Chapter 2. Governing low-energy housing: Assumptions, limitations and alternatives
2.3 Socio-technical systems based approaches to the governance of low-energy housing
2.3.2 Actor Network Theory
In recent years, an increasing amount of research has thought relationally about socio-technical transformations in response to calls for a greater focus on the politics of socio-technical change (Meadowcroft, 2009) and for less techno-centric and linear approaches (Shove and Walker, 2010; Lawhon and Murphy, 2011). As discussed by Longhurst and Chilvers (2013), concepts developed in response to this challenge encompass different notions including: actor networks (Callon, 1986), hybrid collectives (Latour, 2005), socio-technical ensembles (Bijker, 1995), mangle of practice (Pickering, 1995), agencements (Callon, 2007), assemblages (Deleuze and Guittari, 1986) and action nets (Czarniawaska, 2008). Collectively these conceptualisations are known as
‘arrangement’ or ‘assemblage’ theories. Whilst distinct differences distinguish these notions, they all attempt to analyse the complexity of social life (Schatzki, 2002: viii; 2011).
The most notable arrangement theory is Actor Network Theory (ANT) (Latour 1983, 1991, 1992, 1993; Callon 1986; Bijker and Law 1992; Law and Hassard, 1999). Whilst Domestication theory is primarily concerned with the household as a closed system in which new technologies are accepted or rejected, ANT argues that human behaviour co-evolves with different technological systems in heterogeneous socio-technical networks.
It therefore has greater potential for understanding systemic transformations of the residential sector and the energy implications of domestic life. ANT seeks to understand how social and material elements – termed actants – interact to generate knowledge of the world, through processes of social construction and material resistance. The theory suggests that the world is assembled through these interactions and networked connections, and that therefore to make sense of socio-technical stability and change these relations need to be understood. ANT is based on three principles: (1) generalised symmetry between social and material actants; (2) radical relationality between elements;
and (3) association between these actants as a means to achieve change (Farias, 2009: 3).
First, a central tenet of ANT is the commitment to according equal ontological status to natural and social objects. For ANT there is no difference between the social, the natural, ideas, policies, technologies, or infrastructures; according to the notion of generalised symmetry, they are all considered as actants assembled into networks (see Callon, 1986). People form but one element of any assemblage, and as such, the social is ‘nothing other than patterned networks of heterogeneous materials’ (Law, 1992: 2). ‘If human beings form a social network it is not because they interact with other human beings. It is because they interact with human beings and endless other materials too’ (Law, 1992: 3, emphasis added). For instance, keeping comfortably warm in a conventionally heated home involves engaging with an extensive heterogeneous network that includes; the householder(s) themselves, heated air, layers of clothing; the heating system and radiators, the energy source and energy distribution system, and regulation and technical procedures relating to energy supply systems, heating systems and housing.
Second, radical relationality refers to the principle that there is no a priori significance or properties attached to a given object, person or idea. Neither actors, technologies, nor procedures possess fixed attributes that allow them to bring power to play in a given situation. This is because ‘all points in a network are potentially equal in terms of their determining power’ and ‘the ability to determine the shape of the interaction and of the network is produced by the network, by the interaction of its parts’
(Cavanagh, 2007: 34-35, emphasis in original). For example, a cold spell may lower the gas pressure serving the boiler causing the heating system to fail. But similarly, a householder might lower the thermostat and/or choose to turn-on the heating system only at times when they can obtain a cheap tariff.
Third, ANT networks do not simply link up stable constituents but transform the actors, things and procedures held in the network, what ANT refers to as punctualising.
