Receiver Interpretations of Emoji Functions: A Gender Perspective Susan C. Herring and Ashley R. Dainas Department of Information and Library Science Indiana University, Bloomington {herring,ardainas}@iu.edu This paucity of research is surprising, given that gender differences have been reported in attitudes toward, and usage Abstract of, emoji and their antecedents, emoticons. These graphical icons are perceived as cute, feminine, and, in some cultures Previous studies have reported gender differences [e.g., Sug15], inappropriate for males to use. Consistent with in emoji use and attitudes toward emoji. Here we these attitudes, females produce emoji and emoticons more ask whether, and if so, to what extent, females and frequently than males do [CLSAWLM17; Wol00]. males also interpret emoji use differently. We Moreover, the two genders preferentially use different icons conducted an online survey to assess how [CLSAWLM17; Wol00] and use them for different different genders interpret the pragmatic functions pragmatic purposes [Sug15]. Given these findings, it is of emoji in their local discourse contexts, based natural to ask whether, and if so to what extent, females and on [HD17’s] taxonomy of functions. Responses males also understand emoji use differently. (N=523; 352 females, 121 males, 50 ‘other’) 2 Related Works showed few overall differences in how females and males interpreted emoji functions, but the The majority of previous research on emoji has focused on ‘other’ gender differed from the females and their semantic functions. In one of the key studies in this males. Based on responses to demographic and area, [MTCJTH16] examined how people interpret the social media use questions, these differences sentiment and semantics of 22 of the most used emoji as appear related to platform norms (e.g., Facebook rendered by multiple platforms. Mechanical Turk vs. Tumblr). We conclude by discussing the participants were shown an emoji rendering in isolation and implications of these findings for automating were asked to describe the emoji’s meaning in their own words and to rate its sentiment. The authors found within and emoji interpretation. across platform disagreement on both the sentiment and semantic meaning of the emoji. [MKTTH17] followed up on 1 Introduction these findings to see if putting emoji in context would A picture may be worth a thousand words, but graphical decrease the rate of disagreement. The authors took the 10 systems are not a universal language. Research has shown emoji that [MTCJTH16] found to be most prone to that users often disagree about the interpretations of emoji, misinterpretation 1 and collected 200 naturally occurring regardless of whether the emoji are presented to subjects in tweets containing one of these emoji. Contrary to their isolation or with some context [MKTTH17; MTCJTH16]. expectations, [MKTTH17] found that emoji read in the The resultant differences in interpretation have been context of the tweet did not become less ambiguous, and that attributed to several factors, including emoji renderings text can even potentially increase the ambiguity of emoji that differ across platforms, inherently ambiguous forms sentiment. Some emoji are more difficult to interpret than (such as the grimace face emoji), and the receiver’s others, such as the grimacing face [MTCJTH16]. Similarly, familiarity with the sender and the culture of the social [JA17] found that their Mainland Chinese participants media platform where emoji are used [MTCJTH16; agreed on a single mood/emotion for emoji (shown in MKTTH17; Tig16]. One factor that has received little isolation) such as the face throwing [sic] a kiss, the loudly consideration thus far, however, is the receiver’s gender. crying face, the winking face, the stuck-out tongue, and the An exception is [JXLHBA17], who found no overall heart eyes, while at the other extreme, emoji such as the tears differences in the ability of different gender and age groups of joy face, the blushing face, and the grimacing face were to describe and discriminate the dominant emotion associated with multiple unrelated moods/emotions. conveyed by different emoji in a web survey. However, Emoji do not only function on the semantic level, their study examined these interpretations in isolation and however, and the reasons for their use do not derive solely considered only emotion, whereas emoji serve many other from their semantics. Context shapes emoji meaning; thus it communicative functions [HD17], and their interpretation is also important to consider their pragmatic functions in is highly dependent on context [CJT16]. social media discourse. This level has received relatively little attention in the literature, but there are some Copyright c 2018 held by the author(s). Copying permitted for private and academic purposes. 