Variation in Depiction of Same Emoji on Different Platforms Can Lead To Miscommunication

How your device depicts an emoji depends on the operating system it is running. The same "smiley face" emoticon, for instance, appears slightly different when viewed on an iPhone, an Android-powered handset, and a Windows Phone-powered handset. This variation can cause miscommunication between people (PDF), a study by GroupLens Research has found. The research lab in the Department of Computer Science and Engineering at the University of Minnesota said that sometimes this can cause people to misinterpret the emotion and the meaning of emoji-based communication "quite significantly." The conclusion reads: Emoji are used alongside text in digital communication, but their visual nature leaves them open to interpretation. In addition, emoji render differently on different platforms, so people may interpret one platform's rendering differently than they interpret another platform's. Psycholinguistic theory suggests that interpretation must be consistent between two people in order to avoid communication challenges. In this research, we explored whether emoji are consistently interpreted as well as whether interpretation remains consistent across renderings by different platforms. For 5 different platform renderings of 22 emoji Unicode characters, we find disagreement in terms of both sentiment and semantics, and these disagreements only increase when considering renderings across platforms.

Comments

Toldja so, you morons!

[Zontar The Mindless]

This is exactly what I said would happen when I wrote to the Unicode Consortium asking them not to adopt emoji into the standard. Their response was that rendering differences in alphabetic/ideographic symbols with well-defined *objective* meanings never posed such a problem, so rendering differences for entirely *subjective* symbols wouldn't, either. /facepalm

Re:Toldja so, you morons!

[fuzzyfuzzyfungus]

There was arguably a case to be made for the one-time adoption of some 'emoji' in line with Unicode's "Sometimes we do horrible things so that even worse legacy standards and nonstandards can die." policy; but the failure to stop there has been a total clusterfuck.

Even good old Plane 0 is riddled with characters that should never have been allowed to exist; but if the Unicode Consortium had taken the principled stance and refused to hand out the codepoints needed to support migration from various legacy encodings, Unicode would probably still be more or less irrelevant in practice. Cleaning up the mess in the Japanese handset market is at least arguably in line with the same approach.

Once that was done, though, leaving open the invitation to turn Unicode in to a clip-art library was an atrocious plan; and bafflingly stupid(especially since the core mission of rendering actual languages is still pretty deeply unfinished once you wander too far from languages that can be handled with the latin alphabet and a few accents and umlauts.)

Re:Wait, what?

[fuzzyfuzzyfungus]

And it does; within the scope of what Unicode is actually supposed to do,

The problem is every bloody idiot who doesn't understand the 'what Unicode is supposed to do' part. If characters entered on one platform are showing up as different characters on another; you've either got broken software or a unicode problem. If you want absolute certainty that the recipient will see exactly the glyph you associate with a character; you don't have a unicode problem; you want one of those "Image formats" that certain people on the cutting edge of technology are using to encode, store, transfer, and decode things for which visual fidelity is important...

The fact that 'emoji' are still being treated as Unicode's problem, long after the original issue with freaky Japanese legacy handset design has largely been cleaned up, is just so frustrating. If you want your stupid clip-art to reach the recipient intact, we have a variety of (mature and widely adopted) options for that. Full ICC support and absolute color fidelity? Probably not; but good enough? Sure. Stop trying to turn Unicode into the world's most dysfunctional image format!

Do the Japanese have any data?

[swb]

I thought the emojis were kind of a Japanese thing originally. Do they have any data on mixed perceptions of them? I'm assuming that as early adopters they didn't have a uniform version of the characters/icons and may also have suffered from lower resolution depictions of them.

Re:Do the Japanese have any data?

[AmiMoJo]

The early Japanese ones were proprietary, unique to each mobile operator. If you had an AU phone and your friend had a DoCoMo phone you couldn't trade emoji. In short order they got together and merged their sets.

Back then Unicode was totally inadequate for Japanese text as used by people day-to-day, and didn't have any emoji support anyway. Even now the bulk of Japanese text files and plain text email is Shift-JIS rather than Unicode. When Unicode is used in anything but trivial situations needs hacks to render properly. Westerners who run Japanese software need to use AppLocale because most apps aren't Unicode.

It's a total disaster and needs replacing with something better.

Well then maybe

[JustAnotherOldGuy]

Well then maybe, JUST MAYBE....people shouldn't use emojis for actual communication where meaning might be important.

I've heard that there are these things called "words", which, when used properly, have the amazing ability to convey information accurately.

I swear, soon we'll be back to grunting and painting pictures of animals by smearing our feces on cave walls.

Re:Solution: don't use emoji

[JustAnotherOldGuy]

If you can't express yourself with pure text, you are an idiot anyway.

Agreed. Emojis are the confetti of the internet.

Unicode

[fyngyrz]

The problem is every bloody idiot who doesn't understand the 'what Unicode is supposed to do' part.

Oh, you mean like the people who wrote, maintain(ed) and inherited the source code for slashdot.

Two Dots Too Many

[Edward Coffin]

This reminds me of an incident in Turkey back in 2008, described and analyzed in Language Log: Two Dots Too Many. Due to a cellphone being improperly localized, a normal letter i was substituted for the Turkish back unrounded i (which I cannot figure out how to display here, ironically enough), altering the meaning of a text message, leading to a tragic misunderstanding, which resulted in a group attack on the sender who then murdered the recipient and subsequently committed suicide.

Re:Two Dots Too Many

[religionofpeas]

altering the meaning of a text message, leading to a tragic misunderstanding, which resulted in a group attack on the sender who then murdered the recipient and subsequently committed suicide.

Seems to me that it's not the fault of the phone, but rather the hot tempered reaction of the people involved.

Re:University of Minnesota

[dontbemad]

What is the point of such shameless hostility? How is this not a valid research topic? They have shown marked differences in how subjects perceive what should essentially be the same display of emotion by which mobile platforms they use. This has hefty implications on modern sociology, including the ability to predict how subjects may react to the same exact message based solely on their choice of phone.

A joke? Why does something that doesn't interest you have to be labeled as "a joke"?

What you should really be asking is how your comment passes as a valid addition to the discussion. That is the true "joke".