Abstract
State-of-the-art planning theory considers cities as cyborg entities composed by a “natural” part (human beings and their social structures) and an “artificial” part (buildings, infrastructure and other urban artifacts). We contend that this hybrid conception is indissolubly coupled with the ability to discriminate perfectly between the “natural” and the “artificial”. But is this actually the case? We will provide a critical reflection on this ontological issue pointing out that current urban planning theory as well as the general philosophical reflection is not able to produce a rigorous, consistent epistemic criterion to draw this distinction. Long-standing difficulties in this respect are exponentially amplified by recent developments in artificial intelligence, nanotechnology and biotechnology, and their growing relevance in urban environments of the near future risk making the cyborg conception informing the complexity theory of cities obsolete. We will conclude our reflection identifying a possible path for overcoming this dualism toward a more socio-natural conception internalizing the proteiform character of the concept of “nature” itself as well as its inherent cognitive/political element.
Introduction
Are cities natural or artificial? On one hand, the answer may appear straightforward: in popular perception, cities are considered the epitome of the artificial world, the quintessential artificial place. As perfectly summarized by Wirth (1938: 1–2): «nowhere has mankind been farther removed from organic nature than under the conditions of life characteristic of great cities». On the other hand, it is clear that this cannot be the whole story. Consider for instance this citation taken from Davies’ (2011: 109) popular introductory book to architectural theory: «[. . .] the aerial shots of Zanzibar and Marrakesh, or the extraordinary panoramas of Dogon villages in western Sudan, seem to show something alien to the very idea of design as a deliberate act. They look like organic accretions: forests or coral reefs or insects’ nests». We see clearly here the clash between two conflicting perceptions vis-à-vis cities: they are unmistakably man-made entities—yet, at the same time, they seem to display also many of the features humans traditionally associate to nature. So should we consider them natural or artificial systems?
Planning theory confronted at length with this dilemma. The theoretical conceptualization of the city as something natural can be traced back to Geddes (1915), who introduced the city-as-organism concept and, thanks to this preliminary connection, applied biological methods and concepts to urban planning (Batty and Marshall, 2009; Welter, 2002). The city-as-organism paradigm was pervasive also in the thought of Mumford (1970) and Le Corbusier (1947) and, through the latter, influenced the entire modernist approach to urbanism (Batty and Marshall, 2009). In this panorama, Christopher Alexander’s position is undoubtedly original. In his famous 1965 essay A City is not a Tree, he highlighted the fundamental structural difference between planned and historical cities (with the former exhibiting a tree-like structure where the latter being instead characterized by a more complex lattice-like structure), and adopted this inherent structural difference as a demarcation principle for distinguishing between the natural and the artificial. Nowadays, the latest developments in the complexity theories of cities led to the establishment of a conception of the city as a hybrid natural-artificial system, of which humans and the social structures they form are the “natural” part, whereas buildings and infrastructures are “artifacts” (Portugali, 2011, 2016).
Ultimately, the arrival point of this important theoretical reflection carried out within the complexity theories of cities has led to a “cyborg” conception of the city. As Gandy (2005: 28) pointed out: «If we understand the cyborg to be a cybernetic creation, a hybrid of machine and organism, then urban infrastructures can be conceptualized as a series of interconnecting life-support systems» (emphasis added). Here, we want to focus on the fact that the cyborg conception of the city, while stressing the interdependence between the organic/artificial components as well as the cybernetic coordination dynamics taking place between them, nevertheless preserves their ontological autonomy. In other words, in a cyborg, we can distinguish without particular difficulties where the human ends and the artificial begins. We contend thus that the cyborg conception of the city is not per se a solution to the natural-or-artificial issue, but opens up an important ontological issue: can we distinguish between the natural and the artificial?
We argue that this issue has not received enough attention within the current planning discourse. Still nowadays, a large part of the literature uses the terms “natural” and “artificial” without paying enough attention to the unresolved ontological issue they constitute. Nevertheless, leading authors in the field of the complexity theories of cities are well aware of this. In fact, Portugali (1999, 2011) acknowledged that cities fall within the Wittgensteinian “family resemblance” category—displaying a prototypical center (Rosch and Mervis, 1975) but no well-defined boundaries: the natural and artificial categories too would be built on a network of overlapping similarities rather than a single necessary and sufficient condition or set of conditions (Garbacz, 2013). At the same time, however, Portugali and Stolk (2016: xx) stated clearly that «The view that cities are hybrid artificial–natural systems bring with it the need to define the distinction and relationship between the natural and the artificial». This paper aims at reflecting critically on this issue by arguing that we do not have an internally coherent epistemic criterion to draw the distinction between the natural and the artificial.
This may appear as a mere definition exercise, maybe even an highly abstract, not-really-needed philosophical venture. In reality, identifying a coherent epistemic criterion able to ascertain what is natural and what is artificial constitutes a long-standing philosophical quest encompassing the entire development of Western thought from Aristotle on. Nowadays, this question is more timely than ever, since we are said to have entered an entirely new geological age whose main characterization is the complexification of the relationship between the natural and the artificial, the so-called Anthropocene (Crutzen and Stormer, 2000; Lewis and Maslin, 2015; Zalasiewicz et al., 2010). In this context, AI, biotechnology and nanotechnology will have a direct and huge impact on cities, thus for urban planning theory a reflection on the difference between the “natural” and the “artificial” part of the cyborg city—as well as, more broadly, on the very adequacy of our intellectual instruments to categorize them—is more needed that ever.
