 Thanks so much to the first five panelists. You can see there was a great grouping amongst the two. And I really enjoyed seeing the pictures of the electric meters and those crazy, there was a great confluence between the last two talks. So anyway, I'm here to introduce Wendy Chen, who's a professor of modern culture and media at Brown University. I met Wendy when she was a PhD student at Princeton and I was jogging Tom's memory over there, Tom, my husband, because the three of us went together to an IT person who showed us a brand new access to the internet which was called Mosaic. And the three of us kind of stared at the browser which combined images and text, trying to imagine what we were seeing. And I'm sure everything that we saw there and we kept on saying, this is an empty library. It's an empty library. We sort of had no idea how to think about what might potentially happen in that space. But the three of us in very different fields thought about it as a space. There was, even though I was really the only architect there. So specifically though, I remember that day being the one where Wendy told me that her background was in systems design engineering and we proceeded to have a long conversation about how she'd left all that behind and why. And so I think since then we've kind of followed each other's work in different arenas and I guess we're both confirmed internet immigrants, although I kind of hate that term but it really fits the two of us. Although I'm sure neither of us remember it not being there. So I'm thrilled to introduce her today as a keynote. Her work is well known. You heard people quoting her already this morning but also her work is known to cross many disciplinary boundaries which is at the heart of our project over here. She's the author of Control and Freedom, Paranoia in the Age of Fiber Optics, Programmed Version, Software and Memory and my favorite, her recent book, Updating to Remain the Same, Habitual New Media which is MIT Press. She's also edited many books, won lots of awards, some of which are listed in her bio so I encourage you to look at that and Google her, et cetera, if you don't know who she is. So Wendy gave me two titles for her paper. One was discriminating data and the other was ways of knowing cities crossed out as you can see there, networks. I fussed a little about which one of these would look better on the program and really both perfectly respond to the prompt for the conference because cities are networks and data discriminates so let's hear her talk and figure out the best title. Welcome, Wendy. Hello everyone, can everyone hear me? Great, excellent. Let's just get the visuals going, perfect. So thank you, Laura, for that invitation and for inviting me and Dar for amazing organizational skills. It's exciting to be here and the talks so far have been absolutely fantastic and it's appropriate that my graduate class is here given, as has been said, all the resonances between this space and me being a graduate student. So Felicity Scott and I were actually TAs together for Tom Keenan's class who was my advisor and last night in Princeton, I gave a talk and my other advisor, Diana Fuss, was there so I feel like I'm in this really weird time warp. So am I 25 or am I turning 50? Okay, so today we've been asked to consider how technology better to turn 50. Anybody wondering better to turn 50? Okay, so yes. So for today, we've been asked to consider how technology in particular database technologies increasingly mediate the way we know cities. More broadly, the role that technologies have played in changing how urban spaces and social life are structured and understood both historically and in the present moment. Now as this description makes clear, this role isn't simply technological but also political. The struggle to transform cities entails conflict, resistance and delight. It's as much about information and representation as about physics and bodies. And it's about engaging with rather than turning a blind eye to these technologies. So Kergan calls for literacy and data technologies in more robust, creative and far-reaching ways to think about the relationship between the urban and the information systems that enable, engage and express the city. Now to respond to this call, I'm gonna do something that may seem counterintuitive. To foster more robust, creative and far-reaching ways of knowing cities, I want us to consider not only the ways that cities are increasingly networked or surveyed or smart. These are how they're usually discussed. So the ways in which information systems increasingly document and analyze are every move. And therefore undermine the fundamental anonymity of the city. And here you see face recognition technologies. But this kind of tracking happens in a far more efficient and banal way through our phones, by corporations, through our devices which we perversely against all evidence to the contrary, call personal devices. So I want us to consider not only how information systems enable, engage and express the city, but also how the city enables, engages and expresses information systems. For at the core of networks, at the core of current modes of knowing and navigating the city, lies an ideology, a cognitive map, a practice to use David Walksmith's term of the city. A profoundly segregating and segregated practice of the city. So to give you some of the context, very briefly of the talk, this will become part of the book I'm working on right now called discriminating data, individuals, neighborhoods and proxies that explores how identity persists in the era of big data. So how categories of race, gender and sexuality persist and proliferate in and through proxies, in and through algorithms which are allegedly blind to them and are built to blind us to them. So we're constantly recovering what is blatantly obvious and this is supposed