1. A Phone Company Needed Tiny Pictures for Tiny Screens
In the late 1990s, NTT DoCoMo ran "i-mode," one of the first services that let people use the internet on a mobile phone. Messages were short (the Associated Press reported a limit of 250 characters), and the screens were tiny[1]. A young engineer on the team, Shigetaka Kurita, wanted a shorthand that could show things such as a weather forecast at a glance.
In 1999 he drew a set of 176 small pictures. He reportedly built them on a grid only 12 pixels wide and 12 pixels tall[2]. The AP wrote that he collected everyday images such as public signs, weather symbols, the zodiac, and comic-book pictures like a light bulb[1]. Japanese comics often draw feelings as little symbols, for example a drop of sweat for nervousness[2], and Kurita is said to have drawn on that tradition.
The set was a hit. In 2016 the Museum of Modern Art in New York added the original 176 to its collection[3]. But the success hid a problem. Other phone companies in Japan soon made emoji of their own, and the three companies did not store them in the same way.
2. Why an Emoji Could Turn into Junk on a Rival Company's Phone
Here is the surprise: a phone never sends a picture. It sends numbers. Every computer keeps a table that says which number means which character, such as "65 means capital A." That table is called a character encoding, and the receiving phone looks the number up in its own table to decide what to draw.
The Unicode Consortium's technical report on emoji says that each of the Japanese mobile carriers developed different, partly overlapping sets, and that each phone maker used its own extension of the text encoding, so the extensions were incompatible with one another[4]. A 2009 proposal names the three main carriers: NTT DoCoMo, KDDI, and SoftBank[5]. In effect, each company had written its own dictionary, and a number that meant "sun" in one dictionary could mean something else, or nothing, in another.
Emoji were not a toy at that point. The proposal quotes a June 2007 survey of 13,000 Japanese phone users, 80 percent of whom were aged 30 or older: 78 percent of them often or sometimes used emoji in email[5]. Many people used them every day, so the broken pictures were a real annoyance. The fix would need a numbering system that did not belong to any one company.
3. Unicode Gives Every Character One Number for All Computers
Emoji were not the first characters to get scrambled. Before Unicode, there were hundreds of encoding systems, and the same number could stand for different characters in different systems, so text moved between computers could be damaged[6]. Unicode's answer is to provide, in its own words, "a unique number for every character, no matter what the platform, no matter what the program, no matter what the language"[6].
The first published version of the Unicode standard appeared in October 1991[7]. Each character gets a permanent number called a code point, written like U+1F600 (today, the grinning face). The reasoning is simple: if every phone, company, and country agrees on the table, nobody needs to translate between tables.
Unicode did have a place for symbols nobody had assigned yet, called private-use characters, which anyone may define for their own purposes. Before emoji were standardized, the only Unicode version of Japanese emoji was through such private-use characters, and the 2009 proposal warned that these are "subject to misinterpretation and data corruption"[5]. Emoji needed official numbers.
4. Google and Apple Asked Unicode to Add Japan's Emoji
According to the Unicode technical report, Google began converting Japanese emoji to Unicode private-use codes in 2006, and in August 2007 the Unicode technical committee agreed to support encoding emoji, following principles drawn up by a subcommittee[4].
In March 2009, a proposal written by engineers at Google and Apple asked Unicode to add 674 new characters so that every emoji used by the three Japanese carriers could be represented. Another 114 of the carriers' symbols were already in Unicode[5]. The proposal's central argument was not that emoji are cute. It was interoperability: emoji data passes through email systems, search engines, databases, and publishing software, and all of these must handle it without loss[5].
It worked. Unicode announced version 6.0 on October 12, 2010, with 2,088 new characters, including "hundreds of symbols for mobile phones"[8]. The technical report puts the combined list of Japanese carrier emoji at 722 characters, of which 114 were already in Unicode and the remaining 608 were added in version 6.0[4]. (The proposal's 674 and the report's 608 differ; see the note in the sources.)
5. Phones Outside Japan Wanted Emoji Before They Were Official
While the standard was still being written, people outside Japan were already curious. On November 21, 2008, Apple released a free update for iPhones in Japan that included its first emoji font and keyboard[9]. Elsewhere there was no emoji keyboard to switch on.
Some apps added hidden "Easter eggs" that unlocked the keyboard for people outside Japan, although Apple neither supported nor endorsed that[9]. Then iOS 5.0 in 2011 let anyone in the world turn the keyboard on with a quick trip to the settings app[9]. Because emoji were now ordinary Unicode characters, a message with one could be read by phones that had never heard of DoCoMo.
6. Why the Same Emoji Looks Different on Different Phones
Unicode settled the numbers, but not the artwork. The technical report explains that a character's name does not determine its appearance, and that the shape can vary significantly between vendors[4]. Even the charts published by Unicode use only black-and-white sample images; the colored pictures that companies draw are not part of the standard[4].
So the arrangement is like sheet music. The number says which note to play, and each phone company plays it in its own style. That is why the face your friend sends may look slightly different on your screen, and why the number, not the picture, is what really travels.
7. How a New Emoji Gets Approved Today
Adding an emoji is now a formal process. Anyone can submit a proposal, and Unicode says that all submissions are reviewed but "a very small percentage advance for encoding"[10]. Proposers must include evidence that people really would use the emoji, such as search data, and the reviewers favor ideas that fill gaps and avoid ones that are too specific (their example is individual dog breeds)[10]. Logos, brands, and specific people are automatically declined[10].
The path that started with one engineer and a 12-by-12 grid now runs through an international standards committee. The first 176 pictures, drawn in about a month[1], became a shared part of how the world writes.
How to Check the Emoji Story Yourself
- Compare one emoji on two devices. With an adult's permission, send the same emoji from your phone to another phone or a computer, ideally from a different company. The number is the same, but is the drawing? If both screens use the same company's artwork, they may look identical. Try a few and see which ones differ most.
- Look at the original 176. Search for images of Kurita's original set, which the Museum of Modern Art collected. Count how many you recognize, and compare their blocky 12-by-12 look with the emoji on your keyboard.
- Read the argument itself. The 2009 proposal is a short, free PDF on unicode.org. Its "Rationale" section explains in plain terms why a standards body should care about phone pictures.