Facebook’s six reactions are similar to emoji in that they allow users to express emotion nonverbally, but they are more useful to Facebook because they comprise a simpler classification than the thousands of emoji. Viral content master Buzzfeed employs a similar, slightly hipper scheme for the reactions they permit users to post to their articles. Buzzfeed’s scheme is tailor-made for market research: content can be surprising, adorable, shocking, funny, etc.
Bloomberg’s Sarah Frier explained how Facebook formulated its new reactions:
Facebook researchers started the project by compiling the most frequent responses people had to posts: “haha,” “LOL,” and “omg so funny” all went in the laughter category, for instance….Then they boiled those categories into six common responses, which Facebook calls Reactions: angry, sad, wow, haha, yay, and love….Yay was ultimately rejected because “it was not universally understood,” says a Facebook spokesperson.
These primitive sentiments, ironically, enable more sophisticated analyses than a more complicated schema would allow—an important reason why simpler classifications tend to defeat more elaborate classifications. Written comments on an article don’t give Facebook a lot to go on; it’s too difficult to derive sentiment from the ambiguities of written text unless the text is as simple as “lol” or “great.” But a sixfold classification has multiple advantages. Facebook, Buzzfeed, and their kin seek universal and unambiguous sentiments. There is little to no variation in reaction choices across different countries, languages, and cultures.
Buzzfeed’s six possible reactions offered to article readers
The sentiments also make it easy to compare posts quantitatively: users themselves sort articles into “Funny,” “Happy,” “Sad,” “Heartwarming,” and “Infuriating.” From looking at textual responses, it would be difficult to gauge that “Canada stalls on trade pact” and “Pop singer walks off stage” have anything in common, but if they both infuriate users enough to click the “Angry” icon, Facebook can detect a commonality. Those classifications permit Facebook to match users’ sentiments with similarly classified articles, or try to cheer them up if they’re sad or angry. If reactions to an article are split, Facebook can build subcategories like “Funny-Heartwarming” and “Heartwarming-Surprising.” It can track which users react more with anger or laughter and so predict what kinds of content they’ll tend to respond to in the future. Facebook can isolate particularly grumpy people and reduce their exposure to other users, so they don’t drag down the Facebook population. And Facebook trains algorithms to make guesses about articles that don’t yet have reactions. Most significantly, even though these particular six reactions are not a default and universal set, Facebook’s choices will reinforce them as a default set, making them more universal through a feedback loop.*13 The more we classify our reactions by that set of six, the more we’ll be conditioned to gauge our emotions in those terms. The default six smooth out the variations observed when Facebook was conducting tests with a far larger set of emotions, all designed by Disney-Pixar’s Matt Jones:
The full list included admiration, affirmation, anger, anxiety, astonishment, awe, boredom, confusion, contemplation, contempt, contentment, coyness, curiosity, desire, determination, devotion, disagreement, disgust, embarrassment, enthusiasm, fear, gratitude, grief, guilt, happiness, high spirits, horror, ill temperament, indignant, interest, joy, laughter, love, maternal love, negation, obstinateness, pain, perplexity, pride, rage, relief, resignation, romantic love, sadness, shame, sneering, sulkiness, surprise, sympathy, terror, and weakness….Clear patterns emerged in the data. Italians, South Africans, Russians and Brazilians had “Cultures of Love”—sending lots of amorous stickers. The U.S. and Canada were similar in most of their usage patterns—though the Canadians were vastly more “sympathetic,” while the Americans were “sadder.” And the use of “deadpan” stickers predominated across North Africa and the Middle East.
Matt Jones’s sketches for potential Facebook Reactions
Matt Jones’s emotional sketches
A simple classification won out. It is both easier to use and more universal—at the expense of cultural and personal variation.*14 And also, to hear researcher Dacher Keltner tell it to Radiolab’s Andrew Zolli, at the expense of happiness:
Countries that expressed the most “happiness” were not actually the happiest in real life. Instead, it was the countries that used the widest array of stickers that did better on various measures of societal health, well-being—even longevity. “It’s not about being the happiest,” Keltner told me, “it’s about being the most emotionally diverse.”
The diversity comes not just from variety, but from ambiguity. One main appeal of emoji is that they aren’t given to fixed definitions and leave room for interpretation, misinterpretation, and variant meanings across cultures and subcultures. It’s hard to know what to make of this emoji, which goes by the name “upside-down face” (or U+1F643 officially):
Emojipedia describes the upside-down face as suggesting “silliness or goofiness. Sometimes used as an ambiguous emotion, such as joking or sarcasm.” In other words, it can mean many things, depending on context—just like a real expression. If an upside-down smile is not ambiguous enough, there’s the inscrutable “face without mouth” (U+1F636):
Here are some of the most popular emoji used on Twitter, as determined by the website Emojitracker:
“Smirk” looks more sly than superior, while “neutral” and “expressionless” seem almost identical. “Thinking” (U+1F914) has become popularly used to signify sarcastic “throwing shade,” to indicate vacuous disagreement. Providers like Facebook, Apple, Google, and mobile phone companies have their own designs that alter the nuance of the emoji, as with these variously constipated versions of “tired” (U+1F62B).