As such, a whole network ‘acts as a single block’ (Law, 1992: 5) becoming greater than the sum of its constituent parts; ‘it disappears from view as a network and reappears as a unity, as an actor in its own right’ where different parts become ‘locked-in’ to certain roles (Cavanagh, 2007: 34). If actants are de-enrolled from the network and individual components struggle to pursue their individual goals separately, the capabilities of the network collapse; termed de-punctualisation. The overarching processes that allow a network to be represented by a single entity - be that an object, an individual, or another network - is called the ‘sociology of translation’ (Callon, 1986). This concept
was developed in Callon’s (1986) account of an attempt to preserve a population of scallops. He described how the initiative was only made possible as marine biologists enrolled fishermen, the scallops and their scientific colleagues to the project. However, the network failed when these actors dissented, breaking relations that held the conservation effort together. In this article, Callon introduced terms for the phases of translation; (1) problematisation, (2) obligatory passage point, (3) intéressement, (4) enrolment, (5) mobilisation and (6) black-box (see Figure 2.7).
Figure 2.7 – The Phases of Translation
(Source: Adapted from Rodgers et al., 2009)
In ANT, problematisation describes the rationale for actors and actants linking in particular networked configurations. In Callon’s example, the researchers sought to become indispensable by defining the scallop conservation problem and suggesting that it would only be resolved if an obligatory passage point⁸ was collectively negotiated, i.e. the researchers’ programme of investigation. During intéressement, a series of processes strengthen associations between the different entities and support the network’s structure. In enrolment a set of strategies define and relate the roles allocated to the networked entities. The penultimate stage of translation, mobilisation, describes the process of all actors/agents in the network becoming enrolled. Whilst the network can become entrenched and its purpose can become black boxed so that it is not questioned, Callon (1986) explained how translation is a process, never a complete accomplishment, and may fail, as occurred in the case of the scallop research project.
1. Problemisation
⁸ Obligatory passage point: In ANT, these are critical network points, which are often designed by the primary actor to
Whilst such relational thinking about socio-technical aspects of everyday life has been applied to housing (e.g. Røpke and Christensen, 2012), planning (e.g. Rydin, 2012) and a range of energy issues (e.g. Day and Walker 2013; Buzar, 2007; Harrison and Popke, 2011; Schatzki, 2011; Powells, 2009; Longhurst and Chilvers, 2013), ANT has proven a polarising theory. First, the notion of ‘symmetry of agency’ is contested. For Actor Network Theorists, human actors are only able to exercise agency through the ‘effect[s]
of the socio-material networks within which they are entangled’ (Longhurst and Chilvers, 2013: 5). Whilst this extends individualistic, purely cognitive understandings, practice theorists argue that whilst ‘objects…make a contribution, …the nature of that contribution depends on us’ (Schatzki, 2002: 17). Reckwitz holds that,
‘When artefacts can only be effective within practices insofar as they are ‘handled’ by human agents and when they are sites of ‘materialised understanding’, then their status obviously cannot be completely ‘equal’ with that of human agents and their embodied understanding’ (2002b: 214).
SPT, in contrast, pays greater attention to the capacity for people’s actions to influence the trajectory of an assemblage, and highlights the value in labelling the different elements of practice, as opposed to treating the components of a network as equivalent.
Second, whilst ANT adopts a more systems based understanding of social life, the theory is vague and contradictory in its account of what initially prompts a network to develop. Questions remain as to what leads a network to be initiated, and whether recruitment of actors and agents to the network describes goal-directed action. ANT proponents rebut this criticism by downplaying the importance of the inception of a network arguing that these moments can not explain how networks are perpetuated and diffuse; ‘the initial force of the first in the chain is no more important than that of the second, or the fortieth, or the four hundredth person’ (Latour, 1986: 267).
Furthermore, it is suggested that far from the everyday being made up of predictable networks, ‘inconstancy, multiplicity and indefiniteness’ are inherent in real life (Law, 2004: 145; Berker, 2006). Related to this point, ANT can also be criticised for failing to explain why one form of network might be actualised over others, and for not being clear as to how networks are bounded.
Whilst ANT provides useful notions for studying the networked complexity of socio-technical life and how it changes (or remains stable), these limitations require that this research looks outside of this theoretical perspective.