Luhmann,Latour and global petroleum governance

First limitation: unclear relation between states and markets

When international studies scholars once again turned their attention to energy in the early 2000s, initial works focused on potential resource conflicts in an anarchic, state-centric world (e.g. Klare, 2002). Other scholars quickly intervened, highlighting the potential for inter-state cooperation. Either/or debates ensued: In approaching the rather grand theme of Energy and the Transformation of International Relations, for example, Wenger et al. (2009: 5) ask ‘[w]hether the search for energy security will lead to greater international conflict or cooperation’. GEG scholars highlight the unproductive narrowness of such cooperation-vs.-conflict framings. GEG focuses more on pragmatic problem-solving. It stresses the importance of a broader set of issues such as environmental sustainability, energy poverty, and investment. GEG’s most defining trait, however, is its focus on increasing global institutionalization (Van de Graaf, 2013; Wilson, 2014). Based on institutionalization as an indicator of change, scholars identify ‘inadequate and uncoordinated mechanisms’ (Dubash and Florini, 2011: 15) and an institutional ‘landscape of wreckage’ (Victor and Yueh, 2010: 67). Encountering such inadequate institutionalization, scholars often turn back to the image of an increasingly ‘multipolar [inter-state] world’ (Lesage et al., 2010), thereby reproducing orthodox conceptions of the inter-state system and the primacy of politics.

How international politics relates to economic matters remains unresolved. Realist authors, for example, tend to perceive oil firms as mere power instruments of states. Some GEG scholars include economic themes and actors in their elaborations and highlight that markets, too, are governed by rules (Goldthau and Witte, 2010). Others, however, explicitly exclude economic matters from global governance (Florini, 2010: 150; Wilson, 2014: 5). Yet others argue that global energy governance is torn between the poles of ‘market and geopolitics’ (Youngs, 2009) without conceptually clarifying how these poles might interact.

Some scholars have highlighted the potential added value of the field of International Political Economy (IPE) in understanding the interaction of politics and the economy (Kuzemko et al., 2012). Palazuelos (2012), for example, develops a framework consisting of players, scenarios, and exchange mechanisms. This effort is laudable in that it gives equal standing to states and energy firms as well as to (political) institutionalization and market mechanisms. However, in the end, it remains a fairly loose collection of variables.

Enter Luhmann: functional differentiation and structural coupling

In international studies itself, Buzan and Albert (2010) have argued that Niklas Luhmann’s modern systems theory ¹ provides a sufficiently abstract yet precise approach to make sense of global systemic differentiation such as that between state and market. For Luhmann (1995; 1997), modern society consists of functionally differentiated systems, such as politics, the economy, law, science, and religion. The operation of each system is based on a binary code such as power/no power in the political system and payment/no payment in the economic system. Only when elements (in Luhmann’s case: communications) make sense from the perspective of a system’s code do they become relevant for the system. Only if energy can be conceptualized in terms of power, for example, can it enter the political system.

This perspective helps to overcome the limited conceptions of markets in global energy studies. First, functional differentiation allows us to understand the international political system and the global economy as independent function systems. Such symmetry between markets and politics counteracts the above-mentioned asymmetric reduction of complexity by which economic elements tend to be dropped from the ‘big picture’ of GEG. Second, Luhmann’s theory shifts the focus from an ‘either states/or markets’ debate to a ‘both/and’ one. A modern systems theory perspective demands greater emphasis on the parallel, complex, partially contradictory, and often uncoordinated ways in which different function systems structure their relationship to the environment ‘in terms of’ energy (Luhmann, 1997: 132). Third, function systems are structurally coupled to their environment, meaning that systems ‘are oriented by their environment not just occasionally and adaptively, but structurally, and they cannot exist without an environment’ (Luhmann, 1995: 16–17). Structural coupling extends to the relation between function systems. It thus allows the research focus to be directed towards how, exactly, political and economic systems hang together. The added value of systems theory for understanding the interaction between politics and the economy has also recently been highlighted in international relations itself (Pena, 2015).

