return ✕︎

Living in a Plural World

By E. Glen Weyl, Audrey Tang and ⿻ Community

Living in a Plural World

Until lately the best thing that I was able to think in favor of civilization…was that it made possible the artist, the poet, the philosopher, and the man of science. But I think that is not the greatest thing. Now I believe that the greatest thing is a matter that comes directly home to us all. When it is said that we are too much occupied with the means of living to live, I answer that the chief worth of civilization is just that it makes the means of living more complex; that it calls for great and combined intellectual efforts, instead of simple, uncoordinated ones, in order that the crowd may be fed and clothed and houses and moved from place to place. Because more complex and intense intellectual efforts mean a fuller and richer life. They mean more life. Life is an end in itself, and the only question as to whether it is worth living is whether you have enough of it. — Oliver Wendell Holmes, 1900[1]

(A)re…atoms independent elements of reality? No…as quantum theory shows: they are defined by their…interactions with the rest of the world…(Q)uantum physics may just be the realization that this ubiquitous relational structure of reality continues all the way down…Reality is not a collection of things, it’s a network of processes. — Carlo Rovelli, 2022[2]

     Technology follows science. And so if we want to understand Plurality as a vision of what our world could become, we need to start off by understanding Plurality as a perspective on how the world already is. The Technocracy and Libertarianism perspectives, which we critiqued because of their over-emphasis on one particular way of solving social problems (a global expert class in the former case, and entrepreneurs and corporations in the latter case), also have a long history of being tied to what we consider overly simplistic analogies of science.

Technocracy has a long history of being justified by science and rationality. The idea of “scientific management” (a.k.a. Taylorism) that became popular in the early 1900s was justified by making analogies between social systems and simple mathematical models, and logic and reason as ways of thinking about them. High modernism in architecture is similarly inspired by the beauty of geometry. Entrepreneurial sovereignty also borrows heavily from physics and other sciences: just like particles “take the path of least action”, and evolution maximizes fitness, economic agents “maximize utility”. Every phenomenon in the world, from human societies to the motion of the stars, can ultimately be reduced to these laws.

These approaches have achieved great successes that cannot be ignored. Newtonian mechanics explained a range of phenomena and helped inspire the technologies of the industrial revolution. Darwinism is the foundation of modern biology. Economics has been the most influential of the social sciences on public policy. And the Church-Turing vision of “general computation” helped inspire the idea of general-purpose computers that are so broadly used today. But there are also limits to the power of each of these sciences, as we have been increasingly discovering in the past century. Gödel’s Theorem undermined the unity and completeness of mathematics and a range of non-Euclidean geometries are now critical to science. Symbiosis, ecology, and extended evolutionary synthesis undermined “survival of the fittest” as the central biological paradigm. Neuroscience has been reimagined around networks and emergent capabilities.

Plurality similarly looks at social systems from multiple perspectives, and appreciates that any single perspective has limits to its power to explain the world. A corporation can be viewed as a player in a bigger game, but a corporation is simultaneously itself a game, where employees, shareholders, management and customers are all players, and whose outcomes often do not look anything like a coherent utility function. What's more, the abstraction often leaks: individual employees of a corporation are often influenced through their other relationships with the outside world, and not through the corporation itself. Countries too are both games and players, and there too we cannot cleanly separate apart actions between countries and actions within a country: the writing of this very book is a complicated mix of both in multiple ways.

Plurality is thus heavy with analogies to natural sciences: it uses many precisely because it understands the limits in relying too much on any single one. We can give a few examples.


19th century mathematics saw the rise of formality: being precise and rigorous about the definitions and properties of mathematical structures that we are using, so as to avoid inconsistencies and mistakes. At the beginning of the 20th century, there was a hope that mathematics could be “solved”, perhaps even giving a precise algorithm for determining the truth or falsity of any mathematical claim. 20th century mathematics, on the other hand, was characterized by much more uncertainty.

