Self-Aware Science Creation

New Horizons for Second-Order Cybernetics
Series on Knots and Everything, Vol. 60
Riegler, A., Müller, K. H. & Umpleby, S. A. (Eds.)
Singapore: World Scientific, 2017
404 Pages; Print, $168.00

"Sin is believing the lie that you are self-created, self-dependent and self-sustained."
— Saint Augustine

"Objectivity is the delusion that observations could be made without an observer."
— Heinz von Foerster

New Horizons for Second-Order Cybernetics is a collection of essays from an eclectic set of thinkers. Many of these essays comment directly on other parts of the text as often as they comment more generally on the state of the field of second-order cybernetics (SOC). In part II, Ronald R. Kline points out that proponents of a new science "often write histories of their field in order to do what sociologists call 'boundary work,'" a reference to part I (334). Kline is correct to suggest that this is a substantial function of the text. However, the editors and framers of the book have architected it in such a way as to go beyond a simple top-down dictation of a history of SOC. The editors employ a recursive approach, beginning with a prompt (or framing notion) and iterating outward, from commentaries on the prompt to commentaries on those commentaries, in an escalating cycle of reconsideration. At the high end of this recursion the work invites the reader to participate in the ongoing communal processing of the meaning of the SOC revolution. Relatedly, this recursive architecture allows the text to demonstrate and perform on itself the meta-leap to endo-mode science that some contributors describe as characteristic of SOC.

The text consists of two prologue chapters, a part I, a part II, and one epilogue chapter. Part I consists of a series of target articles about the theory of SOC, such as Louis H. Kauffman's formalization of eigenforms, and the practice, such as Ben Sweeting's exploration of the mutual interaction between SOC and design. Each target article is followed by several commentaries on the piece from other second-order cybernetic thinkers and a closing response to the commentaries by the target article's author. The second prologue chapter sets the stage for the exercise of part I and offers a categorization of and reflection on the ideas presented by the authors of part I. The second prologue chapter and all of Part I are noted to have been published previously in 2016 in the journal Constructivist Foundations volume 11, number 3. Part II consists of reflections from another set of thinkers on the contents and conclusions from the originally published set. The first prologue chapter and the epilogue integrate the critiques of part II.

A similar process of soliciting and then synthesizing experiences is used to create each level of the text described above. This process is a function that takes a base idea (prompt) and returns a more refined set of ideas. The function is applied to some unknown prompt (probably generated from many recursive applications of the function to an even more primal prompt idea, at least in an informal way) to generate the target articles of part 1. The same function is then applied to each of the target articles, generating the response commentaries. Together these function applications produce part I and the second prologue chapter, which constitute an organized, if rough, overview of the state of the field of SOC. The function is then applied once more, taking this overview as a prompt to produce the more refined and self-aware picture of SOC that is the entire text. The editors have highlighted this generative function by structuring the text around the prompt and response method, so recursion both generates and structures the text.

This structure fits very nicely with Kauffman's focus on eigenforms, thought objects that remain consistent as the domain around them evolves by acting upon itself. Michael R. Lissack elegantly connects eigenform and "distinction" via the idea of stability (103). Eigenforms provide the stability required to make general distinctions, posit a theory or tell a story in a complex, evolving, self-acting environment. Scientific theories are eigenforms in the domain of the world. Kauffman defines the "world" as an inclusive combination of Wittgenstein's world of facts and his "real" world, and goes on to suggest that SOC can help resolve the conflict between Wittgensteinian worldviews by allowing the world to contain both descriptions of the world and the sources of experiences, as these loci would generate and influence each other (97). A scientific theory is both in the world and describes the world. A theory is an eigenform because it persists like a fixed point in the collective consciousness even as the world evolves. One can see how this new framing is especially helpful in understanding the social sciences, where theories about people may have a more immediate impact on the objects in their domain. If one executes the refinement function under ideal conditions on a true eigenform theory then the function would return the same theory. If the function is applied to a near-eigenform, the result will be a refined theory that is closer to eigenform status.

