It’s the Memes, Stupid
Photo credit: TravisTheGeek, DeviantArt
Donald J. Trump will be the next President of the United States of America.
I have no particular insights about how to deal with what I think is an political outcome as earth-shattering for America as Third Impact was in End of Evangelion. I will be learning as much as everyone else. Like the ending scene of End of Evangelion, we’re likely going to be left with the political equivalent of a kind of desolate wasteland that we have to somehow make something out of despite mutual mistrust, disgust, and hatred. I do have a lot of thoughts as to what happened tonight, much of them stated better by others. There is a lot about this election that is not surprising. You would have to be willfully ignorant not to see it (which was sadly the case for many).
But I’m not going to bore you with my opinions about this. They’re irrelevant now, in any event. And likely to trigger useless arguments and recriminations with friends on social media in an environment of widespread panic, shock, and gloom. This has been a chaotic night in which the electoral results have defied virtually every credible expert’s assessment, to put it mildly. There will be, in the coming days and years, a lot of score-settling. The first question that is often asked when a tremendous upheaval occurs is one of who is to blame for causing it/failing to prevent it. There will be a lot of this. A subsidiary one is either crowing about getting it right or placing blame for getting it wrong. You are going to read a lot of both types of hot take very soon. They are both, however, responses to a general need to make sense out of a world that does not make sense anymore.
I’m writing this mostly as a way of helping myself be able to go to sleep and keep on going tomorrow. To do that, I have to try to understand what happened in my own terms, while the initial feeling of shock is fresh and before the rest of what will likely be a week of various press attempts to make a series of either overly narrow or hindsight explanations color our view of what has been a cataclysmic political upset. Given that this is a first reaction, it is at best semi-coherent and a mixture of comments on various topics. Some kind of accounting for what went wrong is necessary, though I doubt anyone will agree on the specifics (again, I’m biased towards a certain class of specifics).
The broader question is why so many people – even people that called or expertly analyzed much of 2016 – missed what happened tonight, and how it fits into the idea of the world many of us have internalized. The former is going to be debated for a while, and the latter is something only you can determine for yourself. This is my own attempt, however impressionistic and imperfect. If there is an common theme in it, it is that we have far more limitations to our ability to understand the social world than we think and all-too-human factors that mitigate against it. And that any kind of individual or collective response to what happened tonight requires accepting that our assumptions are fatally flawed and that we will need to think about better ones that can somehow replace the ideas that went down in flames tonight.
I’ll start with the polls, as it seems like the most logical place to begin. I doubt that anyone is in the mood for a discussion of polling, but it is also inevitable. What to make of a situation in which virtually all credible outlets blew it? Even FiveThirtyEight – which was the most bullish on Trump’s chances and most skeptical of the information used in the conventional wisdom everywhere else – was a failure. As Chris Arnade observed, FiveThirtyEight treated treated the presidential race as highly volatile – to the point of producing a model that essentially was little different than a snapshot of the polls at time T. If the race was really so uncertain as to justify such an approach, which mirrored the very “horse race” press punditry FiveThirtyEight head honcho Nate Silver so disliked, then it should have – as Arnade argued – simply just said so. Reported the odds via spot polls day-by-day. And said “well, it is what it is.” Instead, as Arnade pointed out:
“[FiveThirtyEight] doesn’t smooth ([v]ery much) the short gyrations in the polls. So, the volatility of the race is high. But…. [t]hat inconsistent with how they look at future where they mean revert, & dampen movement. If they didn’t race would always be close 50/50. ….This is great for page views, because it is always a horse race, but never a dull 50/50 race! Put another way. It is absurd to say, today odds are 68.7, tomorrow 70.1. I mean. You either simply report todays odds via spot polls, and then say, hey shit is volatile, but here it is today. You don’t go next step. The reality is nobody really knows. Giving it a probability with one significant digit is pretending you know shit.”
