Introduction

1Zappavigna and Logi (2024, 617) define emojis as “small picture icons which are standardised by The Unicode Consortium and can co-occur with other characters on any platform that recognises Unicode”. Behind its apparent simplicity, this short definition captures the full semiotic complexity of emojis.

2First, as Zappavigna and Logi’s definition indicates, emojis are “small picture icons”. In Peircean terms, this characterizes them as visual iconic signs, that is, signs whose primary feature is their visual resemblance to the object they represent. According to Peirce’s classification of signs, which divides signs into three main categories based on their relationship to their object, an emoji is therefore primarily an icon, defined as “a sign which refers to the Object that it denotes merely by virtue of characters of its own, and which it possesses, just the same, whether any such Object actually exists or not” (Peirce 1932, para. 247). Yet, as Zappavigna and Logi also note, emojis are integrated into Unicode, the universal standard system for encoding text in digital communication, alongside traditional keyboard characters. As a result, they function as characters embedded in the linear flow of digital text. As Halté (2019) remarks, this integration into verbal language allows them to serve discursive functions such as “indicat[ing] affect or, more broadly, the speaker’s attitude,” and “replac[ing] lexical items, in an illustrative or almost narrative way” (Halté 2019, 366, my translation).

3More importantly for the present study, the fact that emojis operate as visual iconic signs does not imply that their relationship with their referent is exclusively iconic. They may also function as symbols, the third category in Peirce’s classification, defined as “a sign which refers to the Object that it denotes by virtue of a law, usually an association of general ideas, which operates to cause the Symbol to be interpreted as referring to that Object (Peirce 1932, para. 249). However, according to Halté (2019), the symbolic dimension is not the most frequent in emoji use: “most of the time, none of these signs is used, by virtue of a rule, to designate anything other than what it resembles” (2019, 374, my translation). By contrast, it is far more common for emojis to function as indices, the second category in Peirce’s trichotomy, where an index is defined as “a sign which refers to the Object that it denotes by virtue of being really affected by that Object” (Peirce 1932, para. 248). Emojis function as indices when they act as pointers that direct the interpreter's attention to the underlying cause of their use in a particular time and place, which, in the case of emojis, is often the speaker’s affective state or attitudes. Halté provides the example of the smiling face emoji:

If we interpret these signs by their degree of visual resemblance to an object, they also function as indices. For instance, a pictogram resembling a smile, such as 😊, is most often understood as an index of the speaker’s joy, expressing in the immediacy of the speech situation an affect, an emotion, or more broadly, an attitude. (Halté, 2019, 377, my translation)

4In this use, emojis function as indices insofar as they reveal the object they stand for, here, an affect or an attitude of the speaker.

5The skull emoji, 💀, which I selected for this study, illustrates this semiotic complexity particularly well. It was part of the first set of emojis approved in Unicode 6.0 in 2010, which is the first version of the Unicode standard to support emojis, thereby making them available to users in written digital communication worldwide. Initially, according to the reference website KnowYourMeme.com, which is devoted to “researching, documenting and explaining memes”, the skull emoji was “commonplace around Halloween”. It was also used in posts “relating to death or scary things”, where it primarily functioned as an icon with an illustrative function. However, Robertson et al. (2021) note that the skull emoji underwent a major shift in 2013, when it began to be used as an expression of amusement, hyperbolically meaning that the speaker is dead from laughing. This dual function, on the one hand, as an illustrative icon in posts about Halloween, death, and scary themes, and on the other, as an expression of the speaker’s amusement, is now fully acknowledged on Emojipedia.org. On this website, which is the reference online encyclopedia for emojis, the skull emoji is currently featured in two distinct categories: “Halloween” (alongside the Jack-O-Lantern, Ghost, or Spider emojis), and “Laughing” (in addition to the face with tears of joy, rolling on the floor laughing and loudly crying face emojis).

