Emojis: Langue or Parole?

1 Introduction

Emojis spread throughout the world after they became broadly available on keyboards, apps, websites, etc. with the launch of Unicode 8 in 2015; and the emoji lexicon is constantly being enlarged in response to changing needs and trends across languages and cultures. It has become obvious that emojis can no longer be considered an ancillary set of picture words for sprucing up informal written messages. They are being used in all areas of social discourse, from advertising to political campaigning. Their growing importance was acknowledged in 2015 when the Oxford Dictionary chose, as its “Word of the Year,” the “Face with Tears of Joy” emoji:

Figure 1

Face with tears of joy emoji

On its website, the Dictionary justified its choice of an emoji over a traditional phonetic word by claiming that it “captures the ethos, mood, and preoccupations” of the contemporary world. Clearly, the spread of emojis harbors a broad range of implications, from the possible demise of traditional print literacy practices to the evolution of communicative systems and practices that are being shaped more and more by socio-technological forces. A central semiotic question that emojis raise takes us to the doorstep of Saussure (1916): Are emojis a new langue or a new form of parole constrained to specific kinds of digital texts? The emoji phenomenon actually presents us with a “case study” for examining the Saussurean dichotomy and his belief that the two – langue and parole – were not mutually reciprocal systems, but autonomous ones. As will be argued in this paper, emojis appear to collapse this dichotomy – a topic that has become of crucial importance today amid the many revisitations of Saussurean theory and method (Harris 2001; Sanders 2004; Bouissac 2010; Joseph 2012; Thibault 2013; Weber 2017). Emojis also seem to eliminate the traditional distinction between separate verbal and visual modes of representation and communication – that is, between verbality and visuality. Saussure himself (1916: 68, 112) had suggested that verbal language was “the most complex and universal,” and that this was so because “There are no preexisting ideas, and nothing is distinct before the appearance of language.” The sustainability of this premise is clearly testable as well in terms of the emoji movement.

2 Background

An important factor in the rise and spread of emojis is the modern historical context that preceded their emergence (Danesi 2016). It is unlikely that these picture words would have emerged in the first place if modern-day people were not pre-conditioned to accept them as an outgrowth of various trends that took shape in the late nineteenth and early twentieth centuries. One of these is the comic book, which blends images and words in a narrative framework. Artistic movements such as Dada, Futurism, among others, also showed that the printed (phonetic) word, laid out in a linear fashion, was losing its primacy in the domain of written communication. The writer-mathematician Lewis Carroll was already experimenting with visual-iconic writing in the middle part of the nineteenth century, as can be seen in the layout of his “Mouse’s Tale,” which appears in his novel, Alice’s Adventures in Wonderland (1865). The layout resembles the physical form of a “tail,” a word which is a homograph of “tale”:

Figure 2

Lewis Carroll’s “The Mouse’s Tale” (Wikimedia Commons)

Attempts to make writing more visual can also be seen in an 1881 issue of Puck Magazine, a now defunct humor magazine. Its suggestions prefigure modern-day emoticons:

Figure 3

Puck Magazine (1881)

The intellectual and social groundwork for emojis was thus laid by trends such as these. The relevant digital technology came forth simply to provide the physical tools to realize emoji writing in a user-friendly way. Of course, visuality in writing practices existed before the modern era, as can be seen in medieval and Renaissance rebus forms of writing, illustrated manuscripts, and the like. But these did not spread because of the effort, expense, and artistic talent that was involved in realizing them. Digital keyboards, websites, and apps, on the other hand, have made visuality an option easily available to everyone.

It is also relevant to note that emojis as such did not emerge in a representational vacuum. They are an end-point in the evolutions of emoticons and kaomoji. Without going into the historical details here, since they are well known (Danesi 2016; Evans 2017), a linear path can be traced from the latter two to emoji, as shaped by developing technologies. This path can be shown as follows:

Figure 4

Evolution from emoticon to emoji

An emoticon represents a facial expression iconically with combinations of keyboard characters, including punctuation marks, numbers, and letters. Emoticons have been traced as far back as the early 1970s. The word kaomoji refers to Japanese emoticons. These came to prominence in the mid-1980s, migrating to other orthographic systems for a while. Finally, the first emoji was created in Japan in 1999 by Shigetaka Kurita, who was likely using kaomoji as a prototype, creating stock symbols to stand for all kinds of referents, not just faces.

