Emoji: Unicode Sets the Name and Number, Not the Picture

U+1F52B keeps the Unicode name PISTOL, while Twemoji v17.0.3, Fluent Emoji (Flat style) and Noto Color Emoji 2.057 each draw it as a toy gun.

Unicode named U+1F52B PISTOL in 2010, yet three open emoji sets now draw it as a brightly colored toy gun. The name can never change, and the color picture was never Unicode's to fix. Unicode gives each emoji character a number, called a code point, and a character name. The Unicode CLDR project adds a short name and search keywords, and whoever makes the font or image set draws the picture.

The Unicode Standard itself contains only black-and-white images. Each vendor draws its own color version, so the same message can look different on two screens.

Three open emoji sets, side by side

Twemoji, Fluent Emoji and Noto Color Emoji publish their drawings under open licenses, and their pictures differ in shape, color and detail. The table lists who makes each set, its license and how it draws U+1F52B today.

Emoji set	Made by	License for the drawings	Version on this page	U+1F52B drawn as	Source
Twemoji	Twitter, Inc and other contributors; newer releases come from the jdecked fork	CC BY 4.0	v17.0.3, released 2026-06-01	green water gun	LICENSE-GRAPHICS and tag v17.0.3, jdecked/twemoji
Fluent Emoji	Microsoft	MIT License	repository as of 2026-08-24, Flat style	lime-green and red toy gun	LICENSE and file history, microsoft/fluentui-emoji
Noto Color Emoji	Google	SIL Open Font License 1.1 (the font)	2.057	orange and yellow water gun	font file name table; googlefonts/noto-emoji

GitHub labels the Twemoji repository MIT, but that license covers the code. The drawings carry CC BY 4.0 in a separate file, LICENSE-GRAPHICS.

Six code points in three open sets. UTS #51 uses the four foods to show how far drawings can vary; the last two were among the 22 emoji in the 2016 Minnesota study.

UTS #51, the Unicode technical standard for emoji, uses lollipop, custard, honey pot and shortcake as examples that "the shape of the character can vary significantly." It also asks designers to keep the "core" shape: U+1F36F HONEY POT should still read as a pot of honey, not as a sugar cube.

The differences can change how a message reads. In a study published in 2016, Hannah Miller and colleagues at the University of Minnesota collected ratings from 304 people in the United States. The emoji were 22 popular faces and people, each as drawn by Apple, Google, Microsoft, Samsung and LG. When two people rated the very same drawing, they disagreed about whether it was positive, neutral or negative in 25% of pairs.

Across platforms the gap grew. On a scale from −5 to 5, two ratings of the same drawing differed by 1.86 points on average, and ratings of different platforms' drawings by 2.03. For U+1F601, beaming face with smiling eyes, ratings of Apple's drawing sat more than 4.7 points away, on average, from those of each of the other four. The grid above shows today's open drawings, not the ones the participants saw.

That fits the position this site takes on Road Signs: a symbol's meaning is learned, not read off the picture.

What does Unicode decide about an emoji?

Think of an emoji as a message that is encoded once and decoded on every screen. Unicode writes the code, and each font supplies the picture. In literal terms, Unicode fixes two things for good: under its stability policies, an encoded character "will not be moved or removed," and its name "will not be changed." U+1F483, for example, has had its code point since Unicode 6.0 in 2010 and its name, DANCER, ever since.

The short name comes from CLDR, the Unicode Common Locale Data Repository, which also supplies keywords for emoji search in many languages. It is the name a screen reader can speak aloud, and UTS #51 says it "may change over time." The standard's own example is U+1F483, which should be drawn to match its CLDR short name, woman dancing. A separate U+1F57A MAN DANCING arrived later, in Unicode 9.0.

Four layers of one emoji. Unicode fixes the code point and the name DANCER, the CLDR short name woman dancing can change, and the picture comes from a font.

Unicode leaves the color picture to the font. UTS #51 states that "the formal Unicode name of a Unicode character does not determine its appearance." Even the choice between a color emoji and a plain symbol can vary: U+262E PEACE SYMBOL, in Unicode 1.1 (1993) and an emoji since Emoji 1.0 (2015), is set to text style by default. Adding U+FE0F after it asks for the emoji style and U+FE0E for the text style, as Peace Symbol notes.

Emoji also have no script of their own. All 1,447 code points with the Emoji property carry the Script value Common in the Unicode 18.0 data, by this page's count, so none of them is among the 175 scripts on Every Writing System on One Page.

Why did the pistol emoji become a water gun?

Nobody changed the character. U+1F52B is still named PISTOL, and its code point has not moved since 2010; only the drawings changed. The open repositories date each switch to the day. Noto Emoji replaced its revolver with a water gun on March 8, 2018 (UTC), and Twemoji did the same in version 2.6.0, committed on April 11, 2018.

U+1F52B in three open sets, with the UTC commit date of each change shown. The name stayed PISTOL, and the CLDR short name became water pistol in CLDR 38 (2020).

