What is the science of emojis? It is the study of how tiny pictures are made, stored and understood. Emojis began as 12×12 pixel pictures on Japanese mobile phones in the late 1990s. Today they are standardised characters in Unicode, a code that every phone and computer shares. Psychologists and neuroscientists also study how our brains read them, and why the same face can mean different things to different people.
What is an emoji, exactly?
The word emoji comes from two Japanese words: e (絵), meaning “picture”, and moji (文字), meaning “character”. So an emoji is literally a picture-character. It looks as if it comes from “emotion”, but that is a coincidence. The word emoticon, which blends “emotion” and “icon”, is a different thing.
An emoticon is built from keyboard characters, such as :-) or ;-). An emoji is a single character in its own right, stored in a computer’s character set and drawn by a font. That difference matters, and it is the reason the rest of this article has a computer-science half as well as a psychology half.
Emojis also go well beyond faces. The standard includes animals, food, vehicles, weather, flags, hand gestures and symbols. The “Smileys & Emotion” group, which holds the faces, is only 172 of the 3,972 emoji in the latest version. We will count them all later.
The history of emojis: from 1997 to Shigetaka Kurita
People often say that one person invented emoji in 1999. The real story has a few steps, and it helps to get them in order.
1997: the first phone set
The first emoji set on a mobile phone came from the Japanese phone company SoftBank, which at the time traded as J-Phone, in 1997. It had 90 monochrome characters on a 12×12 pixel grid. They showed numbers, sports, the time, moon phases and the weather. It was a practical tool for short messages, not a language of feelings.
1999: Kurita and NTT DoCoMo
In 1999 a designer called Shigetaka Kurita (born 1972) worked on the interface of i-mode, the mobile internet service from NTT DoCoMo. He designed a set of 176 colourful pictograms, again on a 12×12 pixel grid. Think about how little that is. Each emoji had only 144 pixels to carry its meaning, so every pixel had to count.
Kurita drew on several sources. These were manga symbols (called manpu), weather symbols, Chinese characters and street signs. The idea was to give very short messages some of the context that a face-to-face conversation has. In 2016 the Museum of Modern Art (MoMA) in New York acquired the original 176 emojis for its collection.
It is fairer to call Kurita the designer of the best-known early set than the inventor of all emoji. SoftBank’s set came first on mobile phones, and later emoji sets by rival Japanese companies each added their own pictures. This is also why the early history is a story of competing, incompatible sets. A heart drawn for one company’s phone could arrive as a blank box on another company’s phone.
2007 to 2010: the standard arrives
The messy situation in Japan was solved by standardisation. Google began advocating for emoji to be added to Unicode in 2007. In November 2008 Apple released its Apple Color Emoji font for iPhone OS, though at first it was only available to SoftBank customers in Japan. Then, in October 2010, Unicode 6.0 standardised 722 emoji.
That was the real turning point. Once emoji had official code points, any system could use them. A message sent from one brand of phone could be understood by another, as long as both had a font that included the picture.
A short emoji timeline
- 1997: SoftBank (J-Phone) releases the first mobile-phone emoji set, with 90 monochrome characters. (Researchers have since found emoji-like picture sets on earlier Japanese electronic devices from 1988 and 1990.)
- 1999: Shigetaka Kurita designs 176 colourful emoji for NTT DoCoMo’s i-mode.
- 2007: Google begins advocating for emoji standardisation.
- 2008: Apple releases Apple Color Emoji for iPhone OS, first for SoftBank users.
- 2010: Unicode 6.0 standardises 722 emoji.
- 2015: Unicode 8.0 adds five skin tone modifiers, and Oxford Dictionaries chooses 😂 as its Word of the Year.
- 2016: MoMA acquires Kurita’s original 176 emoji.
- 2026: Emoji 18.0 is released in September, with 3,972 emoji in total.
Why were pictures useful in the first place?
Text on its own loses a lot. In a real conversation we use tone of voice, facial expression and gesture to show whether a sentence is a joke, a complaint or a thank-you. Plain text has none of those. Writing “fine.” could mean anything from “absolutely fine” to “I am annoyed”.
Emojis act as a stand-in for some of those missing cues. Scientists who study digital communication sometimes call them non-verbal cues for typed language. They do not replace words, and as we will see, they can be misread. But they add a layer of tone that plain letters cannot easily carry. If you are interested in how real facial expressions are managed in different cultures, our article on display rules of emotion explains why a smile does not always mean the same thing around the world.