1 Smiling face with open mouth and tightly-closed eyes, grinning face with In: S. Wijeratne, E. Kiciman, H. Saggion, A. Sheth (eds.): Proceedings of smiling eyes, person raising both hands in celebration, smiling face with the 1st International Workshop on Emoji Understanding and Applications open mouth and cold sweat, relieved face, see-no-evil monkey, person in Social Media (Emoji2018), Stanford, CA, USA, 25-JUN-2018, with folded hands, smirking face, face screaming in fear, and face with published at http://ceur-ws.org tears of joy. exceptions. Qualitative studies of pragmatic emoji Given such differences, we ask whether, and if so to what functions report that emoji serve as a social tool that can be extent, females and males also understand emoji use used to add personal identity expression or playfulness to a differently. The only study we are aware of that addresses message [Gul16; KW15; CJT16; Sug15], to manage the this is [JXLHBA17], who designed an online survey study conversation [KW15; CJT16], and to maintain to assess the dominant prospective consumer interpretations relationships [KW15; CJT16]. More concretely, emoji, like of the emotion expressed by facial emoji presented in emoticons before them, have been noted to modify the tone isolation on a website. The authors found no overall of the text they accompany [WD01; CJT16; Gul16; HD17; differences in the ability of different gender and age groups NPM17]. Further, [HD17] found that emoji can also in Mainland China to discriminate the dominant emotion function as performative virtual actions, emotional conveyed by different emoji. However, as noted above, reactions, mentions (as compared with uses), riffs, and emoji do not function solely to indicate emotion. The present narrative sequences [see also CJT16; NPM17]. study thus investigates gender differences in receiver In [NPM17], the authors attempted to train a supervised interpretation of emoji function in their discourse context. classifier to identify possible functions of emoji in tweets, including the function “Multimodal,” which aligns with 3 Method pragmatic emoji functions such as tone modification and gesture. However, their classifier struggled with this We created an online survey to assess how internet users of particular classification because there was such low different genders interpret the pragmatic functions of emoji agreement among coders and a small amount of training in their local discourse contexts. We collected the emoji data. While there has been some research on interpretation together with the message in which each occurred 3 and the of emoji and emoticons by humans, that work mainly previous message(s) to which it responded from 14 focuses on how these graphicons [HD17] change or do not graphicon and media focused public Facebook groups. 4 change the meaning of the text with which they appear, These groups were sampled because of their relatively high e.g., [TF96; WD01]. density of graphicon content as compared with other public Researchers have found that factors influencing Facebook groups. Initial items for the survey were selected differences in emoji interpretation include variation in based on cases that the authors found challenging to code in emoji rendering across platforms [MTCJTH16; previous research involving these data [e.g. HD17]. Further MKTTH17], intrinsically ambiguous forms [JA17; items were chosen to expand the selection of emoji types and MTCJTH16], variation in cultural emoji usage norms pragmatic functions represented. Two to five examples for across communities, and the receiver’s familiarity with the each emoji type were included. The included emoji represent sender [BKRS16; Tig16]. However, the gender of the 13 of the most common emoji types, which were rendered in receiver has received little attention in the emoji the survey to match the emoji that appeared in the original interpretation literature so far. This is despite reported Facebook messages. These emoji did not render consistently gender differences in usage of emoji and emoticons, as well across examples. Thus in the survey, to preserve the original as in attitudes toward their use. These graphicons are context, we used a combination of Apple iOS 10 renderings reportedly perceived as cute and feminine [Ma16; Su15]. and screenshots of the emoji as they appeared on Facebook. 