Emoji in social media discourse about working from home

Abstract

This paper applies a social semiotic framework for exploring the functions of emoji in digital discourse about working from home (WFH). This is an important and prevalent discourse in the on-going COVID-19 pandemic due to widespread ‘lockdowns’ aimed at reducing the spread of the virus which have had a profound impact upon how and where people work. The paper focuses in particular on ideational meaning, that is, how experience is construed, exploring how people use language and emoji to express ideational meanings about their daily lives in a large corpus of tweets about WFH. This exploration involves corpus-based discourse analysis focused on how language and emoji coordinate to make meaning, drawing on the concept of intermodal coupling to understand the convergence of meaning across semiotic modes. The paper also considers the role of visual images in this coordination.

Full text

PRE-PROOF VERSION TO APPEAR IN DISCOURSE, CONTEXT & MEDIA
Emoji in social media discourse about working from home
Abstract
This paper applies a social semiotic framework for exploring the functions of emoji in digital
discourse about working from home (WFH). This is an important and prevalent discourse in
the on-going COVID-19 pandemic due to widespread ‘lockdowns’ aimed at reducing the
spread of the virus which have had a profound impact upon how and where people work. The
paper focuses in particular on ideational meaning, that is, how experience is construed,
exploring how people use language and emoji to express ideational meanings about their
daily lives in a large corpus of tweets about WFH. This exploration involves corpus-based
discourse analysis focused on how language and emoji coordinate to make meaning, drawing
on the concept of intermodal coupling to understand the convergence of meaning across
semiotic modes. The paper also considers the role of visual images in this coordination.
Keywords
Emoji, Working from home, social media discourse, Twitter, Work discourse
1. Introduction
The ongoing covid-19 pandemic has resulted in a seismic shift in how and where people
work, and on how they interact and relate to one another. Substantial movement to working
from home (hereafter WFH) is likely to persist and refashion existing work practices in novel
ways (Barrero, Bloom, & Davis, 2021). Unsurprisingly, discourse about WFH is prevalent on
social media platforms. Emoji are pervasive in this communication where they play a
significant role in “the expression of emotion, conveying stances, and negotiating
interpersonal alignments” (Logi & Zappavigna, 2021). Emoji can enact a range of pragmatic
functions relating to modulating tone (Dainas & Herring, 2021; Danesi, 2016), forging
interpersonal alignments (Sampietro, 2016), indicating humorous meaning (Skovholt,
Grønning, & Kankaanranta, 2014; Thompson & Filik, 2016) and negotiating politeness
(Beißwenger & Pappert, 2019). The current study focuses on the ideational function of emoji
in WFH discourse. Ideational meaning is concerned with how experience is represented:
“what kinds of activities are undertaken, and how participants undertaking these activities are
described and classified” (Martin, 2008, p. 17). Some studies of emoji incorporate an
ideational dimension where emoji are involved in the construal of entities or activities in
digital texts, although this is formulated in different ways depending on theoretical
orientation. Dainas and Herring (2021, p. 113); Herring and Dainas (2017) specify a
“mention” function whereby emoji (and other graphicons) have a pragmatic function of
echoing, repeating or illustrating the “textual content”. Gawne and McCulloch (2019)
suggest that object emoji such as food can often be used to ‘illustrate’ the subject matter in
examples such as “‘I love pizza ’ ‘avocado toast forbreakfast ’ ‘[photo of coffee in

mug]     ’”. Other studies note that emoji can “can create experiential meaning
by acting as or forming part of a participant, process, or circumstance within a text”
(Parkwell, 2019, p. 4).
Emoji analysis focused specifically on the COVID-19 pandemic has tended to be
quantitative, employing corpus-driven or sentiment analysis techniques to consider emoji
frequency and density in social media discourse produced during the pandemic (Das, 2021).
Some studies have suggested that emoji analysis can have practical benefits in health
domains, such as providing a novel method for illuminating the gendered impacts of COVID-
19 (Al-Rawi et al., 2020), and also in other domains such as finance, for instance in tracking
relationships between emotional uncertainty and market volatility (Lazzini, Lazzini, Balluchi,
& Mazza, 2021). Most relevant to the present study is emerging work about the role of emoji
in and about working from home discourse. For instance, some studies are beginning to
emerge on emoji use in videoconferencing, for instance in the chat functionality that is
embedded in most platforms (Dürscheid & Haralambous, 2021) and in automatic captioning
for accessibility (Oomori, Shitara, Minagawa, Sarcar, & Ochiai, 2020).
This paper begins with a survey of the key linguistic studies of emoji, before considering in
detail some of the key technical dimensions of emoji. These dimensions are important to
review because emoji, as a special and complex class of non-canonical token, present
substantial barriers to effective corpus-processing. In addition, the characteristics of emoji
encoding are also central dimensions of how they enact meaning as a digital resource. Having
established the technical parameters of emoji, the WFH Emoji Corpus is then introduced. The
analytical method is then explained, specifying the particular model of ideation guiding the
discourse semantic analysis undertaken, and the particular perspective on intermodal
coupling (coordination of resources across semiotic modes) used to explore relations between
emoji, language, metadata, and images in the corpus. The particular patterns of ideational
meaning uncovered through a combination of corpus analysis techniques and close discourse
analysis are then elaborated.
2. Previous linguistic studies of emoji
The major areas of linguistic research into emoji include pragmatic studies of emoji
functions, sematic and corpus-based work on emoji meanings, and studies considering the
social dimensions of emoji use and interpretation. The principles applied in linguistic studies
of emoji have their origin in work on emoticons (c.f. Dresner & Herring, 2010; Skovholt et
al., 2014; Thompson & Filik, 2016). Early research into emoji functions was largely
concerned with their role in the expression of affect but has since expanded to consider their
role in enacting various kinds of pragmatic functions. These functions include indicating
stance and shifts in footing (Sandel, Ou, Wangchuk, Ju, & Duque, 2019) as well as clarifying
intent (Thompson & Filik, 2016). Herring and Dainas (2017, p. 2185) posit that tone
modification is one of the most common emoji functions. Emoji are also involved in
signalling non-seriousness or humour (König, 2019; Yus, 2021), negotiating facework
(Beißwenger & Pappert, 2019) and phatic dimensions such as status (Aull, 2019). Other
studies have focused on the role of emoji in fostering involvement or alignment with an
interlocutor (Sampietro, 2016). Some studies note the “performative value” of emoji in mass
expression of solidarity in situations such as terrorist attacks or disasters (De Cock & Pizarro
Pedraza, 2020, p. 2012). However, due to their proliferation, emoji may be becoming less
pragmatically marked, weakening “their specific pragmatic meanings of playfulness,
emotionality, and indexing social intimacy” (Konrad, Herring, & Choi, 2020, p. 229).

