The History of Emoji: From Japanese Pagers to Universal Language

The word "emoji" means "picture character" in Japanese (絵文字: 絵 e, picture + 文字 moji, character). It does not derive from the English word "emotion" — that's a persistent myth, however etymologically tempting. The confusion is understandable given how thoroughly emoji became associated with emotional expression, but the origins of emoji are more mundane, more commercial, and ultimately more interesting than a story about feelings.

They began with weather forecasts on a pager.

The First Emoji: NTT DoCoMo, 1999

In 1999, NTT DoCoMo — Japan's largest mobile carrier — launched i-mode, an early mobile internet service that allowed users to access email and simple web content on their mobile phones. The service designer was Shigetaka Kurita, a 25-year-old employee in DoCoMo's multimedia division.

Kurita faced a practical problem. i-mode messages were limited to 250 characters, and DoCoMo's weather forecasts needed to communicate conditions quickly. A text string like "cloudy with chance of rain" is 29 characters. A cloud symbol with a rain indicator is one character. The compression ratio was obvious.

But Kurita's ambition was larger than weather symbols. He wanted i-mode to feel warm and expressive — more like a phone call than a telegram. He looked at inspirations ranging from manga facial expressions (the large eyes and exaggerated emotions of Japanese comics) to international road signs (which communicate at a glance without words) to the ASCII emoticons already circulating in text messages.

Working within a 12×12 pixel grid — the constraint imposed by the low-resolution displays of 1999 mobile phones — Kurita designed 176 emoji. The original set included:

• Weather symbols: sun, cloud, rain, snow, thunder
• Technology: phone, envelope, TV, computer
• Nature: flower, cherry blossom, musical notes
• Food: beer, rice ball, hot springs
• Gestures and faces: happy face, sad face, angry face
• Symbols: hearts, stars, arrows, the Japanese currency yen sign

The 176 designs, rendered in 12×12 pixels with only two colors, established conventions that persist in emoji to this day. The 😊 smiley, the ❤️ heart, the ⭐ star — these were among Kurita's originals.

The Carrier Wars: Before Standardization

DoCoMo's emoji were a proprietary system with a proprietary encoding. When SoftBank (then called J-Phone) and KDDI (then called DDI/au) launched their own competing mobile internet services, they created their own incompatible emoji sets.

This created a situation that mirrored the pre-Unicode chaos in text encoding, but compressed into a few years. The three major Japanese carriers each had their own emoji, with overlapping but not identical character sets, and incompatible binary encodings. Sending an email from a DoCoMo phone to a SoftBank phone would either display the wrong emoji or, more commonly, display nothing at all — the character would be silently dropped.

The SoftBank Legacy

SoftBank's emoji, first deployed in 1997 (predating DoCoMo's i-mode), are historically significant for a reason that only became apparent a decade later. When Apple launched the iPhone in Japan in 2008, they partnered with SoftBank as the exclusive carrier. The emoji keyboard on the Japanese iPhone was based on SoftBank's character set.

This meant that when Google and Apple lobbied the Unicode Consortium to standardize emoji in 2007–2009, their proposed character lists were derived partly from the SoftBank encoding. The SoftBank encoding assigned code points in a range that overlapped with Unicode's Private Use Area — and some of those assignments influenced which emoji got standardized and in what order.

The ghost of the SoftBank carrier wars can be seen in the current Unicode emoji list: some emoji that seem redundant or oddly specific (🍡 sweet dango, 🎌 crossed flags, the specific arrangement of Japanese holiday symbols) reflect the cultural specifics of the Japanese market that first demanded them.

Google, Apple, and the Unicode Push (2007–2010)

By 2007, emoji were a major user expectation in Japan but essentially unknown in the United States and Europe. Google was launching Gmail in Japan and discovered that Japanese users expected to be able to use emoji in email. The problem: Gmail used Unicode text, and emoji didn't exist in Unicode.

Google engineers Mark Davis (who had been involved with Unicode since its founding), Kat Haben, and Markus Scherer filed a formal proposal to the Unicode Consortium in 2007 to add emoji to Unicode. The proposal was titled "Emoji Symbols: Background Data" and documented the emoji from all three Japanese carriers, cross-referenced against each other.

Apple joined the effort in 2009, led by Peter Edberg. The combination of Google and Apple lobbying the Unicode Consortium — two of the most powerful companies in mobile computing — was decisive. The Consortium moved quickly.

Unicode 6.0 (October 2010) was the breakthrough: it added 722 emoji symbols to the Unicode standard, covering the core emoji from the Japanese carriers plus additional symbols. For the first time, emoji had official, stable, universal code points. 😀 was U+1F600. 💩 was U+1F4A9. ❤️ was U+2764.

The encoding decisions made in Unicode 6.0 had some quirks that reflect the rushed timeline. Several emoji were encoded in unexpected places — the original 12 emoji that had been in Unicode as "miscellaneous symbols" since much earlier versions (☀️ U+2600, ☁️ U+2601, ☂️ U+2602, etc.) weren't moved; they stayed where they were. This is why the Unicode emoji set is not a single contiguous block but scattered across multiple ranges.