5 [Su15] reports that among Japanese teens, emoji are See Table 1. considered key to girls’ online performance of kawai We anonymized and simplified these messages for the (‘cute’) identities. A number of studies have found that survey. The survey itself consisted of 12 items drawn from females produce emoji and emoticons more frequently than a pool of 46 messages. Four versions of the survey were males do [CLSAWLM17; Wol00]. Furthermore, the two created, for a total of 48 examples (two examples were genders preferentially use different icons [CLSAWLM17; repeated). We designed each block of the survey such that Wol00] and use them for different pragmatic purposes each included at least two examples of all 13 emoji types and [Sug15]. For example, in a study of English language had a similar progression from easier to code items to more newsgroups [Wol00], females used more varied emoticons difficult ones, as determined by the researchers. Assignment and used them (especially smiles) to express solidarity, of respondents to the blocks was random. support, positive feelings, and thanks, whereas males used Respondents selected the best interpretation of the use of emoticons more to express sarcasm and teasing. These each emoji from a list of functions adapted from those findings are consistent with societal stereotypes and identified in [HD17], i.e., tone modification, virtual action, expectations that women express more emotion, especially reaction, and illustration/mention. To these we added the positive emotion, than men [SGDH06]. However, in options softening, decoration, physical action, multiple [CLSAWLM17]’s international corpus, 2 although females functions, other, and “I don’t know”. These 10 options were preferentially used all face-related emoji (indicating a reworded and defined near the beginning of the survey to social orientation), males preferred heart-related emoji make them more accessible to laypersons. Figure 1 (indicating positive emotion). illustrates one of the survey items. 2 4 Principally users from Brazil, Indonesia, Mexico, and the U.S. The Facebook groups that provided examples were: EmojiXpress, 3 CatGIFs, AnimeGIFs, Nihilist Memes, Grumpy Cat Memes, Smiley, We selected messages that contained a single emoji in most cases. In a few messages, the same emoji was repeated two or three times, Stickers, StickersFB, Rise of the Guardians, The Chronicles of and two messages included two different emoji. In the latter case, the Narnia, Star Wars, Percy Jackson, Jared Padalecki, Selena Gomez. 5 survey instructions directed the respondents to focus only on one of Some emoji appeared different (or not at all) to us as Mac and PC the emoji. Emoji reduplication is not considered further in this study. users; we used the images as they appeared on the first author’s Mac for the screenshots, as they were more complete. Table 1: Emoji Types and Emoji Included in the Survey respondents completed the survey, in order to maximize the amount of data available for analysis, we included all surveys in which respondents reported their gender and chose a function code for at least one emoji example. 6 In total, 523 surveys met this criterion (352 female; Mean age: 28.9, Range: 18-70+; 121 male; Mean age: 31.8, Range: 18- 68; 50 ‘other,’ Mean age: 25.2, Range: 18-70+). In the sample of 523, 74.2% of the respondents were The emoji shows that Song Song B is… native English speakers (F: 74.4%, M: 74.4%, O: 72%), followed by German (5.5%), and 75% reported being based o (Virtually) saying “ … may I ask for help here or do you in the U.S. (F: 72.73%, M: 85.12%, O: 66.00%), followed where I should ask pls?”, and then smiling by Canada (4.4%), Germany (4.2%), and the U.K. (2.7%). o (Virtually) smiling in response to the prompt, not We analyzed the function codes the respondents selected necessarily related to the text of their comment by gender overall, as well as broken down by emoji type. o Illustrating the text of their comment Normalized results are presented in charts. Where relevant, o Associating a positive (or some related) tone with their the results of Chi squared tests are presented as p values. comment 4 Findings o Softening their comment o Literally (physically) smiling while typing their Consistent with previous studies, females and males comment reported different amounts of emoji use: 92% of the female o Just using the emoji as decoration respondents reported using emoji, compared to 78% of the males and 79% of the ‘others’. Females more often said they o More than one function is equally plausible used emoji on Facebook ‘often’ (30%) and ‘in every (Specify/Explain your choices) message,’ (2%) whereas males reported using them o Other (Explain) ‘sometimes,’ (38%) ‘rarely,’ (17%) or ‘never’ (13%) more o I have no idea than females (32%, 16%, 6%). More females also reported that they were ‘very confident’ that they understood the Label Emoji Label Emoji intended meaning of emoji when they saw them in social Heart media (58%) as compared with males (47%). Males were “meh” more likely to report being ‘somewhat confident’ (43%) – Eyes Big and, in several cases, ‘not at all confident’ (10%) – than Kiss females (38%, 4%). Respondents who chose “other” for Smile their gender patterned similarly to females, mostly being Blush Smile very (56%) or somewhat (42%) confident in their understanding of the meaning of emoji. Despite these expected differences in usage and attitude, Tears of Crying we found few overall differences in how females and males Joy interpreted emoji function. Both genders chose tone as their Tongue