Emoji evidence errors don’t undo a murder conviction | Hacker News Reader

Emoji evidence errors don’t undo a murder conviction

blog.ericgoldman.org

75 pointshn_acker75 comments

In the same vein was an incident where an improperly localized phone in Turkey caused a sent message to arrive with different characters, with very different meaning, and the fallout was two deaths [1], discussed here [2]

[1] https://languagelog.ldc.upenn.edu/nll/?p=73

[2] https://news.ycombinator.com/item?id=9900758

> There are several lessons to take away from this tragedy. One is that localization is a good thing. Another is that it is best not to kill people who make you angry until you have carefully investigated the situation

wise wise words

The person that was killed was the person who started the violence, the wife. I would posit that instead physically attacking someone who comes over to apologize, you should try listening first.

Also, if your wife and her family are that crazy, give them time to cool down before you put yourself in that situation...

the father was the one who stirred the pot and misinterpreted the text -- blame him.

and this was an estranged wife, which sometimes leads to murders in the US and other places. they were looking for a reason to be pissed at him.

It's interesting because like the article says legal teams may have to get smarter about recreating all the context when evidence like this is used. Even if the emojis rendered the reference implementation of Unicode and what vendors actually represent can vary quite a bit by platform or OS version.

I recently listened to some podcast where they talked about this in the context of threatening texts. Sending someone a gun emoji communicates a very different thing if it shows up as a water pistol vs a realistic looking gun. The court need to see what the sender thought they were sending and what the receiver saw.

https://www.reddit.com/r/coolguides/comments/k1xxio/evolutio...

An obvious problem is Apple renders the firearm emoji as a water pistol, and everyone else renders it as an actual pistol.

According to Wikipedia, now Apple, Google, Microsoft, Samsung, WhatsApp and Facebook all use a water pistol. X was the only platform to roll that back.

Also IIRC there are some renderings where the firearm is aimed to the left and others aim it to the right. I'm having trouble sourcing this though - maybe this was true in the past or it was some other emoji that implies a direction.

All vendor's guns have always aimed at the left, however there are plenty of discrepancies, collectively referred to as "emoji fragmentation"

https://blog.emojipedia.org/2018-the-year-of-emoji-convergen...

Theoretically, there's a mechanism in Unicode allowing you to aim left and right: https://unicode.org/reports/tr51/#Direction

But I don't think it's implemented widely, if anywhere.

Android changed in 2018 -- which adds even more of an issue since serious cases can take many years to go to trial, what it looks like on a phone today might be totally different than what it looked like on a phone at the time of the crime.

All the big tech companies except Twitter now render it as a squirt gun afaik

Almost everyone renders a water pistol now: https://emojipedia.org/pistol (click the "designs" tab)

Twitter/X is the only major exception, and they only changed it to represent a realistic gun recently.

Interestingly, this seems to be a pattern — in cases where the other providers are depicting an emoji “per the spec” but one provider is doing something unusual, the unusual depiction almost always spreads to fixation among the other providers.

For example, “loudly crying face” (https://emojipedia.org/loudly-crying-face) was literally supposed to just be a kind of mouth-open bawling expression. And that’s how everyone did it… except Apple. Apple gave their version a weirdly mixed expression that people sometimes interpret as “crying while laughing because something is so funny” (even though there’s already a separate emoji for that.) And iOS users kept using this emoji to mean that, while everyone else was confused. But instead of Apple fixing their emoji, everyone else gradually changed their depictions to conform to Apple’s non-standard interpretation. (Seriously, look at the history for each provider in the above link.)

I dunno, the SoftBank emoji from 25 years ago is pretty much the same face as Apple's is now.

https://emojipedia.org/softbank/2000/loudly-crying-face

"The SoftBank emoji designs heavily influenced Apple's original emoji font which was designed to be compatible with this set when launching in Japan, due to iPhone being a SoftBank-exclusive phone when first released."

Would be why.

> Apple gave their version a weirdly mixed expression that people sometimes interpret as “crying while laughing because something is so funny” (even though there’s already a separate emoji for that.) And iOS users kept using this emoji to mean that, while everyone else was confused.

This sounds like a small social media niche being interpreted as representative overall. The first 10 search engine results were people primarily interpreting it as a crying/sadness emotion; some used it as an expression of general intense emotion. This seems consistent with actual crying, which can be a reaction to many different intense emotions, not just sadness.

In a criminal defense scenario you should take into account how the sender sees it, potentially how the recipient sees it, and whether the sender was aware the recipient would see it differently.