The structure of this paper goes as follows. The following section aims at reconstructing with more detail how the complexity theories of cities overcome its initial neo-organicism toward an hybrid conception. If the city is a hybrid, how can we distinguish between its natural and artificial part? The next step will focus on this unresolved question by exposing the main conceptions of “natural” and “artificial” underlying current planning debate. Then, we will problematize these conceptions revealing their pitfalls and the consequent impossibility to be used as an epistemic criterion to draw the distinction between the natural and the artificial. Finally, we will reflect on this result and we will propose a novel approach, maintaining that the complexity theories of cities can benefit from the reflection about socio-nature carried out within urban political ecology and trying to identify potential complementarities between the respective literatures. After that, we will draw the conclusions of our research.
The complexity theories of cities approach to the hybrid
The complexity theories of cities emerged as a result of the application of complexity theories to cities science and urban planning (Allen, 1997; Batty, 2005; Batty and Marshall, 2012; Portugali, 1999, 2006). If on one hand complexity sciences provided urban planning theorists with a set of novel and sophisticated theoretical tools—such as fractals, autopoiesis and self-organization (Chettiparamb, 2013, 2018; Moroni, 2015)—that proved to be extremely useful to meaningfully interpret urban phenomena, on the other hand it was accompanied by the resurgence of an organicist conception of the city (see for instance Botkin and Beveridge, 1997; Miller, 1973). The “neo-organicist” conception permeating first-generation complexity theories of cities ultimately derived directly from the central assumption the complexity theories of cities rely on, that is, the idea that the city is a complex system (Bettencourt, 2015). Since, at the early stage of development of complexity theories, also many bio-physical systems were considered complex systems, this assumption led to the establishment of a triangular identity linking both natural systems (organic and inorganic) and cities to the notion of complex system. This triangular identity is the “theoretical bridge” allowing for the direct transposition of models coming from far-from-equilibrium chemistry, ecosystems studies and evolutionary biology to the domain of urban planning and to the broader field of social sciences. Consequently, we had a surge in “hard” mathematical modeling approaches to cities and urbanism (see for instance Allen, 1997; Allen and Sanglier, 1979; Batty, 2005; Bettencourt et al., 2007; Portugali, 1999; Wilson, 1970). This attitude, defined by Byrne (2003) as “complexity reductionism,” can be explained by taking into account two elements. Firstly, the enthusiasm for the discovery that the same complex phenomena (such as phase transitions, feedback loops or emergence) appear to take place in very different substrata (encompassing the chemical, biological and social domain) led complexity scholars to overlook the ontological differences between the fundamental elements of each field, leading thus implicitly to kind of a holistic or vitalist or panpsychist derive. Secondly, the intrinsic limitations of the models of complexity being available at that time, whose lack of definition corroborated the above-mentioned naive enthusiasm for overcoming the hyperspecialization paradigm informing modern science. As Byrne (2003: 173) observed, «this regrettable limitation has been transformed into an assertive dismissal of structures and a “simplistic complexity” assertion that we can understand the social world in terms of simple rules describing and delimiting agents’ behaviours from which the whole of the socially complex can be derived».
Portugali (2011) vehemently criticized the “quantitative drift” complexity theories of cities were undergoing. In fact, «The notion of agents as currently used in urban simulation models [. . .] is a kind of automaton that mimics the behavior of urban agents in a predetermined way. [. . .] this is not sufficient: [. . .] we have to treat each urban agent as a complex self-organizing system too» (Portugali, 2011: ix). This important theoretical turn allowed him to problematize the naive city-as-organism conception underlying “first generation” complexity theories of cities. Assuming that a “simple complex system” is a system characterized by a global complex behavior emerging out of the interaction of simple deterministic elements (such as for instance cellular automata or chemical clocks), Portugali exposed the double fallacy of the triangular identity linking cities, complex systems and organic systems: (1) organic complex systems are not simple complex systems because their parts are complex systems too. (2) cities are not simple complex systems nor organic complex systems because they have an added layer of meta-complexity deriving from the fact that not only their components are organic complex systems, but also that these are subject to a cultural evolution that is way faster than the Darwinian one, enabling virtually instantaneous feedback effects (Portugali, 2011, 2016). However, Portugali’s critique did not only focus on the neo-organicist approach of first-generation complexity theories of cities, but involved also Alexander’s structural criterion. Portugali pointed out that the opposition between “natural” and “artificial” cities is pointless, because all cities are inherently artifacts, i.e. man-made objects. Complexity, in fact, is not a property of cities per se: it arises because of the human beings that populate them—and they are the same both in planned and non-planned cities (Portugali, 2011, 2016). Summing up, overcoming both the “complexity reductionist” approach of the first-generation complexity theories of cities (which equaled cities to natural complex systems) as well as Alexander’s approach (discriminating cities with respect to their origin), Portugali set forth a novel “hybrid” paradigm: «It is in this respect that cities differ from natural complex systems; they are hybrid, artificial–natural complex systems [. . .] composed of human agents as individuals and collectivities, both of which are by their nature complex systems, and artifacts of various forms and scale, built by the human agents, which are essentially simple systems» (Portugali and Stolk, 2016: xvii; emphasis added).
Can we distinguish between the natural and the artificial?