FiveThirtyEight may have, out of all of the pollsters, called the election most accurately by anticipating polling errors. But to me at least the site’s credibility is frankly gone. Models are useful not because they predict well, but because of how and why they predict. As Arnade suggests, the logical implications of the assumptions that the FiveThirtyEight team took concerning 2016 are at variance with the choices that the site made about how to use and explain the model.
The honorable thing to do would have been to simply admit that the journalists were sometimes right about the limitations of long-term forecasting and acknowledge the validity of viewing a year like 2016 as one of simply (from the standpoint of the observer) a neck and neck horse race. Silver’s rival Sam Wang of the Princeton Election Consortium was dead, dead wrong about the race but his model also transparently and consistently reflected (very wrong) assumptions about the race and how it was measured. If Wang learns from the experience, I’ll pay attention to whatever else he does. As for Silver, FiveThirtyEight’s credibility is deader to me than a certain Western Lowlands Silverback gorilla that used to inhabit the Cincinnati Zoo. But frankly Silver can claim a lot of company in having squandered so much credibility.
During the aftermath of FiveThirtyEight’s prediction failure during the primary, complex systems scientist Yaneer Bar-Yam wrote that Silver had acknowledged errors in his specific statistical analyses and not in the overall assumptions guiding the analysis:
Nate Silver is one of the most highly regarded statisticians of sports, politics and other domains . During the 2016 presidential campaign, his early analysis of the chances of Donald Trump becoming Republican nominee stands out—he estimated only a 2% probability. Even though statistics are not about actualities but probabilities, subsequent events do not appear to be consistent with those predictions, as he later acknowledged [2–4]. He has explained the problem with the analysis as due to political factors , and in terms of the difficulty of analysis , but not why the model he used is essentially flawed. Here we point out fundamental problems with the statistical ideas he uses. Statistics begins from an assumption of independence, which is generally not valid. In this case, the assumptions lead to mathematical inconsistencies. ……
…..Statistical assumptions are used because they make calculations possible. But if the assumptions are wrong, so are the calculations. What should be done? Silver has written a thoughtful lessons learned  pointing to the importance of complexity, feedback loops and chaotic dynamics. Incorporating the mathematical frameworks that these processes refer to will advance analyses be- yond statistics to enable better mathematical prediction. Being concerned about interdependence, like the con- cerns about Brexit causing problems for Europe, is not enough. We need to understand interdependence  in order to make correct assumptions, and derive correct conclusions.
As obvious as these ideas may seem to readers, they represent an “out of equilibrium” view of the world that is a minority one in the social sciences. On that note, it is difficult to overstate how much the 2016 electoral results – and the entire year, really – represent a fundamental rebuke to the consensus view of the world held by so many institutions and individuals that produce ideas about it. This includes myself, because while I am not as shocked about a Trump victory as some of my peers I still nonetheless did not believe it would happen and acted as if it would not happen.
Many people will invent a new pet theory to describe how the results were obvious in retrospect. But I would be skeptical of such explanations. Everything is obvious once you already know the answer. Nor should you take much stock in people who got the result right due to metrics that owe more to thinking as wishful as that of those who believed Trump would lose. I continue to believe that none of the competing explanations for 2016 make sense, and its difficult to create explanations for social upheavals in general without tautology. So what to make of 2016, then?
Decoherence Events and Illumination
David Auerbach describes something he calls an Decoherence Event in the novels of Thomas Pynchon:
And looking back, the two-thirds marker seems to have always held a special liminal significance for Mr. Pynchon as a place where, instead of coming to a climax, his narrative signals its dispersal with what could be called a Decoherence Event, in which our working models of reality cease to function together. ((Decoherence is sometimes identified with quantum wavefunction collapse, but decoherence is merely an explanatory model for, among other things, the observance of wavefunction collapse, generated by the loss of a local system’s information to its surrounding environment. In this analogy, we lose the coherent superposition of our (conflicting and incomplete) accounts of the world to a decoherence event. The difficulties in avoiding unwarranted causal assumptions in interpretations of quantum mechanics—as well as the sheer abstraction of the underlying models—may be why Mr. Pynchon has avoided using quantum mechanics in his novels.)) The realms of death and life blur during these events. There are ghosts or other undead entities (TV zombies, Anubis, deaf-mutes, brainwashed narcs), devoid of personality and “scattered” like Slothrop. The final third of each novel then deals with tentative attempts to piece back together the tenuous order (any order, really) which this moment shattered.
