The holistic tribes

This blog post is hopefully the beginning of the lecture notes for an upcoming course. Ultimately, I want to rectify the story of the development of ideas around complex systems, which has neither been a steady and well-informed progression nor a succession of Kuhnian paradigm shifts, but rather something messy and disconnected: a story of reinventions, ignorance, and believers trapped in the cul-de-sacs of their minds . . . and, of course, brilliance.

For now, I’ll leave out stuff based on structuralist ideas—i.e., most network science, all kinds of semiotics, ekistics, structural geography, applied information theory, etc. The problem with including these elements in the history of complex systems is that they blend much better with mainstream, reductionist science, making the edge blurrier.

On the to-do list is also to link up the participants by influence (citation?), but we already know that it will be a sparse and disconnected graph. Finally, the dates refer to the publication of the book or paper popularizing the topic or some similar iconic event. Really finally, the ethnoreligious allegories are tongue-in-cheek—most of what I’m talking about is serious science.

The tribes

In our academia, divided by reductionist legions into disciplinary dominions, there are nomadic tribes, bound by no borders, roaming the realm of knowledge. These tribes are united by a recognition that reductionism cannot lead to all the answers they seek, but, in general, little else. Sometimes, they are guided by the worship of holy scriptures or praying to the same temple. They could be complete civilizations, followers of a patriarch, or mere phantoms of the mind; eclectic congregations wandering across the hinterlands of thought, seemingly to the end of time. Yet, whispers stir among the sages: One day, a leader may be born to unify the tribes, stake their claimed sanctuary, and secure their long-sought retribution. [0]

Tektology 1912

The “universal organization science.” An extinct tribe led by the eccentric Bolshevik Alexander Bogdanov. Largely forgotten, with little direct impact on today’s science, yet impressively early, as this quotation shows:

[T]he experience and ideas of contemporary science lead us to the only integral, the only monistic understanding of the universe. It appears before us as an infinitely unfolding fabric of all types of forms and levels of organization, from the unknown elements of ether to human collectives and star systems. All these forms, in their interlacement and mutual struggle, in their constant changes, create the universal organizational process, infinitely split in its parts, but continuous and unbroken in its whole.

Gestalt psychology 1912

“The whole is different from the sum of its parts” was the Gestalt school’s version of Aristotle’s description of emergent properties (changing “more” to “different”). Their theories of visual perception are still alive and well. Indirectly, the Gestalt psychologists influenced concepts like emergence—especially in the sense of new perceptions emerging from configurations or relative arrangements of units. The Wolfgang Köhler quotation below even forebodes autopoiesis, and similar process-oriented notions of complexity:

[W]herever a process dynamically distributes and regulates itself, determined by the actual situation in a whole field, this process is said to follow principles of gestalttheorie. In all cases of this type the process will have some characteristic which exists in an extended area only, so that a consideration of local points or local factors as such will not give us full insight into the nature of the process. From this viewpoint, even the segregation of circumscribed wholes becomes one more or less particular, though highly important, case among the various possibilities which are included in the most general idea of self-distribution and self-regulation, and in consequence the concept of gestalt may be applied far beyond the limits of sensory fields. According to the most general definition of gestalt, the processes of learning, of reproduction, of striving, of emotional attitude, of thinking, acting, and so forth, may be included as subject matter of gestalttheorie insofar as they do not consist of independent elements, but are determined in a situation as a whole.

The noosphere 1922

The noosphere, the sphere of reason, is the third and newest in the sequence starting with the geosphere and biosphere. First theorized by the French Jesuit priest and paleontologist Pierre Teilhard de Chardin, geologist Vladimir Vernadsky further developed the idea. In recent years, the noosphere has primarily inspired parapsychology and other pseudosciences. However, indirectly, it helped to promote the idea that the connected humanity creates a higher state of being, consciousness, etc., which is an idea that, unfortunately [1], still lurks in the academic discourse of complex systems.