In the present section, we focus on exposing the most relevant definitions of “natural” and “artificial” underlying the current planning debate. When surveying planning literature, one is struck by the vast heterogeneity displayed by the—explicit or implicit—definitions of “nature” in use. Nevertheless, to a closer look, this superficially huge array of conceptions can be traced back to a limited number of “fundamental” definitions. These are, in turn, the expression of some of the most important philosophical schools of thought attempting to draw the distinction between the natural and the non-natural. In the complexity terminology, thus, these fundamental definitions can be thought as the “attractors” of the topology of philosophical semantics for the natural/artificial issue. By informally reviewing literature, we identified three of these “attractors,” each of them linking the natural-versus-artificial divide to a subordinate opposition: non-man-made versus man-made (ontological approach); spontaneous versus planned (ontogenetic/Hayekian approach); complex versus linear (complexity-based approach). The next subchapters are devoted to present, discuss and problematize these three basic conceptions. Note that, as a consequence of the inductive nature of our search, we cannot deductively prove their uniqueness or universality, nor we claim so (even if there are strong reasons to believe so, since these are rooted in those being by far the most diffused and important philosophical conceptions of nature). Finally, one may wonder that, even if none of these definitions is strong enough to constitute alone a coherent epistemic criterion, maybe their linear aggregation/overlapping does. In the final subchapter we will briefly discuss this possibility, showing that even an “aggregate” definition is not likely to work.
The ontological approach
The first conception we want to focus on defines the artificial as “what is man-made”. Consequently, the “natural” is identified as a residual category encompassing everything being non man-made. This criterion is almost trivial and very close to the popular conception as well. This conception is built on a ontological hiatus between natural and artificial objects and substances (Schummer, 2001). Science historian Reyer Hooykaas (1948) pointed out that it can be found in Galen’s thought and in Middle-Age alchemy, tracing it back to Avicenna and Aristotle—albeit the attribution of this divide to Aristotle has been questioned (Schummer, 2001). Portugali (2011) used this criterion when he attacked Alexander’s approach. In that venue, in fact, he pointed out that: «This terminology is to my mind misleading for the simple reason that unlike the tree which is by definition a genuine natural identity, Siena, Liverpool, Kyoto, Manhattan as well as Levittown, Chandigarh and the British New Towns are all artifacts» (Portugali, 2011: 105; emphasis in the original). Portugali, overall, called for an ontological clarification: the buildings and infrastructures that constitute the city do not grow spontaneously—they are inherently and irreducibly man-made objects. Thus, and quite straightforwardly, humans constitute the natural part of the cyborg city and buildings and infrastructures are its artificial part.
While appearing quite satisfactorily in its elegant simplicity, however, the latest developments in technology, in particular in the field known as nanotechnology, revealed the remarkable shortages of this approach. In fact, if at the macro-level we can identify without particular difficulties the difference between non man-made and man-made elements, we cannot say the same at the nano-level. Here, any ontological difference between the two tends to disappear: at this minimal level of aggregation, nature and man have to play with the very same building blocks, so to say. Consider a given molecule built atom-by-atom in a top-down nanotech procedure: it would be simply indistinguishable from the “naturally”-produced one. The situation would be only apparently less complicated for nanomachines and other entities that do not originally exist in nature. Take, for instance, the molecular motor of bacterial flagella: it appears as an exceptionally complicated structure (see Xue et al., 2015), frequently defined “nanomachine” (and the very use of this word expresses the complexity of their perceived ontological status). In this respect, Schiemann (2006: 91) proposed a criterion according to which a nano-object is considered natural if it is impossible to prove that it was produced by human action. However, even adopting this method, he notes that «Given the multifariousness of the relationship between nanotechnology and nature, there are cases where it becomes problematic to distinguish between the two». No ontological criterion appear to be strong enough to work at the nano-scale: probably, here, a more fruitful approach could be recovering the original Aristotelian conception according to which, due to the inherent ontological identity of the substrate, technology (techne) can only imitate nature (physis), but that approach would merely lead to other kind of problems (Engelmann, 2017; Schummer, 2001). One may object that the inapplicability of this criterion to the nanoscale does not seriously harm its usefulness with respect to cities. This argument can be easily rejected if one considers that the pervasiveness of nanotechnology in the near future will have a significant impact on cities, altering both its perceived “human” component as well as its “artifactual” component. With respect to the former, Gandy (2005: 34) has already pointed out that the diffusion of nanotechnology will make the conceptualization of the body-city interface more and more problematic, leaving unresolved the issue about the ontological divide between sentience and non-sentience. As for the latter, nanotechnologically-realized materials will be soon available to builders—especially considering the latest developments in the field known as Construction 3D Printing—and urban planning will need thus to confront with an entirely novel range of artificial/natural building materials (Hager et al., 2016).
The Hayekian approach
The second criterion can be synthetized as follows: something can be defined as “artificial” if it was “planned,” whereas the “natural” is the “unplanned” by definition. Whereas the preceding conception is built on an ontological evaluation, here the approach adopted is more ontogenetic (Gecow, 2011). As Baker (2008) puts it, «artifacts are objects intentionally made to serve a given purpose; natural objects come into being without human intervention». This is the criterion underlying Alexander’s distinction between “natural” and “artificial” cities: «I want to call those cities which have arisen more or less spontaneously over many, many years natural cities. And I shall call those cities and parts of cities which have been deliberately created by designers and planners artificial cities» (Alexander, 1965: 1; emphasis added). This criterion is also present in more recent approaches accruing to the complexity theories of cities. For instance, it kind of intersect with the ontological qualification set forth by Marshall (2012: 200).
This approach has quite an important history in Western thought. It lies behind the distinction made by Christian theology between the categories of the genitum (begotten) and factum (made) as they appear in the Niceno-Constantinopolitan Creed. This same conception underlies the creative method of the most important precursor of what Galanter (2003) defined—quite appropriately—generative art and architecture, Antoni Gaudí, who aimed at replicating the élan vital permeating the natural world by using ante litteram generative design to have the shapes of his buildings to develop in a spontaneous fashion (Bassegoda, 2002). Also Kant’s take on the natural-or-artificial issue can be reconducted to this basic vision: «Kant stresses that an artifact such as a watch has an external maker who designs the internal parts to function for the sake of the whole. An organism, on the other hand, is self-directed. Thus, its organic parts are the efficient cause and for the sake of the whole, while the whole itself is for the efficient cause and sake of the parts» (Engelmann, 2017: 118).
