• Subscribe

The not-so-secret language of 😀

Speed read
  • Emoji use gaining in popularity
  • A few emoji are used a lot, but the majority appear infrequently
  • Most popular emoji have unambiguous meaning and add emotional context

You know what 🤔 means. And you can probably think of a few uses for 💩. But what about 📛? Or 🔏?

Today, ninety-two percent of the online population use emojis, sending 6 billion 😀, 😺, and 💃 a day through text, email, Twitter, Facebook, and other platforms. 😂 was even named word of the year by Oxford Dictionaries in 2015.

<strong>A picture is worth a thousand words. </strong> Researchers at the University of Michigan combed through over a billion messages in an attempt to plot the semantics of the nascent emoji language.

The whole idea of using symbols to enhance text-based communication arguably began in 1982 when computer scientist Scott Fahlman proposed using :-) to identify jokes on the Carnegie Mellon university digital message board.

In the late 1990s, NTT DoCoMo, a Japanese mobile provider, took Fahlman’s emoticons one step further and invented the first emojis, a set of 180 images that could be used on NTT’s mobile internet platform.

Emojis add nuance and emotional resonance to typed communication, compensating for a lack of body language and other social cues. They also facilitate communication across international cultures and languages.

Or, rather, they do if we agree about their meaning.

Some emojis, such as  are used a lot, while others, like  and 🎰 much less frequently. How do people assign meaning to images? What makes one emoji popular and others overlooked? Why is 🌕 used more than 🌙 and 🌛?

To get to the bottom of all this, Wei Ai, a member of the Foreseer research group at the University of Michigan’s School of Information, and his collaborators at Peking University have used machine learning to analyze the semantics of emojis.

“I use emojis a lot every day,” says Ai. “It fascinates me how much information a single emoji can convey. I couldn’t help being curious about how people use and interpret them.”

Voted most popular

Ai and his colleagues collected 1.22 billion anonymized messages from 1.03 million English-language users of the Kika Emoji Keyboard in September 2015. Because Kika is a system-wide keyboard, the collected data is not restricted to use within a single app.

<strong>Wei Ai </strong> and his team use machine learning to chart a theory of emoji language. Courtesy University of Michigan School of Information.More than nine percent of the collected messages included at least one emoji, but the popularity of individual emojis wasn’t evenly distributed. A small portion of emojis were frequently used, while the majority of the 1,281 available on the Kika keyboard were not.

The most commonly used emoji was 😂, described by Unicode (a computer industry standard for representing the world’s writing systems) as ‘face with tears of joy’. Ai speculates that its popularity is related to its association with popular internet slang such as ‘lol’ and ‘lmao’.

“Emojis are more likely to be used if they are semantically close to a popular word,” says Ai. “Popular emojis have clearer meanings and less ambiguity and provide a good substitute to their word counterparts.”

Emotion FTW!

Ai and his team used Large-scale Information Network Embedding (LINE) to project emojis in semantic space, where they could then measure the distance between individual emojis and between words in close proximity.

All of the data for the analysis was hosted on Amazon Web Service’s (AWS) S3 cloud storage. Apache Spark clusters were used to process the data in parallel, and the LINE embedding model was trained on a single EC2 server, using 32 CPU cores.

They found two different relationships between emojis and words: Complementary and supplementary. A complementary emoji doesn’t change the meaning of a message, but adds an extra layer: ‘That cat is adorbs! 😻’. Supplementary emojis actually replace the words themselves. If the emoji were removed it would affect the message’s meaning: ‘Can your 🐩 come over to play?’<strong>What's it all mean?</strong> A graphic representation of emojis and their accompanying words. Machine learning tracked these spatial relationships to derive semantic values for popular emojis. Courtesy Wei Ai.

Complementarity was positively associated with emoji popularity, and many of the most complimentary emojis are sentimental — they add emotional context.

Clear meaning was also correlated with popularity. Emojis with unclear meanings are less likely to be used, though an emoji that can fit into many different contexts is more likely to be popular.

While Ai admits that emoji research may not have a direct impact on society, he says that “as a new type of nonverbal cue, emojis facilitate communication and spice up our chat. We need knowledge of emoji usage to correctly understand users.”

And when we understand each other, we’re less likely to end up in a 💩.

Join the conversation

Do you have story ideas or something to contribute? Let us know!

Copyright © 2023 Science Node ™  |  Privacy Notice  |  Sitemap

Disclaimer: While Science Node ™ does its best to provide complete and up-to-date information, it does not warrant that the information is error-free and disclaims all liability with respect to results from the use of the information.


We encourage you to republish this article online and in print, it’s free under our creative commons attribution license, but please follow some simple guidelines:
  1. You have to credit our authors.
  2. You have to credit ScienceNode.org — where possible include our logo with a link back to the original article.
  3. You can simply run the first few lines of the article and then add: “Read the full article on ScienceNode.org” containing a link back to the original article.
  4. The easiest way to get the article on your site is to embed the code below.