General semantics 1933

Nominally, Alfred Korzybski’s general semantics was a theory of mind and language, but, more importantly, it popularized the notion of levels of abstraction central to systems theories invoking emergence. As the other tribes worthy of a mention, general semantics attracted followers far from its intended application domain. The catchphrase “the map is not the territory” is another legacy of Korzybski.

Rashevsky’s Mathematical Biophysics 1938

Nicolas Rashevsky and collaborators had a very futuristic lab at the University of Chicago, producing several papers close to the core of today’s holistic tribes. Initially influenced by Gestalt theory, JBS Haldane, and D’Arcy Thompson, their “mathematical biophysics” used a phenomenal range of mathematical techniques applied to problems ranging from cell division, chemotaxis, and neuroscience to cognitive and social phenomena. While anti-reductionism was not their main idiosyncrasy, they were definitely a wandering tribe.

General systems theory 1945

Once a visitor to Rashevsky, Ludwig von Bertalanffy’s research considered similar topics but sought universal laws and expressions rather than a parade of specific analyses. Similar to cyberneticists, he articulated the need for other ways of explaining phenomena in systems and emphasized that living patterns are self-sustained and in constant change in open systems.

On a historical note, after the Anschluss, von Bertalanffy joined the Nazi party, which gave privileges during the war years and equally challenging times in the post-war denazification.

Cybernetics 1948

One of the most influential tribes, cybernetics is built on the eternally new [2] insight that explaining a system with feedback loops must be structurally different from identifying cause-and-effect relations. Cybernetics, especially in its British incarnation, managed to become more applied than most other tribes.

Here is a quotation from Ross Ashby’s An Introduction to Cybernetics that captures cybernetics’ tribal identity as an opposite reductionist causal and correlational inference:

Science stands today on something of a divide. For two centuries it has been exploring systems that are either intrinsically simple or that are capable of being analysed into simple components. The fact that such a dogma as ”vary the factors one at a time” could be accepted for a century, shows that scientists were largely concerned in investigating such systems as allowed this method; for this method is often fundamentally impossible in the complex systems. Not until Sir Ronald Fisher’s work in the ’20s, with experiments conducted on agricultural soils, did it become clearly recognised that there are complex systems that just do not allow the varying of only one factor at a time—they are so dynamic and interconnected that the alteration of one factor immediately acts as cause to evoke alterations in others, perhaps in a great many others. Until recently, science tended to evade the study of such systems, focusing its attention on those that were simple and, especially, reducible.

Another hallmark of cybernetics, especially in its later incarnations, is the attention to the observer’s role in the system.

Warren Weaver’s Science and complexity 1948

A contemporary with Wiener and Shannon, Warren Weaver was interested in similar questions about communication, information, and the limitations of mainstream science. His 1948 Science article “Science and complexity” summarizes thoughts similar to Ashby’s (newer) quotation above. 

What meaning is to be assigned to the question: Is a virus a living organism? What is a gene, and how does the original genetic constitution of a living organism express itself in the developed characteristics of the adult? Do complex protein molecules “know how” to reduplicate their pattern, and is this an essential clue to the problem of reproduction of living creatures? All these are certainly complex problems, but they are not problems of disorganized complexity, to which statistical methods hold the key. They are all problems which involve dealing simultaneously with a sizable number of factors which are interrelated into an organic whole. They are all, in the language here proposed, problems of organized complexity.

This article did not create much stir and was rediscovered much later. (In George Cowan’s memoirs, he mentions that the founders of the Santa Fe Institute were ignorant of Warren’s article, and their brand of “complexity” just happened to be very similar to Warren’s “organized complexity.”)

Systems science 1958

It’s a long story, but here I identify “systems science” with the congregation venerating Jay Forrester as their patriarch, making liturgical use of his systems dynamics modeling paradigm of coupled ODEs with delays. Briefly stated, systems science is cybernetics that simulates first and thinks later. That might not be such a bad idea since none of the other tribes can boast such a public recognition as systems science got after their 1973 doomy World3 simulation, giving humanity 80 years or so before pollution-induced famines would decimate even the industrialized world.