This conception has a central role in the thought of Nobel laureate economist and social philosopher Friedrich von Hayek, who devoted to the analysis of the natural-or-artificial issue a significant part of his theoretical effort. He maintained that the nature-society dualism originated with the philosophy of the ancient Greeks, in particular with the Sophists and Aristotle (Hayek, 1967). This duality was reaffirmed by rationalist constructivism, since «[. . .] the new rationalism of Francis Bacon, Thomas Hobbes and particularly René Descartes contented that all useful human institutions were and ought to be the deliberate creation of conscious reason» (Hayek, 1967: 85). This theoretical attitude will lead to the establishment of what Prigogine (1984) would have later defined the “order out of order” paradigm, a central dogma of the reductionist/mechanistic approach informing modern science that posited that order can only be the result of other order—i.e., social order can only emerge out of conscious design or planning. Consequently, «For more than half a century the belief that deliberate regulation of all social affairs must necessarily be more successful than the apparent haphazard interplay of independent individuals has continuously gained ground [. . .] To bring order to such a chaos, to apply reason to the organization of society, and to shape it deliberately in every detail according to human wishes and the common ideas of justice seemed the only course of action worthy of a rational being» (Hayek, 1948: 119). The idea that complex social orders are not the result of conscious design but emerged spontaneously—what Hayek famously defined spontaneous order recovering the original intuition of Scottish and English social thinkers of the 18th century—radically attacked the theoretical foundations of this approach. In order to build on this, in the first volume of Law, Legislation and Liberty, Hayek (1973) reintroduced the Greek ontogenetic distinction between taxis, representing planned order, that is, “order out of order,” and kosmos, defining instead spontaneous order, that is, “order out of chaos”. Hayek contented that this transversal approach was more suited to draw the distinction between the natural and the artificial (Beck, 2018: 24–25).
Neither this conception is strong enough to constitute a valid epistemic criterion. The reason is quite simple: kosmos and taxis are not suitable to be used to discriminate between the natural and the artificial because they never exist in a “pure” form: in other words, they do not act as impermeable elements—highly mixed but nevertheless always to a certain extent discernible—but are more as an inextricably bundled couple. We will briefly elaborate on this point. According to Hayek, the foremost example of the spontaneous order informing society is markets: the prices mechanism enables a spontaneous self-coordination between the plans of a large number of relatively ignorant agents (Hayek, 1967). In this way, the nature/society dualism Hayek attributed to Cartesian constructivism is transcended but a novel dualism arises, the State/economy one. In this conception, thus, we could be tempted to say that market forces constitute the “natural” part of the city, whereas centralized State interventions or discretionary policy tools constitute its “artificial” part. The question thus becomes, can we structure this relationship in such a way to be completely non-distortive of the “natural” development of the socio-economic system? Or, in Portugali’s (2012) words, can we conceive a regulatory framework being not dependent on perfect prediction, i.e. a purely passive frame able to sustain (without interfering with) spontaneous order? Hayek maintained that teleocracy (“control,” goal-oriented planning) is not suited to be used in spontaneously organized systems such as economies due to the lack of the determinism needed to link consistently causes with effects and to the—closely related—“impossibility theorem” defining an absolute limitation to the amount of knowledge available to the central planner (Lavoie, 1984; McCloskey, 1994; Moroni, 2010); instead, we need to focus on nomocracy, i.e. a more passive, non-distortive framework-like kind of regulation. Consequently, Hayek pointed out that the paradigm of “control” informing planning and planning-related sciences should be substituted by the paradigm of “cultivation,” «[. . .] cultivation in the sense in which the farmer or gardener cultivates his plants, where he knows and can control only some of the determining circumstances, and in which the wise legislator or statesman will probably attempt to cultivate rather than to control the forces of the social process» (Hayek, 1967: 19). The biological metaphor adopted by Hayek is indeed highly revealing about the underlying idea: the State is something radically external to the “natural” functioning of the economy/garden, and thus its actions must be limited and nomocratic. Of course, this reasoning is highly reliant on the (quite radical) assumption that the “natural” is always good—or, in any case, always better than the “artificial”.
While the Hayekian reflection on planning proved to be extremely fruitful and still nowadays continues to be explored in the field of complexity, its use as an epistemic criterion to distinguish between the natural and the artificial fell into disgrace. In the specific case of urban planning, for instance, the debate about patterning- and framework-instruments (representing respectively teleocratic and nomocratic approaches) largely emancipated from the natural versus artificial narrative (Alexander, 2012; Moroni, 2010, 2015). In the broader economics debate, the research program of complexity economics inherited many Hayekian intuitions, and the whole Austrian economics itself is now considered a precursor of this approach. However, there are also important differences (see for instance Kilpatrik, 2001; Montgomery, 2000), the most relevant one, for the purposes of our study, is that, whereas Hayek—a market-fundamentalist position that actually radicalized with following Austrian thinkers—considered the State as an exogenous element almost threatening the “natural” self-organization of the economic system, complexity economics strives instead for a systemic reintegration of the State in the picture, seeing it too as a result of the self-organization of the socio-economic system. As perfectly summarized by Colander and Kupers (2014: 46) in Complexity and the Art of Public Policy: «[. . .] market fundamentalists see the state as attempting to do the impossible—to “control” the evolution, and almost as a gut reaction see any state intervention as undermining the system [. . .] What modern complexity theory does is to further develop that story and to bring government into the evolution. [. . .] the market and the government are coevolving and form a symbiotic pair». Complexity economics adopts the Hayekian nomocratic approach (advocating avoiding top-down, command-and-control policies, focusing instead on “setting the eco-structure” enabling bottom-up self-organization), but with the important difference of considering the State able to creatively “alter” the “fitness landscape” on which the open-ended evolution of the economy takes place by manipulating the eco-structure (in order, for instance, to exit industrial lock-ins). These interventions cannot be considered “artificial,” because the self-structuration of the economic system intervenes ultimately on the State itself in terms of democratic processes or other feedback mechanisms, that is via a cybernetic operational closure. In this perspective, the “artificial”—“natural” narrative of the State/economy dichotomy is definitely obsolete, since the «Government is as natural as the market» (Colander and Kupers, 2014: 21).