Second limitation: a deterministic image of energy and a partial solution by Luhmann

While Luhmann’s functional differentiation and structural coupling provide a fresh look at the role of markets in energy governance, a second limitation remains. In global energy studies, energy almost exclusively features as a structure that is external to social relations, one to which the latter has to adapt. This has made it easy for realist scholars to summon the ‘inescapable constraints of geography’ and to assert that growing scarcity ‘will inevitably intensify competition between states over access to vital materials’ (Klare, 2002: 44, 23). Cooperation-focused approaches, too, state that while the prospect of ‘shrinking energy resources contains the potential for conflict,…the current situation may also present opportunities for global cooperation’ (Perovic and Orttung, 2011: 211). GEG scholars also buy into this deterministic image of energy and tend to resort to normative claims in their attempts to overcome it: new forms of energy governance ought to be established to overcome increasing scarcity (e.g. Lesage et al., 2010; Victor and Yueh, 2010).

Luhmann’s work also helps to overcome this second limitation. He conceptualizes system environments as ‘the system-internal correlate of all references that extend beyond the system’ (Luhmann, 1989: 22). The specific environments of sub-systems such as politics or the economy comprise not only material features. They also include communications that are external to the specific sub-system. Since the environment is (sub-)system-specific, furthermore, every sub-system has its own conception of energy. The international political system might have a tendency to reduce environmental complexity by conceptualizing energy as a scarce resource that can be employed as a power asset. This conception, however, is internal to the system itself. At the same time, the economy might see energy as a resource that needs to be distributed effectively, based on the willingness of market participants to pay. Both (and potentially more) conceptions of energy exist in parallel – a clear broadening of perspective compared to most GEG approaches.

Luhmann’s account remains limited in another regard, however: Luhmann sticks to the dichotomy between society and an external, physical environment. His theory aims to conceptualize the social as consisting of communication and communication alone. While a sub-system’s environment can consist of both communication and material aspects, there is a clear hierarchy. For sub-systems, the ‘entire system [of society] then acquires the function of an “internal environment”’ (Luhmann, 1995: 14, 189–90), as opposed to an external material world. Even if politics and the economy were each to create their own environments, at the end of the day, they stand united against an external, material environment. Only communication about energy can enter society, not energy itself. Such an understanding opens the door to slipping back into simplistic dichotomies between society and an external, and potentially scarce, energy resource.

Luhmann and the refinery

Luhmann’s conception of environment only goes so far in overcoming the limitations of global energy studies. This can be illustrated by a more detailed look at the 2008 oil price crisis. In contrast to the simplistic narratives provided by realist energy authors, the 2008 price spike was the result of a great variety of material but also societal factors (cf. Fattouh, 2011). One factor was the interplay between political regulations, crude oil sulphur content and the refinery infrastructure. In the early 2000s, industrialized countries had instituted lower sulphur content regulations for transport fuels. Transport fuels are refined from crude oil in refineries. To comply with regulations, these refineries began to buy more low-sulphur crude and adapted their refining processes, which resulted in decreased refinery output. The price of both low-sulphur crude and transport fuels rose substantially.

With Luhmann, regulations can be understood as political communication. Refinery organizations were irritated by this communication and adapted by changing their production plans in favor of lower-sulphur fuels. These organizations, social systems themselves, were structurally coupled very closely to technical refinery systems. It was the inadequacy and subsequent re-arrangement of this refinery infrastructure that resulted in lower production volumes. Any refinery communication to the market – in the form of production volumes – was thus limited by the materiality of the refining infrastructure. In the market, communication resulted in rising prices. In the political system, in turn, these price communications were interpreted in terms of energy security and resource scarcity.

One can go down this Luhmannian road. However, caveats apply. First, focusing on the role of a concrete refinery system is unusual for Luhmann’s systems theory that tends to discuss system environments at rather high levels of abstraction (Luhmann, 1989). Second, even if the material refinery system’s importance is acknowledged, acknowledgement remains indirect. Systems theory would usually focus on general patterns of communication rather than on the material preconditions of specific acts of communication. Function systems normally come first, followed by organizations, and only then, at the end, is there the refinery itself. One of the major explanatory factors, in other words, is theoretically sidelined. Third, the Luhmannian macro-perspective is stretched even further when considering that the above example is but one of a variety of rather concrete material-social interactions that contributed to the price crisis of 2008. Other such interactions include disruptions between financial oil futures markets and the physical oil trade, as well as wasteful consumer behavior motivated by fuel subsidies. The empirically most interesting aspects of global petroleum governance, in other words, occur in the structural coupling of systems with their material/infrastructural environment rather than in Luhmann’s major units of analysis: functionally differentiated communication systems. Latour’s work provides the means to tackle this problem head-on.

Integrating energy: Latour’s network ontology

Bruno Latour’s empiricism and opposition to macro-theories are of little help in approaching the relation between states and markets. His work, however, provides ample resources to re-conceptualize the relation between society and energy. ² Latour’s (1993; 2004; 2005) default position is one of symmetrical anthropology: the distinction between society and Nature is a modern illusion. There are not only human actors ‘in here’ and non-humans ‘out there’. An increasing number of hybrids, such as climate change or genetically modified organisms are stretching and crossing this distinction. In order to come to terms with these new entities, Latour argues, we need to treat humans and non-humans equally. He decouples agency from its sole reference to human action. Anything that makes a difference to the way networks unfold is a mediator, is an actor. From Latour’s perspective, humans, material objects, ideas, and a variety of other non-humans intermesh and form networks – any of them can potentially be a mediator (Latour, 2005: 128). ³ Societal and physical aspects of these networks cannot be separated in any meaningful way.

This approach provides a substantially different and more nuanced perspective on the role of energy in GEG. It calls for an explanation of how allegedly material phenomena – such as future demand or supply – are produced in concrete networks (Watts, 2012). A conception of energy as a unified and external material structure to which society has to react becomes impossible. There is nothing ‘natural’ and nothing inalterable about energy. It becomes a mix of alleged material elements and social factors that is already ‘in here’. Latour’s ambition is to explore complexity and to search for concrete configurations of networks in empirical practice. This approach provides a much more direct access point to the above case of the refinery, since it would suggest starting right there, with the pipes of the refinery and exploring how these hang together with a variety of other aspects such as ‘human resources’, production plans, pricing schemes, environmental regulations, and finally political and economic practices. It calls for an exploration of the 2008 price crisis ‘from within’.

Luhmann > Latour: hybrid systems on the periphery

In Luhmann’s social systems, personhood is not to be confounded with actual human beings. Personhood and agency are qualities that a system ascribes to elements it observes in its environment. Traditionally, these qualities have been ascribed to humans. Noe and Alroe (2006) aim to extend the ascription of agency to non-humans. They discuss farms as functionally specific, heterogeneous systems. Besides humans, they see cows, technologies, and so forth as actors in this context. They integrate such non-humans by supplementing Luhmann’s person/actor-human differentiation with an analogous actant-object/animal differentiation. In both cases, the former constitutes a representation of the latter within a social system. The authors fall short, however, of elaborating on how these new actants might actually partake in social relations/communicate.

Teubner (2006) goes one step further when he discusses the role of domesticated animals and machines as actors in law and politics. In general, elements have to fulfill certain conditions to qualify as actors within social systems: they have to be systems themselves, capable of communication. Teubner (2006: 513) turns to hybrids, i.e. mixtures of humans and non-humans. He understands hybrids as groups of very different, albeit tightly coupled, systems ‘within a common structural drift [meaning that they co-evolve]’. He then argues that these systems can indeed form a common social system when humans treat actants (non-human systems) ‘as if’ they were engaging in reciprocal communication. According to Teubner, the resulting social system is similar to a prayer which works as an interaction system, only because humans act ‘as if’ God was actually communicating. Under these conditions, hybrids constitute social systems and can thus become actors in functional social systems.

Importantly, Teubner goes one step beyond Luhmann and the refinery. He moves from tight social-material coupling to actual system formation. However, three limitations remain. First, Teubner continues to sideline hybrid systems analytically and instead quickly shifts his attention to function systems. A second problem concerns the prayer analogy. Many non-humans appear to be more inclined to respond than God. It seems more likely, for example, to be bitten by a dog when opting for questionable communication than to be punished by the Almighty. Teubner’s conception of non-human communication thus remains somewhat unconvincing.

Third, while systems theory in general is conscious of irritations and conflicts between social systems, this point is underdeveloped in Teubner’s account. Similar to the above argument regarding multiple energies, he calls for multiplying the multiplicity of actors within multiple function systems (Teubner, 2006: 516). He thereby overlooks situations in which function system-specific ascriptions of agency collide. These situations might be important, however, particularly in the case of potentially scarce resources, such as oil or space. Stichweh (2008) discusses the latter example when arguing that function systems construct their own physical spaces. In the economic system, for instance, even online shops need to deliver their goods. In sports, it is the Tour de France that is particularly popular. What happens, however, when two systems try to use the same space at the same time? Online delivery vans and cyclists show that competing constructions of physical space might – quite literally – collide.

Latour > Luhmann: network formation around hybrids

I will now turn to Latour-inspired readings of functional differentiation and argue that these have the potential to answer Teubner’s limitations. First, as argued above, a Latour-inspired perspective shifts the focus away from the macro-sociological form towards the concrete intermeshing of humans and non-humans. Hybrids are explored from within. Second, Latour’s concept of mediation provides salvation from the prayer analogy. Instead of focusing on humans communicating with things ‘as if’ they were proper communication partners, mediation maintains that any component of a network that makes a difference to how the network evolves – by biting, for example – can be understood as actor.

Finally, the third issue mentioned above – potential conflicts between function systems in referring to the same material element – calls for further elaboration. A good starting point is Randalls’ argument regarding climate change. He argues that it ‘is not a singular entity that is then made governable in different ways; rather climate change is inherently multiply constituted’ (Randalls, 2013: 240). He seems to share Teubner’s position on the multiplication of agency. He departs from Teubner, however, by highlighting a tension between a system-specific and a system-transcending existence of hybrids. The goals and objectives of different assemblages, he argues, ‘are held in tension…by the way they conflict materially and discursively with each other [over climate change]’ (Randalls, 2013: 240).

Guggenheim’s (2009) discussion of historical buildings elaborates this tension more clearly. Guggenheim argues that when non-humans are allowed into the social and when this social is functionally differentiated, the incorporation of objects into specific function systems might make them inaccessible to others. He highlights two strategies employed by society to deal with this situation. The first is isolation. Artworks, Guggenheim argues, can be placed in museums to withdraw them from the market. Isolation restricts access to concrete objects and knits them more tightly into the relevant functional network. Isolation is also key to avoiding the above-mentioned collisions between cyclists and logistics firms. Portions of space are isolated from the economic system so that they can be used by the sport system. The second strategy is multiplication plus standardization. Prefabricated houses, for example, are heavily standardized and, because of that, they can easily be multiplied to serve political as well as economic or other purposes.

Guggenheim argues that both strategies fail in the case of historical buildings. Such buildings are locally fixed and thus cannot be isolated from their environment. Furthermore, given their genuine history, they cannot be readily standardized and multiplied. As Disneyland and Neuschwanstein Castle illustrate, the purely physical form can be multiplied, but historicity cannot. Under these conditions, the multiplication assumption of systems theory becomes implausible.

[To use] a building always means to use it partially, to use it in the presence of others [inhabitants, researchers, tourists] and to interact with others in, through and with the building. The multiplication of uses creates networks that are outside of anyone’s control, and specifically outside the control of any functional system. (Guggenheim, 2009: 46–7)

Network ontology, operational closure and hybrid function systems

To arrive at this approach, the first step is to combine network ontology and operationally closed systems. Some might complain that these ideas are incompatible, since networks and systems are constituted of fundamentally different operations: systems are differentiated from their environment, networks by chains of associations based on mediation and translation. ⁴ This criticism, however, overlooks Luhmann’s insight that internal association and distinction from the environment are essentially two sides of one operation. While Luhmann highlights system differentiation based on a binary operational code, this differentiation can only persist in concrete and continuous chains of communication (see below for a discussion on communication and mediation). And although Latour explores networks from within, this exploration depends on a distinction between network and what he calls plasma: ‘that which is not yet formatted, not yet measured’ (Latour, 2005: 224). To decide on the network position of a road block, for example, we first need to know if we are observing a network of traffic management or bird migration – we have to determine system/network boundaries. The continuous internal production of networks/systems as well as their differentiation from an external environment/plasma are thus integral parts of both theories.

What, then, would a combination of network ontology and operationally closed systems look like? Operational closure implies that systems are not primarily concerned with transforming inputs into outputs. An operationally closed system cannot access its environment directly (Luhmann, 1997: 78–120). Instead, it produces itself from itself (autopoesis). Everything necessary for perpetuating operations (as well as system structures) stems from within the system. In other words, only communication communicates. Nevertheless, Luhmann is very aware that systems remain open in another sense: causation does not stop at system boundaries. Without the structurally coupled environment, including psychological systems and material objects, ‘absolutely nothing can happen’ in social systems (Luhmann, 1997: 96).

Latour’s networks bear surprising similarities to this conception. In short, Latour overcomes the asymmetry between humans and non-humans by creating another asymmetry, that between both of the former and communication. In order to regard humans and non-humans as equal in his symmetrical anthropology, Latour (2004: 83) turns to ‘propositions instead of and in the place of the earlier subjects and objects’. These propositions, however, are essentially a form of communication (also note the etymological roots of Latour’s Parliamentary analogy). This insight has further consequences. Latour (2005: 59) argues that networks are constituted of ‘mediators triggering other mediators’. If networks consist of propositions alone, however, only these can become mediators; and if all propositions are communications, then mediator becomes but another word for ‘important communication’. Latour’s above network-definition thus translates into the following: networks are constituted of communications triggering other communications – Luhmann’s approach to communicative autopoesis. As the above roadblock argument has shown, furthermore, the important in ‘important communication’ can only be evaluated by a differentiation between system/network and environment/plasma. Of course, Latour’s (2004: 71, 83) key concern is to show that communication is tightly structurally coupled to non-communicative and not exclusively human elements, since the latter constitute important speech (!) impedimenta. Just as in Luhmann’s case, causation is not limited to communication: A mediator (read: important communication) ‘is not the source of an action but the moving target of a vast array of entities swarming toward it’ (Latour, 2005: 46).

Given these parallels, the major remaining point of differentiation between the authors then appears to be where the analytical divide is made: Do we include the environment of communication in our system definition or not? Even with regard to this question, however, the authors’ opposition is not as clear as one might expect. It is Luhmann (1997: 92) who reaches out by calling for an ‘important though very narrow exception’ to the rule of operational closure: a system’s structurally coupled environment. It is precisely this exception that Latour is interested in. His network ontology transforms Luhmann’s exception into the rule.

The above argument allows for a new angle on social systems. Without having to let go of operational closure, it suggests understanding systems as hybrid systems. Although communication remains at a system’s core, the system also comprises tightly structurally coupled, non-communicative elements. Mediations, oriented by a binary functional code, become the major operations of hybrid systems. From such a ‘Luhtourian’ perspective, function systems look quite different from Luhmann’s original ones. Politics comprises power plays, political messages, politicians, ballot boxes, legislative procedures, executive orders, and the electorate. The economy, in turn, comprises price signals and trade, but also traders, stock exchanges, production sites, logistics partners, and sales personnel. Such theoretical propositions might obviously provoke a flurry of questions. I will address the ones that are most important for the theme of issue-specific governance, in the following.

System formation, symbolically generalized governance objects, and resistant hybrids

Can ‘Luhtourian’ hybrid systems help to explain the formation of a petroleum governance system? To find an answer, a closer look at system formation is needed. When discussing (function) system formation, Luhmann highlights a specific sequence: problem of reference (PoR) – symbolically generalized communication media (SGCM) – operational closure. PoRs are ‘trivial, everyday situations’ (Luhmann, 1997: 358; see also Farias, 2013) between alter and ego, both of which either take a passive (experiencing) or active (acting) role. ⁵ SGCM emerge in response to PoRs and help to routinize their treatment. Money, for example, has helped to routinize situations in which ego experiences that alter is acting towards a good which also interests ego. While Luhmann argues that SGCM are important in function system formation, he also finds systems which lack such SGCM: religion, education, and medicine. He thus concludes ‘that the functional differentiation of the societal system cannot simply follow the media scheme (Medienschema)’ (Luhmann, 1997: 407–8).

According to Luhmann, religion, education, and medicine are special in that they have a function ‘that consists in a change in the environment’ (1997: 407–8). Although this similarity might suggest a more systematic treatment of non-SGCM system formation, Luhmann leaves this explanatory gap open. Indeed, exploring it further would collide with communicative autopoesis. Network ontology, by contrast, suggests a more systematic way ahead. While Luhmann considers PoRs between subjects only, network ontology requires one to also acknowledge PoRs that might occur between subjects and objects. Luhmann argues that PoRs can be routinized by SGCMs. A similar routinization should be expected in the case of subject-object PoRs. I would like to call its result symbolically generalized governance objects (SGGO). Oil is a good example of an SGGO. Humans continuously face the challenge to extract, transport, refine, and sell crude oil. These situations constitute a subject-object PoR. Single manifestations of this PoR vary, depending, for example, on different geologies or crude grades. From an overall societal perspective, however, such variation is secondary. Oil as such is instead rendered governable in a symbolically generalized way.

SGGOs fill the gap in Luhmann’s Medienschema discussion. They can be found in all three cases that defy this scheme of system formation: God (religion), knowledge (education), health (medicine). The formation of these systems follows a ‘PoR–SGGO–operational closure’ pattern. But why do some objects/hybrids provoke system formations and others not? After all, it appears rather odd to discuss global road block governance. With Guggenheim, I would like to suggest that system formation is the effect of certain object/hybrid properties. Guggenheim mentions two such properties: a hybrid’s resistance to isolation and to standardization/multiplication. I would like to add a third property here: the resistance to substitution. If one or more of these properties occur in a hybrid, it becomes a resistant hybrid that might provoke the ‘PoR–SGGO–operational closure’ pattern. In the next section, I illustrate this argument with regard to petroleum governance.

Petroleum as a resistant hybrid

To conceive of petroleum governance as a system in the above-mentioned hybrid sense, it is crucial that petroleum qualifies as a resistant hybrid. We thus need to examine its resistance to isolation, standardization/multiplication, and substitution. First, there is no way in which a single function system can meaningfully isolate oil. Oil is a notoriously ambivalent resource, one to which neither politics nor the economy has ever managed to claim exclusive access (e.g. de Graaff, 2012: 288–9; Yergin, 1991). The point here is not whether function systems can treat oil in their own ways; they will certainly try. When assuming hybrid function systems, however, such treatment presupposes the actual incorporation of oil into the system. One specific barrel of oil cannot be integrated and isolated by the political, economic, and legal system at the same time. Instead these systems have to struggle over this specific barrel and how to handle it.

This situation could be overcome if oil was a standardized commodity, readily multiplied to meet the needs of all function systems. Road blocks and bikes are good examples of such goods. Important aspects of oil are indeed standardized: the barrel as a basic unit of measurement, the characterization of oils based on the America Petroleum Institute’s (API) gravity standard, the oil price structure based on benchmark crudes and price differentials, as well as standardized products on oil futures markets. Standardization alone, however, is of little help as long as it is not supplemented by sufficient multiplication. Petroleum is a finite resource. Although it is possible to expand production, this is complicated by notoriously long lead times, high investment volumes as well as geological and technical complexities.

Also, the third strategy – substitution – appears unrealistic, at least in the short and medium term. Of the fossil resources, oil is special in that it is the prime fuel in the transport sector. Substitution will not occur until there is a significant technological shift in this sector, towards electro-mobility, for example. Hence, since petroleum cannot easily be isolated, standardized/multiplied or substituted, it appears reasonable to assume that it might indeed spark the formation of an operationally closed system.

The formation of a global petroleum governance system

I will now explore the operationally closed governance system that the SGGO of oil has given rise to. First and foremost, a system needs to differentiate itself from its environment. While the operational codes of many systems are closely related to their SGCM (see: power/no power), global petroleum governance does not have such a medium. SGGOs, however, fulfill a similar function in system formation. Sick/healthy, the health system’s binary code, is closely related to its SGGO ‘health’. Following the same logic, the SGGO of oil thus suggests a ‘oil/no oil’ operational code. All mediations that are designated as ‘oil’ become part of the petroleum governance system, while all ‘no oil’ mediations do not. Based on this differentiation, a system can build up complexity and develop its own structures, such as roles and programs.

Programs (Luhmann, 1995: 314–17) specify the operations of a system and thus reduce the diversity of legitimate mediations that the binary code allows for. While petroleum governance’s code ‘oil/no oil’ has remained constant for the last century and a half, the programming of what petroleum actually means has changed substantially. While oil was seen as an ‘abundant resource’ in the early twentieth century, it turned into a ‘scarce resource’ during the two World Wars and, particularly, in the 1970s. In the 1980s and 1990s, the system’s governance object transformed into a ‘global oil market’ and, in the early 2000s, into a ‘volatile oil market’ (Richert, 2014). Institutions, highlighted by GEG scholars, are an important integrating force in the system and contribute to stabilizing specific conceptions of the governance object. The International Energy Agency (IEA), for example, reduces the complexity of the hybrid system into figures about present and future demand, supply, stocks and transportation. This integrative function, this formation of system identity, is also assisted by journalists as well as by the above-mentioned energy experts and scholars.

