Emoji have become important. They’ve permeated our conversations and our messaging apps and our popular culture to a degree that no one could have anticipated just a few years ago, and when your phone or computer gets an update, new emoji are often featured prominently in the release notes or even announced in their own press releases.
That the “language” is so universal and recognizable is due in large part to the Unicode Consortium, the group of major tech companies in charge of defining and approving new emoji (and dozens of other character sets, besides). Every year, it proposes, discusses, and approves new additions to the language, and that heavyweights like Apple, Google, and Microsoft have become so diligent about supporting new versions is a rare victory for standards in an age where every tech company on Earth is trying to lock you into its own proprietary silo.
But the Unicode Consortium can only do so much to influence the way any given emoji looks and is interpreted. Every new version of the Unicode spec includes a description of each character, a sample image, and other broad recommendations for implementation, but companies implementing the spec are free to represent the emoji pretty much however they want. And as the language’s range of expression continues to grow, so do the opportunities for misunderstanding.
“Zero-width joiners” and compound emoji
The Unicode Consortium has been working hard over the last couple of releases to fix the language’s representation problem—multiple skin tones were proposed in 2014 and began to see widespread use in 2015, and this year’s update ensures that both male and female images exist for all the emoji occupations. And before we continue, let’s pause to acknowledge that representation is important, and it’s great that the Unicode Consortium is prioritizing it, full stop.
Traditionally, every emoji character has had an alphanumerical code associated with it, and creating a new emoji meant coming up with a new code to associate with it. For skin tone and gender, the Unicode Consortium has come up with more creative solutions, choosing instead to combine multiple codes together to create a single emoji.
The gender-diverse emoji for occupations, for example, combine a standard “man” or “woman” emoji with a second emoji to represent the occupation, and they’re glued together by a special invisible character called a “Zero-width joiner” (ZWJ). Platforms that support the new emoji for occupations, including both the upcoming iOS 10 and Android Nougat releases, see the ZWJ and know to display a single emoji rather than two separate ones.
It’s a smart way to continue expanding the number of available emoji while keeping the number of unique codes under control, and different sequences can be used to create emoji that are much more representative and specific than they were a couple of years ago. For example, ZWJ sequences are why the “kissing” and “love” and “family” emoji now have so many possible combinations. In Unicode 6.0, there was one emoji for family; now there are many more.
The Unicode Consortium tracks common ZWJ sequences (any “that are supported on at least one commonly available platform”) so that companies building software can easily support them, but here’s where we start running into potential problems. Since emoji cobbled together with ZWJ characters don’t need to be approved by the Unicode Consortium like an all-new emoji character does, it means that companies can build their own unique emoji without waiting for the Unicode Consortium to approve them. And since these ZWJ emoji will simply show up as two or more separate emoji on platforms that don’t support them, adding new ones doesn’t technically break anything or result in an ugly question mark block.
This can make developing new emoji faster; without a months- or years-long approval process, Apple or Google could go ahead and add flags or fix representation problems quickly without breaking compatibility with other platforms. On the other hand, it can make it more difficult to communicate between platforms based on which ZWJ sequences are supported and which aren’t.
“If this is taken to extremes, it could result in entire sets of incompatible emojis that are indecipherable on other platforms,” Jeremy Burge, founder of Emojipedia, told Ars. “That would be a problem.”
We’re already starting to see some fragmentation as platforms add their own specific ZWJ emoji, characters that the enthusiasts at Emojipedia can keep on top of but that the official Unicode list can’t. Twitter has a pirate flag, the Windows 10 Anniversary Update adds a bunch of “ninja cat” emoji as an in-joke for Windows Insiders, and WhatsApp just added an Olympic Rings emoji that shows up as five plain circles on any other platform.
Sometimes the broken-down sequences have a relatively clear meaning (the pirate flag), but often the intended meaning is vague (Olympic rings, Ninja Cat). And software makers are being left to police themselves when it comes to creating these emoji and implementing emoji created by others.