Closed sets such as Apple's have no public file history, so their dates here come from Wikipedia's article on the emoji. It says Apple announced in August 2016 that iOS 10 would draw a water pistol. By mid-2019, it says, the Apple, Google, Samsung, Twitter and Facebook sets, among others, drew a squirt gun. The article sets these changes against the US debate over gun violence; none of the commit messages in the open repositories gives a reason.

Generated illustration

A toy water gun. Noto Emoji and Twemoji switched U+1F52B from a revolver to a water gun in 2018, but the character is still named PISTOL.

CLDR's name changed later. The English short name for U+1F52B was "pistol" in CLDR 37 and "water pistol" from CLDR 38, released on October 28, 2020. UTS #51 notes that a CLDR short name can come to reflect the preferred drawing better than the formal name does.

Drawings can also change back. Wikipedia reports that X, formerly Twitter, returned to a handgun in July 2024. In the twitter/twemoji repository, a commit dated January 10, 2026 (UTC), draws a dark handgun. As of October 2026 that commit is in no tagged release, and the jdecked fork still ships the 2018 water gun in v17.0.3.

How do several emoji combine into one?

The black cat emoji is not a character of its own but a sequence of three: U+1F408 cat, U+200D ZERO WIDTH JOINER and U+2B1B black large square. A font that knows the sequence draws one black cat, and the joiner, usually shortened to ZWJ, stays invisible.

U+1F408 cat, the invisible joiner U+200D and U+2B1B black large square make the black cat emoji. A font without the sequence shows the two parts side by side.

When a system has no single picture for a ZWJ sequence, UTS #51 says "the ZWJ characters are ignored and a fallback sequence of separate emoji is displayed." The reader then sees a cat next to a black square. UTS #51 asks that a ZWJ sequence be defined only where that split form would also make sense to the person reading it.

Unicode 18.0 lists 1,614 recommended ZWJ sequences, by this page's count. Of those, 314 contain U+2640 FEMALE SIGN and 314 contain U+2642 MALE SIGN, each giving a female or male form. Male and Female Symbols traces where the two signs came from.

Five skin tones, five code points

Skin tone works the same way, with a modifier instead of a joiner. Unicode 8.0 added five modifier characters, U+1F3FB to U+1F3FF. UTS #51 says they are "based on the six tones of the Fitzpatrick scale," a standard from dermatology, so types 1 and 2 share the first modifier. Placed right after a person or body-part emoji such as U+1F44B waving hand, a modifier makes a pair that a font can draw as one tinted picture.

The five skin-tone modifiers alone and after U+1F44B waving hand, as Noto Color Emoji 2.057 draws them. Fitzpatrick types 1 and 2 share the first modifier.

On its own, a modifier shows as a color swatch, so a reader whose font lacks the combined picture can still tell which tone was meant. UTS #51 adds that "the exact shades may vary between implementations."

By this page's count, 136 characters in Unicode 18.0 can take a modifier. The recommended list pairs 135 of them with all five tones, for 675 emoji-plus-modifier pairs. Only U+1F46A FAMILY has no skin-tone form on the list. Another 1,365 ZWJ sequences carry a tone.

How many emoji are there?

UTS #51 warns that the question "does not have a simple answer." By this page's count of the Unicode 18.0 data files, there are 3,963 fully-qualified emoji. ZWJ sequences make up the largest group, 1,614, ahead of 1,400 single emoji, 675 emoji-plus-modifier pairs, 259 country and region flags, 12 keycaps and 3 subdivision flags. That total matches the fully-qualified lines of emoji-test.txt, the file UTS #51 calls "the best definition of the full set."

The 3,963 fully-qualified emoji of Unicode 18.0 by kind, by this page's count of the Unicode data files. Leaving out every skin-tone variant leaves 1,923.

Other counts are just as valid when their rule is stated. By this page's count of the same files:

• 3,972 entries in the recommended lists, if the 9 components (5 skin-tone modifiers and 4 hair components) are included;
• 1,923 emoji, if every skin-tone variant is left out;
• 1,447 code points with the Emoji property, a total that includes the digits 0 to 9, the # and * signs and the 26 regional indicator letters that are used in pairs for flags.

All of these figures are for Unicode 18.0, released in September 2026.

Picture signs that spell consonants are the subject of Egyptian Hieroglyphs.