  • Gödel's theorem: a number of mathematical results from the early 20th century, most notably Gödel's theorem, showed that there are fundamental and irreducible ways in which key parts of mathematics cannot be fully solved. Similarly, Church proved that some mathematical problems were “undecidable” by computational processes. This dashed the dream of reducing all of mathematics to computations on basic axioms.
  • Computational complexity: Even when reductionism is feasible in principle/theory, the computation required to predict higher-level phenomena based on their components (its computational complexity) is so large that performing it is unlikely to be practically relevant. In fact, in some cases, it is believed that the required computation would consume far more resources than could possibly be recovered through the understanding gained by such a reduction. In many real-world use cases, the situation can often be described as a well-studied computational problem where the “optimal” algorithm takes an exponentially large amount of time, and so good-enough “heuristic” algorithms often get used in practice.
  • Sensitivity, chaos, and irreducible uncertainty: Many even relatively simple systems have been shown to exhibit “chaotic” behavior. A system is chaotic if a tiny change in the initial conditions translates into radical shifts in its eventual behavior after an extended time has elapsed. The most famous example is weather systems, where it is often said that a butterfly flapping its wings can make the difference in causing a typhoon half-way across the world weeks later. In the presence of such chaotic effects, attempts at prediction via reduction require unachievable degrees of precision. To make matters worse, there are often hard limits to how much precision is feasible, as precise instruments often interfere with the systems, they measure in ways that can lead to important changes due to the sensitivity mentioned previously.
  • Fractals: many mathematical structures have been shown to have similar patterns at very different scales. A good example of this is the Mandelbrot set, generated by repeatedly squaring then adding the same offset to a complex number:

     Geometry and topology, once the province of Euclidean certainties, turned out to admit endless variations, just as the certainties of a flat earth vanished with circumnavigation. Axiomatic systems went from the hope for complete mathematical systems to being proven, by Kurt Gödel, Paul Cohen, and others to be inherently unable to resolve some mathematical problems and necessarily incomplete. Alonzo Church showed that other mathematical questions were undecidable by any computational process. Even the pure operations of logic and mathematics, it thus turned out, were nearly as plural as the fields of science we discussed above. To illustrate:


Figure 1: The Mandelbrot Set (characterizing the chaotic behavior of simple quadratic functions depending on parameter values in the function) shown at two scales. Source: Wikipedia (left) and Stack Overflow (right).

  • Relationality in mathematics: in mathematics, different branches are often interconnected, and insights from one area can be applied to another. For instance, algebraic structures are ubiquitous in many branches of mathematics, and they provide a language for expressing and exploring relationships between mathematical objects. The study of algebraic geometry connects these structures to geometry. Moreover, the study of topology is based on understanding the relationships between shapes and their properties. The mix of diversity and interconnectedness is perhaps the defining feature of modern mathematics


At the end of the 19th century, Lord Kelvin infamously proclaimed that “There is nothing new to discover in physics now.” The next century proved, on the contrary, to be the most fertile and revolutionary in the history of the field.

  • Einstein's theories of relativity overturned the simplicity of Euclidean geometry and Newtonian dynamics of colliding billiard balls as a guide to understanding the physical world at a very large scale. When objects travel at large fractions of the speed of light, very different rules start describing their behavior.
  • Quantum mechanics and string theory similarly showed that classical physics is insufficient at very small scales. Bell's Theorem demonstrated clearly that quantum physics cannot even be fully described as a consequence of probability theory and hidden information: rather, a particle can be in a combination (or “superposition”) of two states at the same time, where those two states cancel each other out.
  • “Heisenberg’s Uncertainty Principle” puts a firm upper limit on the precision with which the velocity and position of a particle can even be measured.
  • The three body problem, now famous after its central role in Liu Cixin's science-fiction series, shows that an interaction of even three bodies, even under simple Newtonian physics, is chaotic enough that its future behavior cannot be predicted with simple mathematical problems. However, we still regularly solve trillion-body problems well enough for everyday use by using seventeenth-century abstractions such as “temperature” and “pressure”.

Perhaps the most striking and consistent feature of the revolutions in twentieth century physics was the way they upset assumptions about a fixed and objective external world. Relativity showed how time, space, acceleration, and even gravity were functions of the relationship among objects, rather than absolute features of an underlying reality. Quantum physics went even further, showing that even these relative relationships are not fixed until observed and thus are fundamentally interactions rather than objects. Thus, modern science often consists of mixing and matching different disciplines to understand different aspects of the physical world at different scales

     The applications of this rich and plural understanding of physical reality are at the very core of the tragedies of the twentieth century. Great powers harnessed the power of the atom to shape world affairs. Global corporations powered unprecedented communications and intelligence by harnessing their understanding of quantum physics to pack ever-tinier electronics into the palms of their customers’ hands. The burning of wood and coal by millions of families has become the cause of ecological devastation, political conflict, and world-spanning social movements based on information derived from microscopic sensors scattered around the world.


