by Clay Shirky
The fact that people are all talking to one another in these small clusters also explains why bloggers with a dozen readers don’t have a small audience: they don’t have an audience at all, they just have friends. In fact, as blogging was getting popular at the beginning of this decade, the blogging software with the most loyal users was none other than LiveJournal, which had more clusters of friends blogging for one another than any other blogging tool. If blogging were primarily about getting a big audience, LiveJournal should have suffered the most from disappointed users abandoning the service, but the opposite was the case. Writing things for your friends to read and reading what your friends write creates a different kind of pleasure than writing for an audience. Before the internet went mainstream, it took considerable effort to say something that would be heard by a significant number of people, so we regard any publicly available material as being offered directly to us. Now that the cost of posting things in a global medium has collapsed, much of what gets posted on any given day is in public but not for the public.
Fame Happens
It’s also possible to make the opposite mistake: not that conversational utterances are publishing, but that all publications are now part of a conversation. This view is common, though, and is based on the obvious notion that the Web is different from broadcast media like TV because the Web can support real interaction among users.
In this view, the effects of television are mainly caused by its technological limits. Television has millions of inbound arrows—viewers watching the screen—and no outbound arrows at all. You can see Oprah; Oprah can’t see you. On the Web, by contrast, the arrows of attention are all potentially reciprocal; anyone can point to anyone else, regardless of geography, infrastructure, or other limits. If Oprah had a weblog, you could link to her, and she could link to you. This potential seems as if it should allow everyone to interact with everyone else, undoing the one-way nature of television. But calling that potential interactivity would be like calling a newspaper interactive because it publishes letters to the editor.
The Web makes interactivity technologically possible, but what technology giveth, social factors taketh away. In the case of the famous, any potential interactivity is squashed, because fame isn’t an attitude, and it isn’t technological artifact. Fame is simply an imbalance between inbound and outbound attention, more arrows pointing in than out. Two things have to happen for someone to be famous, neither of them related to technology. The first is scale: he or she has to have some minimum amount of attention, an audience in the thousands or more. (This is why the internet version of the Warhol quote—“In the future everyone will be famous to fifteen people”—is appealing but wrong.) Second, he or she has to be unable to reciprocate. We know this pattern from television; audiences for the most popular shows are huge, and reciprocal attention is technologically impossible. We believed (often because we wanted to believe) that technical limits caused this imbalance in attention. When weblogs and other forms of interactive media began to spread, they enabled direct, unfiltered conversation among all parties and removed the structural imbalances of fame. This removal of the technological limits has exposed a second set of social ones.
Though the possibility of two-way links is profoundly good, it is not a cure-all. On the Web interactivity has no technological limits, but it does still have strong cognitive limits: no matter who you are, you can only read so many weblogs, can trade e-mail with only so many people, and so on. Oprah has e-mail, but her address would become useless the minute it became public. These social constraints mean that even when a medium is two-way, its most popular practitioners will be forced into a one-way pattern. Whether Oprah wants to talk to each and every member of her audience is irrelevant: Oprah can’t talk to even a fraction of a percent of her audience, ever, because she is famous, which means she is the recipient of more attention than she can return in any medium. These social constraints didn’t much matter at small scale. In the early days of weblogs (prior to 2002, roughly) there was a remarkable and loose-jointed conversation among webloggers of all stripes, and those with a reasonable posting tempo could count themselves one of the party. In those days weblogging was mainly an interactive pursuit, and it happened so naturally that it was easy to imagine that interactivity was a basic part of the bargain.
Then things got urban, with millions of bloggers and readers. At this point social limits kicked in. If you have a weblog, and a thousand other webloggers point to you, you cannot read what they are saying, much less react. More is different: cities are not just large towns, and a big audience is not just a small one cloned many times. The limits on interaction that come with scale are hard to detect, because every visible aspect of the system stays the same. Nothing about the software or the users changes, but the increased population still alters the circumstances beyond your control. In this situation, no matter how assiduously someone wants to interact with their readers, the growing audience will ultimately defeat that possibility. Someone blogging alongside a handful of friends can read everything those friends write and can respond to any comments their friends make—the scale is small enough to allow for a real conversation. Someone writing for thousands of people, though, or millions, has to start choosing who to respond to and who to ignore, and over time, ignore becomes the default choice. They have, in a word, become famous.
Glenn Reynolds, a homegrown hero of the weblog world, reports over a million unique viewers a month for Instapundit .com, a circulation that would put him comfortably in the top twenty daily papers in the United States. You can see how interactivity is defeated by an audience of this size—spending even a minute a month interacting with just ten thousand of his readers (only one percent of his total audience) would take forty hours a week. This is what “interactivity” looks like at this scale—no interaction at all with almost all of the audience, and infrequent and minuscule interaction with the rest, and it has implications for media of all types. Weblogs won’t destroy the one-way mirror of fame, and “interactive TV” is an oxymoron, because gathering an audience at TV scale defeats anything more interactive than voting for someone on American Idol.
