Late in January, the trial of Ross Ulbricht, the 30-year-old alleged leader of the Silk Road black market website, led to a debate over a piece of eyebrow-raising evidence: a smiley face emoticon Ulbricht had typed at the end of a post.

You know the one:


The question was whether the two-character emoticon mattered to the case. The prosecutor didn't mention its appearance while reading aloud an excerpt from the chat conversation in which it appeared, so Ulbricht's lawyer objected to its omission. And at first glance, it's easy to see why the prosecutor avoided it: A colon and a close parenthesis make up a common emoticon—":)"—that simply meant Ulbricht was smiling and approved of what he was saying in his post.

Or did it? The judge in the trial ordered the jury to take note of symbols like the smiley face in the messages, but the case poses an important question over how we define textual communication. Anything presented in text, whatever the type—characters, pictograms, onomatopoetic expressions—can be interpreted in some way. And court cases involving unconventional, unpronounceable type aren't rare. As The Marshall Project found, emojis (not emoticons) have already appeared in court:

It was probably inevitable that emojis (and their more basic predecessor, emoticons) would get their day in court... Over the last year and a half, these little online signifiers have figured, peripherally, in at least nine cases, including six in the last two months.

In none of these cases did the result hinge on how emojis were interpreted. But the courtroom exchanges indicate that at some point in the near future, the court system will have to determine whether these symbols express such specific and unambiguous meaning that, like words, they can be introduced into evidence.

The question at the moment, therefore, isn't how emojis are defined; it's whether they're vital to a case in the first place.

But the definitions of emojis will raise questions soon, because emojis are essentially fuzzy in definition. They're embedded in complicated social context, making them open to interpretation. The prayer hands emoji, for example, may represent gratitude in some cultures, straightforward praying in others, or even, oddly enough, a high five. While most emojis involving straightforward objects (office supplies, animals, vegetables, weather patterns, etc.) can be interpreted as is, some prove harder to read—especially when they're used as stand-ins for facial expressions.

That's where it gets thorny for any textual communication, says Tyler Schnoebelen, the linguist and chief analyst at language-data company Idibon. The expressions on a person's face can't be translated perfectly in text, even if emojis include a set of varied faces. "Face-to-face, you get intonation, speech tempo, hand gestures, furrowed brows," he tells me. "We're trying to make sense of an impoverished channel of communication. We're trying to make ourselves understood." So, a wide smile in emoji form may not mean joy, but could instead mean a variety of expressions not easily interpreted by the receiver of the message. Which makes sense. Actual facial expressions can be hard to read without proper context, too.

To begin to understand why emojis—and specifically the ones with faces—can be so hard to read, it helps to look at the many ways people express laughter in text. The linguist Ben Zimmer tells me laughter allows for different gradations of emotion and phatic expression—the use of language as a social function instead of a mode to convey information—which makes it perfect for seeing the malleability of definitions behind textual expressions.

The first aspect—the gradation—of interpretation is easy to see with laughter. With just two letters, 'h' and 'a,' we can combine them to produce the sound "haha," but also involve capitalization, length, and repetition to tweak its meaning. One "ha" could mean a chuckle; an almost never-ending, Channing Tatum-like string of the all-caps "HAHAHAHA" means... something else entirely. It works the same way with "lol," with 'h' and 'e' for "hehe," and so on. Each combination means laughter, but each also points to different intonations and different levels of laughing. And repetition has become a common tool to show expression online. In his study of repetition on Twitter, Schnoebelen found that more than half of Twitter users repeated strings of words ("no no no," "omg omg omg," etc.) to get their expressions across, even when constrained to 140 characters a tweet. In other words, repetition can be as meaningful as the word in and of itself. As he explains it in his post:

Bursts of emotion are one of the hardest aspects of speech to capture in written text. It is difficult to capture exuberance and immediacy when you are allowing your readers to read your utterances at their own pace. But these sudden explosions of sentiment are also often the most interesting to track and analyze, giving us insight into a writer’s emotional expressions. Repetition is one of the simplest ways that people overcome these limitations on writing, especially when it comes to expressing laughter.

As for the phatic expression of laughter, "haha," "lol," "hehe," etc. can serve as placeholders as well, regardless of whether the sender of the message is laughing in real life. In these cases, letters stand in for the nods, smiles, or utterances (e.g. "uh huh") that we'd otherwise use in face-to-face conversations to signal to the other person that it's okay to keep talking and reassure that person we're listening. When we don't utter little noises or gestures of understanding, we worry about the silence. "We acknowledge each other all the time through the conversational turn-taking," Zimmer says, "so when it's not there in text, there's social anxiety."

That's the basic idea behind how people interpret emoticons and emojis, too. "The introduction of emoticons and emojis just gives you more tools at your disposal," Zimmer tells me. "Emoji allow you to play with language in a way, and to express emotional state in a new way."

They won't solve everything of course, because no text communication can fully replace how we represent ourselves in speech. Some expressions, like sarcasm, often require in-person intonation. We've tried to create punctuation and emoticons that can do the job, but it's not the same as hearing a sarcastic remark out loud. Others, like shock and frustration, require stage directions like "facepalm" and "headdesk" to convey exactly how the writer is feeling.

So, the answer to the question of whether emoticons are relevant in Ulbricht's case is a resounding yes and no. The characters ":)" could represent a specific emotion on a spectrum of emotions—or mean nothing at all. Each person who reads the emoticon could interpret it differently. And being a part of the conversation—rather than just someone reading it later—offers potentially crucial context.

All this is familiar territory, of course. Words have all kinds of connotations and nuances, too. Adding a visual layer to the way we talk to one another can only help better mirror the way we communicate in person. As Zimmer puts it, "All of these social functions, even if it's not conveying information the way we think language is supposed to, they're the glue that's holding the conversation together. Without it, it falls apart."