The case in question does not hinge upon the crying emoji. This was a tangential discussion.

I mean, it is how my own SO sees/understands this emoji (despite my insistence that that's not "what it means.") I just found out that she uses this emoji that way a few days ago, and then this article reminded me of that.

But I didn't mean to argue from social proof. Rather the opposite!

My argument was more—how else should you read the actual evidence? That evidence being that all the other providers' emojis started off depicting an "anguished" down-turned mouth and scrunched eyebrows; but that all of them then gradually reworked their depictions to instead include a neutral 'O' mouth and raised eyebrows, which removes the signifiers of anguish from the expression.

Why else would they all do that, except to cohere with the expectations of users who somehow communicated to them that they expected the emoji's expression to not be read as "anguished"?

What is it these days that perfectly normal words and things get censored? As far as harmful social phenomena go this is very small, but it's so silly.

Because there's a concerted effort to "denormalize" and stigmatize the perfectly legitimate activity of owning a firearm for lawful purposes like hunting, collecting, and sport shooting.

Just one more front in the culture wars.

I don't agree. It's more like an extreme fear of causing offence, which is one of the greatest crimes any public official or company can commit in todays day and age.

You can definitely fit it in to the ongoing war against culture, but you are reaching a bit.

I can't even blame Big Tech for it. It probably makes them more profit to be proactively timid. When that stops being true so will they.

>You can definitely fit it in to the ongoing war against culture, but you are reaching a bit.

Why would it be a reach when you yourself acknowledge it.

>It's more like an extreme fear of causing offence, which is one of the greatest crimes

Its pretty clear that the left leaning tech world is using it either way for there agenda.

Yes. I wouldn't call that emoji "tears with joy" I would calling it "laughing so much you have tears".

To say nothing of the eyes and smile on U+1F4A9 PILE OF POO which kind of made it more disturbing.

> instead of being followed by two emojis, the message is followed by four closely-spaced rectangles. Neither the text of Delarosa’s in limine motion, nor anything said during the in limine hearing would have informed the trial court that the four rectangles represented two emojis

So I know nothing about this trial and have limited knowledge of the US legal system, but didn't one party just misrepresent evidence here? They would probably argue that it wasn't intentional and thus not perjury, but it still sounds pretty serious. The emojis are just as much part of the message as the latin characters

If I understand TFA right, it was the defense that made the mistake, which is why the appeal failed. You don't get a do-over when the mistake is on your side.

I thought you would if you prove your lawyer was incompetent and sue them?

The defendant could raise an "ineffective assistance of counsel" argument, asserting his counsel was incompetent and it prejudiced him. You can make that claim in your direct appeal, but for procedural reasons, defendants usually make them in separate petitions for writs of habeas corpus. That's not a lawsuit against the lawyer, though. Instead, it's a lawsuit seeking your freedom from the state.

Ineffective assistance of counsel is a thing (and does not require suing the lawyer).

However, failing to properly object to how some emojis were entered into evidence is no where near the standard of being ineffective.

IANAL, but I can imagine the prosecutor's pushback on that: "at what point did you attempt to straighten out your own lawyer about their misinterpretation of your evidence? Can you point us to the line in the transcript where you tried to explain the correct interpretation in court?"

Like, if my life was on the line, and my lawyer was screwing up my evidence, I think I'd try to point that out to someone.

Thank god that isn't the bar for ineffective assistance of counsel.

A vast majority of defendants don't know all the ins and outs of law... that's why we have a profession to deal in this domain. Asking the average defendant to check their counsel's work is ridiculous.

I disagree. Clients shouldn't be expected to know the ins and outs of legal procedure and rules of evidence and a million other things like that. I think they should be reasonable expected to say something when their attorney is making factual mistakes, like "when my client wrote this, they meant...", when the client knows that's not at all what they meant.

To give an example that might resonate to HN folks, suppose your attorney says "I know my client had a bunch of hacking tools on his computer, and that looks bad". Now suppose the laptop in question is your work-issues computer, and you work in pen testing, so you possess all of those for legitimate work-related reasons. I think it'd be hard to appeal on those grounds. "My attorney should have said something?" "You just sat there and went along with it, though." Appeals aren't meant to be an infinite series of do-overs were you get to relitigate every single thing you wish you'd said differently.

If your attorney messes up on legal issues, that's on them. When they mess up on the basic facts, and you don't say anything about it, I think that's kind of on you.

For starters, most attorneys won't allow their clients to utter a single word to law enforcement or on the court record.

Your proposal is that the client has to have enough of a grasp on the legal strategy related to self-incrimination that they could challenge and disregard their attorneys advice, in order to save themselves from something the attorney misinterpreted.