to be insightful. And I'd be happy to talk about this more during the Q and A. But part one, city network as virtual proxy. And I know we're a little behind in time, so I'm gonna talk really quickly, but if it goes too fast, someone wave and I can slow down, okay. So there is increasingly and remarkably a consensus that networks define cities. A growing conviction that cities can be analyzed through the networks that traverse and constitute them. And I say this is remarkable because it's probably the only consensus that exists across the very divided field of urban science and urban studies. So to offer you two examples from the very opposite ends of the spectrum, here you see an excerpt from networks, borders and differences towards the theory of the urban from implosions, explosions by Christian Schmidt who's a critical urban studies scholar and professor of sociology. He argues urban space can be understood by means of the networks that run through it and determine it. And then other from urban characteristics attributable to density-driven tie formation which was published in Nature. This article emerged from the MIT Media Lab's City Science Initiative and the lead author is Wei Pen, a PhD candidate who's currently on leave to launch a big data-driven finance startup. And he argues that understanding the mechanisms of tie formation in cities is the key to the development of a general theory for a city's growth described by its economic indicators and population. And so for all of you unfamiliar with nature, imagine a cross between the National Enquirer in October. And then imagine that this is one of the most powerful publications in science and it will determine whether or not people get tenure. And letters from Nature are regularly picked up by mainstream news outlets so this was picked up by the economist. There's a ton to say, of course, about the differences between these two visions, especially in terms of experience and I'll talk a little bit about experience at the end. But what's important is that they share this common emphasis on the importance of networks to understanding urban space. Why? Why and what can we do with this confluence? What can emerge from this confluence? Now to link cities and networks, networks to geography is hardly profound. One of the first maps of the early internet, when it was ARPANET, represented server clients geographically. This of course recalls earlier transportation network maps and here you see an excerpt from an 1854 telegraph and railroad map of New England. Look at it closely if only Amtrak worked this well right now, right? The modern definition of the network, the notion of networks as something more than things that deliberately are built to look like nets is linked to movement, circulation, both within and across bodies and this is something that Shannon Mattern, of course, has pointed out throughout her work. In this sense, modern networks are profoundly metaphorical. They're about transport, they're about transfer, networks map and abstract movement, they map and abstract space. Yes, but even as they do so, they're haunted by cities, they're haunted by itineraries, they're haunted most importantly by dreams of navigating urban spaces. Neuromancer and snow crash, the science fiction Bibles of the dot-com era make this clear. So as we were looking at the Mosaic map, all the kids in Silicon Valley were reading Neuromancer and snow crash. So William Gibson coined the term cyberspace and here you see his description of it in his 1984 novel Neuromancer. Cyberspace was a consensual hallucination like city likes receding the color of Chiba sky. Neal Stephenson's snow crash outlined the metaverse. Stephenson's vision was very profound for the emergence of things like second life and also the notion of avatars. But his vision of the metaverse mimicked the deeply divided and ethnically segregated physical world outside of it. The term cyberspace, of course, combines cybernetics in space. Cybernetics itself stems from the Greek word, kyberneti, meaning steersman or governor. So it's intimately linked to concepts of control and navigation. Both these visions evoke navigation and control, albeit in very different ways. Cyberspace evokes outer space and expanse, a bodyless exaltation. The metaverse evokes boundaries, walled and gated communities, ethnic enclaves and fears over a triumphant globalization and a drive for a monopoly communications capitalist network that seems more prescient than ever before. Because at its core lies a vision of segregation gone wild that seems to encapsulate the current internet. For our imaginary of networks has moved from this fundamental anonymity, again a different vision of the city to this, right? Neither of which, none of which are outside planetary urbanization. For at the core of social networks, networks and their modes of navigation and agency are theories of small worlds, neighbors and friends. Theories of weak links, social holes, social capital, k nearest neighbor methods to determine identity, k means testing to figure out clusters or neighborhoods within data, homophily factors to determine likeness. So embedded within these are practices of the city, experiences and representations of a deeply segregated space in which space, in which geography, in which physical distance comes to magically define, to magically encapsulate urban centers. So to make this point, I'm gonna engage this article in the move towards urban science in some detail. Because I want to outline how the move towards universal equations for cities assumes and covers over and as well as perpetuates the history of US segregation. So it transforms historical and institutional racism into so-called random defaults. It regenerates racism in order to verify its truth. Because in these models, what's true is what repeats the past. The likely future equals the abstracted past. And because of this, and this is the really perverse part of my argument, I'm gonna say we need to engage this kind of work as proof of segregation. So we have to realize and unpack the ways these models are profoundly metaphorical. So data technologies which construct a ground truth and which seek to end the humanities and social sciences by making urban theory quantitative. And here you see a rather notorious excerpt from Jeffrey West 2010 interview with The New York Times called A Physicist Solves the City. And he's most infamous for discovering that innovation in the cities grows at a super linear rate of 1.12. So data technologies in this attempt to create these principles and magic numbers, I'll argue, actually creates a deeply metaphorical world. A world dominated by proxies, stand-ins, representatives, autonomous and wandering agents. Which brings me to part two, navigating magic numbers and lines. So urban studies has been recently sideswiped by a bunch of physicists mainly coming from Santa Fe who have sought to quantify urban studies. And importantly, it's not just urban studies that are in the crosshairs. So West has also just as controversially taken on biology, in particular, organismic biology. And we have to realize that he's controversial across all disciplines. And biologists, especially organismic biologists who somehow believe the structure of the organism is important, that biology somehow profoundly biological have pushed back and against these theories. Now, the most well-known and controversial article in urban science is this one by Benton Court in West from 2010 in PNAS. And PNAS is kind of like the B side of nature, right? Again, it shows that urban development is super linear, that quantities that reflect wealth creation innovation have an exponential growth of 1.12% versus economies of scale which are less than one. And they compare the growth of cities to the growth of organisms which have a beta as well of less than one. So they argue that as population grows, major innovation cycles must be generated in order to avoid collapse, okay? So this is the theory behind the coming singularity. Now, this article received a lot of attention by the mainstream press, politicians, but has largely been ignored by urban studies and for good reason. But ignoring it hasn't made it go away. In fact, as Brendan Gleason has argued, it's become emblematic of a contemporary decline in the influence of urban social science and the rise of urban revolutionaries who view the city as a natural rather than a social artifact. So to combat this decline, Gleason calls on the social sciences to offer the conceptual that is theoretical means for human urban aspiration. Yes, but to do so, we need to engage rather than ignore this other work because it's not as simple as natural versus social. And to make this point, I'm gonna look closely at this other influential article because it's far less of a strawman. Urban characteristics attributable to density-driven tie formation. Now, this article actually seeks to one up West's claim that by showing that the super linear relationship is actually a special case of another general equation, which is row, natural log, row. According to the authors, this article provides a robust and accurate fit for the dependency of city characteristics with city size, ranging from individual level dyadic interactions to population level interactions. Okay, so what's key about networks is they link the micro to the macro without the need to appeal to heterogeneity, modularity, specialization, or hierarchy. So it does so they argue through a simple bottom up robust model. And their model consists of two essential steps. The first is the creation of a simple analytic model for the number of social ties formed between individuals with population density as its only single parameter. Again, for some reason I wrote row, but Microsoft Word insisted on writing foe. So it says foe, natural log foe, but it's row, natural log row. Okay, so this they argue predicts super linear scaling. Simulations, and secondly, their model creates simulations of the diffusion rate along these ties. And this diffusion rate they argue is a proxy for the amount and speed of information flow and data adaptation. And because this proxy accurately reproduces empirically measured scaling of urban features such as the rate of AIDS HIV infections, as well as communication and GDP in European cities, they conclude that this surprisingly similar scaling exponent across many different urban indicators suggests a common mechanism behind them. Social dot show social tie density and information flow therefore offer a parsimonious generative link between human communication patterns, human mobility patterns, and the characteristics of urban economies again without the need to appeal to hierarchy, specialization or social constructs. So this model, like all network science models, assumes and entails abstraction at two levels. First, the abstraction of real world interactions into social ties, diffusion rates and graphs. And secondly, the mathematical reproduction of the abstractions in the first part as truth. And this model they argue is better than the other models because it's parsimonious. So simpler is always better. And again, this model is true because it reproduces the past, it reproduces past abstractions such as the super linear relationship. This model is true because things that repeat are identical. This model is true because repetition is identity and because repetition of abstractions is truth. Because the likely future is equals the abstract past. So in this article, the reproduction of both the abstract super linear relationship, which again was another theory, and empirical data is taken as proof. So in order to show that row natural log row is true, they show that this relationship produces a curve that matches the simulation of tie formation. They then extrapolate from social tie density to a model of contagion based on this tie formation. So to test the hypothesis that a city's productivity is related to how far information travels and how fast its citizens gain access to innovation or information, they examine how this information flows scales with population density and quantify the functional relationship between linked topology and the speed of information spreading. And they do so by simulating two different models of contagion for information diffusion on networks generated by their model and comparing simulations with their theoretical paradigm. And lastly, to prove their hypothesis is true, they study the prevalence of AIDS HIV infections in cities in the United States and plot the prevalence of AIDS HIV in 90 metropolitan areas in 2008 as a function of population density. They also plot the overall GDP per square kilometer in cities in the European Union as a function of population density as well as population size. Now again, all of this relies on a proxy. On the diffusion rate along these ties as a proxy for the amount and speed of information flow and idea adaptation. AIDS as a proxy as a metaphor for all infection and contagion. And it's this reliance on proxy that's absolutely key. As Vera Tolman and Boaz Levine have argued, proxies are fundamentally ambivalent. They both elucidate and obscure the relationship between action and cause, correlation and causality. They are a pharmacon. Proxies are spaces for approximation. They're empirically known variables used to infer the value of otherwise unobservable or immeasurable others. There are approximations or substitutions. There are concessions to imprecision, forays into unknown terrain. They're speculative. They're spaces for speculation. And as Kathy O'Neill has argued, proxies are used when we don't have access to the truth, to the thing we want to measure. And because of this, they open up the possibility of duplicity. Proxies, in other words, aren't just stand-ins. They're also agents. They can take a cut. And proxies we have to remember aren't simply bad. Global climate change models rely on proxies. Proxies are necessary. And we can't simply attack any model for their use simply of proxies. But what's key in this urban sciences work isn't just that they use as a proxy simulated diffusion rates as a proxy for the amount and speed of information flow. What's key and unacknowledged is how geography and the city become proxies. As noted earlier, the move towards geography enables, they state, them to create models for social diffusion without considering hierarchy, heterogeneity, et cetera. So the question we need to ask whenever we come across such elegant mathematical equations with claim to obviate the need for hierarchy is why. Why is geography such a good proxy for hierarchy? And to get at this, let's probe their model a little more. So this model takes as its ground truth this equation for rank friendship. And what's remarkable is that this is taken from an analysis of live journal users by Lib and Noel. And Lib and Noel came up with this equation in order to relate geography to social networking friendships, which brings me to the first part three, part three repeats, who is your neighbor? So a fundamental concern of Lib and Noel in his article is how to navigate small worlds. And network science is obsessed with small worlds in the milligram male experiment. And how many of you are unfamiliar with the milligram male experiment? Okay, so the idea is you need to send a piece of mail to someone you don't know in a different city. You just send it along to someone you do know. And milligram allegedly showed that it will reach the person within six hops, right? This is why we allegedly live in a small world. We're all six degrees apart. Now, John Kleinberg, who's done some of the best work in small worlds, has used this cover from the New Yorker to describe a navigable small world. So in a navigable small world, each person has random connections that can move beyond their immediate neighborhood. In his model, each person has immediate neighbors and then longer distance friendships where distance could either be geographical or organizational. But the key thing is you have, this is the view from 9th Avenue. One block takes as much as three blocks takes as much as the rest of America, right? Now, Liv and Noel modeled the probability that these longer distance friendships, as rank friendships, based solely on distance. And to do so, they did the following. They first engaged in a thought experiment to prove that the live journal network was indeed a navigable small world. So they simulated a version of the message forwarding experiment using only geographic information to choose the next person who would get the message. And through this thought experiment, they showed that yes, live journal is a small world because most simulated messages get to where they're going within four hops. They then came up with a distance-based equation to explain how this navigability works. And what's key is that their equation, unlike others which were based on organization and hierarchy, only uses population density. So Liv and Noel and his co-authors chose geography because they argued that although interests and occupations might be naturally hierarchical, geography is far more naturally expressed in 2D Euclidean space. So in fact, they argued, embedding geographic proximity within a tree hierarchy is not possible without significant distortion. So what's key is that geography against all experience to the contrary enables the imagining of a flat system. In these systems, architecture does not exist. Their model is basically like Gibson's cyberspace. It divides the Earth's surface into small squares. Each person U in the network has an arbitrarily chosen neighbor in each of the four adjacent grid points. In addition to these four neighbors, person U has a long-range link to a fifth person chosen according to rank-based friendship. That is to the probability that U chooses V as her long-range link is inversely proportional to the rank U V. What's never questioned is why these neighbors exist and why neighbors are likely to be friends and why the probability of connection is linked to geography. Because deeply embedded at the heart of network science is the principle of homophily, the axiom that similarity breeds connection that birds of a feather flock together. That like acts like like. Homophily underlies collaborative filtering, collaborative filtering groups, users, and objects into neighborhoods based on similar characteristics and opinions. So based on your intense likes and dislikes, you're segregated into neighborhoods. And these neighborhoods are fundamentally segregated because you're placed into them based on the intensity of your likes and dislikes. So it creates segregated clusters. So the fact that network analytics produces real-life echo chambers should surprise no one because a really retrograde identity politics drives homophily. And here you see the definition of homophily from one of the most respected textbooks in network science. And I have only the deepest respect for John Kleinberg and his work, but look at the assumptions that are embedded in the statement. Now the links between homophily and segregation are deep and profound. Homophily was actually first coined in 1954 by sociologist Lazarfeld and Merton. And in that same chapter, they also coined the term heterophily. And everybody refers both to the earlier review article that I showed as well as this article as a source for homophily. But not surprisingly, but erased in every single reference to this work is the fact that Lazarfeld and Merton's text itself studied segregation within the United States. So at the inception of homophily are two case studies of segregation. And as well as terming both homophily and heterophily they asked not is homophily always present but at what conditions do we have homophily or heterophily although they did assume that equilibrium equals homophily. But the current version of network science in which homophily has moved from problem to solution has forgotten this history. And homophily is no longer something that needs to be accounted for but rather something that naturally accounts for and justifies the persistence of inequality across facially equal systems. It's become axiomatic that is common sense. But homophily as a starting place cooks the ending point it discovers. Segregation is what's recovered if homophily is assumed. And homophily as a grounding principle imposes naturalizes and projects the segregation it finds. So consider what's taken as evidence as the naturalness of homophily. U.S. segregation. And to explain how segregation arises Kleinberg and easily turned to Thomas Shelling's famous model of segregation which allegedly shows how local homophily can drive global patterns of segregation. So how robust the forces of leading to segregation are. So in this model there are basically two types. There you can be an X or an O. X's and O's differ according to some immutable characteristics such as race, ethnicity, country of origin or native language. In this model each agent wants to have at least what's assumed fundamentally is that each agent wants to have at least some other agents of its own kind as neighbors and they care whether they have at least some neighbors of the same type. So at each stage unhappy agents are moved. So here you see the first move at T equals three. And now at T equals three people are willing to be a minority. What happens though Kleinberg and easily argue via Shelling is that integration is extremely difficult to produce. Here you see what happens when T equals four. So this allegedly proves that in a process not based on segregation agents again are willing to be in the minority. Segregation is the most likely outcome. The problem therefore as Kleinberg and easily explain is that from a random start it's very hard for the collection of agents to find integrated patterns. More typically agents will attach themselves to clusters of others like themselves and these clusters will grow as other agents follow suit. So this is the agent based justification for segregation. But this only works, this model and this interpretation only works if we erase, if we unknow the entire history of U.S. segregation. There is no random start. And their desire not to be a minority is not innocent. If you look at the history of segregation within the United States, which is a history of forced segregation, the desire not to be a minority which again is the fundamental assumption to move once a neighborhood has changed defines white flight. It's a reaction to and against desegregation. It's a reaction to and against the civil rights movement. Furthermore, homophily redefines hate as love. So what's the proof that you love someone? You flee when someone else shows up. And love as inherently love of the same, love as the desire to flee justifies modern definitions of white supremacism. They claim that they don't hate other people, they just love their own. But the only way they can show they love their own is by fleeing whenever somebody else shows up, right? So this model and interpretation renders invisible institutional infrastructions and responsibilities, which is actually something that Schelling's original publication makes clear. So this all comes from his 1971 classic article, Dynamic Models of Segregation. And again, this was 1971 in Cambridge, Massachusetts. This is right when busing is starting in Boston. And it deals explicitly with white flight in neighborhood tipping. So Schelling in this paper openly