There is, of course, a limit to how far we can take this image without getting into the hippy-dippy view of quantum physics, and I have no desire to do so. Interpreted as a sociological thing, decoherence events occur because, paradoxically, the rules and orderly that govern our working models of social reality are their own undoing. Interpreted one way, American politics and society has certain structures and mechanisms that generate understandable and sometimes predictable outcomes. On the other hand, those structures and mechanisms are subject to Clausewitzian friction, meaning a large and complex series of mechanisms working in concert create more opportunities for real world systems to deviate from idealized plans and visions of the future.
It’s been clear from the beginning that 2016 has been a transformational year. It highlights the reality that all of our predictions are based on assumptions and heuristics that could very well be wrong or incorrect in varying ways. My own view of the purpose of social science has changed a lot, not just as a result of a chaotic election year but also of a slow journey towards seeing the problem in terms of not prediction per but of “illumination.” We’re not in a pristine laboratory where we can manipulate inputs and outputs and experimental controls. Instead, we have an unruly external world that we barely understand. Our own observations and measurements also change the external things we observe or our perceptions of the “data” itself. Having more powerful and precise tools doesn’t change this. In fact, it heightens the problem.
All of this is worsened by the fact that our ideas about how the social world works aren’t rooted in well-known physical forces that we can directly measure. Instead, they’re a product of elaborate folk constructs (“institutions”, “legitimacy”, “nation-states”, “social structure”) that we treat as real and various haphazard ways of trying to somehow measure them. And all of us – myself obviously included – tend to have a mixture of status quo bias and survivorship bias. If we think that the current steady state is OK, we tend to rationalize ways why it is the best of all possible worlds or why things tend to seek some kind of steady, stable state. Why the system ‘works’ according to the way that we think it should. We have all received a very expensive lesson in why this is not the case. What you make of it depends a lot on how you much you preferred the outcome.
So what do I mean by illumination? I mean bringing to light things that we didn’t know about the way the world works. That doesn’t really mean prediction (which seems to be a farce right now) or a nice P-value, although both of those things are appropriate in context. It means that producing greater understanding is good and ought to be valued even if it doesn’t lead to precision or certainty. It also means that the process by which people build models of the world is more important than the outcome. The two most valuable people I followed for 2016 analysis – David Auerbach and Chris Arnade – were not predictive modelers. Heck, both of them were wrong about the boolean (Trump/Not Trump) outcome of the election even if they always took Trump more seriously than other observers. But they were able to see things that others did not even if what they saw was neither parsimonious or useful for statistical electoral prediction models.
Auerbach diagnosed a number of political factors – from the nature of the media to the fundamental lack of legitimacy in the “system” perceived by many in the 2016 election – that were not covered by the vast majority of people writing about 2016. Auerbach’s analysis stands out for its eclectic and wide-ranging nature and willingness to point at multiple overlapping sources of failure. Arnade was willing to do what most writing about 2016 were not – to drive around America and talk to people who – depending on the analyst – were either Rational Peasants out of a Cold War modernization theory textbook or frothing maniacs. Arnade did not gloss over his interview subjects’ less-savory sides or pretend they were perfect victims, but he did try to use what he learned to make sense of the numbers he was trained to analyze as a former Wall Street quant.
I know I will, no matter what happens over the next four years, put Auerbach and Arnade at the top of my reading list. They illuminated something that was poorly understood and challenged my pre-existing beliefs about it. It turns out we were all dead wrong at the end of the day, but the insights that Auerbach and Arnade brought to an otherwise ridiculous and ad-hoc series of folk sociological debates stand out. I hope that people who previously dismissed them pay attention and learn from them. But even Auerbach and Arnade are not a substitute for making a good-faith effort to understand and make sense of the changes we see unfolding in front of us. On an everyday, individual, basis.