Further system elements aim to subject petroleum to economic, political, legal or other rationales. Legal activity is most prominent in the context of the Energy Charter, which aims to provide standards for the investment in and transit of energy resources. Economic activity is conducted by oil traders, such as Glencore, international oil companies, such as Shell and ExxonMobil, as well as by national oil companies, such as Saudi Aramco of Saudi Arabia. The example of Saudi Aramco also shows how economic and political practices overlap. While the company strives for profit, it is also confronted with the political demands of its government. Legal, economic and political practices are all part of the petroleum governance system. The petroleum governance system thus comprises different functional logics. Just as in the case of an historical building, these might relate to each other in different ways, including conflict. They do so, however, with constant reference to the systems ‘oil/no oil’ code. This reference, in turn, is not voluntary. Political practices, for example, cannot simply negate the fact that they relate to oil, precisely because of oil’s resistant character. As a consequence, all these practices become part of the oil governance system.

Issue-specific systems and function systems

Systems theorists are likely to object to this argument, since for them most issues – including petroleum – are polycontextual, treated differently by different function systems (see the multiplication assumption above). On a somewhat broader level, for example, Stichweh (2007) and Albert (2016) understand global governance as a differentiated sub-system within the political system. Fischer-Lescano and Teubner (2007) highlight issue-specific differentiation in law. Politics and law might both deal with challenges such as petroleum. These dealings, so these accounts suggest, remain neatly confined to the respective function systems. Whatever happens between these systems comes under the label of structural coupling.

The above argument instead calls for a different approach. Fischer-Lescano and Teubner (2007) themselves argue that issue-specific differentiation in law is driven by differentiation in other, structurally coupled function systems. One might extend this claim and argue that sub-functional systems, such as energy politics and law, might develop a common structural drift. Teubner himself invokes the possibility of system formation in this case (see above). It is precisely this structural drift that is generated by resistant hybrids.

This argument puts issue-specific governance systems and function systems into a contingent and variable relation. Is a petroleum geologist, for example, practicing geology or finding out about oil reserves? Is an OPEC politician practicing power politics or governing oil? Is an oil trader behaving economically or handling volumes of oil? The phrasing of these questions conveys an important message: We ought to accept that communication is already always doubly differentiated. One observer might group communication in terms of functional logic – energy law and environmental law – another might prefer an issue-driven logic – energy law and energy politics. Systems theory favors the former method, and it mainly does so by theoretical fiat. Commentators have indeed criticized Luhmann for betraying his empiricist approach by focusing on large function systems rather than on more concrete governance systems (Mayntz, 1988). One legitimate way of responding to this strategy would be to once again invoke theoretical pluralism.

An alternative, and I would argue, more useful strategy is to empirically analyze the actual tensions between issue-specific and functional hybrid systems. One interesting way of finding out about the relative importance of hybrid governance systems and larger function systems is to look at professional roles. A manager, for example, might move from managing oil to managing circuses. She might as well, however, move from managing oil to becoming an oil administrator in government. While both have occurred in the history of oil, the latter is not only more interesting analytically, but also more common in global petroleum governance (Yergin, 1991). In this case at least, the ‘oil/no oil’ code thus prevails over its functional alternatives. ⁶

Finally, when exploring this struggle between issue-specific and function systems, it is important to account for developments over time. The reproduction of an oil governance system is contingent on the continuous failure of isolation, standardization/multiplication, and substitution strategies. The recent decline in oil prices might suggest that by now oil has been sufficiently multiplied. This, however, would hardly be enough to question the importance of the oil system. All-important inertia holds the basic system together. Production facilities exist and it makes little sense to simply abandon them. The global diffusion of internal combustion engines will ensure continuing demand. OPEC states will remain reliant on oil income, international oil companies are unlikely to radically change their business models all at once, and also the IEA is hardly going to stop producing oil facts. A more long-term and potentially more serious shift might occur in the area of the substitution of internal combustion engines in transportation. In the long run, meaning several decades from now, this might actually result in a less resistant hybrid of petroleum and a less important oil/no oil code. For now, however, there are good reasons to regard global petroleum governance as an integrated, issue-specific governance system.