6In light of these elements and building on Roberton et al.’s (2021) findings, the aim of this article is to examine the more recent evolution of the skull emoji. It presents the results of a study which explored the following research question: How has the use of the skull emoji evolved in recent years? This overarching question entailed answering several sub-questions: How is the skull emoji currently used in digital discourse? Is its iconic use in posts about Halloween or death still commonplace? In its indexical use, what feelings or attitudes does it signal? Has it developed additional functions or uses online? To address these questions, I have employed a combination of qualitative and quantitative methods to explore patterns of use of the skull emoji in two corpora, drawing both on Halté’s Peircean framework and on a more pragmatic approach focused on what users do with emojis in discourse.

7The next section provides a brief overview of previous research on emoji functions and uses, with particular attention to the skull emoji. The third section introduces the data and methodology of the study and discusses the annotation system adopted. The results are presented and analysed in the fourth section. Finally, the last section offers a conclusion and outlines directions for future research.

Previous research

8CMC (Computer-Mediated Communication) studies, which investigate any form of human communication mediated by electronic devices, are no longer in their infancy (see Herring 2004, for a brief history of the field). Since the early 2010s, particularly following the integration of emojis into the Unicode standard, a growing body of research has examined visual cues in digital written communication, with a strong focus on emojis. As Sasamoto (2023, 395) observes, “emoji have attracted huge scholarly attention in many disciplines”, including linguistics, where a large body of work now exists on topics such as emoji usage (Ljubešic & Fišer 2016; Was & Hamrick 2021, Sampietro 2024), their linguistic, semiotic, or pragmatic functions (Danesi 2017; Herring & Dainas 2017; Riordan 2017; Parkwell 2019; Hasyim 2019; Dainas & Herring 2021) or their interaction with linguistic co-text (Schneebeli 2017; Zappavigna & Logi 2021; Hand et al. 2022; Gibson 2024).

9Of particular relevance to the present study is Halté’s work (2016, 2019), which questions the semiotic and discursive status of emojis and emoticons (emoji precursors formed by character combinations rather than dedicated icons). In both articles, Halté draws on Peircean semiotics to propose a more precise classification of these signs and examines their pragmatic and enunciative functions, with a particular emphasis on the interplay between verbal and non-verbal elements. Another key research for this study is Robertson et al. (2021), who conducted the first longitudinal study of emoji semantic development between 2012 and 2018 by applying computational linguistic tools to Twitter data. Equally important for understanding how emoji meaning operates are the studies by Shardlow et al. (2022), who explore emoji polysemy and the disambiguation of their meaning; and Weissman et al. (2023), who investigate the conventional lexicalized meanings of emojis and how they affect processing.

10As for the skull emoji specifically, it has been attracting mainstream media attention for several years, particularly in relation with supposed generational differences in terms of emoji use (see, for example “Smiley faces are sarcastic and the skull is hilarious: how to speak emoji in 2021”, The Telegraph, 7 September 2021; “What does the skull emoji really mean and how to use it", The Sun, 24 November 2022; “Thumbs up or the Skull? Emojis are widening the generational divide”, The Canberra Times, 16 July 2025). However, to the best of my knowledge, only one scholarly article has focused specifically on the skull emoji: Kostadinovska-Stojchevska & Shalevska (2024), which examines its use to signal humour and awkwardness and as a “makeshift punctuation mark” (2024, 124). The present study therefore addresses a gap in existing research on emoji meaning and use.

Data and methodology

Corpora

11The corpora used for this study are the enTenTen corpora. The TenTen Corpus Family is a set of large web text corpora compiled by the Natural Language Processing Centre at Masaryk University (Czech Republic) in collaboration with Lexical Computing, who makes them available in Sketch Engine, the corpus management and analysis software the company has been developing since 2003. The TenTen corpora exist in more than 40 languages and are prepared according to the same criteria. They can therefore be regarded as comparable corpora.

12For this study, I have compared two editions of the English web corpus (enTenTen): the 2018 edition, enTenTen18, and the 2021 edition, enTenTen21, the most recent version available. Earlier versions of the corpus were not suitable: enTenTen13 and enTenTen15 do not include emojis, while enTenTen12 contains only six instances of the skull emojis across nearly 13 billion tokens (words as well as nonwords such as numbers, punctuation, figures and emojis).