The complexity-based approach
One important consequence of the discovery of complexity is a semantic fracture between two adjectives that have been for long deemed synonims—“complicated” and “complex”. “Complicated” equals mechanistic, thus also deterministic and reducible, whereas “complex” means instead holistic, implying non-determinism and irreducibility. There is a crucial connection between complicatedness and the artificial world: indeed, the very term “mechanistic” directly recalls man-made machines—consider the famous Cartesian image of the “clockwork universe”—and these are in fact the perfect example of a linear, deterministic system. As Portugali (2016) noted, «[. . .] artifacts are essentially simple systems. Some artifacts, such as supercomputers, are very complicated, but are nonetheless essentially simple systems». The divide between complicatedness and complexity appears to be an insurmountable ontological divide ensuring thus an high impermeability between the two qualities that makes it, apparently, a very promising candidate for natural/artificial epistemic criterion. Indeed, this vision encountered particular favor in the field of industrial ecology, which contrasts the linear (cradle-to-grave) flow of matter and energy taking place in “artificial” systems and the circular (cradle-to-cradle) flux taking place in “natural” ecosystems (Ayres, 1989; Frosch and Gallopoulos, 1989). Furthermore, complexity science success in modeling living organic systems led someone to go further and consider complexity as the necessary characterization of natural/living systems versus artificial/material/abiotic systems. For instance, supporters of the so-called strong artificial life position maintain that life is not inherently linked to the organic substrate, but can be thought of an emergent property completely detached from the particular medium (Langton, 1986; Ray, 1992, 1993).
The arguments we are going to expose in our critique to the complexity-grounded conception of nature can also be applied to the Hayekian approach we problematized above. Indeed, these two conceptions are particularly close: complex systems theory appears to perfectly complement the taxis/kosmos approach, framing it inside a sounder conceptual framework. In fact, whereas the latter is a more naive approach focusing on the generic and imprecise notion of “spontaneity,” complex systems theory founds this distinction on a precise structural layout: natural systems are adaptive, non-linear and thus non-deterministic (complex), whereas artificial systems are non-adaptive, linear and thus deterministic (complicated). As Portugali (2008) remarked, complex systems are essentially unpredictable—indeed, the very term “counterintuitive” has been introduced to highlight the fact that complex behavior appear to defy human intuitive capabilities (Prigogine, 1984): their behavior, thus, is by definition “spontaneous”. Conversely, linear systems’ behavior is completely determined by their past, they can be reduced to the interaction between composing parts and, as a result, they are perfectly foreseeable once the initial conditions are given: their behavior, in other words, can never be “spontaneous” or “unplanned”. Thus, complex systems theory also allows to reunite in a coherent frame the notions of “spontaneity,” “reducibility” and “predictability”.
However, the relationship between complex and complicated systems is less straightforward than one could expect, and is worthy of a reflection. This is already true at quite a basic, definitory level: «[. . .] being subject to non-linear dynamics is not a necessary condition for a complex system. [. . .] In general, feedback can give rise to complexity even in linear systems. Neither non-linear dynamics or linear dynamics can be necessary conditions for complexity since complexity scientists also study static structures» (Ladyman et al., 2013: 38). The wickedness theory recently proposed by Andersson and Törnberg (2018)—building on the seminal insight by Rittel and Webber (1975)—is based onto an “anatomical partition” of complexity, which “extracts” from the unidimensionality of the simple versus complex/complicated a continuous bidimensional spectrum, the so-called SOS (Spectrum of Overwhelming Systems)—displaying on one axis the degree of complicatedness and on the other the degree of complexity—and situating a number of systems (societies, organizations, organisms, et cetera) on it. In our opinion, this approach yielded three theoretical results being of the greatest importance for the ontological discussion we are carrying on here and for general planning theory. The first one is that, when we have to locate systems inside the SOS, many organic systems are to be classified as complicated instead than complex (Andersson and Törnberg, 2018: 121, identified in technology and organisms the central examples of those systems belonging to the “complicatedness” category): consider for instance the human body: arms or legs are deterministic systems, being centrally controlled by the nervous system. The second noteworthy result is that complexity and complicatedness appear to coexist in many systems—highlighting thus the existence of “hybrid” zones or trans-qualities. For instance, “organizations with human components” or “biological individuals” belong to the trans-complicated sector, whereas “social media movements” can be labeled as trans-complex (Andersson and Törnberg, 2018: 123–124). This is interesting because it highlights a point that is too often overlooked when analyzing human organizations within the frame of complexity: human organizations have also an important deterministic side—a side that in many cases may even be dominant: think for instance to the command chain in the military or to the governmental hierarchy in the public administration, enabling a direct top-down command. The third—and, in our opinion, also the most important—result from Andersson and Törnberg’s approach is the idea that complicatedness, complexity or the trans-qualities are not fixed ontological qualities of systems—instead, systems can move within the SOS. Take for instance the endosymbiont hypothesis: the first step was a situation of competitive interaction between organisms (wickedness); then an increased level of cooperation (trans-complicatedness); finally, co-adaptation of components led to a complicated system (Andersson and Törnberg, 2018: 121). We have a complete movement within the SOS. The dynamic element characterizing this approach gives us an important insight, and it opens up a promising research agenda pointing toward shedding light on the mechanisms through which complicated systems (such as the State, for instance) emerge out of initially complex systems. The crucial element is the alignment between the component, enabling an originally complex (or meta-complex) system to become a complicated one: in fact, «[. . .] human organizations are highly preoccupied with the problem of internally aligning interests and actions. By contrast with the biological cases [. . .], however, the problem is never solved here. Alignment is an ongoing and often highly costly effort of negotiation, persuasion, monitoring, punishment, reward etc.: a struggle to pull organizations away from wickedness, toward the complicated regime where design and governance is more straightforward» (Andersson and Törnberg, 2018: 123). Whereas in fact biological deterministic systems emerge out of an alignment driven by the search for evolutionary advantage, human society remains composed by autonomous individuals: we could say, thus, that while in the biological case this kind of transition is irreversible, in human societies it is more akin to a hysteresis. Notwithstanding these differences, the conclusion we can draw is that the divide between complex and linear systems is not deep enough to be used to define the natural∼artificial couple.
