People use technological artefacts to perform purposeful tasks at their workplace (Trist 1981). The interrelatedness and complex interaction between the social and technical aspects of an organisation can be conceptualised as a socio-technical system (STS) (Trist 1981). Disruptions arising from complex social and technical interactions have been explored by several scholars who have focussed their attention on organisations which must remain accident free due to the potential for large scale destruction or disruption to essential services brought about by accidents (see, for example, Bain 1999; Boin & van Eeten 2013; Cooke & Rohleder 2006; Gephart 2004; Hollnagel 2014; Leveson et al. 2009;
19 Madni & Jackson 2009; Perrow 1984; Rijpma 2003; Roberts 1990; Roe & Schulman 2012; Sagan 1993; Weick 2010; Wesnser 2015). Studies of accidents and disasters in STS have led to the formulation of two well-known theories – the normal accident theory (NAT) and the high reliability theory (HRT) (Cooke & Rohleder 2006; Lee, Vargo & Seville 2013; Leveson et al. 2009; Rijpma 2003).
Perrow (1984) is credited with laying the foundations of the normal accident theory (NAT) (Bain 1999; Cooke & Rohleder 2006; Leveson et al. 2009; Rijpma 2003). According to Perrow (1984), accidents are inevitable in complex STS such as maritime transport. The inevitability of accidents arises due to interactive complexity and tight coupling of the system (Perrow 1984). The more technology is incorporated into a system, the greater is its interactive complexity (Cooke & Rohleder 2006). In complex systems, independent failures interact unexpectedly, and due to tight coupling, these interactions escalate rapidly into the breakdown of the system. According to Perrow (1984) small failures can spread rapidly to damage the larger system and cause accidents. Essentially, NAT proposes that some system responses to changes cannot be foreseen. As a result, some unforeseen responses can cause accidents, potentially leading to catastrophes (Cooke & Rohleder 2006). Such accidents can be viewed as ‘normal’ because the ‘interdependencies in a system are so great that one small glitch in one place can lead to a large failure somewhere else’ (Roberts, Bea & Bartles 2001, p. 71). In other words, normal accidents occur because multiple component failures in a system interact in unanticipated ways (Gephart 2004).
A competing organisational theory of accidents – high-reliability theory (HRT) – maintains that even though some accidents may be normal, the frequency and severity of disasters can be minimised (Cooke & Rohleder 2006; Perrow 1994; Rijpma 2003; Roberts, Bea & Bartles 2001; Sagan 1994). According to Rijpma (2003), in the mid-1980s
20 HRT emerged from studies of so called high reliability organisations (HROs) by a group of researchers at Berkley University. The Berkley University researchers investigated why HROs such as a company operating both a nuclear power station and electricity distribution network, US Navy aircraft carriers, and US air traffic control centres, were able to avoid catastrophic failure even though they operated hazardous technology (see, for example, La Porte & Consolini 1998; Roberts 1990; Rochlin, La Porte & Roberts 1987). Such organisations are considered reliable because they have the ability to ‘manage hazardous technical systems safely and without serious error’ over long periods (Schulman et al. 2004, p. 14). There is now increasing interest among stakeholders on the reliability and resilience of critical infrastructure that provides essential societal services such as large engineered supplies for water, electricity, telecommunications and transportation (Roe & Schulman 2012).
HROs maintain safety and reliability by anticipating and planning for unexpected disruptions (Madni & Jackson 2009). Research on HROs suggests that certain organisational characteristics and processes can assist resilient organisations prevent small problems from escalating into crises or disasters (Boin & van Eeten 2013). Roberts, Bea and Bartles (2001, p. 71) distil research on HROs to provide the following prescriptions for maintaining reliability:
• HROs aggressively seek to know what they don’t know.
• HROs design their reward and incentive systems to recognise costs of failures as well as benefits of reliability.
• HROs consistently communicate the big picture of what the organization seeks to do, and try to get everyone to communicate with each other about how they fit the big picture.
21 Leveson et al. (2009) raise doubt about the generalisability of findings from HRO studies to organisations operating in competitive industries. Unlike HROs, the mission of most organisations is something other than safety (Leveson et al. 2009). The nature of the mission of HROs means that other goals of an HRO have limited impact on its safety goals, whereas for organisations operating in competitive environments, the organisational mission may not be achieved by designing systems and operations for lowest risk (Leveson et al. 2009). However, Waller and Roberts (2003) provide a counter argument, maintaining that organisations other than HROs can learn from HROs how to manage unexpected events in a fast changing, dynamic and complex environment. In particular, Waller and Roberts (2003) emphasise that HROs are exemplars of making quick decisions based on imperfect data and abandoning business as usual routines for improvisation.
In addition to research on HROs, concern with safety and reliability has led to the emergence of a new discipline within engineering, called resilience engineering (Madni & Jackson 2009). Resilience engineering explicitly assumes that it is possible to build systems that can avoid accidents through adjustment, monitoring, learning and anticipation (Hollnagel 2014; Madni & Jackson 2009). The distinction between HRT and resilience engineering however is unclear. Studies of HROs and analyses of accidents under NAT provides indications on why some organisations may fail and others may survive. Proponents of both HRT and resilience engineering paradigms share the view that exercising control over organisational systems, processes and behaviour is the best way to achieve safety and reliability. This view contrasts with the NAT paradigm which highlights the limitations of human control. NAT suggests that complexity, coupling and bounded rationality diminish human ability to influence how organisations cope with unexpected circumstances. The perceived limitations of human control on organisational
22 behaviour has led some scholars to suggest that organisations should be managed as complex adaptive systems (CAS) to cope with uncertainty and unpredictable change. The following section discusses the conceptualisation of resilience in complex adaptive systems.