The surprise held out by social tools like weblogs is that scale alone, even in a medium that allows for two-way connections, is enough to create and sustain the imbalance of fame. The mere technological possibility of reply isn’t enough to overcome the human limits on attention. Charles Lindbergh couldn’t bear to let anyone else answer his fan mail, promising himself he would get around to it eventually (which, of course, he never did). Egalitarianism is possible only in small social systems. Once a medium gets past a certain size, fame is a forced move. Early reports of the death of traditional media portrayed the Web as a kind of anti-TV—two-way where TV is one-way, interactive where TV is passive, and (implicitly) good where TV is bad. Now we know that the Web is not a perfect antidote to the problems of mass media, because some of those problems are human and are not amenable to technological fixes. This is bad news for that school of media criticism that has assumed that the authorities are keeping the masses down. In the weblog world there are no authorities, only masses, and yet the accumulated weight of attention continues to create the kind of imbalances we associate with traditional media.
The famous are different from you and me, because they cannot return or even acknowledge the attention they get, and technology cannot change that. If we want large systems where attention is unconstrained, fame will be an inevitable by-product, and as our systems get larger, its effects will become more pronounced, not less. A version of this is happening with e-mail—because it is easier to ask a question than to answer it, we get the curious effect of a group of people all able to overwhelm one another by asking, cumulatively, more questions than they can cumulatively answer. As Merlin Mann, a software usability expert, describes the pattern:
Email is such a funny thing. People hand you these single little messages that are no heavier than a river pebble. But it doesn’t take long unti
l you have acquired a pile of pebbles that’s taller than you and heavier than you could ever hope to move, even if you wanted to do it over a few dozen trips. But for the person who took the time to hand you their pebble, it seems outrageous that you can’t handle that one tiny thing. “What ‘pile’? It’s just a pebble!”
E-mail, and particularly the ability to create group conversations effortlessly without needing the permission of the recipients, is providing a way for an increasing number of us to experience the downside of fame, which is being unable to reciprocate in the way our friends and colleagues would like us to.
The limiting effect of scale on interaction is bad news for people hoping for the dawning of an egalitarian age ushered in by our social tools. We can hope that fame will become more dynamic, and that the elevation to fame will be more bottom-up, but we can no longer hope for a world where everyone can interact with everyone else. Whatever the technology, our social constraints will mean that the famous of the world will always be with us. The people with too much inbound attention live in a different environment from everyone else; to paraphrase F. Scott Fitzgerald, the attention-rich are different from you and me, in ways that are not caused by the media they use, and in ways that won’t go away even when new media arrive.
For the last fifty years the two most important communications media in most people’s lives were the telephone and television: different media with different functions. It turns out that the difference between conversational tools and broadcast tools was arbitrary, but the difference between conversing and broadcasting is real. Even in a medium that allowed for perfect interactivity for all participants (something we have a reasonable approximation of today), the limits of human cognition will mean that scale alone will kill conversation. In such a medium, even without any professional bottlenecks or forced passivity, fame happens.
Filtering as a Tool for Communities of Practice
Comparisons between the neatness of traditional media and the messiness of social media often overlook the fact that the comparison isn’t just between systems of production but between systems of filtering as well. You can see how critical filtering tools are to the traditional landscape if you imagine taking a good-sized bookstore, picking it up, and shaking its contents out onto a football field. Somewhere in the resulting pile of books lie the works of Aristotle, Newton, and Auden, but if you wade in and start picking up books at random, you’re much likelier to get Love’s Tender Fury and Chicken Soup for the Hoosier Soul. We’re so used to the way a bookstore is laid out that we don’t notice how much prior knowledge we need to have about its layout and categories for it to be even minimally useful. As the investor Esther Dyson says, “When we call something intuitive, we often mean familiar.”
The hidden contours of the filtering problem shaped much of what is familiar about older forms of media. Television shows, for instance, come in units of half an hour, not because the creators of television discovered that that is the aesthetically ideal unit of time, but because audiences had to remember when their favorite shows were on. A show that starts at 7:51 and goes on until 8:47 is at a considerable disadvantage to a show that starts at 8:00 and goes till 9:00, and that disadvantage is entirely cognitive—the odd times are simply harder to remember. (It’s hard to have appointment TV if you can’t recall when the appointment is.) The length and time slots of television had nothing to do with video as a medium and everything to do with the need to aid the viewer’s memory. Similarly, everything from TV Guide to the rise of content-specific channels on cable like MTV and the Cartoon Network were responses to the problem of helping viewers find their way to interesting material.
Traditional media have a few built-in constraints that make the filtering problem relatively simple. Most important, publishing and broadcasting cost money. Any cost creates some sort of barrier, and the high cost of most traditional media creates high barriers. As a result, there is an upper limit to the number of books, or television shows, or movies that can exist. Since the basic economics of publishing puts a cap on the overall volume of content, it forces every publisher or producer to filter the material in advance.Simply to remain viable, anyone producing traditional media has to decide what to produce and what not to; the good work has to be sorted from the mediocre in advance of publication.
