A Life in Error

From Little Slips to Big Disasters

James Reason

• Éditeur: CRC Press
• Année: 2017
• ISBN:
  9781351962230
• Édition: 1
• Sujet: Technology & Engineering
• Sous-sujet: Industrial Health & Safety

This succinct but absorbing book covers the main way stations on James Reason's 40-year journey in pursuit of the nature and varieties of human error. In it he presents an engrossing and very personal perspective, offering the reader exceptional insights, wisdom and wit as only James Reason can. The journey begins with a bizarre absent-minded action slip committed by Professor Reason in the early 1970s - putting cat food into the teapot - and continues up to the present day, conveying his unique perceptions into a variety of major accidents that have shaped his thinking about unsafe acts and latent conditions. A Life in Error charts the development of his seminal and hugely influential work from its original focus into individual cognitive psychology through the broadening of scope to embrace social, organizational and systemic issues. The voyage recounted is both hugely entertaining and educational, imparting a real sense of how James Reason's ground-breaking theories changed the way we think about human error, and why he is held in such esteem around the world wherever humans interact with technological systems. This book is essential reading for students, academics and safety professionals of all kinds who are interested in avoiding breakdowns that can cause serious damage to people, assets and the environment.

Chapter 1

A Bizarre Beginning

One afternoon in the early 1970s, I was boiling a kettle for tea. The teapot (those were the days when tea leaves went into the pot rather than teabags) was waiting open-topped on the kitchen surface. At that moment, the cat—a very noisy Burmese—turned up at the nearby kitchen door, howling to be fed. I have to confess I was slightly nervous of this cat and his needs tended to get priority. I opened a tin of cat food, dug in a spoon and dolloped a large spoonful of cat food into the teapot. I did not put tea leaves in the cat’s bowl. It was an asymmetrical behavioural spoonerism.

I little realized at the time that this bizarre slip would change my professional life—I was a lecturer in psychology at the University of Leicester and had run out of research topics. I used to work on motion sickness and disorientation, but students did not like to be made sick and the pool of usable experimental subjects was fast drying up. I needed something new to do. I had flirted with clinical psychology, but I was neither trained nor temperamentally suited for it.

After I had washed out the teapot and sworn at the cat, I started to reflect upon my embarrassing slip. One thing was certain: there was nothing random or potentially inexplicable about these actions-not-as-planned. They had a number of interesting properties. First, both tea-making and cat-feeding were highly automatic, habitual sequences that were performed in a very familiar environment. I was almost certainly thinking about something other than tea-making or cat-feeding. But then my attention had been captured by the clamouring of the cat beyond the glass kitchen door. This occurred at the moment I was about to spoon tea into the pot, but instead I put cat food into the pot. Dolloping cat food into an object (like the cat bowl) affording containment had migrated into the tea-making sequence. Even the spooning actions were appropriate for the substance: sticky cat food requires a flick to separate it from the spoon; dry tea leaves do not. It seemed in retrospect that local, object-related control mechanisms were at work (see later).

I did not realize it at the time, but this action slip exhibited nearly all the principal characteristics of absent-minded errors:

•  Both behavioural sequences were highly routine, meaning that I was unlikely to be monitoring step by step—my attention was ‘absent’.

•  Both the cat’s bowl and the teapot afforded containment.

•  Some change introduced into a routine sequence of actions (in this case, the cat’s demands) quite often misdirects actions down the wrong path.

This event was of considerable psychological significance. Although unintended, the individual actions were executed smoothly and skilfully. Among other things, this error suggested a way of investigating the control of highly practiced actions. And these events occurred (in my case) on an almost daily basis and right beneath my nose. After years of laboratory study, I suddenly felt I had access to the fabric of everyday life. Experiments, by definition, are usually carried out in highly artificial (controlled) circumstances—but I felt very strongly that these naturalistic observations could reveal the stuff of which human thought and action were truly made.

I would like to end this story by describing another absent-minded slip that also took place in my Leicestershire kitchen. For once, I was not the error maker. I rarely get a chance to observe a slip in progress. I was idly watching my wife make tea. She was boiling the kettle and had the teapot beside it with the lid off. Then she reached up to a nearby shelf and took down a large jar of instant coffee. After adroitly unscrewing the lid, she put three teaspoons of coffee into the teapot, poured water into it, and was then alerted to her slip by the strong smell of coffee. Now there was nothing particularly curious in this slip. Errors of this nature happen relatively frequently in our kitchen. What made the slip especially interesting, however, was the skilful way she unscrewed the lid of the coffee jar. The tin tea caddy had a pull-off, push-on lid. The question of interest was, how did her hand perform the appropriate unscrewing movements on the lid? Her mind was clearly ‘absent’ from the tea-making sequence at the time. My conclusion was that her hand movements were under the local control of the coffee jar. Tea-making and most cooking are test-exit-test-exit tasks.