Research into emoji semantics has tended to focus on how people interpret emoji as well as
issues relating to semantic ambiguity, where emoji might act as “iconic contextualization
cues that compensate for a lack of paralinguistic resources like gesture and gaze” (Hafner,
2021, p. 288). There has also been discourse analytic work oriented towards understanding
emoji meaning-making, for instance considering how identity is constructed through the
expression of stance and engagement using emoji sequences (Ge, 2019). It is also important
to note that emoji operate in specific multimodal and technical contexts, and that dimensions
such as semiotic mode play a role in meaning-making. For example, Chik and Vásquez
(2017) assert the importance of factoring in the impact (both in terms of multimodal
constraints and affordances) that software/platforms place upon semiotic resources such as
emoji in terms of when and where they can occur and how they are presented to the user.
There has also been interest in understanding how sociolinguistic variables such as age,
gender, race and cultural context impact usage and inflect the meanings that can be made
with emoji (Albawardi, 2018; Herring & Dainas, 2020; Miltner, 2020; Nishimura, 2015).
An analytical strategy for exploring the potential meanings that emoji can make relevant to
the present study is to describe how they make meaning in combination with linguistic
resources. For example, Ge and Herring (2018) adapt rhetorical structure theory to analyse
how emoji sequences relate to their accompanying co-text rhetorically and logically in Weibo
communication. Corpus-based approaches also consider how emoji coordinate with their
linguistic co-text. There is currently emerging work combining pragmatics and corpus
analysis (Li & Yang, 2018) and exploring frequency and collocation analysis (Collins, 2020).
Some emoji datasets have also begun to be developed, such as the Lisbon Emoji and
Emoticon Database (LEED) (Rodrigues, Prada, Gaspar, Garrido, & Lopes, 2018).
Parallels are often drawn between emoji and paralanguage, for instance analogising with
relations between gesture and speech. Like co-speech gesture, emoji may be dependent on
their linguistic co-text to varying degrees. For instance, Kendon’s continuum (as dubbed by
McNeill (1992, p. 174)) has been used as a diagnostic criteria aimed at determining the extent
to which emoji is dependent on its cotext (Gawne & McCulloch, 2019). Gawne and
McCulloch (2019) identify six semiotic functions realised by emoji in their data: illustrative,
metaphoric, pointing, beat, illocutionary and backchanneling and suggest that their overall
properties are akin to gesture since emoji “do not decompose into smaller morphological
units, they do not show predictable syntax, their meaning is shaped by context-specific use,
and there is accepted variation in form”. However, as was seen with early studies of digital
discourse more generally, analogising with other semiotic modes, in particular with face-to-
face communication, has significant limitations.
The approach adopted in this paper is a social semiotic approach in the sense that it is
concerned with how emoji are used in their functional context and how they operate as
semiotic resources working together with the discourse semantics of the co-text in which they
occur. This is a multifunctional approach which considers how emoji realise meaning across
the three linguistic metafunctions defined by (Halliday, 1978): the ideational metafunction,
concerned with experiential meaning, the interpersonal function, concerned with enacting
relationships, and the textual function, describing resources for organising meaning into a
coherent text. The approach that most clearly aligns with this social semiotic perspective is a
study of the linguistic functions of emoji which drew on a metafunctional analysis of hashtag
use developed by Zappavigna (2015) together with social semiotic work on multimodality
(Kress & Van Leeuwen, 2001; O'Halloran, 2006). This analysis explored how a single emoji

(‘toilet’ ) can construe meaning in each of the metafunctions, concluding that “emoji are a
highly contextual, flexible modality, likely to continue to shift and morph with the changing
needs and contexts of social media users” (Parkwell, 2019, p. 9).
Elsewhere (Logi & Zappavigna, 2021) we have developed a framework for undertaking a
metafunctional analysis of emoji that includes how emoji function ideationally,
interpersonally and textually at the level of discourse semantics. In this paper we seek to
elaborate the ideational dimension of this framework in order to develop a more delicate
approach that factors in recent work on how discourse semantic entities and occurrences can
be modelled both in language and paralanguage (as detailed in section 5.1).
3. The encoding of emoji
Emoji are encoded via the Unicode Standard (hereon Unicode), an encoding standard,
defined by the Unicode Consortium. This standard is used across software, platforms, and
protocols worldwide in the processing, storing and interchanging of textual data for most
languages (The Unicode Consortium, 2021). Unicode defines a ‘Common Locale Data
Repository (CLDR) Short Name’ for each emoji and these will be used throughout this paper
when referring to emoji in the analysis, although they should not be confused with
deliberations on the meaning of particular emoji in context. There are at least seven ways in
which emoji can be encoded, including the following:
• A single unique codepoint: e.g. U+2615 =
• A codepoint + variation selector-16 (an invisible codepoint which indicates that the
preceding character is to be displayed with emoji presentation if it defaults to text
presentation) (see Error! Reference source not found.).
• Skin tone modifier: e.g. U+1F935 + U+1F3FD =
• Flags: e.g. Regional Indicator Symbol Letter A + Regional Indicator Symbol Letter
U =
• Tag sequences: e.g. U+E0061 U+E0075 U+E0061 U+E0063 U+E0074 U+E007F U+E0061
U+E0075 U+E0061 U+E0063 U+E0074 U+E007F = Flag for Australian Capital
Territory (AU-ACT)
• Keycap sequences: U+23 U+FE0F U+20E3 =
Accounting for these diverse encodings are important to corpus-based studies as they will
impact on how emoji are ‘counted’ by concordance software. For example, if emoji are
decomposed into their base emoji and any standalone modifiers this will impact frequencies
For instance the symbol for (‘Man Singer’ emoji) is encoded as U+1F468 U+200D
U+1F3A4, where U+1F468 is the ‘Man’ emoji , U+200D is the ZWJ, and the
‘Microphone’ emoji is U+1F3A4. Depending on the software ‘Man Singer’ may appear
as its own category in a the frequency list or it may instead contribute to the count of the
‘Man’ and ‘Microphone’ emoji categories.
Emoji are also presented differently to users depending on platform, for instance Figure 1
presents different renderings of the ‘computer’ emoji across what Unicode refers to as
‘vendors’ (e.g. operating system, platform, software application):

