How Did Emoji Start? A Japanese Pager Feature That Went Global

The first widely adopted set was created by Shigetaka Kurita for NTT DoCoMo's i-mode service, which offered internet access on mobile phones years before that was common elsewhere, with messages limited to a small number of characters. A small picture could carry meaning that would otherwise consume the whole allowance, and could restore the tone that short text strips out, which is the function they still serve. Competing Japanese carriers produced their own incompatible sets, so a picture sent from one network arrived as something else or as nothing on another, a problem that eventually forced standardisation. The word emoji is Japanese, combining the characters for picture and letter, and its resemblance to emotion and emoticon is coincidental. Emoticons are a separate and earlier thing, being faces built from punctuation, with the sideways smiley usually credited to Scott Fahlman at Carnegie Mellon in 1982, proposed to mark jokes on a message board.

The decisive step was incorporation into Unicode, the standard that assigns a number to every character in every writing system so that text can move between devices. Google and Apple pushed for emoji to be included so that Japanese users could exchange messages with the rest of the world, and a large set entered the standard in 2010, after which any conforming system could display them. That decision had consequences that are still being worked through:

• •The standard defines a code point and a name, not an appearance, so each platform draws its own version, which is why the same emoji can look meaningfully different on two phones and has caused documented misunderstandings
• •Adding a new emoji requires a written proposal to the Unicode Consortium demonstrating expected frequency of use, distinctiveness and lack of overlap, reviewed by a subcommittee and voted on, with most proposals rejected
• •Skin tone variation was added in 2015 using modifier characters that combine with a base character
• •Sequences joined by an invisible character allow composite emoji, which is how families, professions and flags are built, and why some combinations render as separate images on systems that do not support the sequence
• •Flags are encoded as pairs of regional indicator letters rather than as pictures, which lets the standard avoid deciding what counts as a country
• •Once encoded, an emoji is effectively permanent, since removing it would break existing text, so the standard can add and cannot subtract

Linguists have examined whether emoji constitute a language and generally conclude they do not, since they have no grammar, no productive syntax and no way to express most propositions, and attempts to translate novels into them produce something readers cannot recover the original from. What they do resemble closely is gesture: the linguist Gretchen McCulloch argues they are the written equivalent of the gestures and facial expressions that accompany speech, supplying tone, emphasis, irony and emotional framing that plain text lacks, which is exactly the job Kurita designed them for. That account explains their distribution in real messages, where they cluster at the ends of utterances and modify tone rather than carrying content, and why a message with and without one can mean quite different things. Usage also stratifies socially and generationally, with particular emoji acquiring ironic or sarcastic meanings among younger users that older users miss entirely, which is ordinary language change happening unusually fast and in public.

Encoding images in a text standard creates difficulties that encoding letters does not. Representation has been contested repeatedly, over skin tone, over gender for professions, over which foods and religious symbols are included and over disability, and each resolution requires a committee to make what is unavoidably a cultural judgement. Political questions arrive through flags and symbols. Ambiguity causes real problems, since emoji have been submitted as evidence in court cases where their meaning was disputed, and research has documented substantial disagreement about the sentiment of common emoji even between people using the same platform. Accessibility depends on the names assigned, since a screen reader announces the official name, which may not match how the sender meant it. And the committee's workload grows continuously, because every group with a reason to want representation has a legitimate case, while the standard's designers never intended it to become the registry of what exists in the world.

Emoji were created for a Japanese mobile service in 1999 to restore tone within tight message limits, and incompatible carrier sets forced standardisation. Unicode encoded them from 2010, defining a name and number rather than an appearance, which is why the same emoji looks different across platforms. New ones require a formal proposal and vote and cannot later be removed. Linguists treat them as the written equivalent of gesture rather than as a language, since they carry tone rather than content.