acknowledges that he's deliberately excluding two main processes of segregation. The first is organized action. Therefore, he doesn't even mention the history of slavery or legally enforced segregation or block busting and as well, economic segregation. So as an economist, he's not considering economic segregation. Even though economic segregation might statistically explain some initial degree of segregation. And what's interesting is that this denial, we see this repeated in bizarre statements. So this is back to the way article, which completely misunderstands white flight within the 70s, right? So white flight isn't due to people not understood. White flight, why it argues, is due to people not understanding the benefits of population density. It's not racism, right? So what to do with this? How can we learn from history so we don't reproduce the same? So very briefly, we need to repeat but in a very different way and in a very different register, the question, who is your neighbor? Since I don't have a lot of time, I'm just gonna throw out some suggestions. The first, as should be clear from this talk, is that we need to engage mathematical models of cities. We can't simply ignore them because it's by engaging them that a different dialogue can emerge. Again, they can be proof of segregation. But again, the question becomes, what does the proof do? But if every city planner believes in this, if this has had this kind of pull from politicians, it's an interesting ground to start from, right? So let's ask, what does row natural log row really prove? What does this insight, which is both blind and acutely obvious that geography encapsulates hierarchy, enable us to analyze? Where to go from here? Well, intriguingly, the way article talks about the difference experience makes and how experience is misleading given the overall structure. And again, he wants to talk about overall structure. Well, how would this look like in relation to Carl H. Nightingale's sweeping examination of the history of segregation across global cities? In terms of his analysis of the threefold factors, governments, networks of intellectual exchange and the institutions associated with modern real estate industry that foster segregation. Or how would it look like in terms of Richard Rothstein's color of the law, which uses the US legal system as the most parsimonious description and indicator and reason behind geographical segregation? What if we took seriously the question, again, who is your neighbor? This question has driven Judeo-Christian ethics as well as political theory. But homophily cuts through this Gordian knot. It makes this ethical dilemma trivial by making your neighbor yourself, right? So to express it very simply in an algorithm, right? So what if we took up instead of this Freud's challenge that the neighbor also invokes hostility? What if we once more looked at the question of experience? And David Walksmith, how do you say his name? Does anyone know? Walksmith? Okay, everyone mispronounces my name, so I don't really care, okay. So, but it's good to know. You can mispronounce my name, that's fine. As he argues, these ideological tropes do not endure simply because elite actors want them to. They continue to correspond to a common experience of urban society and thereby influence scholarly as well as everyday understandings of urbanization in the global North. Which brings us once more to this and to ask ourselves what do these lines actually represent? For what are these lines if not traces of habitual actions? What is information if not habit? What new worlds could we create if we started by interrogating the institutions and other embedded actors within the human and habitual actions, right? So what if we viewed these seemingly isolated habits, these connections as shards of others embedded within ourselves that we constantly recreate so they can seem like simple lines? What if we saw these lines not simply in terms of similarity but also in terms of opposition and indifference, right? So think about it, how many of you people know people who are heterosexual? Okay, this heterosexuality actually doesn't make sense in the world of homophily. It's called reverse homophily. Think of electromagnetism, the idea that somehow negative, that electricity goes from negative to positive. This is all erased in these systems which take similarity as breeding connection as the fundamental precept of connection, right? So what if we again thought through heterophily, mutual indifference as different grounds for networks? And what if we embrace these network models and their poor calculations, their poor predictions of the future as diagnosing what's wrong with the present in the past that these futures are predicted so we can then put in place a radically different future? Thank you. We're getting close to lunch but I definitely want to open up to questions. I can't tell you how many times I've used the algorithm nearest neighbor. I will think about it, and I do think about it hard every time I use it but not in the same way but I know people have questions, yeah. Hi, thank you so much for that talk. I have a question around the proxy of the individual, I guess, I'm thinking of the network graph as inherently you start with nodes, you add edges and I was recently reading David Graber's book, Utopia of Rules where I think one of his essays he talks about the so-called opposition between regulation of the free market, right? And his argument, which is not a necessarily new one but I think a well-stated one is that usually you say the free market, there are these individuals who operate within a free market and regulation constrains it. His model is regulation is what makes a market possible in the first place and makes the conditions of the market possible so that entities such as corporations or individuals can operate. So the regulations