Over time, the meanings of emoji blur and diverge. The original Unicode characters for emoji were created to be compatible with the preexisting emoji that had originated in East Asia, which were then exported irregularly to the rest of the world. Such irregularities hampered data mining and sentiment analysis.
Different versions of the :tired: emoji
If the restricted, unambiguous set of six reactions has the effect of narrowing emotional diversity, social media and advertising companies view this tradeoff as the necessary cost of gathering better data on users. The restricted emotional language employed by Facebook is a language a computer can understand and manipulate at high scale. The simplified language of a core set of emotional reactions bridges the computational-human gap—more successfully than the overcomplicated ad hoc classifications of the DSM did. Instead, these reaction sets are reminiscent of the simpler folk taxonomies of Myers-Briggs, OCEAN, and HEXACO, which also break down complex phenomena onto a handful of axes. Facebook’s Reactions even approximately map to the Big Five:
Like: Agreeableness
Love: Extroversion
Wow: Openness
Sad: Neuroticism
Angry: Conscientiousness
The odd one out is Haha, because as always, laughter eludes easy classification despite being the most universal and nonnegotiable of expressions. Yet for the remaining five, there is an inevitable flattening of cultural differences. We saw how much trouble personality classification efforts had while trying to generalize their models across different cultures, and despite Facebook’s empirical research to generalize their six, it’s unlikely that they’re capturing the same sentiments across cultures—rather, they found sentiments that were recognizable by multiple cultures. If the data miners and user profilers get their way, soon enough we will all be Loving, Wowing, Sadding, and Angrying in lockstep.
The language of Reactions is a primitive vocabulary of emotion, vastly simpler than our human languages. It is far better suited to computers and compu
tational analysis. When I introduced graphical emoticons into the Messenger client in 1999, I didn’t foresee any of this. Around 2015, I began noticing a change on my Facebook wall. There was less discussion happening. People I knew were far more content to respond to posts with monosyllables like “yeah” or “ugh,” or else with simple emoji or Facebook’s six reactions. I caught myself contributing dumbly like this, to my dismay. I went back and checked posts from 2009 and 2010. I had written in complete sentences, arguments, with multiple paragraphs. The shift was obvious and drastic. Diversity, nuance, and ambiguity had declined. If passions were fervent and I disagreed with the chorus of “yeahs” or “ughs,” the crowd was far more likely to pounce on me; the same went for any other dissenters. What had happened? These were my friends. But they no longer seemed like the same people. We had been standardized. We were all speaking a different language now. It was the language of Facebook—of computers.
*1 I still blog at waggish.org. People of my generation romanticize the so-called golden days of blogging, but there’s something to be said for a decentralized content publishing platform, and it is unfortunate that Facebook, Twitter, Reddit, and the like have made the act of reading content on some random website a fringe activity.
*2 All large industrialized economies depend on the interchangeability of labor, so that employment and productivity can be calculated in terms of groups of average individuals rather than collections of unique, incommensurable individuals, which would prove immune to economic analysis. Karl Marx was concerned with the exploitation of labor, but not the commodification of it, which is required in order to allocate work under any centralized system.
*3 Joseph Turow’s The Daily You is a superb overview of the history of advertiser targeting and microtargeting.
*4 In July 2017, Facebook quietly renamed “Ethnic Affinity” as “Multicultural Affinity,” without changing the nature of the category at all. Both are euphemisms for race.
*5 For example, “Hispanic” is not a race, yet we treat it like one. The term is more or less meaningless outside the United States, where it performs a specific political function grouping a wide swath of people together who could never fall under a single “racial” category—except as a marker of nonwhite identity.
*6 Part of this segregation owes to the reinforcement mechanisms of consumer society itself, which are merely being amplified by targeted advertising. But targeting introduces an emergent element of bias.
*7 A substantial portion of social science research has been found to be unreproducible—over half of all studies, according to a 2015 study—and my first reaction to any bold claim is now skepticism.
*8 The field of “sentiment analysis,” which tries to gauge the emotion, positive or negative, behind a particular piece of text or a mention of a particular person or object, is plagued with problems, such as thinking “I am not happy” is a positive and optimistic sentiment because it mentions the word “happy.”
*9 Facebook backed off tracking efforts like its 2007 cross-site tracking program, Beacon, as they generated negative press, but subsequently redeployed similar efforts in subtler and more extensive ways, resulting in the Everest of personal information made use of by everyone from lawyers to drug companies to political consultants.
*10 Like many things we call laws, these are soft laws—heuristics and guidelines—rather than inviolable iron laws.
*11 This is not to say that Facebook or other social engines will be the largest companies. Apple, Google, or other companies may continue to hold greater dominance in technology more generally, through their operating system platforms and other initiatives like self-driving cars and network infrastructure. In the online social sphere, however, the loose aggregative nature of Google has already given way to the current dominance of Facebook. Social networks regiment our humanity online.