If the defining idea of 19th century macrobiology (concerning advanced organisms and their interactions) was the “natural selection”, the defining idea of the 20th century analog was “ecosystems”. Where natural selection emphasized the “Darwinian” competition for survival in the face of scarce resources, the ecosystem view (closely related to the idea of “extended evolutionary synthesis”) emphasizes:

  • Limits to predictability of models: we have continued to discover limits in our ability to make effective models of animal behavior that are based on reductive concepts, such as behaviorism, neuroscience, and so forth, illustrating computational complexity.
  • Similarities between organisms and ecosystems: we have discovered that many diverse organisms (“ecosystems”) can exhibit features similar to multicellular life (homeostasis, fragility to destruction or over propagation of internal components, etc.) illustrating sensitivity and chaos.
  • Higher-level organisms that operate through the cooperation of simpler ones (e.g., multicellular life as cooperation among single-celled organisms or “eusocial” organisms like ants from individual insects). A particular property of the evolution of these organisms is the potential for mutation and selection to occur at all these levels, illustrating multi-scale organization.
  • The diversity of cross-species interactions, including traditional competition or predator and prey relationships, but also a range of “mutualism”, where organisms depend on services provided by other organisms and help sustain them in turn, exemplifying entanglement, and relationality.
  • Epigenetics: we have discovered that genetics codes only a portion of these behaviors, and “epigenetics” or other environmental features play important roles in evolution and adaptation, illustrating embedded causality.

    This shift wasn’t simply a matter of scientific theory. It led to some of the most important shifts in human behavior and interaction with nature of the 20th century. In particular, the environmental movement and the efforts it created to protect ecosystems, biodiversity, the ozone layer, and the climate all emerged from and have relied heavily on this science of “ecology”, to the point where this movement is often given that label.


    Modern neuroscience started in the late 19th century, when Camillo Golgi, Santiago Ramón y Cajal, and collaborators isolated neurons and their electrical activations as the fundamental functional unit of the brain. This analysis was refined into clear physical models by the work of Hodgkin and Huxley, who built and tested in on animals their electrical theories of nervous communication. More recently, however, we have seen a series of discoveries that put chaos and complexity theory at the core of how the brain functions:

  • Distribution of brain functions: mathematical modeling, brain imaging, and single-neuron activation experiments suggested that many if not most brain functions are distributed across regions of the brain, emerging from patterns of interactions rather than primarily physical localization.
  • The Hebbian model of connections, where they are strengthened by repeated co-firing, is perhaps one of the most elegant illustrations of the idea of “relationality” in science, closely paralleling the way we typically imagine human relationships developing
  • Study of artificial neural networks: As early as the late 1950s, researchers beginning with Frank Rosenblatt built the first “artificial neural network” models of the brain. Neural networks have become the foundation of the recent advances in “artificial intelligence”. Networks of trillions of nodes, each operating on fairly simple principles inspired by neurons of activation triggered by crossing a threshold determined by a linear combination of inputs, are the backbone of the “foundation models” such as BERT and the GPT models.

From science to society

     Plurality is, scientifically, the application of an analogous perspective to the understanding of human societies and, technologically, the attempt to build formal information and governance systems that account for and resemble these structures as physical technologies built on plural science do. Perhaps the crispest articulation of this vision appears in the work of the leading figure of network sociology, Mark Granovetter. There is no basic individual atom; personal identity fundamentally arises from social relationships and connections. Nor is there any fixed collective or even set of collectives: social groups do and must constantly shift and reconfigure. This bidirectional equilibrium between the diversity of people and the social groups they create is the essence of pluralist social science.

     Moreover, these social groups exist at a variety of intersecting and non-hierarchical scales. Families, clubs, towns, provinces, religious groups of all sizes, businesses at every scale, demographic identities (gender, sexual identity, race, ethnicity, etc.), education and academic training, and many more co-existing and intersecting. For example, from the perspective of global Catholicism, the US is an important but “minority” country, with only about 6% of all Catholics living in the US; but the same could be said about Catholicism from the perspective of the US, with about 23% of Americans being Catholic.

     While we have emphasized the positive vision of pluralistic social science (a “network society”), it is important to note that beyond its inherent plausibility, a key reason for adopting such a perspective is the impossibility of explaining most social problems using monistic atomism given both complexity and chaos. Even in the social science field, economics, that most consistently aims for “methodological individualism”, it is universally accepted that trying to model complex organizations exclusively as the outgrowth of individual behavior is unpromising.