create the environment in which these nodes of corporations can interact in the first place so it's not this binary opposition. And if I apply that back onto like social network analysis or the kind of work you're thinking about and critiquing I'm curious if there's a way to talk about instead of society is made out of nodes that then we then link together, individuals that we then connect and somehow in the connection between these pre-existing individuals we find society if we apply the same kind of logic that Graber is doing is there a way to talk about how cultures, practices, nation states but also like cultures, craft environments that then constitute groups of individuals after that to like flip the logic or the proxy or the model you're using from or that social network analysis is using from instead of going from individuals and mapping the networks between them are there ways you're thinking or people are thinking of models or proxies that go from contexts first, cultures first, practices first, different models of homophilia first and then map the individuals onto them as a way of in something like flipping the script of flipping the script of social network analysis that starts with kind of the individual actor in the first place. Yes, and so my latest book I actually try to think through the ways in which nodes are connected through contact. So trying to think through contact is actually the primary mode through which nodes are created. There's other people who have been also looking at institutions but part of what my work has been seeking to do is that if it's through contacts that nodes are created, right? If we think through habit and what habit does because at one level people dismiss habit in the individual. It's a neoliberal logic, right? How do you solve the environmental crisis you recycle? So the move towards habit and the individual is completely without political potential, they argue, right? But if you actually consider again, refuse the notion of the autonomous individuals first, right? And think of habit precisely as evidence that humans are open to others, right? Habits are shards of the other encased within the self. There are things that you take up from others and that are linked fundamentally to institutions and others, right? That actually allows you even at the level of the node to show and to bring out the ways in which the node is never encased, right? Yeah, I'm wondering if you could speak about the relationship between the proxy and complexity because what strikes me when you referenced Jeff West, I went to that institute in Santa Fe and they're obsessed with complexity, right? And how is it? And I think if I understand your argument about the proxy, it's not to kind of, it's not for us who wanna critique that. The critique is not the most productive when it's accusing them of having reduced, right? Because the proxy is reductive and that's what it does, but why is complexity then so simple, right? And what is that relationship between complexity and the proxy that's more productive? Right, yeah, that's a great question and that's, so we had this exchange yesterday too, right? No, no, no, it's great. So I think that what's really interesting and that was my talk on global climate change so it was an entirely different talk. Poor Mitch has suffered through two talks in two days. So, so I think that what we were getting at earlier with something like singular value decomposition, right? Is that the ways in which these models work is to figure out what to erase, right? So it's allegedly about complexity but it's also about figuring out what you can ignore, right? And what's interesting about how they come up with something like the super linear number or think of it like if segregation is a proxy for the city and it maps linearly, right? Then you can map all that complexity within that simple system because you're dealing with a system of segregation, right? So that's one thing, you figure out the proxy that encapsulates this kind of complexity. Of course, they ignore it but it's actually what grounds the simplicity of their analysis, right? But again, the proxy is about what to erase or to what to use so you can look at one thing, right? In order to create these models. Hi, so something that I think about a lot is how important it is to historicize the thinking of your own field to know what that is and that's part of what you're doing. And one of the things that not really knowing anything about social network analysis but one of the things that I think about here is it seems to assume what's happening is sort of an eternal present. It's like here, now. And so if you wanted to do something like layer other types of information in graphs that were historicized over that and say here's, I don't know, St. Louis in 1950 whatever or 60 whatever in here were the crimes that were committed by the real estate agents who paid people to black people to walk through those neighborhoods to scare the white people to make them run away and these are the houses that were sold within six month period or whatever and you kind of put these things up against each other some way and this doesn't make a lot of sense but you sort of get my point. Like is there a way to bring historical, empirical factual information and butted up against some of this stuff? Absolutely because what you can do is if you take that kind of empirical analysis, historical analysis, you can show that it grounds the assumptions that they're making here but for me what's important about social networking analysis isn't that it's about an eternal present but that it makes an argument about history. So correlation itself has a deeply eugenicist past. It was created the notion of the Pearson correlation coalition. Pearson himself was a eugenicist as was Galton who came up with linear regression and Pearson was a special