Simon and friends 1962

Herbert Simon was one of the most original thinkers of the 20th century, with contributions to artificial intelligence, cognitive science, economics, decision-making, and complexity science. Despite this diversity, Simon’s oeuvre (including that of his colleague, computer scientist Allen Newell) feels surprisingly cohesive, echoing his early investigations into organizational decision-making. The publication closest to the other tribes is probably the essay “The architecture of complexity” reprinted in his book The Sciences of the Artificial, where he discusses complex systems in Weaver’s sense as typically “near-decomposable,” meaning they can usually be understood at different levels of abstractions, but not always, sometimes do the connection across levels make a functional difference.

The Sciences of the Artificial is a great entry point to the Simonian tribal lore, but don’t cite it without reading it. It has a number of miscitations in the complexity literature. Typically, the artificial stuff Simon talks about is designed under the full control of humans. The book is not about technology beyond our comprehension, like the Internet, large-language models, etc.

The Austrian school of economics 1964

This tribe lived within the nation of economics rather than roaming the Earth. Still, we need to mention the Austrian school of economics because contemporary tribes revere some of its teachings and regard FA Hayek as an early prophet of the doctrines of emergence and self-organization. While that is true (at least in the sense of the economy being self-organized by market forces), Hayek was also firmly against mathematical or computational modeling as it “can only in the rarest of instances be turned into specific predictions of what will happen in a particular case, because we can hardly ever ascertain all the facts which will contribute to determine the outcome.”

Finally, the Austrian school is probably the only tribe linked to Bogdanov’s tektology via Ludwig von Mises’s interest in Tadeusz Kotarbiński’s praxeology that was informed by Bogdanov’s work.

Cellular automata 1970

The early study of self-replicating discrete dynamics on a grid started with von Neumann in the 1940s but took off with Conway’s “game of life,” published in a Scientific American column in 1970. Like chaos theory, the game of life is a paradigmatic example of how simple rules can do complex stuff. Most remarkably, one can set up the game to do whatever computation a computer can do—Turing completeness, you know.

Gaia theory 1972

Similar to the noosphere, the Gaia principle tells us that the Earth is, if not an organism, at least a self-regulatory system, including the bio- and geospheres. I guess the allure that we are all profoundly connected explains the endurance of the Gaia hypothesis even though the Darwinian orthodoxy keeps pointing out that there is no evolutionary pressure on Earth as a whole, and others highlight that it is nothing more than a metaphor—given all we know about the Earth, it doesn’t change anything if we additionally impose a metaphorical system epithet on it, be it machine, organism, cell, economy (haha! “the Earth is an economy,” it actually doesn’t not make sense), or whatever.

Autopoiesis 1972

A tribe worshipping the founding fathers Humberto Maturana and Francisco Varela. Pushing ideas of cybernetics and general systems theory further, originally with applications to cognition, consciousness, and what-is-life questions, autopoiesis refers to autonomous systems capable of creating and assembling their components. This idea had an impact even in manufacturing engineering as the final frontier of biomimetic design. Maturana and Varela’s 1972 book Autopoiesis and Cognition [3] is still revered by modern-day followers, easily recognized by their chanting of the holy words “embodied” and “enactive.” 

More is Different 1972

This essay by Nobel prize-winning physicist Phil Anderson was too far ahead of its time and was soon forgotten, despite the fact it was published in Science. Only well after the exodus of statistical physicists to complexity science was “More is Different” rediscovered. Anderson argues that analogs to symmetry-breaking, as they happen in condensed matter physics, connect the different levels of the academic hierarchy. Nowadays, we use the term “emergence” for an explanation of a phenomenon in terms of interactions between its constituents, but interestingly, that word does not appear a single time in Anderson’s essay.