     The field of Industrial Organization, for example, treats firms rather than individuals as the central actors, while most macroeconomic models assume sufficient homogeneity to allow the construction of a “representative agent”, rather than reducing behavior to actual diverse individual choice. In fact, one fascinating features of economic models is that they tend to feature a range of different forms of organization as either the “central planner” (e.g., a technology platform operator or provincial government) or as the “individual actors” (e.g., a municipality or a manufacturer). This is hardly surprising given that a leading result in game theory (the most canonical approach to economic “reduction” of a group to individual behavior) is the “folk theorem”, a variant on chaos and irreducible uncertainty that states that when interactions are repeated, a very wide range of outcomes can be an equilibrium.

     Yet, whatever level of explanation is chosen, actors are almost always modeled as atomistically self-interested and planners as coherent, objective maximizers, rather than socially-embedded intersections of group affiliations. The essence of understanding social phenomena as arising from a “network society” is to embrace this richness and build social systems, technologies, and policies that harness it, rather than viewing it as a distracting complication. Such systems need, among other things, to explicitly account for the social nature of motivations, to empower a diversity of social groups, to anticipate and support social dynamism and evolution, to ground personal identity in social affiliations and group choices in collective, democratic participation and to facilitate the establishment and maintenance of social context facilitating community.

     While we do not have the space to review it in detail, a rich literature provides quantitative and social scientific evidence for the explanatory power of the pluralist perspective [^Assemblage Theory]. Studies of industrial dynamics, of social and behavioral psychology, of economic development, of organizational cohesion, and much else, have shown the central role of social relationships that create and harness diversity[^SocialDynamics]. Instead, we will pull out just one example that perhaps will be both the most surprising and most related to the scientific themes above: the evolution of scientific knowledge itself.

     A growing interdisciplinary academic field of “Science of Science” (SciSci) studies the emergence of scientific knowledge as a complex system[3]. It charts the emergence and proliferation of scientific fields, the sources of scientific novelty and progress, the strategies of exploration scientists choose, and the impact of social structure on intellectual advance. Among other things, they find that, relative to the most efficient ways of discovering existing knowledge (in chemistry, as an example), scientific exploration is biased towards topics and connections related to social connections and previous publications within a field[4]. It finds strong connections between research team size and diversity and the types of findings (risky and revolutionary v. normal science) developed and documents the increasingly dominant role of teams (as opposed to individual research) in modern science [5]. The largest innovations tend to arise from a strong grounding in existing disciplines deployed in unusual and surprising combinations[6]. It illustrates that most incentive structures used in science (based e.g. on publication quality and citation count) create perverse incentives that limit scientific creativity and has helped produce new metrics that can complement and offset these biases, creating a more pluralistic incentive set [7].

     Thus, even in understanding of the very practice of science, a pluralist perspective, grounded in many intersecting levels of social organization, is critical. To advance science and technology of any flavor, therefore, a pluralist outlook is critical.

A future plural?

     Yet the assumptions on which the Technocratic and Libertarian visions of the future discussed above diverge sharply from such pluralist foundations.

     In the technocratic vision we discussed in the previous chapter, the “messiness” of existing administrative systems is to be replaced by a massive-scale, unified, rational, scientific, artificially intelligent planning system. Transcending locality and social diversity, this unified agent is imagined to give “unbiased” answers to any economic and social problem, transcending social cleavages and differences. As such, it seeks to at best paper over and at worst erase, rather than fostering and harnessing, the social diversity and heterogeneity that pluralist social science sees as defining the very objects of interest and value.

     In the libertarian vision, the sovereignty of the atomistic individual (or in some versions, a homogeneous and tightly aligned group of individuals) is the central aspiration. Social relations are best understood in terms of “customers”, “exit” and other capitalist dynamics. Democracy and other means of coping with diversity are viewed as failure modes for systems that do not achieve sufficient alignment and freedom.

     But these cannot be the only paths forward. Pluralist science has shown us the power of harnessing a plural understanding of the world to build physical technology. We have to ask what a society and information technology built on an analogous understanding of human societies would look like. Luckily, the twentieth century saw the systematic development of such a vision, from philosophical and social scientific foundations to the beginnings of technological expression. While that path (dao) of development is today somewhat forgotten, we will rediscover it in the next chapter.

  1. “Life as Joy, Duty, End” ↩︎

  2. [^MultilevelSelection] Wilson, David Sloan et al. “Multilevel Selection Theory and Major Evolutionary Transitions.” Current Directions in Psychological Science 17 (2008): 6 - 9. ↩︎

  3. See a summary in Fortunato et al. (2018) ↩︎

  4. Rzhetsky et al. 2015 ↩︎

  5. Wu et al. 2019 ↩︎

  6. Foster et al. 2015 ↩︎

  7. Clauset et al. 2017 ↩︎