The notion of visuality was discussed at length by Rudolf Arnheim (1969), who challenged the traditional differentiation between “thinking” (associated with language) and “perceiving” (associated with art), claiming that the two have been artificially separated, since they both co-occur in the processing of messages. Visuality has been studied extensively under the rubric of visual semiotics (Barthes 1977; Krampen 1991; Sebeok and Umiker-Sebeok 1994; Jappy 2013). The focus in this branch has been largely (albeit not exclusively) on how visual signifiers such as points, lines, and shapes cohere into meaning structures such as signs and texts. Arrows, for example, are indexical signs that allow spatial and orientational representation. Shapes represent the outline of something iconically – a circle can stand for the sun or the face; a rectangle can represent the surface of a table; and so on. So, in drawings of scenery, a cloud can be represented as a shape, and the horizon as a line. Other visual signifiers include value, color, and texture. Value refers to the darkness or lightness of a line or shape. Color conveys mood, feeling, and atmosphere. Texture refers to the phenomenon that certain visual signifiers evoke tactile and other sensory modalities – wavy lines tend to elicit pleasant sensations, whereas angular ones tend to elicit opposite reactions. Emojis embody all these representational modalities (to varying degrees). For this reason, they cannot be studied in the same way as images in terms of signifier status. They are meaningful units in themselves – hence their designation as picture words.

It is relevant to note that the earliest pictographs were not true reproductions of their referents; they were mainly outlines or sketches of them, as can be seen in the carvings of animals which cover the roofs and walls of ancient caves. The work of Schmandt-Besserat (1978, 1989, 1992) has shown, in fact, that the first writing symbols were probably made possible from the creation of clay tokens, which seem to be image-reproducing objects, much like molds or type-setting stamps. It is not a stretch to say that the emoji keys of today have retrieved the same kind of function of these clay tablets.

3 Emojis as langue

Saussure defined langue as a system of abstract rules used by a speech community, in contrast to the actual linguistic behavior of the individuals in that community, which he designated as parole. Langue involves what Chomsky (1957) called a linguistic competence and the abstract rules that realize it a generative grammar – that is, a set of rules that people use unconsciously to generate sentences. Leaving aside the theoretical problems associated with the Saussurean-Chomskyan paradigm, for the present purposes the key question that emojis elicit is whether or not they involve a grammar that is analogous to linguistic grammar and thus whether or not they constitute a veritable langue in the Saussurean sense.

The problem with ascribing emojis to the Saussurean domain of langue is that there seem to be no linguistic-type rules underlying their assemblage in texts. Rather, they appear to entail a type of visual sign system that is based on episodic, rather than purely linguistic, representation – that is, a system consisting of a series of connected referents that constitute an unconscious narration. So, rather than a strictly linguistic competence, emoji competence can be called simply, episodic, since it involves the ability to understand visual sequences, as in a comic book, and their interrelations through the practical experience of the events that the sequences encode. Unlike Chomsky’s Universal Grammar (Chomsky 2002), episodic competence is not based on a minimal set of structural rules, such as recursion, but on the experiential visualization and memory of events. An example of how episodic competence might unfold, consider the text below, which is now a famous meme in Internet culture found on many public domain websites. The intent of the text is actually enunciated at the top in words – “Dude, let’s rob a bank.”

The emoji units and sequences, arranged in list form, are laid out from top to bottom in terms of episodic events, providing a response to the exhortation to rob a bank. This vertical layout describes the unfolding of events in a sequential way. The top line portrays the sequence in temporal order: gun (the weapon for robbing the bank) + money bag (the booty) + car (escape vehicle). The grinning face emoji below constitutes an unsure and precarious reaction to the sequence. The third line then describes what might happen next – namely, police cars will be summoned to the scene, followed in the subsequent fourth line by a frowning face as a reaction to their arrival. The fifth line then portrays the likely scenario that will then unfold: guns (the shooting that will take place between the robbers and the police) + fire truck + ambulance (indicating the usual order of the arrival of these emergency vehicles, suggesting a possibility of injury). The last emoji is a double dizzy face, which is an expression of anxiety. The final query – “So that’s a no then?” – is the sender’s conclusion that he draws from the emoji sequences of his interlocutor. This message is decipherable in terms of what has been called an episodic system of understanding shaped by the experiential knowledge of a typical robbery scene, which provides an interpretative frame for decoding the text. In effect, the episodes read much like a television crime program. Without this knowledge, the message would become open to different interpretations or even be completely undecipherable in, say, a tribal culture where guns and robberies make no sense as sign forms.