Where do the pictures live? A first look at emoji data
A phone does not store a picture of a smiling face inside your message. It stores a number. When your friend’s phone receives that number, it looks up the picture in a font and draws it. We will look at that number, the code point, in the next section. It is the key idea that links emoji to the wider world of computer science, and it is a good topic for anyone working through IGCSE computer science.
Every emoji you send is therefore two things at once: a precise piece of data that machines agree on, and a drawing that each company designs for itself. The data is the same everywhere. The drawing is not. That gap between the shared number and the individual picture explains many emoji misunderstandings, and it is the subject of part three.
How emojis are encoded: code points
Computers can only store numbers, so every character on your screen, whether a letter, a Chinese character or an emoji, is stored as a number. Unicode is the worldwide standard that gives each character its own number, called a code point. Code points are written in hexadecimal with the prefix “U+”. The standard is maintained by the Unicode Consortium, a non-profit organisation.
Take the emoji that Oxford chose in 2015, 😂 (“Face with Tears of Joy”). Its code point is U+1F602. The grinning face 😀 is U+1F600 and the thumbs-up 👍 is U+1F44D. Notice that these numbers are large. Early computer codes such as ASCII only had room for 128 characters, so emoji needed the much bigger Unicode space.
From code point to bytes
A code point is an idea. To send it across the internet it must be turned into bytes, using an encoding. The most common one on the web is UTF-8. It stores the plain English letters in one byte, but most single emoji need four bytes. For example, 😂 is stored in UTF-8 as the four bytes F0 9F 98 82, written in hexadecimal. In UTF-16, which many programming languages use inside, the same emoji becomes a pair of 16-bit values, D83D and DE02, called a surrogate pair.
| Emoji | Name | Code point(s) | UTF-8 bytes (hex) |
|---|---|---|---|
| 😀 | Grinning Face | U+1F600 | F0 9F 98 80 |
| 😂 | Face with Tears of Joy | U+1F602 | F0 9F 98 82 |
| 👍 | Thumbs Up | U+1F44D | F0 9F 91 8D |
| ❤️ | Red Heart | U+2764 + U+FE0F | E2 9D A4 EF B8 8F |
The red heart is a good puzzle. The heart symbol U+2764 existed long before emoji and can be drawn as a plain black text symbol. The extra code point U+FE0F, called variation selector-16, asks the computer to show it as a coloured emoji instead. That is why one visible heart can be made of two code points and six bytes.
Skin tone modifiers
The first emoji people were mostly yellow, a neutral colour. Many users wanted to see themselves in the pictures, so in June 2015 Unicode 8.0 added five skin tone modifiers. They are based on the Fitzpatrick scale, a skin-type scale developed by the dermatologist Thomas Fitzpatrick in 1975. The five modifiers are code points U+1F3FB to U+1F3FF.
The clever part is how they work. A modifier is not a separate picture. It is a code point placed straight after a human-shaped emoji, and the font draws the emoji in that skin tone. So 👍🏽 is just two code points: U+1F44D (thumbs up) followed by U+1F3FD (a medium skin tone). If a device does not understand the modifier, it shows the plain thumb and a coloured square side by side. That is a useful clue that two code points are hiding behind one picture.
ZWJ sequences: building new emoji from old ones
Unicode has a limited number of free code points, and adding every combination as a new character would be wasteful. So it uses a trick called a zero-width joiner, or ZWJ. This is an invisible character, U+200D, that tells the computer: “join the neighbouring emoji into one picture if you can”.
The family emoji 👨👩👧 is a good example. It is made of man (U+1F468), ZWJ (U+200D), woman (U+1F469), ZWJ (U+200D) and girl (U+1F467). That is five code points that a supporting device draws as one family. If a system has no picture for that exact combination, it falls back and shows the three separate people. Nothing is lost, because the message still makes sense.
Flags are letters in disguise
Country flags use another trick. A flag is two special “regional indicator” letters. The UK flag 🇬🇧 is U+1F1EC followed by U+1F1E7, which are regional indicator versions of G and B, the country code. The England flag is built differently, from a black flag U+1F3F4, then hidden tag letters that spell “gbeng”, then a cancel tag U+E007F.