13EnTenTen18 comprises almost 26 billion tokens, among which 264 instances of the skull emoji. Despite its label, the corpus includes material collected over several periods: November-December 2016, January 2017, November-December 2017, and October-November 2018. EnTenTen21 is considerably larger, with about 61,5 billion tokens and 1541 instances of the skull emoji. It was compiled between October and December 2021 and in January 2022.

Data selection

14I began the annotation process with enTenTen18, using Sketch Engine’s concordance tool to retrieve all instances of the skull emoji in the corpus and display them in context. As previously noted, Sketch Engine identified 264 instances of the skull emoji. I then manually reviewed these results to exclude duplicates arising from sequences where the emoji appeared multiple times in a row, such as here:

15In terms of meaning, repeated emojis are relevant as they often signal amplification, however, from a purely quantitative perspective, they still constitute a single instance of use of the same emoji. This must be taken into consideration to avoid inflating the frequency count since Sketch Engine registers each single emoji as a distinct token. After removing duplicates, the dataset comprised 198 distinct instances of the skull emoji. To maintain comparability between corpora, I analysed the same number of instances in enTenTen21 as in enTenTen18. For this purpose, I applied Sketch Engine’s “get a random sample” function to enTenTen21, which, as noted above, contains 1541 instances of the skull emoji. From this set, I randomly selected 198 occurrences of the skull emoji (duplicates excluded).

Data annotation

16I annotated the two datasets according to how the skull emoji was used in each instance, and more specifically based on the relationship between the emoji and what it represents. Halté (2019, 374) summarizes the three types of relationship between a sign and an object, as distinguished by Peircean semiotics, as follows:

• an iconic relationship, which corresponds to a formal resemblance between the sign and the object.

• an indexical relationship, rooted in the context of utterance, in which the sign reveals the object, the latter not being perceptible without the sign.

• a symbolic relationship, based on a rule that arbitrarily links the sign to an object.

17I chose not to retain the symbolic dimension as a distinct annotation category because this aspect is present in the vast majority of skull emoji uses and is therefore not sufficiently discriminative. Indeed, it is precisely because of the conventional symbolic association between the skull and death, which predates the emoji, that the skull emoji is used to illustrate posts about murder, Halloween, or to indicate that the speaker is “dead” with laughter or embarrassment. I therefore chose to classify all instances into two broad categories only, which I called “iconic” and “indexical”, depending on the emoji’s primary function, either as an icon or as an index, and which I subsequently subdivided into more specific subcategories.

18The first category, the iconic category, comprises instances in which the skull emoji functions as a direct graphic representation of a skull, without conveying instructions on how to interpret the speaker’s intention or attitude. Its role is therefore primarily illustrative or ornamental. As noted above, the instances classified in this category often carry a symbolic dimension. However, the symbolic dimension of the skull is pre-existing and independent from its use as an emoji, and it is mobilized in these posts precisely because the emoji serves as an iconic representation of a skull.

19I further divided this category into four subcategories, the first two of which correspond to what KnowYourMeme.com describes as the initial use of the skull emoji:

201. Halloween / Día de los Muertos: instances in which the skull emoji is used to illustrate posts referring to these holidays, an American holiday of Celtic origin and a Latin American holiday that both involve remembering and honoring the dead.

21In both instances featured in Figure 2, the skull emoji is used because of its iconic nature: alongside jack-o’-lanterns or ghosts, the skull belongs to the paraphernalia of Halloween and, with roses, candles, or marigolds, to that of Día de los Muertos, both of which commemorate death and the dead. The symbolic nature of the skull is also significant but, as written above, the skull emoji merely conveys this symbolic dimension because it depicts a skull, which already carries a symbolic association with death outside of digital discourse. The same observation applies to the next two categories.

222. Death / macabre / scary themes: instances in which the skull emoji, again in virtue of the conventional association of the skull with death, is employed in relation to death, murder, killing, illness, disease, or other macabre, frightening or spooky themes.