default interpretation slightly more than half the time (F: Frown Out 51.1%, M: 51.8%). Female respondents tended to say that Grimace the emoji were functioning as reactions (p=0.0731) or Wink expressing multiple functions (p=0.0452) more than males. Heart In contrast, males chose the “I don’t know” option more often than females (p=0.006). Figure 2 shows the breakdown Figure 1: Example from the Survey of the functions (excluding tone to display the results for the other functions more clearly) by gender overall. Each participant was also asked to provide information The ‘other’ gender category – comprising 50 people and about their gender (female, male, or other), age, native 584 function codes – differed from the self-identified males language, country of residence, and social media usage. and females in its preference for two code options, multiple Between January 11 and February 20, 2018, we shared functions (p=.0007) and other (p=.087). In addition, the the survey with students and colleagues at our university as ‘other’ genders were somewhat less likely to choose tone well as with friends, family, and strangers via social media (46%; p=.0829 compared to M), and instead selected a wider (Facebook, Tumblr, Reddit, and Ravelry). A total of 658 variety of codes – especially softening, action, and mention surveys were collected, and 628 people reported their – although these preferences were not statistically gender (413 female, 152 male, and 63 other). As not all significant. 6 The average dropout rate after answering at least one question was 11% (F: 11%, M: 13%, O: 4%), ranging from 8% to 15% in the four blocks. 16% 12% 8% p=.073 p=.0007 p=.045 p=.087 p=.0006 4% 0% Action Reaction Mention Softening Decorative Physical Multiple Other I don't Functions Know Female Male Other Figure 2. Functions Selected in the Survey (except for Tone) by Gender (+p ≤.1, *p ≤.05, **p ≤.01, ***p ≤.001) In Figures 3-12, the function results are broken down by (p=.0430) and mentions (p=.0499) more than males did. The type of emoji. Each function has a distinctive emoji profile: ‘others’ were also more likely to interpret the blush emoji as Hearts and kisses were especially interpreted as expressing having multiple functions (p=.0127), and frowns (p=.0809) virtual actions (Figure 4); smiles and winks as softening the and kisses (p=.0452, compared to M) as having other force of a message (Figure 5); grimaces and tears of joy as functions. These findings are consistent with the tendency reactions to a prompt (Figure 6); and kisses and hearts noted above for the ‘other’ gender respondents to offer more (especially) as mentions that illustrate message content varied interpretations of the functions of the emoji in the (Figure 7). Even tone marking, which is the function the survey than the female and male respondents, and to reject respondents selected most often for emoji use overall, is the simple function options provided in the survey in favor associated more with certain emoji (tongue out, crying, of multiple and alternative interpretations. frown) and less with others (e.g., grimace, kiss) (Figure 3). Next we examined the degree to which each gender As for the additional “function” options that we included agreed among itself on its most-preferred (most frequently- to supplement [HD17’s] taxonomy, big smiles and hearts chosen) functions. Table 2 shows the percent agreement of were interpreted as decorative by some respondents each gender on their first choice of pragmatic function by (Figure 8), and some respondents (especially males) emoji type. Table 3 shows the number of questions for which interpreted the heart eyes emoji as describing a physical respondents of each gender chose tone, action, mention, action (described in the survey as “looking adoringly” at softening, or multiple function as their most frequent one’s computer screen) (Figure 9). Finally, the results for choice. 7 The highest rates of agreement for all genders were multiple functions, other, and I don’t know point to emoji for the tongue out, frown, and “meh” emojis and for tone, examples for which the respondents were either not which was the most preferred function for those emoji. satisfied with the specific options provided in the survey or However, ‘other’ gender respondents tended to agree among which were especially difficult to interpret functionally. themselves more (have higher percentages) than females and For example, the tears of joy emoji was said by several males for both emoji types and pragmatic functions, except respondents to have other functions (e.g., laughing in a for blush and tears of joy (Table 2) and action (Table 3). The mocking way) (Figure 11), and the grimace emoji, which ‘others’ even agreed more on tone, although they chose it is known to be ambiguous [MTCJTH16], received the most less often. They also had higher agreement than females and “I don’t know” responses (Figure 