Figure 5

“Dude let’s rob a bank” (EverythingFunny.org)

As this kind of text brings out, understanding and constructing an emoji text is different from a phonetically laid out one. The latter involves knowledge of phonemics (and their graphemic counterparts), grammatical structure, and vocabulary choices. An emoji text does not – it is visual, not phonemic, episodic rather than grammatical, and its vocabulary is multimodal (involving iconicity and indexicality). Needless to say, the same message could be told with words instead. But the emoji counterpart is more broadly understandable than the same message written in some specific language, since it does not require the grammatical and lexical competence of that language. So, the question of ascribing emojis to the domain of langue, as defined by Saussure, is a problematic one. Langue involves abstract (unconscious) knowledge of rule-making principles. Now, while this criterion is not applicable to emojis, the fact that their usage is broadly understandable does indeed seem to make them a kind of langue, but one that is distinctive from a linguistic langue in the Saussurean-Chomskyan sense.

A comparison between emojis, which are picture words, and phonetic words is thus essential in determining whether or not emojis constitute a langue . Phonetic words can be defined by their formal morpholological structure and by their referential purview – as nouns, verbs, etc. A similar kind of classificatory process seems to apply to emojis. A cloud emoji, for instance, is likely to be perceived as a noun corresponding to the word cloud. A sunrise emoji, on the other hand, showing the shape of a sun as it rises up from a background, can suggest either the noun sunrise or a verb describing the rising of the sun:

Figure 6

Emoji types

Now these very same emoji forms could be used to describe something in a written text (such as a cloudy mood or a sunny disposition), thus functioning also as adjectives. In the cloud emoji, the grayish-white color can also be used in messages to suggest various emotions (such as dullness or boredom); on the other hand, the sunrise emoji might suggest an uplifting of emotions or something similar. In effect, specific emojis cannot be assigned monolithically to a part of speech, because they blend several parts, with one or the other being foregrounded on the basis of the context. This makes it difficult to pigeonhole an emoji morphologically. Nonetheless, emojis have the same kind of referential functions of morphemes, albeit in more extensive ways.

Rather than speak of words as morphological classes, it is perhaps more relevant to use Edward Sapir’s (1921) notion of words as manifestations of a “vocabulary blueprint” in the human brain that allows speakers of different languages to communicate the same ideas with the specific vocabulary resources at their disposal. To show this, he got speakers of several indigenous languages of the southwestern United States to render the English sentence He will give it to you in their respective languages (Sapir and Swadesh 1946):

The fact that the English sentence was so easily translated by speakers of the above languages, despite differences in actual vocabulary and linguistic grammar, gives substance to Sapir’s claim of a vocabulary blueprint. A specific vocabulary might include information that may be excluded by others, or else it may eliminate details that others consider relevant to messages. In English, for instance, we must indicate the gender of the actor (masculine he) and the object (neuter it), as well as the number (singular in this case), and tense of the verb (future in this case will give). We do not need to indicate, as speakers of some of the other languages above need to do, the size or shape of the object, whether or not it is visible, or whether the action was observed by the speaker. In sum, Sapir’s notion can be applied to the emoji lexicon, which also seems to realize the brain’s vocabulary blueprint but in visual, rather than strictly verbal, ways

The term emoji code is often used in theoretical discussions of emojis, perhaps to avoid the formal requirements that langue entails. The term was actually introduced by Saussure (1916: 31) himself in order to differentiate it from langue, which can be made up of various codes (spoken and written). For instance, if a verbal text is written in Swahili, the speaker-hearer must know the Swahili language and its grammar (langue) in order to extract any meaning from it (the code). So, calling emojis a code is correct, since they inform us how a text is to be deciphered. The paradox, therefore, is that emojis may constitute both a langue and an interpretive code. This could be the underlying cause of why emojis, which were intended to enhance broader comprehension of written texts, irrespective of language and its orthographic features, have displayed such a high degree of variability. They were designed artificially as a universal langue, but they have ended up being interpretive codes that vary according to users of specific languages. For example, the smiley figure was meant to be as culturally neutral as possible, designed as a simple facial circle colored in yellow as an obvious attempt to remove recognizable facial features associated with race or ethnicity. However, almost right after its spread into universal usage, the code behind it became subject to culturally shaped meanings, leading to new designs. The result has been an attenuation of the desired universality of the code and, consequently, of the universality of the episodic langue it subserves.