Though the filtering of the good from the mediocre starts as an economic imperative, the public enjoys the value of that filtering as well, because we have historically relied on the publisher’s judgment to help ensure minimum standards of quality. Where publishing is hard and expensive, every instance of the written word comes with an implicit promise: someone besides the writer thought this was worth reading. Every book and magazine article and newspaper (as well as every published photo and every bit of broadcast speech or song or bit of video) had to pass through some editorial judgment. You can see this kind of filtering at work whenever someone is referred to as a “published author.” The label is a way of assuring people that some external filter has been applied to the work. (The converse of this effect explains our skepticism about self-published books and the label reserved for publishers who print such books—the vanity press.)
The old ways of filtering were neither universal nor ideal; they were simply good for the technology of the day, and reasonably effective. We were used to them, and now we have to get used to other ways of solving the same problem. Mass amateurization has created a filtering problem vastly larger than we had with traditional media, so much larger, in fact, that many of the old solutions are simply broken. The brute economic logic of allowing anyone to create anything and make it available to anyone creates such a staggering volume of new material, every day, that no group of professionals will be adequate to filter the material. Mass amateurization of publishing makes mass amateurization of filtering a forced move. Filter-then-publish, whatever its advantages, rested on a scarcity of media that is a thing of the past. The expansion of social media means that the only working system is publish-then-filter.
We have lost the clean distinctions between communications media and broadcast media. As social media like MySpace now scale effortlessly between a community of a few and an audience of a few million, the old habit of treating communications tools like the phone differently from broadcast tools like television no longer makes sense. The two patterns shade into each other, and now small group communications and large broadcast outlets all exist as part of a single interconnected ecosystem. This change is the principal source of “user-generated content.” Users—people—have always talked to one another, incessantly and at great length. It’s just that the user-to-user messages were kept separate from older media, like TV and newspapers.
The activities of the amateur creators are self-reinforcing. If people can share their work in an environment where they can also converse with one another, they will begin talking about the things they have shared. As the author and activist Cory Doctorow puts it, “Conversation is king. Content is just something to talk about.” The conversation that forms around shared photos, videos, weblog posts, and the like is often about how to do it better next time—how to be a better photographer or a better writer or a better programmer. The goal of getting better at something is different from the goal of being good at it; there is a pleasure in improving your abilities even if that doesn’t translate into absolute perfection. (As William S. Burroughs, the Beat author, once put it, “If a thing is worth doing, it’s worth doing badly.”) On Flickr, many users create “high dynamic range” photos (HDR), where three exposures of the same shot are combined. The resulting photos are often quite striking, as they have a bigger range of contrast—the brights are brighter and the darks are darker—than any of the individual source photos. Prior to photo-sharing services, anyone looking at such a photo could wonder aloud, “How did they do that?” With photo sharing, every picture is a potential site for social interaction, and viewers can and do ask the question directly, “How did you do that?,” with a real hope of getting an
answer. The conversations attached to these photos are often long and detailed, offering tutorials and advice on the best tools and techniques for creating HDR photos. This form of communication is what the sociologist Etienne Wenger calls a community of practice, a group of people who converse about some shared task in order to get better at it.
John Seely Brown and Paul Duguid, in their book The Social Life of Information, put the dilemma this way: “What if HP [Hewlett-Packard] knew what HP knows?” They had observed that the sum of the individual minds at HP had much more information than the company had access to, even though it was allowed to direct the efforts of those employees. Brown and Duguid documented ways in which employees do better at sharing information with one another directly than when they go through official channels. They noticed that supposedly autonomous Xerox repair people were gathering at a local breakfast spot and trading tips about certain kinds of repairs, thus educating one another in the lore not covered by the manuals. Without any official support, the repair people had formed a community of practice. Seeing this phenomenon, Brown convinced Xerox to give the repair staff walkie-talkies, so they could continue that sort of communication during the day.
By lowering transaction costs, social tools provide a platform for communities of practice. The walkie-talkies make asking and answering “How did you do that?” questions easy. They would seem to transfer the burden from the asker to the answerer, but they also raise the answerer’s status in the community. By providing an opportunity for the visible display of expertise or talent, the public asking of questions creates a motivation to answer in public as well, and that answer, once perfected, persists even if both the original asker and the answerer lose interest. Communities of practice are inherently cooperative, and are beautifully supported by social tools, because that is exactly the kind of community whose members can recruit one another or allow themselves to be found by interested searchers. They can thrive and even grow to enormous size without advertising their existence in public. On Flickr alone there are thousands of groups dedicated to exploring and perfecting certain kinds of photos: landscape and portraiture, of course, but also photos featuring the color red, or those composed of a square photo perfectly framing a circle, or photos of tiny animals clinging to human fingers.