kind of eugenicist. He did not believe in Mendelian genetics which is why he looked at correlation and a lot of the arguments that he makes about correlation eerily resonate with the arguments that are now made about big data and correlation and how they obviate the need for causality. But the reason why he went to correlation and the way he theorizes correlation as well as linear regression was because he literally thought that the British population was regressing to an earlier mean. So correlation was necessary in order to understand how a historical mean haunts the present and therefore determines the future. So for me, it's not that it's ever present but the embedding of certain historical assumptions within the model that are absolutely key. No, Greg Alman died recently. And there were debates about music and the Beatles and Quincy Jones said this yesterday what bad musicians they were. But it was interesting because I had this conversation the day before and how Duane Alman was his great guitar player and was behind a lot of the Aretha Franklin cuts like It Ain't Fair, The Wait. And there's a whole culture around music. I was born and raised in Park Slope. And there were the white kids, i.e. the Italians, the blacks, the Spanish, the Irish boys, and the hippies. And there's still this whole culture, every time they do this homophily thing, it's always like a black, white thing. It's never like, okay, we play fiddles or we wear purple and I knew Adam purple. Stuff like that. So I often wondered, like have you ever thought about doing some other kind of study based on this homophily thing here? Not to mention, you know, homophily itself. But I mean, not that. Just basic, you know. It's very interesting to me because I thought about that and his wish was to keep the music going. And so is mine. It's good old home music. I mean, rock and roll is American music. That's a homophily statement in itself. Now, what's really interesting actually is that a lot of this network stuff claims that it's no longer about race. That in fact, what they've found are factors that aren't racial, but that map onto things in really disturbing ways. So Cambridge Analytica first. So this is the whole thing behind psychogeographics. And what I'm just gonna do is show you just a little clip from this because it shows you how this has both emerged and repeated itself in various ways precisely as a way allegedly of moving away from these issues. Okay, so. In a sense. Clearly, graphics and economics will influence your world view. But equally important or probably more important are psychographics. That is an understanding of your personality because it's personality that drives behavior. And behavior obviously influences how you vote. So how is this possible? Well, at Cambridge, we've rolled out a long form quantitative instrument to probe the underlying traits that inform personality. This is the cutting edge in experimental psychology known as the ocean model. Ocean being an acronym for openness, how open you are to new experiences. Conscientiousness, whether you prefer order and habits and planning in your life. Extroversion, how social you are. Agreeableness, whether you put other people's needs and society and community ahead of yourself. And finally, neuroticism, a measurement of how much you tend to worry. By having hundreds and hundreds of thousands of Americans undertake this survey, we were able to form a model to predict the personality of every single adult in the United States of America. So how does this impact marketing and communications in elections? For a primary, the Second Amendment might be a popular issue amongst the electorate. If you know the personality of the people you're targeting, you can nuance your messaging to resonate more effectively with those key audience groups. So for a highly neurotic and conscientious audience, you're gonna need a message that is rational and fear-based or emotionally-based. In this case, the threat of a burglary and the insurance policy of a gun is very persuasive. Conversely, for a closed and agreeable audience, these are people who care about tradition and habits and family and community. This could be the grandfather who taught his son to shoot and the father who will in turn teach his son. Obviously, talking about these values is gonna be much more effective in communication your message. Okay, so they claim they're not using race, but the face of white neuroticism is a, of a white, sorry. The face of conscientious neuroticism, of course, is a white female, right? And they claim that they've come up with these correlations based on your actions to determine your group. And if I could get this up, what I could show you is what they, the assumptions for IQ, right? The proxies for IQ. This from the talk last night, yeah. And so if you look at IQ, is this, which one is it? There's Princeton over there on the side. Okay, this one, great, thank you. So if you look at the sort of correlations that they've come up with for IQ, they're profoundly linked to gender, race, and class. And I'll just get this one up for you to see. Okay, so this comes from work done by Kaczynski who argues that based on your Facebook likes, they can determine many things about you. And here you see, get that up. The most predictive Facebook likes for IQ. So what's interesting is as they claim no longer to be dealing with Christians of race, gender, and sexuality, embedded within these proxies are precisely these questions. So my question is, how can racism be inscribed in systems which seem to be without race, right? And what happens if we deal with those questions in a different manner, precisely through the proxies that they proliferate through? All right, I'm afraid I'm gonna end the question and answer session, but I'm sure everybody is gonna be referring to your talk in the afternoon. So thanks a lot.