23In Figure 3, users chose it to illustrate a death count, the act of killing, or a location with a gothic vibe, that is, one that evokes gothic culture, characterized by its fascination with death and all things dark and gloomy as well as its distinctive aesthetic, including skulls, bats, and a tropism towards dark colours. It should be noted that Halloween-related posts often incorporate this spooky dimension, since one of the core components of the holiday is having fun while being scared (contrary to Día de los Muertos, which is more of a celebration of death as a natural stage of life). Nevertheless, I chose to maintain posts explicitly mentioning Halloween in a dedicated sub-category to better track the evolution of this specific use.

243. Username: Instances in which the skull emoji is incorporated directly into usernames, generally as an emblem of gothic culture, fantasy, or other dark sub-cultures, again derived from the symbolic value of the skull discussed above.

25In the two examples shown in Figure 4, the skull emoji is again essentially used for what it represents, a skull, serving as a traditional visual symbol of death and the macabre. It is also employed by users as a tool for constructing online identity. The first pseudonym belongs to a member of flightrising.com, a site where users breed, collect, and care for dragons. The second belongs to a Steam user discussing video games and is likely a nod to Daenerys Targaryen, a character called “the mother of dragons” in the Game of Thrones saga. It could therefore be argued that inserting the skull emoji into a username suggests the first hints of indexical functioning, as the skull may also be meant to indicate a personal affiliation with a specific 'dark fantasy' subculture. However, since it does not go as far as to reveal or signal an emotional or attitudinal stance, I have chosen to categorize these instances as primarily iconic.

264. Self-referential: instances where the skull emoji refers to itself or illustrates the word “skull”.

27This last sub-category of the iconic category is arguably the most purely iconic use of the skull emoji, as it literally stands for a skull, without any symbolic or indirect connotations. It merely illustrates the word “skull”, offering a simple visualization for the word already designated.

28The second category, the indexical category, comprises cases where the skull emoji functions as a sign through contextual association, pointing to the underlying cause of its use in this particular context, which may be an affect, an emotion, or more broadly an attitude or stance adopted by the speaker toward something or someone. As such, instances in this category all provide cues for interpreting the speaker’s intention or attitude and, by extension, the meaning of the post.

29Again, these instances often draw upon the skull’s pre-existing symbolic association with death: it is precisely because of this association that the skull emoji has come to indicate that the speaker is “dead” from embarrassment, shame, or laughter. However, this symbolic dimension remains prior to and independent of the skull emoji’s use as an index of the speaker’s affect or attitude, which simply repurposes it to that end.

30I further divided the indexical category into two different subcategories depending on what the emoji signaled, drawing on DuBois’s stance-taking theory (DuBois 2007). DuBois’s theory centers on the idea that taking a stance, that is, expressing one’s attitude, emotions, or position regarding an object, is one of the most important things people do in interaction. He identifies three main stances, two of which are particularly relevant to the indexical use of the skull emoji and form the basis for its two subcategories.

31The first one is the affective stance, in which “speakers position themselves affectively, they choose a position along an affective scale” (DuBois 2007, 143). Prior research has shown that indicating an affective stance is a common function of emojis, (e.g., Walther 2006; Kralj Novak et al. 2015; Herring & Dainas 2017; Schneebeli 2018, Shardlow et al. 2022). Specifically, as Robertson et al. (2021) observe, the skull emoji has been widely used to signal amusement since at least 2013. According to Emojipedia.org, it now frequently expresses figurative death, for instance, dying from extreme laughter, frustration, or affection. The emotion signaled by the skull emoji may therefore be positive (e.g., amusement, mirth, awe, love, eagerness, admiration) or negative (e.g., shock, embarrassment, anger, mortification, disappointment, shame, disgust, anxiety).

32In Figure 6, the skull emoji first indicates amusement, alongside the expression “had me rolling” (as in “rolling on the floor laughing”, which is also part of the Internet vernacular to express amusement). In the second case, it signals the speaker’s joy in a post where an Instagram user recounts how they lost weight after going to the gym.

33On the contrary, in Figure 7, the skull emoji in the first post is meant to indicate the speaker’s annoyance when waking up after a bad dream. In the second post, it indicates how disgusted the speaker is at having to eat a slice of pizza he has been given, which has anchovies on it.