Figure 1 Different presentations of the laptop emoji, U+1F4BB
1
.
Since I am writing this paper in the Microsoft word, when explained as part of the body
paragraph, emoji will be presented using the default encoding which this software applies,
which is similar to the ‘Browser’ representation in Figure 1. However, this is different to the
way emoji appear on Twitter, the platform from which the corpus was derived. Thus,
examples of concordance lines will be presented as figures in order to retain Twitter styling.
4. Dataset: the Covid Corpus
4. 1 Corpus construction
The dataset explored in this study is a corpus of tweets containing any of the following case-
insensitive search terms commonly used in tweets about working from home: WFH,
workingfromhome, workfromhome, "working from home", "work from home", #WFH,
#workingfromhome, and #workfromhome. This selection criteria were aimed at retrieving a
broad range of discourse about working from home. The software ‘Social Feed Manager’ was
used to harvest these tweets from the Twitter API, collecting every 30 minutes for 7 days (1-7
June, 2021). The posts were extracted as JSON files. JSON is a “lightweight, text-based,
language-independent syntax for defining data interchange formats” designed to be easy to
read by humans at the same time as easy to parse by computers (ECMA, 2017). The tweets
extracted in JSON format included the full set of data fields that Twitter uses to represent a
tweet. The harvest produced a set of 262,125 tweets from which the ‘full text’ field was
extracted in order to retrieve the contents of the tweets for the purposes of analysis. This
resulted in a corpus of 123,826 posts following removal of duplicate tweets and retweets, and
tweets in languages other than English (since the detailed discourse analysis required high
levels of understanding for interpreting meaning that were not possible for other languages
due to the researchers’ language proficiency). All tweets that contained emoji were then
isolated, resulting in a corpus of 24,458 posts (620, 038 words).
4.2 Sampling and concordancing
Since the data analysis undertaken on the corpus involved consideration of both quantitative
patterns (e.g. frequency information) and qualitative discursive patterns that were difficult to
automatically identify, texts were sampled for close discourse analysis through a convenience
sampling method. This convenience sampling was augmented by any clues for the most
fruitful analytical pathways that could be gleaned from any quantitative patterns that could be
detected using concordance techniques. Regular expressions and python scripts rather than
concordance software were used for better control of the process, such as ensuring the
scrutability of any patterns detected in light of the difficulties of processing emoji accurately
as explained in the previous section.
4.3 Issues in creating an emoji frequency list
If we generate a wordlist without any processing, the most frequent emoji would be the emoji
shown in Table 2.
1
For full Unicode Standard rendering list see https://unicode.org/emoji/charts/full-emoji-list.html.
Browser Apple Google Facebook Windows Twitter Samsung Gmail

Table 1 Most frequent emoji in the WFH Emoji corpus
This list includes some problematic instances: the ‘Female sign:’󰪟󰪠emoji can both occur in a
tweet as an emoji, or can be part of a more complicated emoji ZWJ sequence. Manual
exploration was unable to uncover any tweets in the corpus using 󰪟󰪠on its own and thus, this
count most likely corresponds to use in ZWJ emoji such as ‘Woman getting massage’ .
For example, consider the two ways of displaying this emoji in Figure 2 via the Twitter app
on an iPhone, and via Twitter accessed via the Chrome browser. In the latter the ‘Woman
Getting Massage’ emoji has been decomposed as the emoji ZWJ sequence (where ZWJ does
not display).
Figure 2 Two visual representations of the same emoji across vendors.
While most modern concordance software has been redeveloped to cope with UTF8, this
does not get around many of the issues raised in this section since concordance software will
not likely be able to recognise sequences of emoji codepoints on unannotated corpora. In
addition, it is also impossible for concordance software to be up to date as emoji sequences
are constantly being extended.
5. Analytical method
The approach adopted in this paper combines methods from corpus linguistics and discourse
analysis. This section explains the discourse semantic systems of ideation (section 5.1) ,
introduced in section 1, and details the approach to accounting for the coordination of
ideational meaning across language and emoji (intermodal coupling, section 5.2). The data
were coded by one of the researchers using these discourse semantic systems and applying
the principle of minimum mapping. Each instance was reviewed by the other researcher and
difficult or ambiguous cases were discussed.
5.1 Ideational meaning

In order to interpret the patterns of meaning observed in concordance lines (lines of text taken
from the corpus), a social semiotic framework for interpreting how language and emoji
cooperate was used. This framework was based on Logi and Zappavigna (2021) which
explained how emoji, in coordination with language, can enact three functions:
• Interpersonal: enacting relationships and values, for instance via an evaluation
e.g. I love working from home

• Ideational: construing experience as processes and participants in a clause e.g.
I bought a new laptop for working from home
󱇯󱇰󱇱󱇲
• Textual: organising discourse into texts, for instance marking the theme in a
clause e.g. WFH
󰛐󰛑󰛒󰛓󰛔󰛕󰛖󰛗󰛘󰛙󰛚
is my favourite way to work.
These are the three metafunctions defined by Halliday (1978) that form the basis of the social
semiotic approach to language manifest in Systemic Functional Linguistics.
This paper focuses in particular on ideational meaning, that is, how experience is construed in
language, acting “as a theory of reality, as a resource for reflecting on the world” (Halliday &
Matthiessen, 1999/2006, p. 7). It explores how people use emoji together with language to
express ideational meanings about their daily lives. The approach draws on the discursive
system of IDEATION, as described in (Martin, 1992) and further developed by Hao (2020, p.
64). This work models ideational meaning at the level of discourse semantics as sequences of
figures made up of elements of different kinds: entities (naming and taxonomizing items:
people, places and things), occurrences (happenings or activities), and qualities (description
or assessments) (Table 2):
stratum
field
discourse semantics
lexicogrammar
terminology
activity, item,
property
sequence
clause complex
figure (state figure, occurrence
figure)
clause
element (entity, occurrence, quality)
Participant, Process,
Circumstance
Table 2 Terminology for describing ideation at each stratum of language (based on Hao (2015, p. 192)).
The discourse semantic analysis of ideation undertaken in this study focuses on the construal
of elements in emoji-language semiosis. It draws on Hao’s (2020, p. 64) systems for
describing linguistic ELEMENTS (ENTITIES, OCCURRENCES and QUALITIES) and FIGURES (STATE
FIGURES and OCCURRENCE FIGURES), considering how these ideational meanings are co-
construed in language and emoji by applying the concept of intermodal coupling. ENTITIES
are the ideational discourse semantic units construing items in a field of experience from a
static perspective, and include sources, things, activities, semiosis, places and time as
illustrated by the language in the examples in Figure 3:

Figure 3 Entity categorisation system, adapted from Hao (2020, p. 64) with linguistic examples from the corpus.
OCCURRENCES, on the other hand, enact a dynamic perspective and are often realised as
mental or behavioural processes, for instance: ‘One of my friends just sent me this .
#WorkFromHome’. QUALITIES describe or assess ENTITIES and OCCURRENCES, for example:
‘Wfh is so boring ’.
The configuration of an occurrence with other elements in a clause is referred to as an
OCCURRENCE FIGURE. In other words, this is a clause where some kind of activity is
happening for instance ‘taking a nap’ as in the first example in Figure 7. The alternative
choice is a STATE FIGURE. These are typically clauses describing or evaluating a situation or
state of affairs rather than construing a dynamic activity sequence. An example is the second
instance in Figure 7 where ‘WFH discourse’ is evaluated as ‘so frustrating’.
Figure 4 Figure system adapted from Hao (2020, p. 64) with linguistic examples from the corpus.
However, these discursive formulations have been developed for language, and the extent to
which they may be able to account for intermodal relations is still under investigation. For a
relevant study considering their application to speech-gesture semiosis see Logi, Zappavigna,
and Martin (in press); Ngo et al. (2021).
5.2 Intermodal coupling
CATEGORISATION
source
thing
activity
semiotic
place
me
They just announced total lockdown in economic
sector & my boss said i still might have to
work from home.
Hah I wish! I work from home
No FORCED INJECTION to work from home � � � �
i love working from home ..i just hate doing
accomplishment reports �
In these pandemic times …”working from home”
depending on Airtel network is such a
nightmarish experience!!� �
entity
I misplaced my coffee while I’m working from
home...HOW?!
occurrence figure
state figure
figure TYPE +Occurrence
+En ty
Some guy working from home is
taking a nap �
Thank you ❤ the whole WFH
discourse is so frustrating on all
fronts, truly.

The core unit of analysis used to understand emoji-language semiosis was an ‘intermodal
coupling’ interpreted using the principal of ‘minimum mapping’ (Zhao, 2010, 2011) to
understand how language and emoji coordinate to make meaning. Minimum mapping was
originally developed for studying the coordination ideational meaning in language and
images in web-based digital learning resources. It offers a principled approach for analysing
the distribution of meaning across modes. The key idea is that if multiple modes co-express
dimensions of social actions (such as an activity or participants) then they can be said to form
a coupling (e.g. a verbiage-image coupling). In terms of ideational relations between images
and verbiage, Zhao (2011) defines three main organising principles for couplings:
abstraction, generalisation and specification, that we interpret as providing useful insights
into ideational convergence. Abstraction involves semantic relations of ‘is’ or ‘standing for’;
generalisation refers to semantic relations of co-classification and exemplification, and
specification incorporates semantic relations that specify circumstances. The specific
hyponyms, hypernyms, meronyms etc. used in the diagrams illustrating semantic relations in
the analysis sections were guided by the Synset (semantic) relations defined in the lexical
database, WordNet (Miller, 1995).
Along these lines, while a particular resource, such as an emoji, has the potential to construe
generalized or multiplied meanings, when construed in a particular text with another
resource, such as a particular linguistic choice, then their meaning potential is constrained to
a shared region of meaning. Another way of putting this is that there is a logically resolvable
relationship between ideational meaning realised as an emoji and ideational meaning realised
in accompanying language. For example, the ‘Woman Technologist: Light Skin Tone’ emoji
in the following post does not necessarily denote the linguistic OCCURRENCE ‘work from
home’ (Figure 5). However, in combination the meaning across these two semiotic modes
construes the more committed entity that might be glossed as ‘woman who is working from
home’.
Figure 5 An example of a shared region of meaning in a language-emoji coupling.
Similarly, in the tweet in Figure 6 the brand of vaccination dose is specified in the language
but not the ‘syringe’  emoji:
Figure 6 An example of a language committing more delicate meaning.
The concept of ‘commitment’ has been used to assist interpretation of the dimensions of
meaning that are contributed by each mode in an intermodal coupling. This involves
considering “the amount of meaning potential activated in a particular process of
instantiation” in terms of the options taken up in particular discursive systems (Martin, 2008,
p. 45). Different modes may converge or diverge across visual and verbal modalities in terms
of how they commit a meaning, a notion which Painter, Martin, and Unsworth (2013) refer to
as intermodal complementarity. As we will see in the analysis which follows, both emoji and
language can commit more or less delicate meanings depending on the co-text and the
context.

6. Analysis
The most frequent emoji in the corpus appear to have an interpersonal function, that is, a
function that “is concerned with enacting interpersonal relations”, for instance through
negotiating attitudes and adopting intersubjective stances and affiliations (Halliday &
Matthiessen, 1999/2006, p. 7). These emoji are likely involved in the expression of
evaluation and negotiation of solidarity or discord (e.g. ‘Face with Tears of Joy’ (n=1,
freq.=4196), Loudly Crying Face (n=2, freq.=2610), and Rolling on the Floor Laughing
(n=3, freq.=1848)). For example, in the following Twitter interaction the ‘Face with
Tears of Joy’ in the response by User 2 suggests a playful jocular meaning where the
emoji appears to function as a laugher token:
User 1: Snooze, snooze again and repeat.
User 2: Now I know why people are late to meetings even while wfh
This playfulness can also be interpreted as aligning with the everyday experience construed
by User 1 of hitting the snooze button on an alarm clock in order to stay in bed longer. This
type of humorous solidarity around the shared frustrations of working from home was a
frequent interpersonal pattern in the corpus.
While interpersonal meaning is extremely important in negotiating affiliation in the corpus,
the focus of the present study is on ideational meaning. The most frequent ‘object’ emoji in
the corpus likely to be involved in construing ideational meanings are shown in
Table 3. Most of these tended to occur in marketing tweets or tweets advertising jobs that
were not detected in the initial corpus cleaning phase. Because of the diverse nature of this
register, it was not possible to remove these tweets automatically.

Table 3 Most frequent object emoji in the corpus.
Approaching the corpus qualitatively, inspection of concordance lines revealed two key
semantic domains or fields in which many emoji converged with language in terms of
ideation: WFH and COVID-19 pandemic (Figure 7). The sections which follow are focused
on entities and occurrences in these two areas, with a particular focus on the ‘Hot Beverage’
emoji (freq. 166) which was the most frequent emoji that consistently occurred in
everyday discourse about workers’ experiences of WFH (as opposed to the discourse of
employers, marketers and others).
Figure 7 Emoji related to WFH and COVID-19 as semantic domains or fields.
6. 1 Ideational convergence and WFH
Emoji co-construed ideational meanings about WFH as both FIGURES (STATE FIGURES and
OCCURRENCE FIGURES) and ELEMENTS (ENTITIES, OCCURRENCES, and QUALITIES) in
combination with language. For example, the tweet in Figure 8 construes a STATE FIGURE (my
wfh office was busy 󱇴󱇵󱇶󱇷) evaluating the WFH office environment that is the projected opinion
of the author.

