A new database introduced in a recent research paper connects online dictionaries of emojis with a semantic network to create the first machine readable emoji inventory called EmojiNet (http://ift.tt/2eQ2dLD).
In April 2015, Instagram reported that 40% of all messages contained an emoji. New emojis are constantly being added. With the rapid expansion and surge of emoji use, how do we know what emojis mean when we send them? And how do we ensure that the person at the other end knows what we mean? It turns out that the meaning of emojis varies a whole lot based on context.
Emojis, derived from Japanese “e” for picture and “moji” for character, were first introduced in the late 1990s but did not become Unicode standard in 2009. Emojis are pictures depicting faces, food, sports, animals and more, such as unicorns, sunrises, or pizza. Apple introduced an emoji keyboard to iOS in 2011 and Android put them on mobile platforms in 2013.
Emojis are different from emoticons, which can be constructed from your basic keyboard, like (-:. The digital use of emoticons has been traced back to as early as 1982, though there are earlier reported cases in Morse Code telegraphs. Earliest uses of emoticons have been later overturned as accidental, including in a poem by Robert Herrick in 1648, which is disputed, and a New York Times transcript of an Abraham Lincoln speech in 1862, also potentially a typo.
The difference between an emoticon and emoji is technically important because computer systems treat emojis like letters from a non-western language, such as Chinese or Japanese. The Unicode Consortium, a nonprofit organization in Mountain View, standardizes emojis and requires that “while the shape of the character can vary significantly, designers should maintain the same ‘core’ shape,” but there is still variability.
As a result, emojis can look different across brands and services like Google, Apple, and Twitter. This adds another wrinkle to an emoji conversation: If you’re an iPhone user texting with a friend who has Android, you might not actually be saying or seeing the same thing.
The authors point out that one of the main problems with understanding emojis–especially for machines– is that they can mean very different things based on the context. They give three examples, including the smiling emoji with tears, a gun, and a bag of money.
Source: Wijeratne, et al.
In an effort to parse the many meanings of emojis, the authors have created the first machine readable sense inventory for emojis called EmojiNet (http://ift.tt/2eQ2dLD), which aims to make sense of emojis, or what researchers term “emoji sense disambiguation.” The system links four emoji resources online, including Emoji Dictionary, a crowdsourced database that organizes the meanings of emojis, with BabelNet, a comprehensive multilingual machine readable dictionary or “semantic network,” which links Wikipedia concepts to dictionary words.
By linking together these resources, EmojiNet provides several important dimensions and ranges of potential meanings to a single emoji. The database also shows how emojis appear on various platforms, defines keywords, related emojis, and relevant nouns, verbs, and adjectives.
Take, for example, the emoji “folded hands,” which the authors cite as the most commonly misused emoji in social media. The emoji is defined as “two hands placed firmly together, meaning please or thank you in Japanese culture. Other common uses for this character include prayer/praying hands, or a high-five.” Keywords include gesture, ask, bow, and folded, and associated nouns are hand, heaven, hallelujah, prayer, Namaste, and thanks. Associated verbs are hope, wish, thank, begging, praying, blessed, plead, and high five; adjectives include heavenly, pleading, and praying.
Another example is the “face with tears of joy” emoji, which is defined as “a laughing emoji which at small sizes is often mistaken for being tears of sadness. This emoji is laughing so much that it is crying tears of joy.” The emoji is associated with the nouns like giggle and laughing, verbs crying and lol, and adjectives hilarious and joyful.
Now, if you have any confusion over what your friend meant in their last emoji text, you can decode it with EmojiNet– or build code that uses EmojiNet to do that for you automatically– but it just might be faster to ask your friend directly.
Video of Siri on CarPlay announcing all emojis: