Does not compute
How social media breaks society
There are days when I think that the history of our era, whenever it is written, will say that we were a society with manageable problems until the internet came along and ruined everything. I’m not completely convinced of this; not yet. But there are days when I think it, and more of them all the time.
I am not alone. In a recent essay, the social psychologist Jonathan Haidt argues that “the past ten years of American life have been uniquely stupid”, owing to the damage social-media has done to social cohesion and trust. It certainly does appear that our capacity for collective reasoning and cooperation has eroded—a point I made in a dispatch last year on “the stupid society”. But if social media, broadly defined, is in fact eating holes in the social fabric, enabling anti-social behavior and ultimately contributing to the destruction of democracy, how can we know?
There are studies which try to assess social media’s effects, as Haidt notes:
Recent academic studies suggest that social media is indeed corrosive to trust in governments, news media, and people and institutions in general. A working paper that offers the most comprehensive review of the research, led by the social scientists Philipp Lorenz-Spreen and Lisa Oswald, concludes that “the large majority of reported associations between digital media use and trust appear to be detrimental for democracy.” The literature is complex—some studies show benefits, particularly in less developed democracies—but the review found that, on balance, social media amplifies political polarization; foments populism, especially right-wing populism; and is associated with the spread of misinformation.
And as Haidt says in comments to Gideon Lewis-Kraus, the author of a new New Yorker piece on the subject, the evidentiary standard here need not be knowledge beyond a reasonable doubt. Given the stakes, a “preponderance of evidence” that social media is harmful is probably enough to start thinking about taking action to save ourselves.
Do we even have that though? As Lewis-Kraus notes, Haidt, along with Chris Bail, has created a public Google doc which gathers together academic research and other resources focused on social media’s effects on individuals and groups. And while there are plenty of worrying results there for those looking to find them, the overall picture is a murky one. Here’s the New Yorker again:
Bail told me that, over all, he was less confident than Haidt that the available evidence lines up clearly against the platforms. “Maybe there’s a slight majority of studies that say that social media is a net negative, at least in the West, and maybe it’s doing some good in the rest of the world.”
So what do we do? Sadly, we don’t get to experiment on a thousand separate but identical macro-societies, which we could conveniently divide into treatment and control groups in order to see how often Facebook leads to social collapse. And it seems to me that one thing to understand about this huge and important question is that it illustrates the limitations of careful empirical work in illuminating questions relating to the operation of complex social systems. That is a lesson with broader applications, for a number of important social sciences.
But if there is some meaningful chance that social media really is quite destructive—and the evidence, murky as it is, is not necessarily inconsistent with that possibility—then simply shrugging our shoulders isn’t enough. We need to grapple with this question as best we can.
It would be helpful if we had a clearer notion of how exactly it is that society does what it does when it functions cohesively. Our understanding of this issue is not all that good, partly because it is a very difficult thing to try to understand, but maybe also a little because the most prestigious and influential social science that we’ve got has made almost no effort to work out a realistic picture of how the disaggregated behavior of individuals adds up to a stable and complex society.
We do have some things to go on. The implicit model that a lot of people seem to be working with when thinking about these questions—in which good societies are those in which rational individuals have access to good information which they use to make good decisions—seems way off the mark. That implicit model seems to inform many of the research questions that one can find in Haidt’s Google doc. The worry many people seem to have is that individuals who had been receiving “good” information instead find themselves routed via social media into echo chambers, or places which are conduits for misinformation, or dens of extremist ideas. If research shows that this is not happening to all that great an extent, then perhaps worries about the effects of social media are overstated.
But society doesn’t work like that—and thank goodness. If it took a population of rational individuals drawing sensible conclusions from good information to make a stable and prosperous society, we’d be truly fucked.
The truth is likely much closer to the model Haidt is working with: in which the secret to social success is cooperation in the carrying-out of complex tasks, and in which social trust and shared narratives facilitate that cooperation. The hard-core reasoning that makes society tick doesn’t happen at the level of a single individual; it happens in interactions among many individuals. This collective information-processing work sometimes happens through markets; economists aren’t wrong to identify market mechanisms as a powerful means through which we facilitate collective cooperation at massive scale. But computational cooperation also happens at a deeper level, helping to create and sustain the societies in which market activity takes place.
It does so through our social connections. We respond to incoming information in the way that we do based on our sense of who we are and what a person like us is supposed to do, based on the social context in which we find ourselves, based on our understanding of what incoming information means—where meaning is in large part a matter of social convention. Our social connections shape our identity and help orient us within society. And the structure of all these connections across all of society represents a system which, in aggregate, handles a lot of complex computation. Stuff happens; our response, as individuals, is informed by our place in the system; all those responses, aggregated, yield society.
This system is an evolving one. People are born; people die; new relationships are forged; new ideas are hatched; new norms develop. We are nodes in an information-processing system, but we’re not automatons. When the world around us is changing, we reason about what’s happening and why and what we ought to do about it, and as a consequence some people amend their beliefs and perhaps also their place in the system, which subsequently functions differently than it did before. And so in some sense, society is adapted to prevailing conditions; given the sorts of stuff that tends to happen to our societies, we’ve self-organized in a way that yields broadly acceptable results. If that weren’t the case, then people would either kick up a fuss in ways which improve those results, or else society would collapse into something less complex.
If society functions in something like this fashion, then the arrival of the internet cannot help but act as a massive and destabilizing shock. Use of the internet crowds out a lot of activity that used to be common, and which formed part of the social information-processing system. We now have different sorts of interactions with different sorts of people in different sorts of contexts. The new sorts of interactions that happen across social media lead to a kind of social disorientation; the social landmarks we used to use to remind ourselves who we are and what we’re doing have vanished. Identities and norms have thus become more fluid, further altering the operation of the information-processing system. Then on top of that, the internet is altering real-world economic and human geography, creating still more disruption. Some industries are thriving while others decline; migration patterns are shifting; work and consumption habits are changing.
These shifts change, in small ways and large, how we respond to incoming information, yielding an aggregate social response which is…well, look around. The problem is not so much one of garbage in, garbage out than one of a misfiring motherboard. We’ve hacked ourselves, by signing up en masse to this radically new way of interacting with other humans.
Are there historical analogues from which we can learn something about our present situation? One, which I’ve mentioned before, is the experience of industrialization and urbanization which unfolded from the late 18th to the early 20th century. The explosive growth of industrial cities disrupted long-established social relationships and patterns of behavior, and created vast new settings in which human interactions took place. These new settings led to the emergence of strange new social problems, and provided a breeding ground for all sorts of contagious things: from deadly pathogens to radical ideologies.
There were lots of good things that came along with the emergence of industrial cities. But it took a long period of social malfunction and an unprecedented wave of social activism and mobilization to turn cities from politically volatile death traps into places which were functional and livable, and which constituted the social and economic basis of stable, prosperous democracies. The technological and economic forces which led to mass urbanization were a shock to the prevailing information-processing system, but over time society rewired itself such that its collective response to incoming information came to yield outcomes that were broadly acceptable.
To the extent that the experience of urbanization is a useful guide to our present circumstances, should we be optimistic or pessimistic? On the one hand, urbanization led ultimately to a place that most people would agree was clearly superior to pre-industrial life. On the other hand, city life was very bad for lots of people for long stretches of time. It created conditions favorable to the emergence and spread of a variety of noxious political ideas, with world historical consequences. It produced social ills for which we continue to lack good solutions today. And to return to an earlier point: we have no way of knowing the frequency with which mass industrialization and urbanization leads to social collapse of one form or another; we don’t get to run the experiment more than once.
Still, I think it is right to imagine that there is in the emergence of the internet the possibility that we eventually arrive at a much better place than we were pre-net. There is scope for optimism. But there is also considerable room for pessimism. If one were inclined to be particularly gloomy, one might observe that the internet is to a first approximation a purely informational sort of shock. That means, first, that it can transform society far more quickly than a change which is at least as intense in the movement of atoms as of bits. We’ve all gotten online far faster than our forebears became urban, in a way which probably makes the task we face harder.
And, second, I worry that this shock, by its very nature, undermines our capacity for adjustment. In the 19th century, people in an urban setting interacted with others face-to-face, just as people had in a rural setting. Trust may have been lower in the urban setting, and identities and norms more fluid, but interactions nonetheless took place within a clear context which the parties to an interaction could perceive and agree upon, and which could form the basis of new conventions and settled meanings. But today? More than a decade into the social-media experiment, context still proves elusive, again and again. This makes it so incredibly easy to miscommunicate, to misconstrue. It is a trust-shattering phenomenon, which makes the task of rewiring formidably difficult. And I worry, on some days, that it has fundamentally broken society.
If we are going to overcome all of this, though, I think we have to understand the nature of the problem. Because the issue isn’t just one in which some individuals receive bad information, but rather one in which we all engage with each other a little differently, we are all part of the problem. And because we are all part of the problem, we could conceivably improve upon the situation by modeling better behavior, working together to establish healthier norms, and exercising self-restraint. We can be compassionate and forgiving. We can remember that these networks want us to perform, and that we can instead choose not to. We can be silent, or gracious, or both. We can give others an opportunity to provide more context, and we can express more of our ideas in places that allow for the communication of more context: in blogs, or even in person.
We can do better. We have to, if we want to get through this.