Figure 8 An example of synecdoche in an ideational coupling
Here the ‘computer; emoji has a metonymic relationship to the ‘wfh office’ PLACE ENTITY,
appearing to enact synecdoche since the WFH office is represented by the computer, just one
aspect of the items routinely found in offices.
An example of a FIGURE co-construed in emoji and language is the tweet in Figure 9. This
post is a reply to a tweet ‘Me 30 seconds after joining a Zoom when no one else has joined’
which included an image of Meryl Streep with the caption ‘Why is no one ready?’. The four
emoji converge with ‘getting all glammed up’ to realise an OCCURRENCE FIGURE about the
activity of getting ready in order to be presentable for a zoom meeting. The emoji construe a
sequence of OCCURRENCES that are metonyms realising different aspects of getting ‘glammed
up’. The ‘Woman Technologist’ emoji appears to indicate the outcome of the activity,
although, interestingly did not occur at the end of the sequence as might have been expected.
Figure 9 An example of emoji and language realising an OCCURRENCE FIGURE.
The tweet in Figure 10 features the ‘alarm clock emoji’ which converges with meanings
about both about ‘time running out’ and ‘times changing’ in the verbiage. From the
perspective of interpersonal meaning, the emoji functions as an evaluative metacomment
similar to meanings achieved by hashtags in this culminative position.

Figure 10 An example of an ‘object’ emoji with a dual ideational/interpersonal function.
These three examples suggest that emoji can converge with language both inside and outside
the clause, in a similar way to hashtags as described by Zappavigna (2015, 2018). For further
work on the evaluative functions of emoji see Logi and Zappavigna (2021), we will explore
syntagmatic relations in the next section.
6. 2 Patterns of meaning involving the ‘Hot Beverage’ emoji
As mentioned earlier, the ‘hot beverage’ emoji was the most frequent object emoji
involved in discourse about everyday work. Indeed, other studies have noted the role that
discourse about coffee plays in ambient affiliation on social media across multiple contexts
involving people talking about their everyday lives (Zappavigna, 2014). Before, considering
the details of the discourse semantic meanings in which this emoji is involved, it is useful to
look at its involvement in syntagmatic relations.
Syntagmatic relations: emoji in initial, integrated and culminative position
Drawing on an approach adopted for considering hashtags in terms of the information flow of
a text (Zappavigna, 2018), we can consider where an emoji appears syntactically in the posts
in terms of whether they appear in an “initial” position at the beginning of a tweet, are
“integrated” into the lexicogrammar of the post, or occur in “culminative” position at the end
of the post, along with other elements such as hashtags and other metadata. The ‘hot
beverage’ emoji most commonly occurred in culminative position (freq.=81; 61%) (Figure
11)

Figure 11 Examples of the ‘hot beverage’ emoji in culminative position
2
.
This included emoji that co-occurred with hashtags in culminative position, as in the final
two examples in Figure 11.
The next most frequent position for the ‘Hot Beverage’ emoji was integrated position
(freq.=51, 38%), where it could perform functions both inside and outside the clause (Figure
12):
Figure 12 Examples of the ‘hot beverage’ emoji in integrated position.
In these cases the emoji can perform a range of functions inside the clause, including
specifying the beverage mentioned in the post ,as in the first two examples in Figure 12, or
engaging in an metonymic relation with ‘morning’ as in the final example (since coffee is a
beverage that is frequently shared in material posted in the morning on social media
platforms (Zappavigna, 2014). Indeed, coffee sometimes co-occurred with a ‘good morning’
greeting. The hot beverage emoji rarely occurred in initial position (freq.=2, 2%).
2
The final example includes the custom Twitter ‘stay home’ emoji that is enabled when a user includes #stayhome
in a tweet.

More generally, emoji can realise ideational clause constituents that appear to be specified to
a greater extend by field and syntagmatic relations than solely by shared ideational taxonomy.
For example, consider the emoji in Figure 13. Here the emoji occupies a position in the
clause that would usually be occupied by a THING ENTITY given the grammatical tendency (in
English) for a Complement (the object) to follow a Finite (the verb). Together the language
and emoji form an OCCURRENCE FIGURE.
Figure 13 An example of emoji realising a clause constituent.
Thus, a speaker is ‘primed’ to interpret the semiosis following drinking as more likely to be
the Complement of the clause for work on semantic priming and emoji see (Weissman &
Tanner, 2018; Yang, Yang, Xiu, & Yu, 2021). The second feature reflected in this
interpretation is an ideational field relation, narrowing the potential ideational meanings
following the instantiation of drinking to things that can be drunk, thus constraining the scope
to meanings that fit this field. The above example is ambiguous regarding whether the ‘hot
beverage’ in question is tea, coffee or some other heated beverage. In some cases, this kind of
indeterminancy is resolved through L1 collocation with a more committed linguistically
realised THING ENTITY (Figure 14):
Figure 14 Examples of  in culminative position and co-classified by L1 lexis.
A more complex example is the post in Figure 15 which bundles linguistically-realised
ENTITIES as adjacent emoji. The activity entity ‘dance party’ and the thing entity ‘coffee’ are
converge with the ‘hot beverage’ and the ‘Woman Dancing: Medium-Light Skin Tone’
emoji. This post also incorporated a video of a mother dancing with her children that
humorously identifies the ‘tiny coworkers’.

Figure 15 An example of emoji realising adjacent THING ENTITIES.
6.1.2 Collocates of the ‘hot beverage’  emoji
Returning to a more global perspective, the most frequent collocate of the ‘Hot Beverage’
emoji was another itself, as the concordance lines in Figure 17 suggest. The repetition of this
emoji in first two posts in Figure 16 upscales quantity, appearing to construe a ‘lots of hot
beverage’ THING ENTITY. The third, example, however, rather than upscaling quantity,
appears to upscale the positive attitude about tea. This in accord with the findings of Logi and
Zappavigna (2021) who note that the repetition of emoji can serve to upscale co-occurring
attitudinal meanings
Figure 16 Examples of repetition of 

Another example involving repetition is the post in Figure 17, a reply to a tweet about
another user’s breakfast choices. In this example the repetition of the emoji echoes the plural
THING ENTITY in the verbiage. In terms of intermodal coupling, ‘home made latte macchiatos’
in the language serves to specify the type of beverage as coffee rather than tea or some other
hot drink. It also upscales the positive attitude in the post realised by ‘advantage’ and ‘enjoy’.
Figure 17 An example of an emoji with a dual ideational and interpersonal function.
The posts in Figure 18 are both examples of how particular conative dimensions of the ‘hot
beverage’ emoji such as comfort and familiarity can be foregrounded by the linguistic co-
text. The coordination of meaning between language and emoji in this example illustrates
how the verbiage and emoji reciprocally specify each other’s meaning potential to a region of
shared meaning. In this case, connotations of comfort and familiarity associated with a ‘hot
beverage’ (realised by the emoji) are foregrounded when cross-referenced with similar
connotations associated with working from home. Thus, even though this region of each
mode’s meaning potential is typically less salient, it is foregrounded when realised
intermodally.
Figure 18 Emoji foregrounding conative meanings – comfort/familiarity