It’s The Memes, Stupid
To put it bluntly, the world that I grew up with is dead. Not just in terms of politics and society, but the fundamental assumptions I have about how and why the world works are now irrelevant.The strength of my beliefs in those assumptions had been, after all, decaying ever since the fateful years of 2001-2003, but now they’re completely kaput. I have to face the reality that they are worthless and more of a harm to me and my loved ones than a benefit. As of right now, I plan to still finish my PhD at some point. But as to what comes afterwards, I’m much less sure. Many of the knowledge-producing institutions that I always dreamed of being a part of and pursued a PhD to be a part of look like a silly joke right now. I’m less sure about other things as well. I had been successively putting Papers We Love in Strategy on the back burner as the craziness of the election combined with other commitments distracted me.
I had planned to seriously start it up when the election was over. Now, I frankly don’t think I am going to bother. I don’t think even strategy – my first and true intellectual love – can be useful to me given the difficulty of applying any kind of organized body of knowledge without adjusting the basic assumptions I have about how the world works. If that is not reliable, then nothing else is. I’d rather spend the time I would have devoted to the venture of the Papers We Love community either reading about new frameworks for looking at the world or working on teaching myself more about how to manipulate computers (a key factor in the 2016 outcome, as Sean Gourley noted).
I’m also not sure, at this point, that the typical ways of formulating and communicating knowledge about the world that I was grown up to idolize – policy analysis and academic research – are as valuable as they were when I first set out on the path I am on now. Over the years I’ve become keenly aware of how much competing forms of representing and communicating knowledge are making these styles of representation and communication irrelevant or at the very minimum making them far less powerful. For the most part, analysis in the media has been replaced by hot takes, clickbait, and tribalism. In the policy realm, its less and less clear that policy analysis gives useful answers or even supplies useful questions. I’ve already said my piece about what my ideal for social science is – illumination – and academia in the US at least has less and less place for that style of inquiry. So what now? If the old ways of thinking about our country and politics are dead, what should replace them?
I don’t know, but it begins from a very different view of ourselves and the environments we inhabit, especially as they are transformed by information and computation. I said earlier that we don’t live in a pristine laboratory that can be perfectly experimentally manipulated. I hinted, however, that we may be a part of some other kind of laboratory. Specifically, we are living in a kind of “world laboratory” that Brian Holmes aptly described here:
What does it mean to be part of a cybernetic system? For a conscious human being it means taking part in an evolving loop, where you are both the subject and the object of experimentation. This is the relation that has developed between scientific inquiry and world-changing technology. Researchers reshape the environment that defines them, and vice-versa. Such self-affecting loops are the vectors of a radical constructivism, an artificialization of existence. Their content and their continuous metamorphosis are what gives form to life in a cybernetic society.
It seems that we live in an era of data and information that has had many profound effects on our society. A society that exists within an enormous, accidental, and globe-spanning information and computation architecture described aptly by the Puppetmaster in Ghost in the Shell:
It can also be argued that DNA is nothing more than a program designed to preserve itself. Life has become more complex in the overwhelming sea of information. And life, when organized into species, relies upon genes to be its memory system. So, man is an individual only because of his intangible memory… and memory cannot be defined, but it defines mankind. The advent of computers, and the subsequent accumulation of incalculable data has given rise to a new system of memory and thought parallel to your own. Humanity has underestimated the consequences of computerization.
Hence, “facts” (if they ever were important in the first place) have been overthrown by data. Overarching cultural narratives have been replaced by a kind of animalistic consumer behavior originally seen in otaku subcultures. There is no difference, Auerbach noted, between politics and attention in such a world:
Trump’s political rise is a product of the commodification of attention. As the ballooning of new media and analytics have facilitated the microscopic examination of consumer attention, the analysis has been performed with indifference to the consequences of that attention. Just as Donald Trump does not care why he is loved, worshipped, and feared—no matter what the consequences—we have seen massed content production turn to clickbait, hate clicks, and propaganda in pursuit of viewer eyes. By mindlessly mirroring fear and tribalism, the new media machine has produced a dangerous amount of collateral damage.