The “aggregate” approach
Finally, let us briefly reflect on the internal coherence of the general epistemic criterion we may derive via a simple linear aggregation of the three main conceptions we highlighted. The wickedness theory explains that the relationship between ontological and ontogenetic approaches is not theoretically strong: the naive correlation between complicatedness and artifacts is due to the fact that components of technical systems (such as machines) are automatically enslaved in a deterministic fashion, that is the parts do not need to pursue actively alignment (either due to negotiation, persuasion, etc. for the social case or due to evolutionary dynamics for the biological one) to be “enslaved” and work linearly. It is true, thus, that man-made objects are complicated and not complex, but the relationship does not work the opposite way: complicatedness does not imply necessarily that an object is man-made, as we have seen in the case of biological complicatedness. Furthermore, and more importantly, notice that this relationship is valid only at the current state of technology, that is as long as human technology is only able to produce complicated systems such as deterministic machines. However, this is no longer the case: what about man-made complex systems such as neural nets? The very latest developments in technology added a further layer of ontological confusion, and made the linkage between the different conceptions of nature more fragile than ever. In 2019, researchers were able to synthesize xenobots—“living robots” made of frog embryonic cells (for an in depth introduction to the issue, see Coghlan and Leins, 2020; Kriegman et al., 2020). As Brown (2020) noted: «They’re neither a traditional robot nor a known species of animal. [They are] a new class of artefact: a living, programmable organism». Xenobots spurred an important ethical and philosophical debate, which, due to the extraordinary recent nature of this breakthrough, is still developing while we are writing the present paper (see Botkin and Kowacki, 2020; Coghlan and Leins, 2020). As it has been pointed out, the creation of the xenobots—while per se extremely challenging from the philosophical standpoint—is merely a highly demonstrative accomplishment whose true revolutionary potential becomes clear only when we see it as a dramatic manifestation of the power of Artificial Intelligence (xenobots were engineered by an AI), thus framing it within the broader debate on AI—and this brings the “theoretical battle” on a field being way closer to cities. In fact, the development of smart cities will make AI more and more a protagonist of urban planning, both at the theoretical as well as at the practical level (for a notable contribution to the seminal literature on this theme, see Kirwan and Fu, 2020). If a city is a cyborg, should an AI be considered part of the “natural” element of the city? Or does it belong to the “artificial” side? Notice the paradoxical nature of these questions, even from the terminological standpoint. The ontological status of AI is tremendously murky (Lavelle, 2020), and this recent developments made the issue even more grave: we are facing a complex autonomous “artifact” (AI) that not only is able to behave adaptively (i.e. in a way not pre-programmed by its human creators, making every criterion founded on linearity/foreseeability pointless) but is even able to design other objects (in this case, even trickier, another quasi-autonomous object composed by organic cells). The conceptual complexity of the issue appears overwhelming.
Beyond “natural” and “artificial”: A possible synthesis?
So far, we demonstrated that none of the conceptions of nature prevailing in the planning literature and in the broader philosophical debate as well is able to express a coherent epistemic criterion to draw the distinction between the natural and the artificial. This last section is devoted to reflect on the consequences of this result. Should we interpret it merely as a sign of the inadequacy of our current theoretical tools? Or should we abandon the very notions of “artificial” and “natural”? In this case, which approach should we adopt?
As for the first question, we believe that our inability to answer the question in satisfactorily terms relates to a broader philosophical conundrum. Notice that, from a purely philosophical standpoint, all these difficulties can be traced back to the fact that, once we abandon the assumption of human exceptionalism to see mankind as an integral part of life on Earth, we perform a reintegration of the observer in the observed domain: due to this reintegration, we have a coincidence between epistemic object and subject, a self-referentiality whose main outcome is the generation of logical paradoxes (Le Moigne, 1985; Morin, 1993).