12). males on reaction, mention, and softening, and they were the Figures 3-12 also show more gender variation than the only gender to prefer multiple functions for one example. overall results in Figure 2. Self-identified females and We also calculated the degree to which the survey males differed significantly in only two interpretations, i.e., respondents agreed with our own code assignments for the females interpreted both frowns (p=.0105) and smiles items in the survey. It should be borne in mind that the items (p=.0394) as having multiple functions more often than were chosen in the first place because we found them males did. We also observed a slight gendered tendency in challenging to code, so our code assignments might not be the emoji they interpreted as softening (smiles were accurate. Respondents agreed with us in their most preferred slightly preferred in this function by females and winks by function codes for 60.4% of the survey items (F: 66%, M: males), although these differences were not significant. In 50.9%, O: 64.2%). Because they chose tone as the default contrast, the ‘other’ gender deviates from the females and for most items, we also took into consideration their second males in more interpretations. For example, the ‘others’ most-frequent choice. This increased their overall agreement were significantly less likely to interpret smiles as tone rate with our codes to 83%. These results are broken down modification than males (p=.0344) or females (p=.0493), by gender in Table 4. Females and ‘others’ had higher rates and they were also less likely to interpret the tears of joy of agreement with our interpretations than males did. emoji as tone (p=.072, compared to M). Instead, the ‘other’ Further evidence of the tendency of the ‘other’ gender to gender interpreted the tears of joy emoji as actions assign multifaceted interpretations to emoji comes from the 7 The other pragmatic functions were never chosen as the most preferred function for any item and thus are excluded from Table 3. 100% 15% 80% 60% + 10% 40% * 5% 20% 0% 0% Female Male Other Female Male Other Figure 3: Tone Figure 8: Decorative 40% 8% 30% 6% 20% 10% * 4% 2% 0% 0% Female Male Other Female Male Other Figure 4: Virtual Action Figure 9: Physical Action 40% 20% 15% * 30% 20% 10% ** * 10% 5% 0% 0% Female Male Other Female Male Other Figure 5: Softening Figure 10: Multiple Functions 20% 10% * 15% 8% 10% 6% + 4% 5% 2% 0% 0% Female Male Other Female Male Other Figure 6: Reaction Figure 11: Other 25% 25% 20% 15% * 20% 15% 10% 10% 5% 5% 0% 0% Female Male Other Female Male Other Figure 7: Mention Figure 12: “I don’t know” Table 2: Within-gender percent agreement on most-frequent choice of pragmatic function by emoji type (with number of survey items for each emoji type). Big Heart Tears Tongue "meh" Blush Crying Frown Grim. Heart Kiss Smile Wink Avg. Smile Eyes of Joy Out F 65.3% 39.5% 50.9% 77.2% 72.3% 36.6% 43.1% 52.8% 45.3% 55.9% 59.8% 83.3% 57.0% 56.9% M 68.9% 38.5% 51.1% 71.0% 75.3% 35.4% 39.8% 51.1% 46.7% 52.3% 64.8% 80.5% 46.4% 55.5% O 70.5% 50.5% 46.4% 82.1% 72.2% 48.5% 47.5% 44.2% 46.6% 56.9% 55.8% 82.3% 55.6% 58.4% # 3 3 3 3 4 5 3 2 5 7 4 4 3 3.8 8 Table 3: Number of questions for which respondents of each gender chose tone, action, mention, softening, or multiple function as their most frequent choice, and their within-gender percent agreement. Tone Action Reaction Mention Softening Multiple Avg. # % # % # % # % # % # % % F 35 61% 4 44% 3 33% 3 40% 5 43% n/a n/a 56.9% M 40 57% 3 56% 2 33% 2 42% 2 44% n/a n/a 55.5% O 29 64% 5 48% 4 39% 6 48% 4 53% 1 46% 58.4% Table 4: Number and percentage of questions for which respondents of each gender agreed in their first- or second-most- frequent choice with the authors’ code assignments for tone, action, reaction, mention, and softening. Tone Action Reaction Mention Softening All Agreements # % # % # % # % # % # % Female 23 100 6 100 4 50 7 100 6 67 46 87 Male 23 100 5 83 3 38 5 71 6 67 42 79 Other 20 87 6 100 4 50 7 100 7 100 44 83 Authors 23 – 6 – 8 – 7 – 9 – 53 – responses to an open-ended question in the survey: “Do you males (26%). They also found the survey “very easy” or have any other comments about emoji use in social media?” “somewhat easy” (64%) more often than the females (59%) The ‘others’ provided proportionately more answers to this and males (52%) did. In terms of their social media use, they question (40% of ‘other’ vs. 29% of males and 22% of were also less likely to have a Facebook account (61% females), and their comments tended to be longer and to compared with 85% F and 93% M) and more likely to have focus on nuances of emoji interpretation. For example, one an account on the micro-blogging site Tumblr.com (32%, ‘other’ gender respondent commented: compared with 19% F and 13% M). We discuss these demographic differences below, together “Emojis are useful as shortcuts, not just in a one- with the finding that the ‘other’ genders often differ in their to-one way (eg, a thumbs up emoji meaning that emoji interpretations compared to the females and males. a person agrees with what the other person suggested) but also more ambiguously. a heart 5 Discussion emoji can be used to express support and care to a friend. it's that i dont know the words to say to The analysis of our survey data in response to actual you right now, but i love you and i care about you, (anonymized) Facebook group messages revealed two main and all the other things i dont know how to say findings: right now bc wow is that overwhelming.” 