4 Emojis as parole

Evidence that emojis are now an intrinsic part of parole in digital media, that is, of everyday communicative practices through the Internet, is now rather abundant (for example, Miller et al. 2016; Moschini 2016; Vidal, Ares, and Jaeger 2016; Alshenqeeti 2016). Two relevant aspects involve the use of emojis for phatic and emotive communicative functions. The former was defined by British anthropologist Bronislaw Malinowski (1923) as formulaic utterances that are important less for their meanings than for their social functions. Greetings, for example, do not carry a semantic load in the normal sense; they are used to make contact or show group adherence. Emotivity is the use of words and phrases to convey tone, point of view, emotional state, etc. (Jakobson 1960). In terms of the phatic function, a smiley used at the beginning of a text message constitutes an opening social protocol such as “Hello,” “Dear so-and-so,” and the like. But it simultaneously encompasses an emotive function, since it provides an opening interpretive frame for imbuing the tone of the message with positivity, thus ensuring that a bond between interlocutors is established.

An analysis of text messages shows that three phatic functions are now part of systematic emoji usage (Danesi 2016):

1. Utterance opener. An opening smiley, or some other emoji, conveys a salutation, allowing the sender to strengthen or establish a friendly bond with the interlocutor even when a message may have some negativity in its contents.

2. Utterance ending. The smiley and similar sentiment emoji (such as hearts) constitute a typical good-bye function in a message.

3. Support. Putting emojis in some locations allows the sender to support the content visually and thus effectively.

Emotivity reveals one’s state of mind or emotional state. In face-to-face communication, people use interjections, intonation, and other prosodic devices, alongside specific keywords and phrases, to convey their feelings, explicitly or implicitly. In most informal digital messages, these are typically supplemented or even replaced by emoji forms.

In addition to such communicative functions, research has shown that emojis vary along social and psychological scales and are thus open to interpretive variability and even misinterpretation – both central aspects of parole. This is remarkable, given that the original emoji lexicon was designed to be free of such fluctuation and variability (Barbieri et al. 2016; Chen et al. 2018). For example, Chen et al. (2018) examined emoji usage in a large dataset of smartphone users across the world. They found that emoji selection and frequency was gender-coded, to the point that a machine learning algorithm was easily constructed that accurately inferred the gender of the user based on the emoji features they discovered. Studies are now also showing that misinterpretation emerges in various domains of the emoji lexicon. A well-known example is the nail polish emoji, which has been found to trigger a whole array of unwanted connotations that users in some non-English speaking countries want to avoid, finding the emoji offensive:

Figure 7

Nail polish emoji

The thumbs-up emoji is another problematic one.

Figure 8

The thumbs-up emoji

This gesture is offensive in parts of the Middle East, Africa, Russia, and South America. In some of these areas, it is the equivalent of using the middle finger in the Western world. The list of such culturally offensive emoji is an extensive one, and need not concern us here. The point is that ambiguity, misinterpretations, and cultural-coding have emerged unexpectedly in emoji usage. In effect, one cannot assume an isomorphic relation between an emoji and its interpretation. Glikson, Chesin, and van Kleef (2018) even found that smileys do not increase perceptions of warmth (as they were designed to do) and in the workplace they may actually decrease perceptions of competence.

Humorous and ironic emoji are also problematic, given the cultural variation that characterizes these strategies. In fact, it seems that the humor or irony is processed according to the language and culture of the users. Benjamin Weissman and Dean Tanner (2018) studied brain wave patterns of people reading texts with emojis in them finding that emojis are processed in the same ways that we process ironic language. This is a remarkable finding indeed, since it suggests that emojis are discourse markers that overlap with langue and parole, constituting devices that are powerful emotively.