If humanity is part of nature, it follows immediately that, as Gecow (2010) observed, «[. . .] artifacts are parts of human civilization, which itself is a part of life evolution. Their status is thereby the same as that of a bird nest, ant-hill, honeycomb, or even a snail shell». Consequently, there is a growing awareness of the fact that active adaptation of the environment to the specific needs of the species is an inherent characteristic of virtually all forms of life on Earth: the reflection on “animal architecture” (for instance beavers’ dams, birds’ nests, termites’ mounds) carried out within complexity theories of cities eloquently testifies this novel post-exceptionalist attitude (Dong, 2016; Turner, 2016). Extremizing this line of reasoning, let us also add that even radical modification of the climate is not a prerogative of the human species: as Ayres (1989: 1–2) pointed out, the carbon dioxide that built up the atmosphere of primordial Earth was a waste product of the metabolism of early fermentation-based life forms and, in a similar fashion, oxygen was added up to the atmosphere as a by-product of the metabolism of anaerobic photosynthesizers such as cyanobacteria. Evolutionary dynamics led novel life forms to use the waste products of others as inputs, spontaneously generating thus a circular energy exchanges layout. We could argue that even the air we breathe, the quintessential “natural” thing, is, to a broader perspective, something “artificial”. Indeed, the entire history of life on Earth is the story of a never-ending, complex overlapping of different “interventions” on the climate and environment: the resulting change in fitness landscape led new species to find new niches in these mutated environments, causing at the same time, in their turn, novel environmental mutations to be exploited by other species (Day et al., 2003; Kauffman, 2000).
We may be tempted to stop here and radically dissolve the artificial into the natural. However, the reasoning we have been conducting so far tells only a half of the history. In fact, at the same time, reintegrating humanity within “nature” calls for a deconstruction not only of the “artificial,” but also of the “natural”. It is useful to recall here part of the critical reflection carried out by Cronon on the notion of “wilderness,” which exposes the internal incoherence of the conservation approach and the hidden exceptionalist conception on which it rests: «[. . .] wilderness is not quite what it seems. Far from being the one place on earth that stands apart from humanity, it is quite profoundly a human creation—indeed, the creation of very particular human cultures at very particular moments in human history. It is not a pristine sanctuary where the last remnant of an untouched, endangered, but still transcendent nature can for at least a little while longer be encountered without the contaminating taint of civilization. Instead, it’s a product of that civilization, and could hardly be contaminated by the very stuff of which it is made. Wilderness hides its unnaturalness behind a mask that is all the more beguiling because it seems so natural» (Cronon, 1995: 69; emphasis added). In a similar vein, in his 2008 book Beyond Nature and Culture, Philippe Descola carried out a Derridean de-construction of the nature/culture dualism, pointing out that “nature” is inherently a cultural concept, neither universal nor eternal, but instead particular to each specific culture and furthermore suitable to change over time to time as a result of an internal dialectic political/philosophical process (Descola, 2013).
We see clearly how the reintegration of humanity inside “nature” generates a theoretical paradox: everything is “natural,” but nature itself is a cultural (i.e., an “artificial”) concept. As Le Moigne (1985) pointed out, however, the “methods of paradoxes” is a fruitful way to explore complexity, as the accomplishments of Gödel and Turing testify—but we need the right theoretical tools. The true question thus is: how can we internalize the self-referentiality of the natural-artificial relationship? How can we go beyond the cyborg-dualism informing current complexity theories of cities?
We believe that the complexity theories of cities could benefit greatly from the reflection carried out in another sub-field, urban political ecology. Urban political ecology can be considered as an attempt at internalizing the self-referentiality of the nature/artificial issue via introducing the idea of socio-nature. Indeed, urban political ecology is the result of a theoretical trajectory being quite similar to the one displayed by the complexity theories of cities: in fact, it emerged as a result of a long-term process leading to overcoming the human exceptionalist paradigm informing urban sociology (Catton and Dunlap, 1980). In fact, both the two main traditional approaches to the theorization of urban metabolism (Chicago school and industrial ecology) were built on a radical society/nature dualism (Wachsmuth, 2012). The Chicago school approach conceived urban metabolism as an exclusively social phenomenon, with no role at all for “nature” (Burgess, 1925; Park, 1915). It is true that the “new ecological paradigm” expressed by industrial ecology was successful in reintroducing “nature” in the frame by highlighting the role of ecosystems services as the vital supporting flows underlying urban metabolism (Foster, 2000; Girardet, 1996; Wolman, 1965)—nevertheless it kept the two domains rigidly separated, with the former being merely the “fuel” of the latter, substantiating an heavily human exceptionalist narrative—according to which “cities swallow nature”—that permeates still nowadays the environmentalist reflection (Clement, 2010; Girardet, 2008). Drawing on Lefebvre’s (1991) observation that the discursive/symbolic element is an irreducible feature of the material/technological production of “nature,” Swyngedouw (1996) maintained that «the “world” is a historical-geographical process of perpetual metabolism in which “social” and “natural” processes combine in a historical-geographical “production process of socio-nature” whose outcome (historical nature) embodies chemical, physical, social, economic, political and cultural processes in highly contradictory but inseparable manners». In this approach, everything mankind does is perceived as socio-natural, included «the production of dams, the re-engineering of rivers, the management of biodiversity hotspots, the transfiguration of DNA codes, the cultivation of tomatoes (genetically modified or not) or the construction of houses» (Swyngedouw, 2006: 27). In fact, every man-made object embeds the entire net of political, economical and ecological relationships that enables its production, and these dimensions are radically irreducible one to another, dismissing thus ab initio any possibility of a binary “reduction”. As Wachsmuth (2012: 517) acknowledged, «[. . .] simply renaming “society and nature” “socio-nature” does not make the binary disappear through force of will, but in substantive terms urban political ecologists have been more successful at dismantling the persistent conflation of society with the city and nature with the countryside than any other research program in the social sciences».