34The second subcategory within the indexical category is dedicated to emojis signaling an evaluative stance, in which “a stancetaker orients to an object of stance and characterizes it as having some specific quality or value” (DuBois 2007, 143). Such evaluations can again be positive (e.g., finding something impressive, great, awesome, beautiful) or negative (e.g., finding something horrible, cringe, not cool, stupid, ridiculous, appalling, or more broadly being critical of something or someone, to mock, be sarcastic, or even insulting).

35In Figure 8, the skull emoji first appears alongside the hundred points emoji and the fire emoji, which both mean something is excellent. In turn, the skull emoji signals a very positive evaluation of Arsenal’s celebration after an impressive goal. In the second post, the skull emoji signals how great it is to be given the chance to sit next to British singer Adele for two hours.

36On the contrary, in Figure 9, the skull emoji indicates general disaffiliation with something or someone.

37In the first case, the user is distancing himself from a text he wrote when he was in ninth grade, and in the second case the skull emoji is used in the context of a discussion criticizing Kanye West’s behaviour.

38Instances classified in the indexical category were therefore further annotated into the two following subcategories:

1. Affective stance: instances in which the skull emoji allows users to signal an affective stance, positive or negative.

2. Evaluative stance: instances in which the skull emoji allows users to signal an evaluative stance.

39A last category was created to group cases in which the meaning or function of the skull emoji could not be determined with sufficient certainty. This third category is labeled “unclear use”.

Observations on data annotation

40Three important observations must be made before turning to the results.

41As is implied above, iconic, indexical, and symbolic relationships are not mutually exclusive. First, as Halté (2016, 219) points out, emojis never fully lose their iconic nature: regardless of their function, they continue to depict something visually. However, those classified as indexical emojis also contribute to signal the speaker’s affective or attitudinal stance. As noted above, emojis can also function as symbols. In this respect, the skull emoji functions in a way that echoes what Halté (2016, 239) observes about another non-face emoticon or emoji, the heart:

Some cases are more complex than simple facial emoticons, involving all three aspects of the relation between sign and object as conceived by Peirce—for example, the emoticon ‘<3.’ This is an icon of a heart, with the heart itself functioning as a symbol (that is, a sign arbitrarily linked by human convention to its object) of love. Through a complex semiotic process, the emoticon ‘<3’ thus becomes an index of the speaker’s feeling of love: it is precisely because it is an icon, an imitation of a symbol, that it can serve as an index of what the symbol signifies. (my translation)

42Similarly, the skull emoji remains an icon insofar as it visually represents a skull, yet, as previously noted, it frequently operates as a symbol of death, to which it is conventionally associated. Through this symbolic function, it may further acquire an indexical dimension, serving as an indicator of affective state—whether positive or negative—as in the expressions “dead from laughter” or “dead from embarrassment”.

43Secondly, distinguishing between the affective and the evaluative stance is not always straightforward, as the two frequently co-occur. Consider, for example, the instance in Figure 9, “It’s kinda cringe 💀”. The adjective “cringe” simultaneously indicates a negative emotional reaction and a negative evaluation of the referent of “it” (a text that the speaker wrote when he was in ninth grade). Accordingly, the skull emoji appearing in this utterance could be interpreted as signaling both evaluative and affective stance. However, in line with DuBois (2007), who defines evaluative stance as object-oriented and affective stance as speaker-oriented, the classification was based on the immediate context of the emoji and the element to which it referred. In Figure 9, the emoji was interpreted as object-oriented, since it appears at the end of an utterance expressing disaffiliation. It was therefore categorized as signalling first and foremost an evaluative stance. Nevertheless, many instances display characteristics of both categories, to varying degrees, so much so that the distinction between them cannot be regarded as absolute.