1. Self-identified females and males mostly agreed in Typical male comments, in contrast, included: their interpretations of the 13 emoji types represented “It is a very interesting development in linguistics” in the survey. “Many overuse it [emoji] for no reason which at 2. The ‘other’ gender respondents differed more from the times is irritating” females and males than the females and males differed from each other. Comments by females were more varied, and resembled ‘other’ comments more than male comments. The lack of gender differences in female and male Finally, the ‘other’ group differed in its responses to the interpretations of emoji functions is surprising in light of the demographic and social media usage questions. It was considerable evidence that females and males use emoji younger; 80% of the ‘other’ respondents were between the differently. At the same time, people can often understand ages of 18 and 29, and 40% were between 18 and 22 years language that they do not themselves produce, as in the case old. The ‘others’ reported being more confident in their of language learners. Moreover, our findings are in line with survey answers (36%) compared with females (30%) and the findings from [JXLHBA17] regarding the interpretation of emoji sentiment, and suggest that females and males have 8 The total is 49, because several of the survey items had multiple versions depending on, e.g., which part of a message the respondent interpreted the emoji as associated with. similar mental representations of emoji. The differences in interpretations for emoji use, and each was selected by the female and male emoji usage reported in the literature are majority of respondents for at least some of the examples. Of thus presumably social in nature, e.g., emoji play a role in the additional function options that we included, softening identity performances, as suggested by [Sug15; Wol00]. was also selected quite often by the survey respondents, Performing gender via emoji can be a double edged sword, which suggests that it should be added to the taxonomy of however, as one female respondent noted at the end of the emoji functions. Of the other functions we added, all were survey: chosen by some respondents for some items; however, the physical action function was chosen only rarely, and the “I don’t really like using them. I’m female and I decoration function was selected only slightly more. These feel that if I didn’t use enough positive emojis, it do not appear to be major emoji functions, at least in would be perceived as rude, unenthusiastic, or Facebook groups. At the same time, the results showed that apathetic by my acquaintances or friends. This is tone modification, in addition to being a very common sort of a pain because I want men who see me on function for which emoji are used [HD17; CJT16; Gul16; social media to take my ideas seriously, and I’m NPM17; WD01], was also the default interpretation that the pretty sure that the more emojis I use, the dumber receivers assigned to most of the emoji. I’ll be perceived by men.” The limitations of the study should be acknowledged. The A systematic investigation of gender differences in use vs. survey instrument contained only two to five examples of interpretation of emoji would be a useful future study. each of the 13 emoji types, and the emoji were not always As for the ‘other’ genders, it is tempting to conclude that rendered in the same way within the survey, but rather were they resist simple interpretations of emoji functions in the rendered according to how they appeared in the source same way that they resist binary gender norms. However, Facebook messages (Table 1). This was done to preserve the we lack specific information about the make-up of the authenticity of the context, but it also introduced an element ‘other’ gender. The category could include internet users of variability that was not controlled for in our analysis. Still, who identify as non-binary, gender-fluid, or who otherwise for the four cases where emoji types were rendered variably reject the gender binary, but it could also include people who (blush, crying, frown, and tongue out), respondents generally identify as female or male but for whatever reason prefer not agreed on the pragmatic functions across cases, except for to provide that information in