The travesty of diffusion theory is not that it has displaced overlay theory and source analysis, but that it has been allowed to fertilize a generation of academics and practitioners who liken its inventor and enthusiastic promoter, Suds Beamershiff, to an Einstein of crowd size analytics. Arch and preternaturally adroit in conversation, Beams (as his grad students and lovers call him) turned the narrow and strenuously academic discipline into a distinct ring in the big top (or lower circle, some might say) of contemporary politics with his recent smattering of talk news appearances where he would shake his warm chaos of bangs gelled up above his blond eyebrows as he raised his left index finger to make and hold a point. The camera was as fascinated as the public was and he found himself quickly voted onto the editorial board of Crowd Demography, Science and Philosophy, the preeminent publication for both the practitioner and the cognoscenti. There was scant support for diffusion, but there was genuine new enthusiasm for Beams that even infected the old-schoolers drowning in their musty beards and tweeds. The most obvious comparison to Beams’ rapid rise was the sudden global fascination with Australian Rules Breaking that was shattering expectations about dance, art, and even crowd-sizing issues during street performances in Wollongong and Perth. The Kangaroo Punch Up was gaining mindshare and the masses followed.
All of that overshadows Crowd Analytics 2024 (Crownal 2024) even as the National Institute of Standards and Technology (NIST) has started promoting a competition to ascertain the accuracy of all known methods for analysis, a shoot-out of sorts designed to evaluate the different approaches and enable better depths of crowd insights. The locals ride the Metro to Reston Town Center in their light professionals while the visiting crowd of academics in polos and hoodies bounces from restaurant to bar and then downtown for shadow vacations mashed onto the end of the conference. The coffee is in tankards and the cookies disappear into pockets for a later surreptitious snack to fend off jet lag or hangovers. The talks drag on but then break into raucous Q-and-As with diffusion always being touted as the way forward for any murky failure of the established methods. Too many free variables? Too many hyper parameters? Too much noise? Diffusion might be the answer. Dr. Julian Sartor, an enviably thin statistician and competitive cyclist, his haunting long face punctuated by imperial blue eyes, snapped in frustration at the constant diffusion whiplash, remonstrating an intense Korean man that diffusion was yet to be demonstrated. The audience breathed visibly and the small man retreated into the murky crowd, swiping an angry tear from his right eye.
This is a summary then from the first day of the conference. I am assigned to cover it by the Atlantic and there is an expectation that a long-form article will emerge, something that builds a shadow thesis tinseled up with fascinating crossbeams of history, personality, and reflection. I began a week ago reading as much as I could about the field. Vaulted from obscurity by the national obsession with crowd sizes, Crownal has become an argument on stilts for those who profit from groups or want to. You can think of it like Nielsen ratings for gatherings and it has overtaken views or likes or stars or fresh tomatoes.
There is nevertheless an aridity to much of the research. Much of it is thermodynamic or information-theory in action, substituting people for gas molecules and statistical distributions of bits. The math carries on and is often itself analyzed to ascertain its limits. At the corners are the philosophers, per the journal’s name, who mix together some cultural criticism and meta-analysis of process, technique, and motivations. Preeminent is politics and how it shapes our expectations of crowd analytics. But this isn’t politics in the sense of parties and candidates, but is instead the matrix of goals that the analysts carry with them like their neck pillows for the long plane rides.
Dr. Lela Ornativ is the loudest and most obscure voice in the philosophy of crowd analytics, long holding that even counting people amounts to a form of quantitative coercion, for isn’t every individual just that, an individual? To quantify is to minimize the one and substitute a lump that collapses the richness of personal experience into a cluster of meaninglessness. Lela, the-one (preferred pronoun because “they” promotes a “massing and othering of the self”), promotes a holism for crowds that is not reductive but instead uses neologisms like “grindividuation” and “countisms” to counteract collectivization in sheer bulk numbers.
Beams loves being on panels with the-one, grinning and waving his long finger into spectral flourishes around the air in front of him, like an air conductor for Shostakovichan excesses.
“But what do we accomplish in all of this relabeling, Lela? Is there any output?”
“Exactly the wrong question, Beams. When the field has such a narrow focus on numerics it becomes othering and constructs a false implication that empirical truths and analytical outcomes are the sum total of crowds, when it is the individual that is greater than the sum of the parts.”
“But isn’t that the goal of the field? We don’t diminish people, we merely count them. The counting and estimating is useful but not in any way at the expense of the people themselves.”
“There are existentialized proto-grammars built into everything you just uttered and what is expensed is precisely the individuated self, collapsed into a number, a graph, shown on TV and captured by social media to coerce the mass of people into an amalgam. A count is a weapon. We poll, count, market-to, and estimate, and the soul dissipates.”
“I’m not much into souls, Lela.”