As the foregoing discussion implies, the Saussurean dichotomy of langue and parole and his belief that the two were separate systems – which Chomsky (1957) later labeled competence and performance – is not sustainable, at least in the domain of emojis. The emoji “test case” thus indirectly confirms the view of linguists, such as M. A. K. Halliday (1975) and Dell Hymes (1971), that the two dimensions are interactive, not autonomous. Hymes argued that speech (parole) is as regulated by meaning-based structures as linguistic competence is by rules of grammar. He put forth the notion of “communicative competence,” defining it as the specific use of a language for purposes of communication and the system of implicit rules of usage and locutionary adaptation that it entails. Moreover, Hymes claimed that parole actually altered langue, which changes gradually as we use it.

Today, it has become apparent that langue subserves broader cognitive processes. George Lakoff (1987) has argued that the foundations of langue are figurative, not syntactic in the Chomskyan sense. He proposed that grammar and communication (linguistic and communicative competence) were intertwined with figurative semantics. To show how even a simple grammatical category had a basis in figurative cognition, Lakoff (1987) gave the example of the Australian language Dyirbal, which, like many other languages, had grammatical gender. Each of its nouns is assigned to one of the available genders. In European languages, the gender is frequently unpredictable from its literal meaning. For example, the word for “table” is masculine in German (der Tisch), feminine in French (la table), and neuter in Greek (to trapézi). Dyirbal has four genders, which are based on the meaning of nouns. One of these includes all nouns pertaining to women, to fire, and to dangerous things (snakes, stinging nettles, and the like). These reflect a perception of the world that is based on metaphorical thinking. In a fundamental way, each emoji is a figure of mind – a type of cognitive metaphor – that should and can be studied under the rubric of “Lakoffian grammar” – an area that certainly needs further investigation.

5 Concluding remarks

The conclusion that can be put forth tentatively here is that emojis constitute a self-contained semiotic system, involving both langue and parole in an integrated fashion, with one dimension affecting the other. This implies that the original Saussurean dichotomy is clearly not sustainable in this domain. Also, if emojis do indeed constitute a langue, it is one that is based on episodic grammar, rather than on any model of linguistic grammar. It is not syntax or morphology that guide the distribution of the emojis in a text, but narrative structure. In this way, they might indeed approach a quasi-universal code of communication, given the high degree of iconicity that visuality entails, which translates homogeneously across languages more so than verbality does. Nevertheless, the research has been showing that even in this domain interpretive variability emerges. Peirce (1931) referred to this type of iconicity as “hypoiconicity.” Unlike Saussure, Peirce viewed semiosis as originating in the perception of some property in an object. Since iconic signs are fashioned in specific contexts, their interpretations are not universal, even though they spring from the same human perceptual (abductive) apparatus. Peirce used the term hypoicon to acknowledge this context-constraining dimension of what he called Firstness. Nevertheless, because it is a sensory-based sign, the emoji referent can often be figured out even by those who are not part of the contextual situation if they are told how it simulates, resembles, or substitutes it.

The emoji movement was sparked by the belief that words set people apart and may bring about misunderstanding and conflict. For this reason, people have often dreamed of creating an artificial, universal langue, which everyone could speak and understand unambiguously. The reason given for such a language is a simple one – if all people spoke the same tongue, cultural and economic ties might be much closer, and good will would increase between countries. René Descartes is believed to have originated the idea of a universal language in the 1600s. More than 200 such languages have been invented since he made his proposal. Today, only Esperanto is used somewhat. Esperanto has a simple, uniform morphological structure – adjectives end in a, adverbs end in e, nouns end in o, an n is added at the end of a noun used as an object; and plurals end in j. The basic core vocabulary of Esperanto consists mainly of root morphemes common to the Indo-European languages. Clearly, this hardly makes it a universal langue.

The appearance and spread of emojis may have taken over from the Esperanto movement. If the movement is sustainable, then the way we read and write messages is not only radically different from the past, but it may also indicate a shift away from a linear mode of processing information, as imprinted in alphabetic-phonetic layouts, toward a more holistic and imaginative mode. However, since emojis depend on specific technologies, it remains to be seen if they transcend these and become a permanent part of the evolution of language and communication.