An important point of the socio-natural approach is its “transcendental” characterization, which is remarked many times by Swyngedouw (1996: 96): «[. . .] we must insist on the need to transcend the binary formations of “nature” and “society” and to develop a new “language” which maintains the dialectical unity of the process of change as embodied in the thing itself» (emphasis added). And again, «[. . .] Every body and thing is a cyborg, a mediator, part social part natural but without discrete boundaries and internalizes the multiple contradictory relations that re-defines, re-works every body and thing» (emphasis added). This emphasis on the need to develop a novel theoretical language overcoming radical duality poses urban political ecology and complexity theories of cities in a largely convergent trajectory. Indeed, cutting edge reflection carried on in state-of-the-art complexity theories of cities is moving in this very direction. In an important contribution, Kelso et al. (2016) applied Kelso and Engstrøm’s (2008) philosophy of complementary pairs to urban planning. This cognitive theory aims at modeling human mind’s innate tendency to dichotomize. “Natural” and “artificial” are the perfect example of one of the Pythagorean-like opposites human mind automatically applies to reality. Building up on the seminal intuition by Nobel laureate quantum physicist Niels Bohr—the idea that contraria sunt complementa, Kelso et al. (2016) maintained that «Complementary pairs are those things, events and processes in nature that may appear to be contraries, due in part to our ubiquitous tendency to dichotomize, but are mutually related and inextricably connected». The theoretical framework on which this theory is based on is metastable coordination dynamics: in this state no fixed points exist at all, and the system keep itself in a “transient stationarity” where the remains of fixed points generate kind of a “ghost attraction” (Kelso, 2012; Kelso and Tognoli, 2014). The interesting point is that, in this state, opposite tendencies appear to coexist: metastability overcomes the contradiction between contraries and appears to be able to harmonize them in a cognitive coniunctio oppositorum. Consequently, Kelso and Engstrøm’s (2008) introduced the visual sign “∼” to signal the inextricable complementarity between aspects: in this novel conceptualization, thus, the apparent natural/artificial contradiction becomes the natural∼artificial complementarity.
Not only the philosophy of complementary pairs constitutes an important sign of the direction undertaken by current complexity theories of cities reflection, but also, due to its generality, enables us to interpret the similarities as well as the differences between urban political ecology and complexity theories of cities. In fact, Kelso and Engstrøm (2008) pointed out that complementary pairs are subject to four basic types of interpretations divided into two main categories: both/and and either/or (Kelso et al., 2016: 49). Both complexity theories of cities and urban political ecology, by introducing hybrid concepts, moved from the latter to the former—at the same time, they are different with respect to the specific interpretation given to the complementary pair within the both/and domain: complexity theories of cities preserves ontological separation (dualist both/and), whereas urban political ecology adopts a monist both/and vision, stressing a more transcendent approach to complementarity. As we noted, in fact, the complexity theories of cities are now built on a cyborg conception seeing cities as hybrid meta-complex systems composed of a “natural” part (humans) and an “artificial” one (buildings and infrastructure). This conception assumes that we can identify the one and the other—and we tried to show that, while this can be hold true at a simple, aggregate level of representation, the actual complexity of this ontological issue has been overlooked. Here is our final provocation/suggestion: given the fact that future technological developments in urban planning (AI, biotechnology, nanotechnology) will more and more blur this distinction making the cyborg conception of the city no more a completely reliable conception, complexity theories of cities should “quantum jump” toward a more transcendent socio-natural conception. In other words, the complexity theories of cities should focus on overcoming the right dualism, which is not the natural/artificial one (which proved to be largely inconsistent) but instead the culture/nature one, toward a novel conception internalizing the paradox derived from the epistemic reintegration that implies that everything is “natural” and, at the same time, everything is seen and categorized through a lens being cognitive, cultural, and political. Swyngedouw (2015) himself suggested an implicit linkage with complexity theories when pointed out that «[. . .] co-produced natures [. . .] are complex, chaotic, often unpredictable, radically contingent, historically and geographically variable, risky, patterned in endlessly complex ways and ordered along “strange” attractors (Prigogine and Stengers, 1985)». After all, after the “cognitive turn” that initiated when Portugali (2011, 2016) remarked the limitations of modeling the city as a simple complex systems composed by cellular automata and called for taking into account the cognitive peculiarities of the human agents populating the city, the complexity theories of cities moved more and more toward a reintegration of this semiotic/cognitive/behavioral element. The passage from a rigid, ontological to a dynamic, cognitive conception of the natural/artificial issue was already present in nuce when Portugali (1999) applied to cities Wittgenstein’s “family resemblance” categorization, overcoming a logical atomist perspective and acknowledging instead the more complex network structure defining the genesis of categories. Furthermore, Haken (2006) already attempted at analyzing the natural/artificial divide from a purely cognitive vantage point. The next step may involve coming back to this intuition and expand it in order to include its inherent dynamical aspect by introducing also the political element and the synergetic feedbacks informing the formation of ideas in human social systems.
Conclusions
This paper aimed at reflecting critically on the ontological issue that derives directly from the fact that nowadays the city is typified as a cyborg system. We pointed out that no conception of nature is strong enough to constitute a rigorous epistemic criterion able to draw the distinction between the natural and the artificial, and this inability may potentially put in crisis the very conception of the city as a hybrid. In fact, it is true that, nowadays, cities can be thought as entities composed only by human agents being complex systems and artifacts such as buildings or computers being complicated systems, but in the near future AI, nanotechnology and biotechnology will dramatically impact urban environments, and thus urban planners and theorists will necessarily have to include them in the picture: to which side? As we demonstrated, the difficulties are remarkable: the old dualism artificial/natural, thus, will appear more and more obsolete when typifying cities as cyborgs. We observed thus that, for complexity theories of cities to continue being a valid interpreting framework for cities, it will need to overcome the dualist both/and approach toward a more socio-natural vision internalizing the self-referentiality deriving from the reintegration of mankind within “nature,” the impermanence and cultural relativity of the concept of “nature” itself and its continuous transformation as well, modeling it in terms of complex cognitive dynamics taking place both at the level of the individual and at the social level as well.