I personally think that's way too much to ask of a defendant. You even see the accused who are lawyers themselves just hire other lawyers and promptly stfu. Counsel always decides and implements the strategy.

No one would expect the defendant to address the court, but if you’re on trial and your counsel is screwing something up, you need to let them know. “Psst, hey, that’s not what I said.” So is the defendant claiming that their lawyer overrode their correction, to their disadvantage? Because, if not, they should have said something.

And how did you come to find out that the defendant didn't do that? That would explicitly be not included in the record.

If I understand TFA then the defendant is arguing that his message about owning a gun was made less glib by the verbatim inclusion of a tears-of-joy emoji plus a smiling-devil-horns emoji at the end.

That is... an unusual argument to make.

Lawyers in getting paid to debate their guess of meanings of emojis shocker.

I have noticed that men and women tend to use different emoji and ascribe different meaning to them. Ex. I see the skull emoji used to indicate laughter more often used by women and crying face used more by men.

There are some tribes where men and women have completely different languages, I wonder if we will end up that way with emojis.

I used to use the :-P emoji on AIM as sort of "that was a joke" emoji, and if I had a nickel for every girl who thought I was flirting with them for using it, I'd have 3 nickels. Not a lot but still weird it happened 3 times. Also how I got my first girlfriend. I wasn't flirting, but she seemed to like what I was doing it so I figured let's give it a shot.

Try changing ":" to ";" and that number is going to get increased by quite a lot. :D

> men and women tend to use different emoji

I went through a brief emoji phase where I had to scroll through a large list of emoji to find just the right one. I eventually realized this was just stupid and use words now.

As an aside, I dislike the juvenile Apple renderings. They look like stickers that come on sheets you'd give a second grader.

I think you're talking about https://en.wikipedia.org/wiki/Ubang_language

It does actually seem very applicable here - it's not two different languages, but there's a seemingly random subset of words in the language that have completely different male and female versions. If emojis are part of the language, and men and women use different ones to refer to the same concepts, that's basically the same thing.

The issues with emojis go much deeper than this. Even if we agree on how exactly they displayed, their social meaning is highly dependent on the context of a conversation. Instead of allowing outside investigators to divine their own meanings and introduce them as evidence, courts should insist on testimony from the person or people those communications were meant for. If said people give wildly differing testimony to what investigators think is truthful, then they can go down the rabbit hole of how the codepoints were displayed and whatnot.

One tricky aspect of this is if the messages are from the defendant then the defendant is almost certainly going to use their right not to testify (especially if they are actually guilty) as they'll be asked all sorts of other difficult questions they likely won't want to answer.

The state has been using this trick since forever. They don't even need to have written correspondance to misconstrue, they do it all the time with just testimony.

Same with everyday language. We just got used to its nuances.

Yeah, the novel bit here is only that the lógos displayed to the different parties can be materially different for technical reasons unrelated to the intent of either party. The rest is bog-standard ambiguity of language.

Exactly. For example, there's an enormous difference between someone saying "I'd like to choke that guy" in exasperation, and someone saying "I'd like to choke that guy" while discussing the weekend's plans with their friends. People say exaggerated things all the time and it's generally understood as not expressing their genuine intent.

A popular cultural example: "I shot the clerk?". https://www.youtube.com/watch?v=7SbI_lGPahg

> exactly

There are many different words that have different shades of meaning. For example, bird, duck, pigeon. None of that works with emojis.

Google's "Messages" client on Android has a feature where, in RCS chats, if you send or receive a message that solely consists of a single applicable emoji (and I believe in response to certain emoji reacts as well?), a little animation plays. The :joy: and related emojis trigger a "haha so funny"-type animation, :cry: and similar trigger a sad, rainy cloud, and so on.

An interesting thing I noticed recently is that :skull: triggers the same "haha so funny" animation as :joy: does! Which kind of surprised me, because I was using the skull to convey "lol I'm dead", so it fit here, but I wouldn't think that's the primary use for it.

Because "lol I'm dead = something is hilarious = :joy:". "Emoji evidence" is crazy. :skull: can mean a million other things, same with a (water) pistol emoji.

Theoretically this is probably where good jury selection comes in.

If a juror is presented a message with an explanation that is obviously “out of touch” with its intended meaning, the juror loses some trust and applies more scrutiny.

> For example, my software renders the smiling face with horns as red [], but often the depiction is purple.

It seems the article is ironically falling for the same problem. This would be worked around by including images of the emoji variants, rather than relying on Unicode.