an online survey. Still, clues one blush item where the ‘others’ disagreed with the females as to why the ‘other’ gender respondents differ in their and males. A fine-grained analysis at the level of the understanding of emoji functions can be inferred from their individual example was not feasible for this study due to age and patterns of social media use. They are young, and insufficient data. Future gender-based research should they are less likely to have a Facebook account and more include multiple examples of a function associated with a likely to have a Tumblr account than the other survey particular emoji (e.g., a big smile serving as softening). respondents. Consistent with this, Tumblr users skew It is also likely that the respondents’ interpretations were young, and the platform has many LGBTQ users [BR17]. influenced by the specific examples we included. For This suggests that there are different norms of emoji example, the three items containing heart emoji each had a interpretation associated with different social media different code that was favored by all three genders (i.e., platforms (e.g., Facebook vs. Tumblr). Interestingly, action, mention, and tone). The included examples were not emoticons and GIFs are more common than emoji on necessarily representative; some were chosen because they Tumblr, suggesting that the ‘other’ genders encounter emoji posed challenges when we coded them in earlier studies [e.g., elsewhere (79% of the ‘other’ respondents who finished the HD17]. Moreover, there are many emoji that we could not survey indicated that they use emoji), or possibly that the include in the survey, given the need to keep the survey to a other graphicons on Tumblr influence their perception of reasonable length. Future research should investigate what emoji. (Cf. [HD17], who suggest that the different functions are associated with other emoji, and whether and if graphicon types interact as a system.) Unfortunately, the so how the interpretations of those functions vary. number of ‘other’ gender respondents in our survey is too Another limitation concerns the ‘other’ gender category. small to allow for reliable examination of differences in their No information was available about why respondents makeup on variables such as platform use and age. There is selected that category; some people may just not have wanted a need for further research on this interesting category, to share their gender information, while others may be gender perhaps incorporating one-on-one interviews. nonconforming. That uncertainty coupled with the relatively Another finding of this study is that different emoji types small population of ‘others’ (N=50; 584 function codes) specialize to some extent in expressing different pragmatic makes the ‘other’ gender results less reliable and more functions. Although tone modification was the most challenging to interpret than those for self-identified females common interpretation assigned to the emoji examples and males. Nonetheless, a number of indicators suggest that overall, and all of the emoji were interpreted by at least some the ‘other’ category possesses internal coherence, such as the respondents as indicating tone, the blowing a kiss emoji, for fact that its members tended to agree with each other on their example, was more often interpreted as an action or mention emoji interpretations. In future research, users of different (illustration of the accompanying text), and the grimace social media platforms could be interviewed to gain further emoji was often interpreted as a reaction. This finding, if insight into their norms of emoji use and interpretation. supported by further research, offers a more nuanced Finally, the findings of this study are limited in their understanding of emoji function than in previous studies. generalizability in certain respects. The survey items included Last, the survey results validate [HD17]’s functional challenging instances of emoji use drawn from particular categories, in that all those functions were selected as Facebook groups where graphicon use is common, and the survey respondents were not randomly or systematically comment!" Graphicons in Facebook comment selected. The ways in which emoji function on other social threads. Proceedings of HICSS-50, 2185-2194. IEEE. media platforms may differ. However, the findings should [JA17] Jaeger, S. R., & Ares, G. (2017). Dominant meanings of generalize to other similar Facebook contexts, and since facial emoji: Insights from Chinese consumers and Facebook is extremely popular, many people likely have comparison with meanings from internet encountered the kinds of emoji instances used in our survey, resources. Food Quality and Preference, 62, 275-283. which may make the findings more generally applicable. [JXLHBA17] Jaeger, S. R., Xia, Y., Lee, P-Y., Hunter, D. C., Finally, despite the survey including difficult-to-interpret