Hierarchy theory 1973

The reason reductionist science works so well is that the world is fairly hierarchically organized. This blog post is about tribes who recognize the need to explain the rest, so it might seem somewhat funny that there was an attempt to create a unified theory of hierarchies. This tribe was gathered by biosemiotician extraordinaire Howard Pattee in a series of conferences in the 1970s. The conference notes—the only archeological evidence of the tribe’s existence—show traces of tensions with Pattee urging the participants to look beyond subtle differences in how to define the concept, yet a majority of participants spent a considerable time arguing for their particular definition.

Synergetics 1978

First, note that there are two synergetics, by Buckminster Fuller and Hermann Haken. Here, we discuss the latter. OK, maybe I should mention that Fuller’s Synergetics books are fun reads, but to call them scientific would be a stretch. Like Marshall McLuhan, Fuller was an oracle producing a dizzying mix of original profundities, platitudes, and nonsense, and it’s up to the reader to sort out the gems.

Back to Haken’s synergetics books. These cover a range of the physical side of what is now a complex systems canon, like pattern formation, bifurcations, and the fluctuation-dissipation theorem. Haken’s most original contribution is a fuzzification of the order-parameter concept originally from Ginzburg and Landau’s theory of phase transitions, which he applies to relatively simple ODE models of lasers. Today, it seems unclear what we gain by sacrificing the precision of the original concept and applying it to a very different, less complex context. Anyway, Haken was a source of material for several early complex systems textbooks, and thus, his teachings still attract worshippers.

Prigogian thermodynamics 1979

Traditional thermodynamics mainly discusses processes that are approximately in equilibrium. This is a helpful limit for designing thermal machinery (steam engines, etc.) but not for discussing the non-equilibrium world around us. Thus, Nobel laureate Ilya Prigogine concluded that the rest of the world (basically all of it, that is) should be described by non-equilibrium thermodynamics. After his award, Prigogine published a series of controversial popular science books where many readers saw blatant logical fallacies and delusions of grandeur, but the tribespeople saw poetry and the dawn of a new science (where Prigogine’s most spectacular claim was that physics is fundamentally time-irreversible).

Chaos & fractals 1982

In the mid-20th century, mathematical models described a world of smooth curves and geometric shapes unfamiliar to anyone with the window open to reality. This changed in the 1960s when Lorenz and coworkers discovered chaotic dynamical systems—simple deterministic systems of equations where almost identical initial conditions could send the solutions off into very different trajectories. Trajectories that, even though bounded, never return to where they have been before.

Chaotic dynamical systems usually have to share the “chaos theory” umbrella with fractals. These are geometric objects that look similar at different scales—keep zooming in, and you’ll start recognizing patterns from a moment ago. Like chaotic dynamical systems, fractals also show that even simple math can create complex patterns. But otherwise, the main reason these topics are treated together is probably that they were developed in the same era using similar methods (where computer methods played an important role).

It’s an arbitrary choice whether one should count chaos theorists as a holistic tribe. On the one hand, they fit pretty well in a reductionist worldview. Lorenz considered the complex problem of atmospheric convection and then reduced it to his system of three equations that happened to be chaotic. On the other hand, chaos theory heavily impacted later holistic tribes, including forming the social identity of complexity science. Furthermore, fractals brought model-analogies of the real world as scientific explanations (common throughout the tribes) to the public attention. 

Finally, Benoit Mandelbrot’s 1982 book The Fractal Geometry of Nature had a pivotal role in popularizing chaos theory. Not only the general public but also scientists from other fields caught the fractal bug. Leo Kadanoff reported only four years later that Physical Review Letters (the most iconic physics journal of the later 20th century) “complains that every third submission seems to concern fractals in some way or another.” Even corporate research labs invested heavily in fundamental research on fractals. [4] This hype bubble has deflated (rather than burst like the related catastrophe theory), but the fundamental insights are still essential for a mathematical modeler.

Complexity science 1984

This notably large and persistent tribe is defined by their all-embracing, Bahai-like faith that sprung out of a recognition that the ultimate problems of very different disciplines have much in common. Perhaps inevitably, the scientific content has varied with the times. For some time, complexity science seemed to converge on emergence as the leitmotif, but, as of late, it has lost a bit of focus on holistic themes. Members pray towards the temple of Santa Fe Institute.

