Emoji — From Japanese Mobile Phones to Unicode and the World

Emoji were born on Japanese mobile phones in the late 1990s. The best-known early set is NTT DoCoMo's, which appeared with i-mode in 1999: 176 characters on a 12-by-12 pixel grid, developed under Shigetaka Kurita and taken into the collection of the Museum of Modern Art in New York in 2016. But DoCoMo's were not the first emoji on a phone. The Pioneer-made SkyWalker DP-211SW, released by J-Phone in November 1997, already had 90 of them.

The carriers' separate, incompatible sets were brought into the international character standard by Unicode 6.0 in October 2010, and from there, through the iPhone and Android, they became a worldwide vocabulary. The English word "emoji" is simply the Japanese 絵文字: e (picture) plus moji (written character).

Timeline: from a pager heart to Unicode

J-Phone's 90, before DoCoMo

For years, emoji were said to have started with DoCoMo's 1999 set. The correction came from Emojipedia, the emoji reference site, in an article of 8 March 2019. Using a press release and a manual, it showed that J-Phone — later Vodafone's Japanese arm, and now SoftBank — had released the SkyWalker DP-211SW with 90 emoji in November 1997. Today's U+1F4A9 PILE OF POO, the article notes, originated in that 1997 set.

There was a catch, as Emojipedia also notes: the DP-211SW did not sell well, and its emoji were not supported on any other J-Phone model. Unless the person you were writing to had the same phone, they were close to pointless.

Kurita, who led the development of DoCoMo's emoji, said much the same in a post of 3 January 2019, as reported by Emojipedia: the first emoji on Japanese mobile devices were on pagers, and on phones the first were not DoCoMo's but J-Phone's Pioneer DP-211SW.

Meanwhile the Unicode Consortium's technical standard UTS #51 still says that emoji became available on Japanese mobile phones in 1999. Sources do not agree on what counts as "first," and the answer depends on whether you mean a pager symbol, a single phone or a set that people could actually exchange.

How DoCoMo's emoji were made

What made DoCoMo's set spread, Emojipedia points out, was that far more phones supported it, and every manufacturer shipped the same images.

According to a December 2016 article in Keitai Watch, Kurita had the idea in March 1998, prompted by how a heart symbol had softened text messages on pagers. i-mode was originally due to launch in December 1998, which left about a month for the work. He had intended to let the handset makers draw the pictures, but they did not know what to do, so Kurita drew them himself on graph paper, 12 by 12 dots at a time — about 200 in all, mixing symbols for i-mode services such as banks, convenience stores and weather with hearts and faces meant for mail.

Looking back in 2025, Kurita told followers (as reported by ITmedia) that the concept took a few months and the drawing less than a month, and that he chose each emoji's colour himself from six available.

From incompatible sets to one standard

UTS #51 sets out how standardisation happened. Emoji were popular in Japan, but each carrier developed a different, partly overlapping set, with incompatible text-encoding extensions. Characters that another carrier's phone could not display turned into 〓 (U+3013 GETA MARK), and they were easily corrupted or dropped. A proposal in 2000 to encode DoCoMo's emoji went nowhere, because it was unclear whether they would catch on and the carriers did not support it.

The push came from outside Japan, from companies whose email services had to exchange messages with Japanese phones. In 2006 Google began converting Japanese emoji to Unicode private-use codes, building internal mapping tables in 2007. That year a proposal to extend Unicode's scope to emoji was approved, and in January 2009 engineers from Google and Apple jointly submitted a proposal to encode them.

Unicode 6.0, released on 11 October 2010, added 2,088 characters, including more than 1,000 symbols, chief among them emoji. Of the 722 characters in the union of the Japanese carriers' emoji, 114 were already in Unicode 5.2; the remaining 608 arrived in 6.0. A data file released with it, EmojiSources.txt, maps each Unicode character to the Shift-JIS codes used by DoCoMo, KDDI and SoftBank.

Going global

Apple added an emoji font with 471 glyphs in iPhone OS 2.2 in November 2008. At first it was limited to Japanese iPhones, and some designs are said to resemble those of SoftBank, the iPhone's carrier in Japan. With iOS 5 in 2011, users anywhere could pick emoji from the keyboard.

In November 2015 Oxford Dictionaries chose the Face with Tears of Joy emoji as its word of the year — the first time an emoji had been selected.

On 26 October 2016 MoMA announced its acquisition of DoCoMo's original 176 emoji, a gift from the company, and put them on show in its lobby from that December.

The end of DoCoMo's emoji

On 21 May 2025 NTT DoCoMo announced that, in view of how emoji were now used on phones, it would stop providing its own DoCoMo emoji on models released from late June 2025 (July for Samsung Galaxy). Those phones use Google's Noto Color Emoji or the manufacturer's set, and old DoCoMo emoji are converted. The notice was later updated to say that existing phones would also lose DoCoMo emoji after upgrading to Android 17.

i-mode itself closed on 31 March 2026; that story is on the i-mode page. DoCoMo's own pictures are disappearing from phones, but the emoji they helped start live on in Unicode.

For the wider story, see the mobile timeline and the internet and web timeline; for the companies, Google and Apple.