Beresford, M. K., & Ares. G. (2017). Emoji examples, respondents mostly agreed among themselves questionnaires can be used with a range of population and with the researchers’ interpretations. Agreement on segments: Findings relating to age, gender and frequency of emoji/emoticon use. Food Quality and other, less challenging emoji uses should be even higher. Preference. https://doi.org/10.1016/j.foodqual.2017.12.011 6 Conclusion [KW15] Kelly, R., & Watts, L. (2015). Characterising the This study found that female and male social media users did inventive appropriation of emoji as relationally not differ appreciably in their interpretations of emoji meaningful in mediated close personal relationships. In Experiences of Technology Appropriation: functions in Facebook messages. This suggests that Unanticipated Users, Usage, Circumstances, and researchers and designers need not be unduly concerned Design. http://opus.bath.ac.uk/46780/1/emoji_ about misunderstandings of emoji due to user gender; both relational_value.pdf females and males appear to understand (or misunderstand) [Lid98] Liddy, E. D. (1998). Enhanced text retrieval using emoji in similar ways. However, research findings for one natural language processing. Bulletin of the Association platform do not necessarily transfer to other platforms. The for Information Science and Technology, 24(4), 14-16. norms of graphicon usage on different social media [Ma16] Ma, X. (2016). From Internet memes to emoticon platforms need to be taken into account in research on emoji engineering: Insights from the Baozou comic interpretation. phenomenon in China. In HCI 2016, Part III, 15–27. The findings of this study have implications for automating emoji interpretation. Identifying pragmatic [MKTTH17] Miller, H., Kluver, D., Thebault-Spieker, J., Terveen, L., & Hecht, B. (2017). Understanding emoji ambiguity usage is a challenging task in Natural Language Processing in context: The role of text in emoji-related [LID98]. We propose that some version of the emoji miscommunication. In Proceedings of ICWSM 2017, function taxonomy could be used to train a classifier to pp. 152-161. AAAI Press. recognize emoji functions in public Facebook groups. The [MTCJTH16] Miller, H., Thebault-Spieker, J., Chang, S., Johnson, associations found in this study between emoji types and L., Terveen, L., & Hecht, B. (2016). Blissfully happy or functions, if validated by further research, could assist ready to fight: Varying interpretations of emoji. In greatly in identifying those functions. Proceedings of ICWSM 2016. AAAI Press. http://www- users.cs.umn.edu/~bhecht/publications/ 7 References ICWSM2016_emoji.pdf [NPM17] Na'aman, N., Provenza, H. & Montoya, O. (2017). [BKRS16] Barbieri, F., Kruszewski, G. Ronzano, F. & Saggion, Varying linguistic purposes of emoji in (Twitter) H. (2016). How cosmopolitan are emojis?: Exploring context. In Proceedings of ACL 2017, Student Research emojis usage and meaning over different languages Workshop, pp. 136-141. with distributional semantics. In Proceedings of the 2016 ACM on Multimedia Conference, 531-535. [SGDH06] Shields, S. A., Garner, D. N., Di Leone, B., & Hadley, ACM. A. M. (2006). Gender and emotion. In J. E. Stets, & J. H. Turner (Eds.), Handook of the Sociology of Emotions [BR17] Byron, P., & Robards, B. (2017). There's something (pp. 63–83). New York: Springer. queer about Tumblr. The Conversation. https://theconversation.com/theres-something-queer- [Sug15] Sugiyama, S. (2015). Kawaii meiru and Maronaka neko: about-tumblr-73520 Mobile emoji for relationship maintenance and aesthetic expression among Japanese teens. First Monday, [CLSAWLM17] Chen, Z., Lu, X., Shen, S. Ai, W., Liu, X., & Mei, 20(10). Q, (2017). Through a gender lens: An empirical study of emoji usage over large-scale Android users. arXiv [Tig16] Tigwell, G. W. (2016). Oh that's what you meant!: preprint arXiv:1705.05546 Reducing emoji misunderstanding. MobileHCI 2016. [CJT16] Cramer, H., de Juan, P., & Tetreault, J. (2016). Sender- [TF96] Thompsen, P. A., & Foulger, D. A. (1996). Effects of intended functions of emojis in US messaging. In pictographs and quoting on flaming in electronic mail. Proceedings of the 18th International Conference on Computers in Human Behavior, 12(2), 225-243. Human-Computer Interaction with Mobile Devices [WD01] Walther, J. B., & D’Addario, K. P. (2001). The impacts and Services, pp. 504-509. New York: ACM. of emoticons on message interpretation in computer- [Gul16] Gullberg, K. (2016). Laughing face with tears of joy: mediated communication. Social Science Computer A study of the production and interpretation of emojis Review, 19(3), 324-347. among Swedish University Students. Bachelor Thesis [Wol00] Wolf, A. (2000). Emotional expression online: Gender in General Linguistics, Lund University. differences in emoticon use. CyberPsychology & [HD17] Herring, S. C, & Dainas, A. (2017). "Nice picture Behavior, 3(5), 827–833.