Other posts, such as the example in Figure 19, expressing how much the user missed face-to-
face interactions, manifest a coordination of similar conative meanings. In this example the
use of the ‘Hot Beverage’ emoji references its common usage in texts where the author is
describing ‘having a coffee’ with someone else. This is an example of how the ‘fuzziness’ of
an individual emoji can be used to simultaneously construe multiple aspects of meaning – in
this case, the simultaneously refers to the THING ENTITY ‘coffee’, the classifier in the
THING ENTITY ‘coffee mates’, and the OCCURRENCE FIGURE ‘have a coffee’.
Figure 19 Emoji foregrounding conative meanings - conviviality
We might think of the degree of intermodal convergence of emoji and language as a cline,
spanning more ‘localised’ co-textual convergence to more ‘global’ meanings requiring more
activation of the context. For instance, a mapping of ‘coffee’ or ‘tea’ with might be the a
maximally localised meaning. However, some relations involve more extensive knowledge of
the context. For example, consider the tweet in Figure 20 which included a gif of a child on a
bike circling rapidly around a playground, converging with meanings about coffee causing
‘speediness’ and ‘busyness’. In this example, there is a connotative relation between the ‘Hot
Beverage’ emoji and the concept of high-speed activity or excessive (‘manic’) business
(‘ManicMonday’ also references a popular phrase and song). Here the convergence of
meaning relies on more this contextual knowledge and knowledge about the cultural
meanings that coffee and caffeine have accrued over time, such as facilitating productivity,
wakefulness, and creativity. These meanings, due to “the powerful cultural impact of
iconization, these apparently positive properties of caffeine have great influence on the daily
practices of many people” (Zappavigna, 2014, p. 143) and are echoed in the meanings made
in the gif. We will explore relations between the post and images/video/gifs in the next
section.

Figure 20 An example of an emoji, language and a gif activating connotative meanings.
6. 2 Factoring in the role of images
A consistent pattern in the corpus was the embedding of images of coffee into posts that
included the ‘Hot Beverage’ emoji, together with discourse about dimensions of the
user’s visual perspective or sensory experience. The concordance lines in Figure 21 are
examples of posts that were accompanied by images in which tea coffee or are salient, which
Zappavigna and Zhao (2020) term ‘coffee selfies’:
Figure 21 Examples of posts accompanied by ‘coffee selfies’.
These images were often inferred selfies (e.g. the photographer’s hand depicted holding a
coffee cup) or implied selfies, and often included a laptop (e.g. Figure 23). For instance, the
image in Figure 22 is an implied selfie where the compositional structure suggests the
presence of the photographer, in this case of the photographer seated at a laptop at a desk. In
order to interpret the role of these images, we return to Zhao’s (2011) framework for image-
verbiage couplings referred to in the explanation of minimum mapping earlier. Zhao defines
three main organising principles for couplings: abstraction, generalisation and specification
(Table 2).

Depending on the level of granularity of analysis, the tweet in Figure 22 can be interpreted as
involving 3 key intermodal relations in terms of ideational meaning as it is distributed across
the image, text, emoji and tags. The first is the circumstantiation of an OCCURRENCE FIGURE
(1), realised in the body of the post as ‘WFH on a Sunday’ with the ‘laptop emoji’ and the
‘hot beverage’ emoji construing meanings about drinking a hot beverage at a laptop. The
‘Hot Beverage’ and ‘Laptop’ emoji co-exemplify the image of the laptop (if we
assume part of a laptop stand is visible in the top right corner) and the mug containing a dark
liquid that is most likely coffee (2). The linguistic THING ENTITY ‘magic mug from
@amojinph @amojinprints’ identifies the mug by naming and tagging the designer (3).
Figure 22 Ideational intermodal relations in a tweet containing an OCCURRENCE FIGURE.
If we consider relations internal to the body of the post the laptop and hot beverage emoji
converge in an OCCURRENCE FIGURE. In addition, in the second clause, the mug is specified as
‘magic’ which, after inspecting the designer’s site reveals that the ‘Sparkles’ emoji
functions as a classifier rather than the other likely reading of positive evaluation, since this
user sells mugs that reveal an image when hot water is added.
Another example of ideational intermodal relations is a post containing an implied selfie
featuring coffee is Figure 23. In this case the image functions to circumstantiate a STATE
FIGURE about the authors opinion on the view from her working location. The sequence of
‘Hot Beverage’, ‘Laptop’, and ‘Books’ emoji co-exemplifies these entities in the image (with
the books appearing on the laptop screen as well as in the verbiage ‘Books for Topics’). The
‘Sunglasses’ emoji implies that the photographer is wearing sunglasses.

Figure 23 Ideational intermodal relations in a tweet containing a STATE FIGURE .
7. Conclusion
This paper has explored the coordination of ideational meaning in tweets about working from
home, adopting a social semiotic perspective. It focused on how language and emoji
converge to construe meanings about workers’ experience of working away from the office
during the COVID-19 pandemic. The study forms part of a broader project on the social
semiotics of working from home, which both considers how worker’s frame their experience
of remote work and how they communicate through videoconferencing platforms. Emoji are
resources often used in the chat function of such platforms (Dürscheid & Haralambous,
2021). Future work might draw on the analytical approach developed here to explore how
emoji are used to negotiate the complex meanings that are made people collaborate and
maintain social affiliations in their online work interactions.
The approach adopted in this paper refines the ideational dimension of a broader
metafunctional framework, that also accounts for interpersonal and textual (organising
discourse) functions of emoji (Logi & Zappavigna, 2021), and aims to make the modelling of
ideational relations more delicate. The approach has synergies with the analytical methods
applied in Zappavigna (2015, 2018) for understanding the function of hashtags in social
media discourse. The multimodal approach adopted is also congruent with work informed by
similarly intermodal frameworks such as the adapted rhetorical structure theory employed by
Ge and Herring (2018) and the grapholinguistic approach of Dürscheid, Meletis, and
Haralambous (2019). It also aligns with the mention, restatement and illustration functions
identified by (Dainas & Herring, 2021).
Emoji raise interesting theoretical issues about what counts as a ‘unit of meaning’ and a
semantic ‘relation’ in the multimodal analysis of social media texts. Throughout this paper
we have drawn on discourse semantic relations that were originally developed for work on
language, to see how far we could push this modelling to account for emoji-language
semiosis. We have explored how emoji and language can function as semiotic resources that