Sources

1. 1

   The Unicode Consortium, Unicode emoji data files 18.0 (emoji-test.txt, emoji-sequences.txt, emoji-zwj-sequences.txt, emoji-data.txt, emoji-variation-sequences.txt) and Unicode Character Database 18.0.0 (UnicodeData.txt, DerivedAge.txt, Scripts.txt), 2026.

   https://www.unicode.org/Public/18.0.0/emoji/ (full data files)

2. 2

   Mark Davis and Ned Holbrook, eds., Unicode Technical Standard #51: Unicode Emoji, Version 18.0, revision 31 (The Unicode Consortium, 2026-08-28).

   https://www.unicode.org/reports/tr51/tr51-31.html (full text)

3. 3

   The Unicode Consortium, Unicode Character Encoding Stability Policies: Encoding Stability and Name Stability (applicable from Unicode 2.0).

   https://www.unicode.org/policies/stability_policy.html (Encoding Stability and Name Stability sections)

4. 4

   Unicode CLDR, release 48.2, common/annotations/en.xml (English emoji short names and keywords), 2026-03.

   https://github.com/unicode-org/cldr/releases/tag/release-48-2 (English annotations file; German and French files for the cross-language names)

5. 5

   Unicode CLDR, releases 37 and 38, common/annotations/en.xml; release 38 published 2020-10-28.

   https://github.com/unicode-org/cldr/releases/tag/release-38 (English annotations file in both releases; release record)

6. 6

   Twemoji, version 17.0.3 (2026-06-01), jdecked/twemoji repository on GitHub; graphics under CC BY 4.0 (LICENSE-GRAPHICS), code under the MIT License.

   https://github.com/jdecked/twemoji (license files, README and the SVG files shown on this page)

7. 7

   Twemoji, twitter/twemoji repository on GitHub: tags v2.5.1 and v2.6.0 (commit 7c1d3e9, 2018-04-11) and commit a6f943b (2026-01-10); commit dates in UTC.

   https://github.com/twitter/twemoji (commit history of the U+1F52B files, tag list, license files and README at the pinned commits)

8. 8

   Microsoft, Fluent Emoji, microsoft/fluentui-emoji repository on GitHub (commit 1ffb34c, 2026-08-24), MIT License.

   https://github.com/microsoft/fluentui-emoji (license file, README, metadata files and the Flat-style SVG files shown on this page)

9. 9

   Google, Noto Emoji, googlefonts/noto-emoji repository on GitHub: commits 617db97 (2014-08-14), 2c7bdf1 (2014-12-19), f931bea (2017-08-25), cf666e6 (2018-03-08) and d3e0359 (2021-05-15) for svg/emoji_u1f52b.svg; commit dates in UTC.

   https://github.com/googlefonts/noto-emoji (commit history of the U+1F52B files; LICENSE, AUTHORS and README at commit f931bea)

10. 10

    Google, Noto Color Emoji, version 2.057 (font build 2026-09-11), SIL Open Font License 1.1.

    https://github.com/googlefonts/noto-emoji (font file: name, cmap, GSUB and SVG tables)

11. 11

    Hannah Miller, Jacob Thebault-Spieker, Shuo Chang, Isaac Johnson, Loren Terveen and Brent Hecht, "'Blissfully Happy' or 'Ready to Fight': Varying Interpretations of Emoji," Proceedings of the Tenth International AAAI Conference on Web and Social Media (ICWSM 2016), 259–268, https://doi.org/10.1609/icwsm.v10i1.14757.

    https://ojs.aaai.org/index.php/ICWSM/article/view/14757 (full text)

12. 12

    Wikipedia, "Pistol emoji," revision 1380262827 of 2026-10-14.

    https://en.wikipedia.org/w/index.php?title=Pistol_emoji&oldid=1380262827 (lead and the section on glyph changes)

Download the data: emoji counts by kind (CSV) (Data: Unicode emoji data 18.0 and Unicode Character Database 18.0.0 (https://www.unicode.org/Public/18.0.0/), © Unicode, Inc., Unicode License v3 (https://www.unicode.org/license.txt), retrieved 2026-10-16) · 25 sample emoji with names and versions (CSV) (Data: Unicode emoji data 18.0 and Unicode Character Database 18.0.0 (https://www.unicode.org/Public/18.0.0/), © Unicode, Inc., Unicode License v3 (https://www.unicode.org/license.txt), retrieved 2026-10-16; CLDR 48.2 annotations (https://github.com/unicode-org/cldr/releases/tag/release-48-2), © Unicode, Inc., Unicode License v3 (https://www.unicode.org/license.txt), retrieved 2026-10-16) · U+1F52B drawing history (CSV) (Data: Unicode emoji data 18.0 and Unicode Character Database 18.0.0 (https://www.unicode.org/Public/18.0.0/), © Unicode, Inc., Unicode License v3 (https://www.unicode.org/license.txt), retrieved 2026-10-16; CLDR 37 and 38 English annotations (https://github.com/unicode-org/cldr/releases/tag/release-38), © Unicode, Inc., Unicode License v3 (https://www.unicode.org/license.txt), retrieved 2026-10-16; Commit histories of googlefonts/noto-emoji, twitter/twemoji, jdecked/twemoji and microsoft/fluentui-emoji (https://github.com), retrieved 2026-10-16)

Corrections

No corrections yet. Report an error through the corrections address in the page footer; every change is logged here as YYYY-MM-DD — what changed and why.

Decode next:

• Egyptian Hieroglyphs: Picture signs that also spell consonants, and the 5,105 code points Unicode gives the script.
• Male and Female Symbols: The two signs that mark 628 emoji sequences as female or male, and the planets Linnaeus took them from.

Cite this page