The Skin Tone Problem and Its Solution

When emoji were standardized in 2010, all human figures had a single "default" appearance — effectively a cartoonish yellow, chosen to be ethnically neutral in the same way a smiley face emoticon is neutral. This was immediately criticized as inadequate, particularly as emoji moved from a Japanese technical context into global mainstream use.

The Unicode Consortium's solution, introduced in Unicode 8.0 (2015), was elegant in its simplicity: skin tone modifier characters. Five code points — U+1F3FB through U+1F3FF — were defined as Fitzpatrick scale modifiers, corresponding to the six dermatological skin tone classifications (types 1 and 2 were merged into a single modifier). When one of these modifiers follows a human emoji, it overrides the default yellow with the specified skin tone.

The modifier approach had practical advantages: it didn't require assigning five new code points for every human emoji, and it was backward compatible — systems that didn't understand the modifier would just display the default yellow emoji. The same principle was later extended to multi-person emoji where each person can have an independent skin tone.

ZWJ Sequences: Building Complex Emoji

The most technically sophisticated development in emoji encoding is the Zero Width Joiner (ZWJ) sequence. ZWJ (U+200D) is a Unicode control character originally designed for use in Arabic and other scripts to control joining behavior between letters. Emoji repurposed it to create "compound emoji" — sequences of multiple characters that platforms are expected to render as a single image.

The family emoji is the canonical example. 👨‍👩‍👧‍👦 is not a single character — it's a sequence of seven characters:

U+1F468 (man) + U+200D + U+1F469 (woman) + U+200D + U+1F467 (girl) + U+200D + U+1F466 (boy)

The ZWJ between each emoji tells the rendering platform: "treat these adjacent emoji as a single unit." Platforms that support ZWJ sequences render one combined image. Platforms that don't support a particular sequence display the individual emoji separately — a graceful fallback.

This mechanism was extended to: - Profession emoji: 👩‍💻 (woman + ZWJ + laptop) creates "woman technologist" - Couple emoji with skin tones: two person emoji with skin tone modifiers, joined by ZWJ and heart - Rainbow flag: 🏳️‍🌈 is white flag + ZWJ + rainbow - Transgender flag: 🏳️‍⚧️ is white flag + ZWJ + transgender symbol

The flexibility of ZWJ sequences means emoji can be extended almost indefinitely without adding new Unicode code points for every variant. However, it also means that the number of "emoji" is not a fixed number — it depends on how you count sequences, and new sequences can be introduced without a new Unicode version.

Emoji Governance Today

The Unicode Consortium formed a dedicated Emoji Subcommittee in 2015 to manage the growing complexity of emoji standardization. The subcommittee reviews proposals, which can be submitted by anyone but must follow detailed guidelines.

New emoji are added once a year (typically in September with each new Unicode version). The selection criteria are deliberately conservative: - High expected usage: Will many people use this frequently? - Image versatility: Can it work at small sizes and in different contexts? - Distinctiveness: Is it clearly different from existing emoji? - Completeness: Does adding it complete a logical set?

These criteria famously excluded a "dumpling" emoji for years because the subcommittee wanted to wait until a single image could represent the concept across Chinese, Japanese, and other Asian cuisines. 🥟 (dumpling/potsticker/gyoza) was finally added in Unicode 10.0 (2017).

The annual process is not without controversy. The selection of which objects and concepts deserve emoji immortality reflects whose proposals make it through the subcommittee, and critics have pointed to an underrepresentation of non-Western, non-Japanese cultural objects in the emoji canon.

Emoji 16.0 and Current State

Emoji 16.0 (accompanying Unicode 16.0, released 2024) added a modest set of new emoji, continuing the trend toward smaller annual additions as the most common use cases have been covered. Current focus areas include:

• Additional gender-neutral and gender-inclusive variants
• Further nationality flag support
• Profession and activity variants to improve representation
• Occasional new objects and nature symbols

The total number of emoji in Unicode 16.0 is over 3,600 when counting individual code points, or approximately 1,872 when counting recommended sequences (RGI — "Recommended for General Interchange" emoji). The discrepancy reflects the difference between the raw code points and the combinations that platforms are expected to support.

Platform Rendering Differences

A persistent source of confusion is that the Unicode standard specifies which emoji exist and their semantic meaning, but not what they look like. Apple, Google, Microsoft, Samsung, Twitter/X, and other platforms each draw their own emoji designs. The result is that 😂 on iOS looks noticeably different from 😂 on Android, which looks different from 😂 on Windows — same code point, different glyph.

This has led to documented miscommunication: researchers studying emoji use have found cases where the same emoji conveys different emotional valence on different platforms. The "grinning face with smiling eyes" 😁 was found to read as slightly negative or uncomfortable on some platforms due to the specific illustration style used.

Search and browse all emoji with their Unicode code points in our Unicode Lookup tool.

A Timeline of Emoji History

Next in Series: Emoji are the newest members of a much older tradition of embedding symbolic meaning in written text. Discover how punctuation marks, from the paragraph sign to the ampersand, evolved over two millennia in The History of Typography Marks: From Gutenberg to Unicode.