How many emojis are there? Emoji 18.0
The number depends on what you count. Unicode’s emoji-counts page for version 18.0 lists 3,972 emoji in total, and this figure includes sequences, such as skin-tone and gender variations, not only single characters. Emoji 18.0 was released in September 2026, together with Unicode 18.0. According to Emojipedia, the new version adds 19 emoji, and they are expected to appear on devices through 2027, as each company designs its own pictures and releases software updates.
| Category | Emoji in version 18.0 |
|---|---|
| People & Body | 2,430 |
| Flags | 270 |
| Objects | 268 |
| Symbols | 224 |
| Travel & Places | 221 |
| Smileys & Emotion | 172 |
| Animals & Nature | 161 |
| Food & Drink | 132 |
| Activities | 85 |
| Component | 9 |
Look at the biggest category. People & Body has 2,430 entries, far more than any other group, because every hand gesture and person comes in several skin tones and often several gender versions. The old claim that there are “over 3,000 emoji” was true for years, but it is now out of date.
Who decides which emojis exist?
New emoji are not invented by phone makers alone. The Unicode Consortium publishes a process in which people can write a proposal, explaining why a new emoji would be widely useful. A Unicode committee for emoji reviews proposals. If one is accepted, it receives a code point, and then each company, such as Apple, Google, Microsoft and Samsung, designs its own version. That means the number is shared, but the artwork is not.
The Emojipedia notes on version 18.0 show that this process can change late in the day: one proposed emoji was removed from the draft list in January 2026 and another was added in its place. Drafts are not guaranteed until the version is released.
Test your knowledge: reveal the answers
What is a code point? (1 mark)
A unique number, written in the form U+XXXX, that Unicode gives to each character.
Explain how the single picture 👍🏽 is stored. (2 marks)
It is two code points: U+1F44D for the thumbs-up, followed by the modifier U+1F3FD, which tells the device to draw it in a medium skin tone.
What does a zero-width joiner do? (2 marks)
It is an invisible character (U+200D) placed between emoji so that a supporting device combines them into one picture, such as a family. If unsupported, the separate emoji are shown instead.
If you want to see how other kinds of language compare, the way writing systems squeeze meaning into symbols is explored in our article on the language with the most words.
Same code, different picture: emoji rendering
Here is the catch that comes from part two. Unicode defines which emoji exist and gives them names and code points, but it does not draw them. Each company designs its own artwork. So when you send U+1F600 from one phone, the receiver’s phone looks up its own font and draws its own version of the grinning face. The meaning is shared, but the design can be quite different. Eyes, mouth, teeth and even the colours vary from one platform to another.
The data stays identical, which is the whole purpose of the standard, but the rendering changes. This is similar to sending the word “hello” in two different fonts. The word is the same, but one font might look friendly and another stern. With a face, those small changes in the shape of a mouth can change the mood.
The research on misreading
A research team at the University of Minnesota, led by Hannah Miller with colleagues including Brent Hecht, tested this in a 2016 paper called “Blissfully Happy” or “Ready to Fight”. Participants rated popular emoji drawn in the styles of several platforms. When people looked at the very same rendering, they disagreed about whether it was positive, neutral or negative 25% of the time. When the renderings came from different platforms, the disagreement increased. The authors concluded that there is significant potential for miscommunication.
Notice what this does not prove. It does not say that emoji are unreliable in general. It says that single faces are open to interpretation, in the same way that a raised eyebrow can be.
When designs change
Companies also redesign their emoji over time. A well-known example is the pistol emoji, which Apple changed to a water pistol in 2016, and which other major platforms later followed. Before that, the same code point could look like a toy on one phone and a real weapon on another. The change shows that designers know a drawing can carry a different tone from the standard’s name for it.
How the brain reads emojis
When you see an emoji, your brain does not wait for you to translate it into words. Our brains are tuned to recognise faces very quickly. The infographic below sums up the idea in four steps, from seeing a symbol to understanding the message. It is a simplified model, so the studies that follow give the more careful picture.
Emoticons and the face response
One of the best-known experiments used electrodes on the scalp to record a brain signal called the N170. This is a rapid electrical response, about 170 milliseconds after seeing a picture, that is strongest when we look at faces. In 2014 a team led by Owen Churches in Australia, with colleagues including Michael Nicholls and Hannah Keage, published an event-related potential study of the emoticon :-) in the journal Social Neuroscience. They tested 20 people with upright and upside-down faces, emoticons and meaningless strings of characters.