44A third and final observation concerns the use of emojis to signal humour or laughter. It is relatively common for emojis, particularly those depicting laughter, to function as illocutionary force-indicating devices, a role first identified by Dresner and Herring (2010). Such emojis may, for instance, serve as mitigation devices, or as markers indicating that an utterance is intended humorously (Schneebeli 2017, Herring & Ge-Stadnyk 2024). The two examples in Figure 10 exemplify this function:

45In the first example, the context is a friendly exchange among viewers of The Challenge, an MTV reality competition show, who are commenting on a particular scene. In this post, the skull emoji allows the user to signal laughter and thus mitigate the potentially negative comment “I’m going to have nightmares tonight”, which is not intended literally but instead follows some playful banter about the scene. The context for the second example differs somewhat but the role of the skull emoji is comparable. Here, the emoji appears in an introductory disclaimer on a Wattpad account dedicated to fictional online conversations between celebrities. In this setting, the skull emoji functions as a marker of laughter, reinforcing the notion that the author’s content is not to be interpreted seriously. It serves to defuse potential tension or hostility by signaling the author’s benign intentions. Therefore, the skull emoji operates not only as an indicator of humor but also as an index of the speaker’s general attitude and emotional stance. This dual function complicates the task of distinguishing between laughter that reflects genuine amusement and laughter that frames an utterance as intended to be amusing. Provided the emoji conveyed laughter, and even when it served as a marker of illocutionary force, these instances were classified as manifestation of a positive affective stance. For this reason, a separate subcategory of indexical use was deemed unnecessary.

Results

Context : evolution of emoji use online

46To more effectively analyse the results of the study, it is necessary to situate them within the broader context of the evolution of emoji use online during the periods of enTenTen18 and enTenTen21. Although a span of four years may appear relatively short, developments in digital communication are characterized by rapid change (Sindoni & Moschini 2021; Jones 2023). Emoji usage in particular has undergone a significant and rapid shift, as documented in Emojipedia’s report on Emoji use on Twitter:

47At the beginning of 2018, a little over 14% of tweets contained at least one emoji, whereas by early 2021, this figure had risen to just over 20%, a substantial increase. Although it is not possible to calculate the absolute number of emoji occurrences in enTenTen18 and enTenTen21, the skull emoji itself appears more frequently in the latter. Specifically, its proportion in enTenTen21 is 144% higher than in enTenTen18. This increase is not unique to the skull emoji but reflects a broader trend affecting many other emojis: compared to enTenTen18, the frequency of 😂 (face with tears of joy) in enTenTen21 rose by 101%, the frequency of 😭(loudly crying face) by 66%, and ❤️(red heart) by 43%.

Analysis

48The following table presents the distribution of the skull emoji in enTenTen18 and enTenTen21, showing the number of instances across the predefined categories and subcategories of use. Percentages for iconic, indexical and unclear use are calculated relative to the total number of instances of the skull emoji, while percentages for affective and evaluative stances are calculated relative to the total instances within the indexical category.

Table 1. Number and proportion of instances in each of the predefined categories and subcategories

Type of use of💀	enTenTen18	enTenTen21
Iconic	110 (55.6%)	76 (38.4%)
Halloween / Día de los Muertos	49	17
Death / macabre / scary things	44	48
Username	4	6
Self-referential	13	5
Indexical use	83 (41.9%)	112 (56.6%)
Affective stance	53 (63.9%)	53 (47.3%)
Positive emotion	40	41
Negative emotion	13	12
Evaluative stance	30 (36.1%)	59 (52.7%)
Positive evaluation	12	8
Negative evaluation	18	51
Unclear use	5 (2.5%)	10 (5%)

49These figures yield several noteworthy observations that provide insights into how the skull emoji has evolved recently.

50First, in enTenTen21, the skull emoji is still frequently used with a purely iconic function, serving primarily as an illustrative or ornamental element. This is particularly striking because such iconic uses are often overlooked in both media articles and scholarly studies of emojis, which typically focus on the role of the skull emoji in indicating amusement or embarrassment. The data presented here, however, demonstrate that the function of the skull emoji as an expression of figurative death, meaning dead from laughter or embarrassment, should neither be treated as its sole role nor be overstated, as is often the case in the existing literature.