In structuring the text around the refinement function, the editors illustrate the natural iterative process by which new scientific facts or narratives are converged-upon. Kauffman brings up a mathematically convergent sequence to illustrate that "the fixed point [of a reflexive domain] does not depend on initial conditions" (94). One will always approach the fixed point with infinite recursion, but there are many paths to the fixed point depending on starting conditions. Within the domain of science, the path to the fixed point will likely be messy and heavily dependent on initial conditions. Intellectually destabilizing events such as von Foerster's initial jump to the inclusion of "the observer in science" created (or created the need for) the new subdomain of science called SOC (30). As the community processes and makes sense of these disruptions, it settles into stable uses of the new vocabulary and a reliable narrative built on top of that by iteratively refining the collective experience of this new field and converging to a fixed point in the new domain. Therefore the text can be seen as a documentary-style re-telling of how the SOC revolution disrupted various pillars of thought and continues to coalesce towards its end state. At the same time, the composite text remains especially cognizant of the fact that this telling and re-telling is what actually constructs the next generation of SOC. The presenters and the reader lose their illusions of objectivity and are nudged towards engaging with the material in an "endo-mode" as they are all part of the system of collective knowledge. The phrase "endo-mode" refers to "science from within," which means knowing that a theory must sit in the world it describes in a stable way, gracefully handling the subjectivity of an observer who is not separate from the observed system (10).

Operating in the endo-mode, the text does perform seemingly subjective boundary work including, as Kline points out, the selective inclusion of certain thinkers over others. The boundary work also includes the delineation of ideas represented by dichotomies of: endo- versus exo- mode, SOC versus second order science, and theory versus application. Critically, the work done by several contributors to move past the unresolvable radical constructivism versus realism debate and focus on practical applications of SOC also impacts the delineation of boundaries. The subjectivity is apparent at the synthesis level, but is also implicit in the selection of writers and basis topics. A few influential thinkers dominate. However, as the theory develops further, the effects of the initial conditions should dampen as the field converges. Only time will tell if building a mythos around certain names like von Foerster's is helpful in situating SOC within the world — or if that piece of the narrative will fade. Being transparent about the underlying recursion and including the reader turns what would be subjective foci into natural tools for situating SOC within its broader context.

The world of knowledge does not exist in isolation but is an eigenform of a larger reflexive domain. Therefore, it is impossible to know where the convergence of SOC will lead. It seems clear that eigenforms, the endo-mode and other ideas have broad and deep usefulness across the sciences, Arts and Humanities. What is less clear is how SOC will fit into the overall eigenform of the hierarchy of knowledge, as reflected in university departments. One relevant example is von Foerster's lab, which lost its funding after the passage of the Mansfield Amendment (31). Now, work that is similar to what von Foerster's lab might have ultimately produced is categorized as part of artificial intelligence research and is therefore approached with a different set of priorities and timelines. This outcome may seem to be due to external factors, but can also be seen as the internal behavior of a larger system settling into an intellectual narrative that fits its complex needs.

SOC could still become a full-fledged research program, or more specialized researchers might continue, as Scott puts it, to take "from cybernetics what they find valuable" (123). While one may serenely wait for stability in the long run, in the short term the structure of knowledge reflexively affects the direction of the path of convergence (see the von Foerster story above) and is therefore worthy of attention and influence. Laurence D. Richards proposes measuring the usefulness of SOC by how much it contributes to the goal of "the transformation of society to a more just and equitable one" (362). Given the importance of making timely progress towards this goal, there may be a moral argument to be made in recommending every thinker give this text a chance to challenge their conception of science.

---

Copyright © 2020 The Johns Hopkins University Press. This review first appeared in American Book Review, Volume 42, Number 1, November/December 2020, pp. 15–16 (DOI: 10.1353/abr.2020.0120). The version of record is available at Project MUSE.