Georg Simmel wrote in 1900 that money serves as the abstract form of valuation, making all values commensurable while emptying the metric of any specific content. So it is today with attention: we have moved on from commodifying value to commodifying human attention. In the metrics of pageviews and clicks, the reasons for the attention and its consequences fall away. And so a fame generator like Donald Trump becomes not just a symptom but a catalytic attractor: the news media turned him into a phenomenon in pursuit of attention to their properties, even as the “serious” members of the press denied he could ever become a candidate. After all, he was a strategy for attention, devoid of any political program. Alas, such a distinction between politics and attention is no longer meaningful.
I’d extend Auerbach’s notion further: there is no difference between politics and memes in the world we live in today. Richard Dawkins originally saw memes as a cultural unit of reproduction, a vessel embodying certain forms of cultural ideas, symbols, or practices that are transmitted from one mind to another via social intercourse.
….[O]ne of the most popular of all Twitter hashtags (the “hashtag” being a genetic—or, rather, memetic—marker) was simply the word “#Viral.” In the competition for space in our brains and in the culture, the effective combatants are the messages. The new, oblique, looping views of genes and memes have enriched us. They give us paradoxes to write on Möbius strips. “The human world is made of stories, not people,” writes the novelist David Mitchell. “The people the stories use to tell themselves are not to be blamed.” Margaret Atwood writes: “As with all knowledge, once you knew it, you couldn’t imagine how it was that you hadn’t known it before. Like stage magic, knowledge before you knew it took place before your very eyes, but you were looking elsewhere.” Nearing death, John Updike reflected on A life poured into words—apparent waste intended to preserve the thing consumed.
The idea of an “Internet meme” takes it one step further. Limor Shiffman identifies these kinds of memes as:
a piece of digital content that spreads quickly around the web in various iterations and becomes a shared cultural experience….. digital content units with common characteristics, created with awareness of each other, and circulated, imitated, and transformed via the Internet by many users.
Both definitions also imply larger algorithmic mechanisms of variation, mutation, competition, and inheritance governing how memes (broadly or thinly defined) work. And that they can emerge in unexpected and often highly novel ways, like the glitchy patterns of online culture in general. Crucially, memes and other forms of digital culture such as political bots now rival traditional forms of knowledge and culture and in some ways replace or supersede them. In 2016, the memes won. Not in the sense of Trump winning. But in the way in which memes of varying sorts became the dominant way that we communicate with each other and understand the world we live in.
Any response to this year has to take this into account. And do so without yielding to the temptation to pine after an era in which was this was not so or demonize cyberculture in general a wretched hive of scum and villainy. People will argue for the next four years about whether it was the economy, cultural resentment against elites, or racism that elected Trump. I wish anyone sure of a mono-causal answer luck in defending it. But at the risk of yielding to one myself (and to paraphrase a popular saying in American politics) – “it’s the memes, stupid.” It will be hard to explain 2016 as a sociological experience to children and grandchildren without talking about memes. I have a opinion, as does Auerbach, about what the simplest explanations for 2016 are. And to some degree neither of us were completely surprised. Again, I also think my opinion is probably irrelevant at this point because the best time for it to have been taken seriously was at least a year ago. But even with this opinion in mind you can’t talk about the endless nightmare that was 2016 without talking about memes.
Subcultural memes, bots, and other forms of technology that represent, shape, distort, mutate, select, reproduce, combine, or generate information are not only sources of political power, they are also significant and under-analyed features of contemporary society. Memes and bots are both alike in that they are forms of automation – memes (in the Dawkins telling) almost robotically replicate themselves, and computer programs of varying degrees of complexity or simplicity also increasingly outnumber humans in social forums like Twitter. The Puppetmaster said in Ghost in the Shell that humankind has underestimated the consequences of computerization. This was a gross understatement. If there is no distinction between politics and memes (or other forms of cyberculture), we have a long road ahead in which we have to adapt to the consequences.