51Secondly, a comparison of enTenTen18 and enTenTen21 reveals an almost complete reversal in the relative proportions of iconic and indexical uses. In enTenTen18, the iconic function is predominant, accounting for 55.6% of the corpus. By contrast, in enTenTen21, the indexical function becomes dominant, with 56.6% of uses. This shift indicates that the skull emoji is increasingly used to indicate a stance, that is, to convey information about the speaker’s emotional state and attitude toward the interlocutor or toward the propositional content of the message.

52Changes within the indexical category are equally revealing. In enTenTen18, indexical uses are predominantly affective, with approximately 64% indicating emotions, most frequently positive emotions such as amusement or mirth. In enTenTen21, however while affective stance remains significant, with a stable proportion of positive and negative emotion (40 to 13 in 2018, compared with 41 to 12 in 2021), the majority of indexical uses now fall into the evaluative category (52.7% of indexical emojis, compared to 36.1% in enTenTen18). In this evaluative function, the skull emoji is now typically employed to convey a critical tone and is frequently associated with messages that include reproaches, criticism, attacks, insults, or other forms of negative evaluation. It is difficult to identify a clear and tangible explanation for this shift towards signaling a negative opinion. It could simply reflect a gradual return to the skull’s more negative connotation, as a symbol of death, danger, or mortal risk (as on the label of dangerous household cleaning products, pesticides, and laboratory chemicals) as if speakers framed the judgement they voice as potentially (and/or meant to be) harmful or damaging.

53This double shift—first, toward predominantly indexical use, and, secondly, toward a predominantly evaluative stance—may suggest an ongoing process of pragmaticalisation of the skull emoji. This means that the skull emoji may be moving toward a predominantly pragmatic function of providing cues for the interpretation of utterances, for example by signaling the attitude of the speaker. Aijmer (1997), in her study of “I think”, describes pragmaticalisation as the tendency of certain lexical items to evolve into “elements which cannot be analysed in terms of truth” (1997, 3), but which instead convey the speaker’s attitude toward the hearer or the message. Similarly, in departing from its initial iconic function, the skull emoji has taken on a pragmatic role as both an indicator of affective and evaluative stance, toward which it is shifting in a gradual and predominant manner. While the pragmaticalisation of emojis has received limited scholarly attention, recent research has begun to explore the crucial role emojis play in shaping the interpretation of a message or social media post. In particular, Yus (2025) provides a comprehensive account of this emerging line of inquiry from the perspective of pragmatics.

Discussion and conclusion

54The main research question of the study presented in this article was: How has the use of the skull emoji evolved in recent years? A relatively clear answer emerges from the collected figures and their analysis: the skull emoji has shifted towards a more indexical and less iconic usage, although the latter is far from having disappeared. In its indexical use, it also appears to have taken a more negative valence and be increasingly used to signal a negative, critical, sarcastic, or even sardonic judgment.

55Naturally, there are some limitations to this study which need to be considered. The primary limitation concerns sample size. Although the source corpora, enTenTen18 and enTenTen21 contain billions of tokens, they include only a limited number of instances of the skull emoji. However, at present, no large-scale corpus of emojis is publicly available, making it difficult to work on larger datasets. A second limitation is that I was the sole annotator. Given the potentially ambivalent nature of the skull emoji, it is possible that other annotators might have interpreted certain occurrences differently. To address this, some studies (Was & Hamrick 2021; Shardlow et al. 2022; Weissman 2023) rely on multiple annotators. In this study, however, strict analytical criteria were applied, including consideration of both the narrow co-text and the broader context of each occurrence, for example to determine whether each instance was more object-oriented or speaker-oriented and vice-versa.

56This study opens several interesting perspectives for future inquiry. One question is whether the skull-and-crossbones emoji is used in the same way as the skull emoji. In particular, it would be worth examining the claim made by the instructional website WikiHow that the skull-and-crossbone emoji is “typically used literally”, “when discussing death, danger, or bones in a literal sense”, whereas the skull emoji “can be used literally and metaphorically”, and both can be used “for the spooky things”, depending on personal preference. Another interesting line of inquiry would be to explore how the skull emoji compares to traditional face emojis expressing laughter, such as the face with tears of joy emoji, and to assess whether the laughter conveyed by the skull emoji carries a more negative undertone or implication.