🤌 and its meanings: An emojical study
Disclaimer: this paper uses emojis from the latest standard and it might not work well in some browsers. You can find a PDF version here.
Non-verbal communication is not easy, but Italians are somehow better at it. See this video of a little Italian girl as an example: https://www.youtube.com/watch?v=Z5wAWyqDrnc. In this paper of the Computational Social Science Lab Holiday Paper Series, we set out to learn about non-verbal, non-facial communication online through the use of one of the most iconic Italian expressions: 🤌. We study its meaning in multilingual Twitter datasets and its association with sentiment to develop a tool to help using hand emoji in online communication. We further study the role of skin color in hand emoji in Reddit and the change in their frequency of use since the beginning of the COVID-19 pandemic, as an example of an emojical study focusing on various aspects of the use of 🤌.
Emojis offer the possibility of new non-verbal expression online. Without having to rely on pictures or videos, we can encode our facial expression and hand gestures in text. Over recent years, the frequency of emojis in online communication has been growing steadily, for example with tweets containing emojis reaching a 20% rate in 2020 . As new emojis are added to include offline gestures, some cultures use them with different meanings as others. Furthermore, gestures can have a meaning in a culture and be adopted by another culture through online text communication, creating a novel kind of online assimilation of non-verbal signs.
One of the newest and most confusing emojis is 🤌, a typical Italian gesture but with other meanings and uses in other languages. The pinched-fingers emoji, as it was named by the Unicode Consortium when debuted in March 2020, was supposed to graphically represent the exasperated expression “what do you want?”, since this is the common meaning of this hand gesture when made by Italians. Its creators submitted a proposal  supporting the need for such an emoji, basing their arguments on the uniqueness and completeness of such a gesture, together with its high frequency of usage in the large group of Italian speakers and people with Italian cultural affinity or ancestry. Finally, their proposal was accepted in 2020, together with another hundred new emojis. However, a conversation around other possible meanings of this gesture rooted in other cultures and languages pointed out that it can also mean “wait a minute” in Israel and in the Arab world or represent an iconic gesture of the famous K-pop star, Yuri.
Here1, we study the meanings of by analyzing Twitter data to identify the contexts in which 🤌 appears. We consider various languages to detect the drivers of its use in the events of recent years. We continue by studying the sentiment associated with and design a NLP tool to help non-native-hand-waving speakers to learn how to express themselves in emoji hand language. We then analyze how the use of general hand emojis has changed during the pandemic and how skin modifiers of hand emojis are associated with negative votes to Reddit comments.
Using the brand new historical search endpoint of the the Twitter API v2, we retrieved an exhaustive dataset of tweets in English that contain 🤌 up to 2021-11-28, ignoring retweets. To obtain a set of topics that can give a general intuition of the different uses and contexts of the emoji, we applied topic modelling with the Mallet Latent Dirichlet Allocation (LDA) algorithm in Python’s Gensim package wrapper .
We preprocessed the text by removing stopwords, detecting bigrams, and studying the most and least frequent words. Since the quality of preprocessing tools differs across languages, we do this first exploration only with English data. Additionally, we assume that English, in contrast to other languages, can be seen as an approximation to the different uses of the emoji in other languages when the speakers of a language communicate online with general audiences by tweeting in English. As a consequence, English Twitter text might also contain the meaning which Italians associate with 🤌.
Another important parameter in the use of the topic model is the number of topics. We explore a large range of values and compute the coherence score for each number of topics, obtaining a maximum coherence score for a model with 24 topics. For the obtained topics, reported in Figure 1, we group and describe the diverse contexts in which we encounter the “unique” (or not so much) Italian emoji. We also take into account examples of tweets that represent each topic, and compare with a smaller model with 15 topics that summarizes better some groups of topics.
The first three topics are related to Italian culture: Topic 1 is a direct representation of Italian culture and language and its relation to hand gestures and emojis. Topic 2 and Topic 3 are related to food and sports, specifically to Italian gastronomy and sport events in which Italy was heavily involved (e.g., they won the UEFA Euro 2020). Secondly, Topic 4 shows a strong connection to politics, which could be linked to complaints about regulations and political decisions. The next broad group of topics (5-9) is related to the concept of quality of different objects (art, movies, music...). The fact that the expression the "chef’s kiss" appears in these topics hints that 🤌 may express appreciation for good quality and excellence in these tweets. Next, we surprisingly find a topic related to cryptocurrency, in which 🤌 could mean holding a investment in cryptocurrency for future profits. We further observe a topic that we interpret as spam-related tweets, which make use of a large number of emojis to attract attention. An example of a fictional tweet that combines these two topics would be the following: "We just minted an NFT of 🤌 and put it for sale for the affordable price of 100 ETH!! Buy it now here: https://bit.ly/3q4IEUQ"2.
Another surprising finding is that K-pop fandoms (or fan groups), such as Army (BTS fan group), dominate part of the usage of this emoji. Studying some of these tweets individually revealed that 🤌 can be used to replace the finger-heart gesture. The K-pop industry popularized this gesture in South Korea around the 2010s. It still does not have an official emoji, despite the strong presence of K-pop topics online. However, a new emoji for the finger-heart gesture will be released in the upcoming 2022 official emoji update. It would be interesting to study whether this context of 🤌 still appears a while after this release, or if its meaning finally returns its original one. Two further well-defined topics relate to negative emotions. The first includes insults in general and the second one is particularly linked to stress due to school or university deadlines, and possibly includes complaints. The other topics found in this classification were not easily interpretable, and dissolved into the topics described above when the model was set to 15 topics. We chose the version of 24 topics due to its higher coherence score and because we like it and this is a Holiday Paper Series article after all.
Thanks to the new counts endpoint of the Twitter API v2, we can track the frequency of an emoji within a language over time. We calculated the daily count of tweets in six languages (English, German, Spanish, Italian, Japanese, and Korean) that contain 🤌, dividing the count over a baseline that we estimated with the daily count of tweets in those languages that contain at least one stopword. We excluded retweets in all these counts to measure frequency of expression and not in sharing text. We then smoothed the daily counts with a seven-day rolling average. Figure 2 shows the time series for Italian, English, and Korean with spikes highlighted. Using our access to Brandwatch, we generated word clouds including retweets3 on the peaks to get an idea of the words driving the increase in use of 🤌.