*12 This law is an analogue to systems engineer and Go creator Rob Pike’s fourth and fifth rules of software engineering, which emphasize why software engineering tends toward simplicity whenever possible: “Rule 4. Fancy algorithms are buggier than simple ones, and they’re much harder to implement. Use simple algorithms as well as simple data structures. Rule 5. Data dominates. If you’ve chosen the right data structures and organized things well, the algorithms will almost always be self-evident. Data structures, not algorithms, are central to programming.” Linus Torvalds expressed a similar sentiment: “Bad programmers worry about the code. Good programmers worry about data structures and their relationships.”
*13 Wired cheerfully observed that Facebook’s Reactions “may not reflect the world in which we live, but they’re a good deal closer to the one we want.” This assumes that the world you want is one of standardized, superficial reactions to content devoid of nuance and sophistication.
*14 On the other hand, Keltner and Facebook’s research indicated that stereotypically happy Americans used the “Sad” Reaction more than nearly any other country save for those of the Middle East, Afghanistan, Pakistan, Mexico, Cambodia, Peru, Ireland, Azerbaijan, and a few others. Historically downbeat Russia used “Sad” Reactions significantly less than average. It seems hazardous to generalize from the use of the Reactions to people’s actual emotional states. The types of responses may just not be as unified and universal as the Reactions make it seem.
EPILOGUE: THE REDUCTION OF LANGUAGE, THE FLATTENING OF LIFE
I bought translations of all kinds of my own existence.
—TOM PHILLIPS, A Humument
COMPOSER Alvin Lucier’s masterpiece of minimalism, “I Am Sitting in a Room,” begins with the following text, spoken into a tape recorder:
I am sitting in a room different from the one you are in now. I am recording the sound of my speaking voice and I am going to play it back into the room again and again until the resonant frequencies of the room reinforce themselves so that any semblance of my speech, with perhaps the exception of r-r-r-rhythm, is destroyed. What you will hear, then, are the natural resonant frequencies of the room articulated by speech. I regard this activity nnnnnot so much as a demonstration of a physical fact, but more as a way to s-s-smooth out any irregularities my speech might have.
Quickly, Lucier’s words become incomprehensible. He uses the language of combat—“reinforce,” “destroyed.” What occurs is a battle in which the feedback process overcomes the spoken words. The room slowly triumphs, rendering soft, metallic feedback in place of speech. What survives are the aspects of Lucier’s speech that were most compatible with the room. Lucier’s words, his voice, and his stutter are lost to those aspects that the room most strongly reflects. Instead of words, there are frequencies, just as computers take our words and reduce them to coded numbers, grouped en masse.
Attuned to the new digital spaces in which we exist, we are slowly building a hybridized language that bridges human and machine. This language empowers us to better analyze the resulting data, but it also eliminates our irregularities—that which is lost in the translation to its computational representation. Machines play back to us smoothed and regularized versions of our lives and interactions, precisely encoded and quantified. These are inexact replicas of more complicated phenomena, and yet for the benefits of technology, we accept computers’ version of us as the real thing. Irregularities are treated as pathological exceptions, or else they’re ignored altogether.
The playwright Richard Foreman declared that humans are becoming “pancake people…spread wide and thin as we connect with that vast network of information accessed by the mere touch of a button.” I think that this flattening, or smoothing as Lucier would have it, is indeed taking place, but it is not caused intrinsically by the glut of information surrounding us. Rather, it is caused by a feedback process: we see approximated digital images of ourselves and we take them to be who we truly are. We begin to think we are the room. Our information today is very different from the information we had fifty or even t
wenty years ago, because it is more compatible with computers. It is the language of labels and classifications, a lingua franca for computers and humans alike.
I am, in general, cautiously positive toward technology. I don’t have faith that technology, broadly speaking, is guaranteed to produce a greater good. Rather, I believe that we signed a Faustian contract at the dawn of the modern age. I believe we can address the wide-scale catastrophes threatened by global warming, unsustainable ecology, and the possible failure of our increasingly complex infrastructure only through the development of further, more advanced technologies. We may blow ourselves up before we save ourselves, but I don’t consider halting or reversing technological progress to be a plausible option. Science and technology are terrifying yet amazing, but they also produce a dangerous amnesia as change accelerates. I’ve felt it myself.
I would not trade my life for the life of the most well-off person in ancient Greece, Renaissance England, or Enlightenment France—three periods with which I’m familiar enough to have affection for. But in even the most barbarous and oppressive of historical moments, there remain aspects of past human experience that are edifying, beautiful, and even occasionally uplifting. This too is knowledge, and it is the sort of knowledge that is most likely to be lost in the ever-growing move toward capturing human experience in computation. This knowledge can be only approximately communicated through human language, so difficult for computers to grasp.
In fearing this loss, I am a conservative in the way that the Marxist philosopher Gerald Cohen described himself:
Bitwise Page 27