrieved from
13. Hannah Miller, Jacob Thebault-Spieker, Shuo http://www.unicode.org/faq/emoji_dingbats.html
Chang, Isaac Johnson, Loren Terveen, and Brent 23. Joseph B. Walther and Kyle P. D’Addario. 2001.
Hecht. 2016. “Blissfully happy” or “ready to fight”: The Impacts of Emoticons on Message
Varying Interpretations of Emoji. International Interpretation in Computer-Mediated
AAAI Conference on Web and Social Media, Communication. Social Science Computer Review
Icwsm, 259–268. 19, 3, 324–347.
http://doi.org/10.1089/cyber.2011.0179 http://doi.org/10.1177/089443930101900307
14. Midas Nouwens, Carla Griggio, and Wendy E 24. Sanjaya Wijeratne, Lakshika Balasuriya, Amit
Mackay. 2017. “WhatsApp is for family; Sheth, and Derek Doran. 2017. EmojiNet : An Open
Messenger is for friends”: Communication Places Service and API for Emoji Sense Discovery. the
in App Ecosystems. CHI ’17 - Proceedings of the 11Th International Aaai Conference on Web and
2017 CHI Conference on Human Factors in Social Media (Icwsm-17), May, 437–446. Retrieved
Computing Systems, 727–735. from
15. Oxford Dictionaries. 2015. Word of the Year 2015. https://www.aaai.org/ocs/index.php/ICWSM/ICWS
Retrieved from M17/paper/view/15551
https://en.oxforddictionaries.com/word-of-the- 25. Sanjaya Wijeratne, Lakshika Balasuriya, Amit
year/word-of-the-year-2015 Sheth, and Derek Doran. 2017. A Semantics-Based
16. Jaram Park, Vladimir Barash, Clay Fink, and Measure of Emoji Similarity.
Meeyoung Cha. 2013. Emoticon Style : Interpreting http://doi.org/10.1145/3106426.3106490
Differences in Emoticons Across Cultures.
Proceedings of the 7th International AAAI
Conference on Weblogs and Social Media
(ICWSM), 466–475.
17. Ann Senghas, Sotaro Kita, and Asli Özyürek. 2004.
Children Creating Core Properties of Language:
Evidence from an Emerging Sign Language in
Nicaragua. Science 305, 5691, 1779–1782.
http://doi.org/10.1126/science.1100199
18. Selina Sutton and Shaun Lawson. 2017. A
Provocation for Rethinking and Democratising
Emoji Design. Proceedings of the 2016 ACM