Artificial life 1987

Manufacturing replicas of ourselves is a part of the profoundly human quest for the truth of who we really are. This desire has taken many expressions over the centuries. With the advent of the desktop computer, simulating characteristics of life, self-replication, etc., became a boom even outside of science. On the scientific [5] side, the social identity of the ALife tribe was formed at a conference in Los Alamos in 1987. The Santa Fe Institute cosponsored this conference, and the link between ALife and complexity science was initially very strong.

A peculiar thing about the ALife tribe is that early on, it agreed on “open-ended evolution” as its holy grail. It didn’t take long, however, before someone let the computer run overnight and woke up to the sound of Windows95 . . sorry, that was a silly joke . . before someone claimed open-ended evolution. But the ALifers didn’t go: “Hey guys, we’re done! Pack up!” Oh no! They are still around, running their conferences and looking for open-ended evolution.

Self-organized criticality (SOC) 1989

A tribe born out of statistical physicists trying to catch on the chaos-theory boom of the 1980s. At its core is a class of mechanistic models to generate power-law fluctuations (which may still be relevant for some specific systems). In the mid-90s, SOC spiraled out of scientific sanity with ever bolder claims (cf. the name of their holy scripture—How Nature Works) justified by a special plea for uncritical thinking (“details don’t matter”) based on a misapplication of the universality concept of the theory of critical behavior.

Complex networks 1998

The study of systems that can be represented as graphs has a history as rich and disconnected as this blog post in general. However, the vast majority of it follows the structuralist premise that things are what they are because of how they are connected to others, so let’s go on and decode the information about how things are connected. I.e., stuff for another day. However, around the turn of the millennium, there was a brief boom of network science motivated by a search for general laws in a similar spirit to the early years of SOC.

The “new” kind of science 2002

A peculiar case of outsider science backed by big dollars. Philosophically, it follows the “But, hey! What if everything actually is computation?!” tradition of Edward Fredkin (another wealthy and ostracized computational scientist) and some of the ALife crowd. Before you go on pondering whether there actually is a grain of truth there, read Cosma Shalizi’s delightfully heretic review.

Data science 2008

With the corporate hype of companies making money from the digital breadcrumbs we drop from our tables came the idea of theory-free science, focusing on prediction alone. While this shocked many conservative scientists, it should have been embraced by the holistic tribes. If people can’t do science without their epistemic preconceptions bringing us a hegemony of reductionism, then why not let the machines do it? But so far, data science has been almost a return to reductionist science, via causal inference and all that jazz.

Snarky comments

[0] Sorry, I can’t keep up this fantasy style much longer 🙂

[1] It’s unfortunate because whatever (if anything) that emerges from a connected population would be something other than the human mind (which we also don’t understand), invalidating explanations based on imposed concepts from human psychology. I.e., the fallacy of explanations by analogy.

[2] It seems like people never learn this, and every once in a while, someone presents it as a new finding (when it was only new for that someone). Take economics as an example. Around 20 years ago, the complexity economics of Brian Arthur et al. made this point fairly recognized, but now it seems forgotten, and causal inference (by nature more simplistic) is all the rage. It’s just a matter of time before we will hear it again.

[3] Personally, I find it almost unreadable because they were so hellbent on pointing out that everything ever is a part of grand, self-referential recursions. Throughout the book, they point it out; again, again, and again, in a while-true loop (and maybe that was the point, but, my god, please).

[4] Was there any return on that investment? For some time, mobile phone antennas were inspired by fractals, but as far as I can tell, they aren’t anymore.

[5] Or rather “philosophical”—you would never see an ALife claim falsified by a “meat life” (as the tribespeople call it) experiment. Some (at least back in the day) hardcore ALifers would respond that the simulations are reality and thus this is not philosophy but biology.