coordinate to make these discourse semantic meanings together (e.g. ‘coffee’ and the ‘Hot
Beverage’ emoji converging to construe a THING ENTITY).
The exploration of the ‘hot beverage’ emoji in particular, offers an insight into the ways
in which emoji enact a ‘multiplication of meaning’ (Lemke, 1998) in social media discourse.
In addition to its role in construing ELEMENTS and FIGURES, this emoji was involved in more
complex meanings, particularly when we factored in relations to the images in the tweets,
drawing upon Zhao’s (2011) principles of generalisation, abstraction and circumstantiation.
This emoji also appeared to function as a bonding icon (Stenglin, 2008), due to its accrual of
positive associations relating to productivity and alertness. This resonates with the
conclusions of other studies of social media language and images about coffee (Zappavigna,
2014), and about the kinds of intersubjective relations that are realised in particular kinds of
‘coffee selfies’ (Zappavigna & Zhao, 2020; Zhao & Zappavigna, 2018).
Like hashtags, emoji are very flexible in the kinds of relations they can enact in coordination
with the verbiage of a social media post. This paper has explored examples where emoji
coordinate within and outside the clause, where they function in tandem with the linguistic
syntax, and where they forge relations with the represented objects and participants in
images. We have leant toward referring to these relations as ‘intermodal couplings’ since the
visual dimension
3
of emoji is so important, and have drawn upon Zhao’s (2011) concept of
minimum mapping, originally develop for image-language relations, in order to investigated
the distribution of shared meaning in emoji-language semiosis. At the same time, we
acknowledge that emoji can be thought of as operating as paralanguage, supporting meanings
in the verbiage, a point more apparent in the role of emoji play in interpersonal meaning-
making. Teasing out how to best model these complex relations in larger datasets across more
contexts and registers will likely require substantial caffeination .
3
Indeed the aesthetic function of emoji is an aspect worthy of investigation that was beyond the scope of this
paper.

8. References
Al-Rawi, A., Siddiqi, M., Morgan, R., Vandan, N., Smith, J., & Wenham, C. (2020). COVID-19 and the
Gendered Use of Emojis on Twitter: Infodemiology Study. Journal of Medical Internet
Research, 22(11), e21646.
Albawardi, A. H. (2018). Digital literacy practices of Saudi female university students. University of
Reading,
Aull, B. (2019). A study of phatic emoji use in WhatsApp communication. Internet Pragmatics, 2(2),
206-232.
Barrero, J. M., Bloom, N., & Davis, S. J. (2021). Why working from home will stick Working Paper 28731.
In: National Bureau of Economic Research.
Beißwenger, M., & Pappert, S. (2019). How to be polite with emojis: a pragmatic analysis of face work
strategies in an online learning environment. European Journal of Applied Linguistics, 7(2),
225-254.
Chik, A., & Vásquez, C. (2017). A comparative multimodal analysis of restaurant reviews from two
geographical contexts. Visual Communication, 16(1), 3-26. doi:10.1177/1470357216634005
Collins, L. C. (2020). Working with images and emoji in the Dukki Facebook Corpus. In S. Rdiger & D.
Dayter (Eds.), Corpus Approaches to Social Media (pp. 175-198). Benjamins: Amsterdam /
Philadelphia.
Dainas, A., & Herring, S. C. (2021). Interpreting emoji pragmatics. In C. Xie, F. Yus, & H. Haberland
(Eds.), Approaches to Internet Pragmatics: Theory and practice (pp. 107-144). Amsterdam /
Philadelphia: John Benjamins.
Danesi, M. (2016). The semiotics of emoji: The rise of visual language in the age of the internet:
Bloomsbury Publishing.
Das, A. (2021). How has the coronavirus (COVID-19) pandemic affected global emoji usage? Journal of
Human Behavior in the Social Environment, 31(1-4), 425-434.
doi:10.1080/10911359.2020.1838383
De Cock, B., & Pizarro Pedraza, A. (2020). Any #JesuisIraq planned?: Claiming affective displays for
forgotten places. Pragmatics, 30(2), 201-221. doi:https://doi.org/10.1075/prag.18059.dec
Dresner, E., & Herring, S. C. (2010). Functions of the nonverbal in CMC: Emoticons and illocutionary
force. Communication theory, 20(3), 249-268.
Dürscheid, C., & Haralambous, Y. (2021). Emojis are everywhere. How emojis conquer new contexts.
Grapholinguistics and its applications(4), 501-512.
Dürscheid, C., Meletis, D., & Haralambous, Y. (2019). Emojis: a grapholinguistic approach.
Grapholinguistics and its applications(1), 167-183.
ECMA. (2017). ECMA-404 The JSON data interchange syntax. Retrieved from https://www.ecma-
international.org/publications-and-standards/standards/ecma-404/
Gawne, L., & McCulloch, G. (2019). Emoji as digital gestures. Language@ Internet, 17(2).
Ge, J. (2019). Emoji sequence use in enacting personal identity. Paper presented at the Companion
proceedings of the 2019 world wide web conference.
Ge, J., & Herring, S. C. (2018). Communicative functions of emoji sequences on Sina Weibo. First
Monday.
Hafner, C. A. (2021). Discourse and computer-mediated communication. In K. Hyland, B. Paltridge, &
L. Wong (Eds.), The Bloomsbury Handbook of Discourse Analysis (Vol. 2, pp. 281-294).
Halliday, M. A. K. (1978). Language as social semiotic: The social interpretation of language and
meaning. London: Edward Arnold.
Halliday, M. A. K., & Matthiessen, C. M. I. M. (1999/2006). Construing Experience through Language:
a language-based approach to cognition. London: Continuum.
Hao, J. (2015). Construing biology: An ideational perspective. (Linguistics). University of Sydney,
Unpublished PhD thesis.

