Doing Normal Work – Processes at Level 1
In Chapter 2,1 proposed an organisation that will serve as a framework for structuring the rest of this book. I have offered the view that what we call crew resource management (CRM) is actually the application of expertise to the task of controlling the action. Aviation is a collaborative, distributed process that aims to achieve the desired goal within the bounds of tolerable risk. Finally, humans are frail and our capacity to do work is moderated by a suite of individual differences. In this chapter, we start to look at the behaviours associated with the systems model. When I outlined the model (or, possibly, my organising framework), I listed some of its characteristics, one of which was emergence. Each level in the hierarchy will exhibit properties that cannot be explained simply based on the functioning of the individual components at that level.
The chapter starts with an elaboration of activity as thought. All observed behaviour is the output from mental processing but rarely do we have the opportunity to trace a pilot’s line of thought as events unfold. Next, I elaborate on the concept of goals, boundaries and margins because this hypothetical ‘space’ is the arena in which individuals act. Sense-making is fundamental to action, and this is examined in some detail before we look at decision-making as option selection. A key theme of this chapter is that individual action is both shaped by and, in turn, shapes the environment. The problem we face, though, is captured by Johnson (2004), who suggests that things fail in ways that we simply cannot imagine (remember non-ergodicity and radical uncertainty from previous chapters). As an aside, because of this, ‘safety management’, Sisyphus-like, is forever doomed to fail. Chater (2018) argues that, in any case, because the future is unknowable all we can ever do is respond to events as they happen. In Chater’s view, because humans are inveterate storytellers, we create narratives that rationalise the relationship between the actions and outcomes and identify ‘causes’ that are no more than fictions. From this perspective, the concept of ‘proaction’ becomes absurd. These two positions imply that, even under normal conditions, any intervention we make to achieve a goal will have a degree of uncertainty attached to the outcome and, thus, we will be constantly having to adjust to circumstances as events unfold.
Controlling complex systems is an act of collective computation. Fundamentally, individuals engage with the world and then respond to outcomes. They share cognitive resources to make choices about future actions. Expertise makes us more sensitive to the cues that exist in situations as they develop. Research now suggests that brain activation precedes conscious thought and, from an evolutionary perspective, it makes sense for the brain to be leading consciousness rather than lagging. But the fact remains, we are still largely simply responding rather than predicting. This idea that we are always reacting to the present drives the rest of this chapter.
Having said that aviation is a collaborative and distributed process, work, by definition, is shared across individuals. As a result, it can be hard to separate those specific properties of a situation that are exclusive to a single person. For this chapter, though, I will try to limit the discussion to those key internal processes that are the necessary precursors to shared activity. Ironically, the idea of single-pilot CRM has seemed to be an oxymoron ever since the CRM was first specified as a training requirement. However, if you look at single-pilot accident reports you will see many features that have parallels with large commercial aircraft accidents. In all probability, then, it is in single-pilot commercial operations that we might see some clues to functions at Level l.
Doing a job of work involves initiating activity, monitoring the conduct of that activity and assessing progress towards the desired goal. A key part of work is making decisions about which activities to modify, and how to modify them, to ensure continued progress towards the goal. Once the goal is achieved then the job can be considered complete and we can terminate the activity. From a systems model perspective, at this level, control is manifested in action. Individuals require a schema of some sorts to control action or a process for resolving uncertainty. Feedback is provided by the responses to inputs observed in the world. Effective evaluation of feedback requires the individual to have adequate mental models of the world and the processes being managed. Figure 5.1 illustrates Level 1 in my hierarchy. Part of the task is to correctly configure the aircraft to cope with disturbances while ensuring progress towards the operational goal but, of course, work involves more than just manipulating the aircraft. We will look at errors in more detail in the next chapter but, typically, things are deemed to have gone wrong when pilots either fail to execute an action plan within acceptable boundaries or adopt a plan that was inappropriate for the situation.
The focus of this chapter, then, will be on the individual operator engaged in a task. Although some of the discussion will involve multi-crew operations, I will not be addressing the social nature of work but will focus, instead, on what individuals do. That said, we will touch on the idea of distributed processing: two or more actors combining their cognitive attributes to get the work done. I propose that safety is an emergent property at this level. It is only as a result of the way we act that outcomes are either of increased risk or most likely to succeed. Safety is a product of how individuals act in the world.