Upright emoticons produced a large N170, as if the brain treated them like faces. But there was a twist. Turning a real face upside down makes the response bigger, while turning an emoticon upside down made it smaller. The authors concluded that an upright emoticon is processed in face-sensitive brain areas because it has a familiar face-like layout, but the brain does not handle it in exactly the same way as a real face. If you want the wider background on the brain regions involved, our guide to neuroscience and the brain is a good place to start.
Reading irony: the wink study
Another study looked at how emoji affect understanding of meaning. In 2018, Benjamin Weissman and Darren Tanner published three brain-wave experiments in PLOS ONE, with 106 adult English speakers aged 18 to 35. Volunteers read sentences word by word that ended in a smile, a frown or a wink emoji. A wink at the end of a sentence produced brain responses called the P200 and P600, linked to people reading the sentence as ironic. The larger the P600, the more likely a person was to give an ironic reading.
The study also found that people differed. Some read the wink as irony and some took the sentence literally. That is an important point for everyday use: an emoji can add meaning, but the reader still has to decide what it means.
Emoji and emotion: what psychology shows
Are emoji mostly positive?
A 2015 paper in PLOS ONE by Petra Kralj Novak and colleagues built a “sentiment map” of emoji. They used over 1.6 million tweets in 13 European languages, which human annotators labelled as positive, neutral or negative. From these they ranked 751 frequently used emoji. They found that most emoji are positive, especially the most popular ones. Tweets with emoji had an average sentiment score of +0.365, compared with +0.106 for tweets without them.
This does not mean that emoji cause happiness. It shows that people tend to use them with positive messages, and that the labellers agreed more when an emoji was present. It is a measure of how emoji are used, not of how they change anyone’s mood.
Do smileys help at work?
A smile in real life makes you look warm and capable. Does the digital smiley do the same? In a 2018 paper called “The Dark Side of a Smiley” (online 2017) in Social Psychological and Personality Science, Ella Glikson, Arik Cheshin and Gerben van Kleef ran three experiments on work-related first impressions. Their finding was that smileys did not increase perceived warmth, and they decreased perceived competence. People who saw the other person as less competent also shared less information with them.
Science also checks its own results. In 2023 a preregistered replication with 847 participants, by Lai and Mayiwar in the journal Collabra: Psychology, used the same design as one of the original experiments. It confirmed a negative effect on perceived competence. But it found a positive effect on warmth, whatever the level of formality, which differs from the original report. So the careful summary is that smileys in formal work messages may cost you some credibility, while the effect on warmth is less clear.
Can emojis change your mood?
Honest answer: the studies above show that emoji change how a message is read and sometimes how a sender is judged. They do not prove that sending a smiley makes you or your friend feel happier in the long term. The infographic above summarises reasonable ideas: a positive emoji can make a message feel friendlier, a negative one can add emotional intensity, and context changes everything.
That last point is the most reliable finding of all. The same emoji means different things depending on the people, the words around it and the situation. Part four looks at why.
If this has made you curious about the organ doing all the reading, our collection of awesome facts about the brain is a fun next stop.
Why people read emojis differently
If the code is the same and the pictures are similar, why do people still disagree? The answer is that an emoji is not a word with a dictionary definition. It is closer to a gesture, and gestures depend on who is using them. The infographic from earlier lists five factors: culture, age, context, platform and personal experience. We have already covered platform. Let us look at the others.
Culture and age
Different communities develop their own habits. A gesture that is friendly in one place may be rude in another, and emoji inherit that. Age matters too: groups of people who grew up with different apps often have different “default” meanings for the same picture. Researchers have repeatedly found that interpretation varies, which is why a message from a grandparent and a message from a classmate may use the same emoji for different purposes.
It is a bit like spoken language. In our article on the most common languages in the world you can see how words travel and change. Emoji do the same on a much smaller scale. Many people call them a “universal language”, but a better description is a shared set of symbols whose meaning is still negotiated by each community.
Context: the thumbs-up problem
The thumbs-up 👍 is a classic. In one conversation it means “good”, in another “OK”, “thanks” or “I have read this”, and sometimes it is used with sarcasm. The words around it, the relationship between the people and even the time of day all change the reading. The Weissman and Tanner study shows the same idea at the level of brain waves: a wink changes how a sentence is understood, but not everybody takes it the same way.
The infographic’s example with the word “Okay” is a nice experiment you can run yourself. “Okay.” with a neutral face, “Okay!” with a smile and “Okay…” with a sideways look can feel neutral, positive and possibly negative. Same word, three different tones, all produced by tiny changes in punctuation and a small picture.