Hao, J. (2020). Analysing scientific discourse from a systemic functional linguistic perspective: A
framework for exploring knowledge building in biology: Routledge.
Herring, S. C., & Dainas, A. (2017). “Nice picture comment!” Graphicons in Facebook comment threads.
Paper presented at the Proceedings of the 50th Hawaii International Conference on System
Sciences, Hawaii, USA, January 4-7.
Herring, S. C., & Dainas, A. R. (2020). Gender and Age Influences on Interpretation of Emoji Functions.
ACM Transactions on Social Computing, 3(2), 1-26.
König, K. (2019). Stance taking with ‘laugh’ particles and emojis – Sequential and functional patterns
of ‘laughter’ in a corpus of German WhatsApp chats. Journal of Pragmatics, 142, 156-170.
doi:https://doi.org/10.1016/j.pragma.2019.01.008
Konrad, A., Herring, S. C., & Choi, D. (2020). Sticker and emoji use in Facebook messenger: implications
for graphicon change. Journal of Computer-Mediated Communication, 25(3), 217-235.
Kress, G. R., & Van Leeuwen, T. (2001). Multimodal discourse: The modes and media of contemporary
communication. Edward Arnold: London.
Lazzini, A., Lazzini, S., Balluchi, F., & Mazza, M. (2021). Emotions, moods and hyperreality: social media
and the stock market during the first phase of COVID-19 pandemic. Accounting, Auditing &
Accountability Journal, https://doi.org/10.1108/AAAJ-1108-2020-4786.
Lemke, J. L. (1998). Teaching all the languages of science: Words, symbols, images, and actions. Paper
presented at the Conference on science education in Barcelona.
Li, L., & Yang, Y. (2018). Pragmatic functions of emoji in internet-based communication---a corpus-
based study. Asian-Pacific Journal of Second and Foreign Language Education, 3(1), 16.
doi:10.1186/s40862-018-0057-z
Logi, L., & Zappavigna, M. (2021). A social semiotic perspective on emoji: How emoji and language
interact to make meaning in digital messages. New Media & Society,
https://doi.org/10.1177/14614448211032965.
Logi, L., Zappavigna, M., & Martin, J. R. (in press). Bodies talk: Modelling paralanguage in Systemic
Functional Linguistics. In D. Caldwell, J. R. Martin, & J. Knox (Eds.), Developing Theory: A
Handbook in Appliable Linguistics and Semiotics. London: Bloomsbury.
Martin, J. R. (1992). English text: system and structure. Amsterdam: Benjamins.
Martin, J. R. (2008). Intermodal reconciliation: Mates in arms. In L. Unsworth (Ed.), New literacies and
the English curriculum: Multimodal perspectives (pp. 112-148). London: Continuum.
McNeill, D. (1992). Hand and mind: What gestures reveal about thought. Chicago: University of
Chicago press.
Miller, G. A. (1995). WordNet: A Lexical Database for English. Communications of the ACM, 38(11), 39-
41.
Miltner, K. M. (2020). “One part politics, one part technology, one part history”: Racial representation
in the Unicode 7.0 emoji set. New Media & Society, 1461444819899623.
Ngo, T., Hood, S., Martin, J. R., Painter, C., Smith, B., & Zappavigna, M. (2021). Modelling Paralanguage
Using Systemic Functional Semiotics: Theory and Application. London: Bloomsbury.
Nishimura, Y. (2015). A sociolinguistic analysis of emoticon usage in Japanese blogs: Variation by age,
gender, and topic. AoIR Selected Papers of Internet Research.
O'Halloran, K. L. (2006). Multimodal discourse analysis: Systemic-functional perspectives. London:
Bloomsbury Publishing.
Oomori, K., Shitara, A., Minagawa, T., Sarcar, S., & Ochiai, Y. (2020). A Preliminary Study on
Understanding Voice-only Online Meetings Using Emoji-based Captioning for Deaf or Hard of
Hearing Users. Paper presented at the The 22nd International ACM SIGACCESS Conference on
Computers and Accessibility, Virtual Event, Greece.
https://doi.org/10.1145/3373625.3418032
Painter, C., Martin, J. R., & Unsworth, L. (2013). Reading visual narratives. London: Equinox.

Parkwell, C. (2019). Emoji as social semiotic resources for meaning-making in discourse: Mapping the
functions of the toilet emoji in Cher’s tweets about Donald Trump. Discourse, Context &
Media, 30, 100307.
Rodrigues, D., Prada, M., Gaspar, R., Garrido, M. V., & Lopes, D. (2018). Lisbon Emoji and Emoticon
Database (LEED): Norms for emoji and emoticons in seven evaluative dimensions. Behavior
Research Methods, 50(1), 392-405. doi:10.3758/s13428-017-0878-6
Sampietro, A. (2016). Exploring the punctuating effect of emoji in Spanish Whatsapp chats. Lenguas
Modernas(47).
Sandel, T. L., Ou, C., Wangchuk, D., Ju, B., & Duque, M. (2019). Unpacking and describing interaction
on Chinese WeChat: A methodological approach. Journal of Pragmatics, 143, 228-241.
doi:https://doi.org/10.1016/j.pragma.2018.08.011
Skovholt, K., Grønning, A., & Kankaanranta, A. (2014). The communicative functions of emoticons in
workplace e-mails. Journal of Computer-Mediated Communication, 19(4), 780-797.
Stenglin, M. (2008). Olympism: how a Bonding icon gets its “charge”. Multimodal Semiotics: Functional
Analysis in the Contexts of Education, 50-66.
The Unicode Consortium. (2021). The Unicode Standard. Retrieved from
http://www.unicode.org/versions/latest/ last accessed 7 June 2021
Thompson, D., & Filik, R. (2016). Sarcasm in written communication: Emoticons are efficient markers
of intention. Journal of Computer-Mediated Communication, 21(2), 105-120.
Weissman, B., & Tanner, D. (2018). A strong wink between verbal and emoji-based irony: How the
brain processes ironic emojis during language comprehension. PloS one, 13(8), e0201727.
Yang, J., Yang, Y., Xiu, L., & Yu, G. (2021). Effect of emoji prime on the understanding of emotional
words–evidence from ERPs. Behaviour & Information Technology, 1-10.
Yus, F. (2021). Pragmatics, humour and the internet. Internet Pragmatics, 4(1), 1-11.
doi:https://doi.org/10.1075/ip.00067.yus
Zappavigna, M. (2014). Coffeetweets: Bonding around the bean on Twitter. In P. Seargeant & C. Tagg
(Eds.), The language of social media: Communication and community on the Internet (pp. 139
-160). London: Palgrave.
Zappavigna, M. (2015). Searchable talk: the linguistic functions of hashtags. Social Semiotics, 25(3),
274-291.
Zappavigna, M. (2018). Searchable Talk: Hashtags and Social Media Metadiscourse. London:
Bloomsbury.
Zappavigna, M., & Zhao, S. (2020). Selfies and recontextualisation: A social semiotic perspective on
the visual structure of Instagram images. In M. Miles & E. Welch (Eds.), Photography and its
Publics (pp. 205-227). London: Bloomsbury.
Zhao, S. (2010). Intersemiotic relations as logogenetic patterns: the time factor in hypertext
description. In M. Bednarek & J. R. Martin (Eds.), New Discourse on Language: Functional
Perspectives on Multimodality, Identity, and Affiliation (pp. 195-218). London: Continuum.
Zhao, S. (2011). Learning Through Multimedia Interaction: The Construal of Primary Social Science
Knowledge in Web-based Digital Learning Materials. (Doctor of Philosophy). University of
Sydney, Sydney.
Zhao, S., & Zappavigna, M. (2018). Beyond the self: Intersubjectivity and the social semiotic
interpretation of the selfie. New Media & Society, 20(5), 1735-1754.