Emoji history is often compressed into one inventor, one company, and one date. The real story is more interesting. It includes punctuation jokes on university networks, Japanese text faces, pager and phone interfaces, incompatible carrier encodings, museum design, and a standards project whose main goal was not cultural fame but reliable interchange.
Before emoji, punctuation learned to smile
People had arranged type into faces long before the internet, but networked communication gave the practice a durable social role.
On 19 September 1982, Carnegie Mellon computer scientist Scott Fahlman proposed :-) and :-( on an online bulletin board to distinguish jokes from serious posts. The original thread was later recovered from backup tapes and published by Carnegie Mellon.
These Western-style emoticons are usually read by tilting the head. They work within a small ASCII repertoire, which made them easy to type and transmit on early systems.
Kaomoji turned the face upright
Japanese kaomoji—literally “face characters”—are normally read without rotating the head: (^_^), (T_T), or m(_ _)m.
Their design emphasizes the eyes and can use a much larger character repertoire: Japanese punctuation, kana, Greek and Cyrillic letters, mathematical signs, and fullwidth forms. The sides may act as the outline of a face, while characters outside the face become arms, tears, ears, or gestures.
Kaomoji are not merely interchangeable drawings of emotion. Research on Japanese online interaction shows that they can help writers balance formality and informality and perform relationship work. A bowing kaomoji, for example, can soften thanks or an apology in a way that depends on conversational context.
Mobile interfaces needed compact pictures
During the 1990s, Japanese pager and mobile services experimented with small pictographs. The best-known early collection is Shigetaka Kurita’s set of 176 designs for NTT DOCOMO, released in 1999 and drawn on a 12-by-12-pixel grid.
MoMA’s collection notes that Kurita drew from manga, public pictograms, Zapf Dingbats, and existing emoticons. The set included weather, transport, hearts, devices, and expressions—an interface vocabulary, not simply a gallery of smiley faces.
Calling this the single birth of emoji hides the surrounding ecosystem. DOCOMO, KDDI, and SoftBank developed overlapping but incompatible carrier sets. Unicode’s own source charts preserve those multiple lineages, and some symbols now treated as emoji came from still older standards such as Zapf Dingbats, Wingdings, or Japanese broadcast symbols.
The compatibility problem changed everything
Carrier emoji were treated like text by users, but companies assigned them private codes. A symbol sent from one network could turn into a different symbol or disappear on another.
The 2009 proposal to encode emoji in the universal character set described 674 additions needed to represent the core sets of Japan’s three major carriers. Its central argument was interoperability: email, search engines, databases, and phones needed a stable representation that could survive conversion without data loss.
Unicode 6.0, released in 2010, incorporated the major set. That did not give every vendor the same artwork. It gave systems shared character identities and sequences, so the intended item could travel between them.
Emoji became text without becoming a language
Modern emoji can behave like punctuation, gesture, tone, illustration, reaction, or a substitute for a word. Their position in a message changes what they do. A face at the end may soften a request; an object may clarify a topic; a repeated symbol can become emphasis or rhythm.
But emoji do not form a single universal language with fixed grammar. Meanings vary with platform artwork, age, culture, community, and conversational history. The eggplant, folded hands, or slightly smiling face can acquire meanings far beyond their official names.
Standardization makes exchange possible. It does not standardize every interpretation.
New emoji are selected, not simply uploaded
Unicode now maintains a formal proposal process. A candidate must be visually distinctive, broadly useful, likely to be used frequently, and not an overly specific or transient variation of something already represented. Logos, specific people, and exact copyrighted images are not suitable characters.
Once encoded, a character cannot be casually removed without breaking existing text. That permanence is one reason the process asks for evidence rather than treating an online petition as sufficient.
Three systems still live side by side
Emoticons, kaomoji, and emoji did not replace one another in a clean sequence.
:-)remains quick, typographic, and deliberately plain.(^_^)offers expressive construction from text characters.😊offers a standardized identity with platform-specific artwork.
Stickers, GIFs, custom server emoji, and reaction images add further branches. Digital expression evolves by accumulation: each medium keeps useful habits from the previous one while inventing new conventions.
Emoji’s decisive transformation was not from punctuation into pictures. It was from local, incompatible pictographs into interoperable text.
Explore the construction vocabulary in the kaomoji library and compare it with the themed emoji sets. The difference between “built from visible punctuation” and “encoded as an emoji sequence” becomes immediately apparent.