A friend passed along this article recently which describes research by Tyler Schnoebelen that explores the rules people use when writing with emoji/emoticons. When people hear that I work on “visual language” they often think of things like this. However, Schnoebelen actually did study these things, scouring through roughly 500,000 tweets and analyzed the types of productions that people made.
For example, he noted that people tend to place emoji at the end of clauses and sentences rather than the beginning. This makes sense, because people won’t want to divide up their syntactic clauses. Instead, they’d prefer to use emoji like punctuation to divide up those clauses.
There also seems to be at least some linear causality when people are trying to communicate in emoji alone. The article notes that a story told through emoji would have to go in sequence, like this:
But, rearranging the order creates a different story, possibly less comprehensible:
This type of ordering could possibly use the rules in my visual narrative grammar, but in fairly stripped down and choppy form. Scrambling the order of units though is a common technique we’ve used before in studying the structure of sequential images.
These are particularly interesting for me because this type of things is what people often mistake my research for. However, emoji are not actual “visual language” as is laid out by my theory. Rather, they are at the intersection of visual language with writing—a conversion of sound into graphics (a learned synesthesia). They can act like “visual gestures,” supplementing or enhancing the expressions of the text. Or, they can be like a pidgin, a hybrid communicative resulting from the intersection of rules of different systems, often using combinations based on basic rules of meaning without a complex grammar.
Nevertheless, emoji are particularly interesting because they still can reveal various preferences that people have for ordering meaningful information that is likely shared across various domains. For example, another thing discussed in the article is that people will generally show an emotional state first, then show the cause of it, as in the laughing-monkey-poo example above and this one, where sadness precedes the reason why (a broken heart):
These examples use a structure common across communicative systems, which is to express the Agent (the doer of an action) prior to the Action. In this case, the Agent is sad, and the sadness is caused by the action of a broken heart. Often, this pattern will have a Patient (a receiver of an action) as well, in a canonical Agent-Patient-Act order. This pattern occurs very frequently in both full languages (such as Subject-Object-Verb ordered grammars), and in communicative systems that lack a full grammar (such as when non-sign language speakers are asked to communicate with only gestures).
I think there could likely be some interesting work done using experiments that ask people to communicate with emoji alone. That might be a nice follow up to Schnoebelen’s work. He’s done the corpus analysis of looking at what people have done in naturalistic expression. An interesting next stage would be to test various rules in experiments.
So… are emoji “visual language”? Not really in any full blown sense. However, they can tell us about verbal and visual languages, and the rules of communicative systems more generally.
(Wayback machine note: A long time back I did this related post about emoticons and the brain)