Given it was not directly brought up in the motion in limine, it sounds like there were other concerns with the message anyway?

> it’s possible/probable that the trial outcomes would have been the same with or without the Facebook message evidence.

It seems like defense counsel was primarily concerned that the message indicated the defendant owned/possessed a gun. The emoji argument seems to be a secondary concern that he raised only on appeal.

As a side note, the FACE WITH TEARS OF JOY emoji doesn’t actually depict tears of joy, but laughing so hard that tears come out of your eyes.

Try LOL instead.

LOL isn't quite the same.

How is it different?

Illustrating another facet of problems with emojis and icons. For example, an image of a duck. Is it meant to be a duck, a goose, a pigeon, or a bird? Don't you love that ancient oil lamp icon that is supposed to mean "oil pressure"? Unlike words, you cannot even look up the definition of an emoji or icon.

Icon based written languages have been replaced over time with phonetic ones. There's a good reason for that. Icons and emoji don't work.

Constantly inventing new ones and adding them to Unicode is simply retarded.

There, I said it!!

i love how after having actual law professionals take a look at his appraisal, he has to retract all the things he said about law, leaving you with how he noticed emjois werent rendered on a printout. typical internet know nothing

> he argues the court should have excluded a Facebook message that

OT: Don't use Facebook or anything by that company.

The year is 2025 and the courts are debating "emoji evidence." That is where we are as a species.

I’m not sure I understand this comment. Emojis are a form of communication. Communications can and are evidence used in court. If someone drew pictures related to guns, and then was accused of a gun crime, that evidence would be used. If someone communicated non-verbally to someone by drawing their finger across their throat and then pointing at the person, who later alleged they were attacked by that person, that would be evidence. Emojis are simplified pictograms used as shorthand to communicate, like acronyms or initialisms are simplified representations of multiple words, like someone saying “RIP to you for what you did” could be a threat.

If someone sent an email threatening someone else, the court should not present that email incorrectly as raw HTML code. If a WhatsApp message was sent with text bolded for emphasis, it shouldn’t be shown to the jury in plain text. So I don’t understand this derisive attitude towards "emoji evidence."

> If someone sent an email threatening someone else, the court should not present that email incorrectly as raw HTML code.

If there was no text/plain alternative, showing the HTML would be acceptable evidence of their crime. Sentencing guidelines for sending mail without a text/plain alternatives aren't established, but I think 1 year for the user and 10 years for the software developer is fair.

I am just despairing the state of the species. We're communicating through comic pictograms like we're reverting to a neolithic state.

Pictographic communication continued into, and separately came about, far more recently than the neolithic era. And history isn't a linear movement of only progress. There is no obvious reason to think the pictographic communication is a degradation of communication.

> There is no obvious reason to think the pictographic communication is a degradation of communication.

Phonetic alphabets evolved from pictograms, not the other way around. For example, A is an upside down bull's head.

Pictograms are definitely a degradation.

Why is using alphabets and pictograms a degradation? Explained in a way that is something other than "pictograms were used long ago so they are bad".

It also seems that the assumption is that alphabets were an improvement in every way. Why is this a given?

Because you cannot look up a pictogram to divine its meaning. There are a million words in English. You can trivially look any of them up.

> "pictograms were used long ago so they are bad"

Please don't put words in my mouth that I never wrote.

Pictograms evolved into phonetic alphabets over and over in history, because phonetic alphabets are objectively better.

Take a chapter from "Lord of the Rings" and redo it with pictograms. Good luck with that.

How is using comic pictograms as one of the many ways we communicate some sort of reversion? We use different vocabularies when talking to different audiences, for instance I speak much more casually with friends than with my boss. We often specifically use vocabulary and word choice to provide context to the nature of the conversation. Like using formal and respectful wording to highlight professionalism, or using casual slang to highlight a joking or lighthearted tone.

As we have moved more informal conversations to written form (texting everyday with friends is a lot more casual than sending paper letter correspondence through the mail to friends), we have added ways to provide tonal context that is lost by not hearing someone’s voice or seeing their body language. Adding “LOL” or “haha” to indicate your statement is meant to be a joking tone, for instance. Emojis are just another way to do that and to reinforce the casual nature of the communication. Someone might use the turtle emoji when messaging their girlfriend about how long they have been waiting in line to give the message a cute playful tone, where they wouldn’t use it when talking about a production slowdown in a message to their coworkers.

Its fine not to like emojis, but it is eyerollingly pretentious to act like it is some indication of the de-evolution of society.

Reply on news.ycombinator.com