Do emojis replace words?
No, and the research does not suggest they do. The infographic “The Science Behind the Smile” puts it neatly: emojis do not replace language, they add another layer of meaning to digital communication. Words carry the main message. Emojis add tone, such as friendliness, humour or irony. The wink study is a good illustration, because the brain used the emoji to adjust the meaning of a sentence that was already written in words.
Oxford’s Word of the Year 2015
In 2015 Oxford Dictionaries made a surprising choice. Its Word of the Year was not a word at all. It was the “Face with Tears of Joy” emoji, 😂, U+1F602. It was the first time that an emoji had won the award. Oxford explained that traditional alphabet scripts were struggling to meet the demands of 21st-century communication, and that emoji were filling a gap in how people express themselves.
The choice was debated. Some people said an emoji is not a word, and they had a point, since it has no dictionary spelling or sound. Others said that language is whatever people use to communicate, and that millions of people were using this face. Either way, it marked the moment when emoji were taken seriously as part of how we write. For the wider picture of how we write and speak, see oral and written communication.
Using emojis well: a quick checklist
The research gives us some practical guidance. None of it is a rule, but it will help you avoid the most common misunderstandings.
- Let the words carry the meaning. Treat the emoji as tone, not as the whole message.
- Think about the reader. A formal email to someone you do not know may be judged differently if it contains a smiley.
- Remember that your picture may look different on their phone.
- Avoid ambiguous emoji for serious news, such as bad results or apologies.
- If a reply seems cold or strange, ask rather than guess. A misread emoji is a very common cause of a misunderstanding.
- Use skin tone and gender variants that suit the message and the people in it.
The questions about being careful with online communication also come up in our essay on the importance and disadvantages of social media.
Test yourself
Key points to remember
Frequently asked questions about the science of emojis
Who invented the emoji?
Shigetaka Kurita is best known for designing the first widely recognised set of 176 colourful emoji for NTT DoCoMo’s i-mode in 1999. However, SoftBank (then J-Phone) released an earlier set of 90 monochrome characters in 1997, so the history involves more than one company and designer.
What does emoji mean in Japanese?
Emoji combines the Japanese words e (絵), meaning picture, and moji (文字), meaning character. So it literally means “picture character”. It looks similar to “emotion” and “emoticon”, but that resemblance is a coincidence rather than the origin of the word.
How many emojis are there in 2026?
Unicode’s emoji-counts page for version 18.0, released in September 2026, lists 3,972 emoji in total. This counts sequences such as skin tone and gender variants as well as single characters. Emojipedia reports that version 18.0 added 19 new emoji.
How are emojis stored in a computer?
Each emoji has a Unicode code point, a number such as U+1F602. It is stored as bytes using an encoding like UTF-8, where this emoji takes four bytes. The device then uses a font to draw the picture. Some emoji are sequences of several code points.
What is a ZWJ sequence?
A ZWJ sequence uses the invisible zero-width joiner, U+200D, to join several emoji into one. For example, the family emoji is man, ZWJ, woman, ZWJ and girl. If a device lacks a combined picture, it displays the separate emoji instead.
How do skin tone emojis work?
Unicode 8.0 in June 2015 added five skin tone modifiers, U+1F3FB to U+1F3FF, based on the Fitzpatrick scale. A modifier follows a human-shaped emoji, and the font draws that emoji in the chosen tone. So a medium-skinned thumbs-up is two code points.
Why do emojis look different on different phones?
Unicode defines the character and its name, but not the exact drawing. Each company designs its own artwork. A 2016 study of popular emoji found that people disagreed about the sentiment of the same rendering 25% of the time, and more across platforms.
Do emojis activate the same part of the brain as faces?
Partly. A 2014 brain-wave study found that an upright emoticon produced a large N170 response, a signal strong for faces. But inverting it had the opposite effect to inverting a real face. So emoticons seem face-like, but are not processed identically to faces.
Why was an emoji Oxford’s Word of the Year 2015?
Oxford Dictionaries chose the Face with Tears of Joy emoji, 😂, because of its huge use and cultural impact. It was the first emoji to win. Oxford said that traditional alphabet scripts were struggling to meet the needs of 21st-century communication.
Does using emojis make you look less professional?
It depends. In a 2018 study of work-related first impressions, smileys did not increase perceived warmth and lowered perceived competence. A 2023 replication confirmed the competence effect but found higher warmth. For formal messages to people you do not know, it is safest to use words.