Beams brushes it away, but becomes awkward shortly later when Silas Permian of Rice offers up a novel critique of diffusion theory. A mathematical physicist who jumped fields, Silas is convinced that there is an inherent instability in the analysis, some kind of zeros or “holes” that undercut the entire effort. Beams crinkles his eyes down through his nose at the suggestion and barks that the shoot-out will show Silas wrong. Silas recuperates briefly and retorts there might be unknown consequences if his analysis is correct.
I am rapt at the interchange. Another something to write about. I’m always chasing hints, foreshadowing, intrigue. The world is boringly complete in some ways, with exotic personalities abraded down by the serious matters of daily living. I have to build up these narratives, circle around for justifications and translate them into a train heading for a clever mountain resort. The-one is right at least about me. I can’t just graft together numbers and make a chart. Each person is a story punctuated by their physical traits, intellectual predilections, and minor quirks. There is an inward intensity to Silas that is like a closed room, while Beams is airily transparent. The-one is a semantic gnat buzzing everywhere the-one happens to be. I imagine the-one at a grocery store requesting the produce not be separated by type to create a holism of fruit.
I intercept Silas as he marches towards the men’s room. He seems puzzled at the journalistic interest but briefly sketches out the problems in the math. It all comes down to division-by-zero I think he is claiming. Certain crowd configurations don’t fulfill the equations in the imaginary plane, so if there is an annular ring of people moving smoothly in that circle, like Shakers in an olden dance, the count will move towards infinity. It has to do with the smoothness of the manifold that makes the diffusion self-annihilating. People are diffusing into one another, he says at one point, but it appears to be a reference to tiny slices of that annulus through which the people are flowing. I let him proceed to the loo as he starts shifting on his feet, clearly in distress.
After updating my notes following an awkward group dinner around a disk table with a stained red tablecloth, I email my editor that I think I see a bundle of theses emerging. Between the fracturing of expectations of science, reductionism as an ever-present put-down, and this breakdown of the mathematics of diffusion, there is enough to work with when peppered with a bit of inventive character development. But the shootout and the panel afterwards will finalize the arc. Still, I have my doubts that there will be anything more than inconclusiveness from the NIST-driven affair. It’s always this way in the social sciences, just more and more questions.
I have a few notes that I gathered, along with the rules PDF given to me, on the basic format of the Crowd Analytics Flow Competition at Crownal 2024, as it is formally known. A matronly woman by the name of Claire Wishby heads up the effort, though she has a few Deloitte contractors hanging around and managing things. I briefly interview her before the competition and she is forthright that moving the needle on analytics requires some kind of shared baseline. If every research group is analyzing different crowds, from space or drone footage, or from door access counts, or cell phone connection patterns, it is impossible to separate out the wheat.
But to what end, I rhetorically query as we get through some of the background material and history of crowd analytics. She arches an eyebrow and complains that the press is far too critical about matters that they don’t understand. There is informational gold in crowd analytics, she insists, and the government drone transforms into an advocate. From designing stadiums to stopping famines, knowing how crowds behave is foundational. I apologize and lament that I am most defiantly not writing a hit piece. I just need all the perspectives. We part with her distrust perhaps a bit mollified.
In the end I have to stand on the mezzanine level as the participants break up into groups in the massive lobby of the hotel. The staff moved couches and tables to the carpeted walkways adjacent the central marble floor. The analysis teams are set up near me at mildly decorated tables. I see Beams crack his knuckles as a graduate assistant tips a camera slightly more towards the floor below.
Then they are off, with participants moving in blocks like massing sheep, leaders emerge then step back as the crowds thin and break up. They form starfish shapes for two minutes, then collapse back into chaos. The exact motions and instructions for the participants are unknown to the analytics teams. I glance around the tables and see the Tokyo University team suddenly raise their arms in a joint flourish of success, then return to their laptops.
Minutes pass and there is murmuring concentration at the competitors’ tables as the crowd shifts again, then marches out the front doors of the hotel, then returns again but in piecemeal fashion through side doors and from the room wings. I catch Beams grinning and pointing at his laptop, then swatting his hair back from his eyes.
The-one is in the crowd, I now notice, but keeps changing direction with a stubborn randomness, then is righted back into the pattern by the-one’s nearest neighbors. Trying to play along, but can’t stop deconstructing even the action of crowd acting.
There is a shuffling and suddenly they are all rotating in a disk that starts to thin in the middle until they are in a ring that diffuses out, larger and larger, and then, there in the exact middle, I see the most incredible thing, a brief flashing form, maybe from the clerestory windows or skylights, maybe a reflection off a jet landing at Dulles, maybe off a satellite in orbit. Maybe. And the crowd has stopped in their motion and are all staring at it too as it morphs and reflects, and then there is nothing but the cool veined marble floor and the world is silent.
Edit!