“I can see a real dilemma for vendors if ZWJ proliferate in unexpected ways,” Burge told Ars. “In particular, sequences such as the Olympic Rings emoji introduced by WhatsApp don’t follow conventions that other vendors use (it’s made of non-emoji characters), and to top it off it’s a copyrighted image which no other platform could introduce!”
I would expect Apple, Google, and the rest to be relatively conservative about introducing their own platform-specific emoji even with ZWJ characters, but going forward they could introduce more potential for confusion when communicating across different platforms or services. This is what Unicode was created to avoid, but there have always been inconsistencies even without ZWJ characters.
Apple vs. the world
The Unicode Consortium provides names and sample images for all emoji, but, as that page and any Emojipedia article can show you, the pictures you’ll see on actual phones and computers all look slightly different. Different OS and app developers want the emoji to match the rest of their software’s aesthetic rather than using one universal character set.
Typically, this isn’t a problem. A smiling or frowning face will look different on iOS and Android, but they both convey the same information. At worst, a handful of characters are mildly ambiguous or just a little weird—the “grinning face with smiling eyes” emoji looks more like a grimace on Apple devices. Samsung devices render the cookie emoji as a pair of snack crackers rather than the usual chocolate chip-style. Android 4.4 inexplicably rendered the yellow heart emoji as a hairy heart instead.
But every once in a while the difference is big enough that it could cause problems—case in point, the iOS 10 pistol emoji, which will look like a bright green squirt gun rather than a handgun. All other current platforms render it as a handgun, and the Windows 10 Anniversary update even changed the icon into a handgun rather than the ’50s-style sci-fi ray gun that Windows used previously. Apple similarly lobbied (successfully) to have a new rifle emoji removed from Unicode 9.0. Regardless of your position on gun control and whether you think these changes will move the needle at all (for my part I think the changes are well-intentioned but not likely to make much of a difference), making an emoji weapon look so different from the Unicode sample and every other platform creates a real potential for serious misunderstanding.
It’s particularly problematic when Apple goes its own way because the company is often seen as the “default” for emoji designs—when you see emoji in the real world, they usually look like Apple’s. Apple didn’t invent emoji by any stretch of the imagination, but by including an easily accessible emoji keyboard in iOS 5, by enabling that keyboard by default in iOS 8, and by adopting new emoji in iOS 9.1 and iOS 10, the company has been instrumental in introducing the character set to new users and helping it spread.
Ideally, all parties using Unicode would do their best to stick to the established examples laid out by the Consortium; doing otherwise introduces too much potential for miscommunication. In practice, the major companies using Unicode are mostly maintaining similar character sets by self-policing, looking around at each other and tweaking their emoji to be more in line with what others are doing (often, this means changing them to look more like Apple’s).
“The varied appearance of emoji on different platforms I see as a lesser issue in 2016 than it has been in the past,” Burge told Ars, “and an issue that is slowly resolving itself as platforms look to one another for cues on consistency.”
In cases like the pistol emoji, it would ultimately be better for Apple to voice its concerns to the Unicode Consortium rather than going off and doing its own thing. But given past behavior, the ecosystem may shift back to a more uniform representation of the pistol emoji by itself without the need for strict oversight or external pressure.
As the language grows and becomes more expressive, the potential for misunderstanding and fragmentation increases, but hopefully the desire to maintain a consistent, universal, easy-to-understand character set will outweigh any particular company’s desire to make sweeping changes based on its whims.
“I truly believe one of the reasons for the success of emoji has been the cross-platform nature,” Burge told Ars. “We have had graphical emoticons for years, but these tended to be for one app or platform only. The universal nature of emoji is a key part of what people enjoy. You can send an emoji and know the other person will see it, even if some platforms vary exactly how these look.
“It’s in the interests of each platform to be responsive to users, so if enough people demand cross-platform emoji support to continue as it does today, vendors should take that into account,” he added.
Listing image by Andrew Cunningham