This raised the interesting notion that well-used familiar objects develop a local control zone. Once the hand enters that zone it is automatically directed to perform object-appropriate actions. These control zones become particularly evident when we are mooching around the house in a state of reduced intentionality—as when we are waiting for a visitor or phone call. We have no immediate goal except to pass the time. Under those conditions, we stray past a fruit bowl and pick up an apple; or we stray past the open door of a bathroom and clean our teeth, even when that has recently been done, or we front up to the toilet and take a leak that we don’t really need.

These aimless periods reveal that much of our behaviour is under the control of the immediate environment, often resulting in the unintended activation of action programs appropriate for the circumstances. I have even wandered into the bedroom and started to undress in the middle of the day, standing beside my bed.

Both of these slips—the cat food and the tea-making—give us important insights into the way we control our routine actions in a familiar environment. When actions are under-specified—by reduced intentionality, forgetting, misdirected attention and the like—they tend to default to behaviours that occur frequently in that particular context. Such ‘strong-habit-intrusions’ occur in a variety of mental contexts.

In tip-of-the-tongue states, for example, we struggle to retrieve a word that we know we know. When this happens, we often find ourselves calling to mind a word we know, but we also know that it’s not the one we seek, although it often feels ‘warm’. We call these words ‘blockers’. Here’s an example: I was struggling find the title of the film Deliverance, but I kept retrieving another single-word film title: Intolerance (the D.W. Griffiths film was much more familiar to me than Deliverance). Both words also had the relatively unusual ‘-erance’ ending. Which brings us to St Augustine’s question: if we keep retrieving the wrong word, how do we know it’s the wrong word if we can’t retrieve the sought-for word? Clearly some part of my brain knew the answer because of the ‘-erance’ endings. This suggests that mental life is lived on several levels, and consciousness is only one smallish part of it.

Chapter 2

Plans, Actions and Consequences

Despite the fact that most people understand what is meant by the term error, there is no universally agreed scientific definition. Indeed, some of my distinguished colleagues would wish to dispense with the term altogether. I do not agree, so I will spend the rest of this chapter trying to put together a working definition of error that will serve as a framework for what follows.

It is generally accepted that errors entail some kind of deviation of human performance from an intended, desired or ideal standard. Although such discrepant actions may have adverse effects, they can also be inconsequential or even benign, as in trial-and-error learning or serendipitous discovery. How many of you, for example, have accidentally stumbled upon better ways of working with a computer application (e.g. a word processing package) by clicking on the wrong window and then trying out a useful command you had not used before?

Errors are not intrinsically bad, though their consequences and the workplace local conditions provoking them are often undesirable. It is not so much the psychological processes that determine the nature of the outcome; rather it is the circumstances of their occurrence that shape an error’s consequences. Activating the kettle rather than the toaster in your kitchen can be mildly irritating, but the same kind of switching error committed in the control room of a nuclear power plant or on the flight deck of an aircraft can be disastrous.

Nearly all human actions, correct or otherwise, entail three basic elements: plans, actions and consequences. As such they seem a good starting point for constructing an error definition—and, as we shall see later, also an error classification.

Plans

Stop and think for a moment. What are you going to be doing for the rest of the day? Tomorrow? Next week? In six months’ time? Two years hence? It is likely that you have some plans for these periods, although as the time in question extends further into the future it is probable that your answers will become increasingly more vague and uncertain. Much of what occupies our thoughts is the making and refining of plans. They are central to our understanding of error.

Consider the following imaginary example as our starting point: It is noon, and you are hungry. It’s been a wretched morning and you feel like treating yourself to a decent lunch. Do you want to eat French, Indian, Italian or Chinese? You favour Italian because you have a fancy for a large plate of spaghetti carbonara. Will you go to Luigi’s or La Dolce Vita? You favour Luigi’s. It’s further away but the parking is easier. But there is a problem: you are trying to lose weight, and spaghetti carbonara is rich in calories.

This plan begins commonly enough with the need to alleviate a state of tension, in this case a combination of hunger pangs and dissatisfaction with the frustrating morning. There are many ways of achieving this, but a specific intention is quickly formed to have a decent lunch. Again such a goal could be achieved by a variety of means, but one particular plan is favoured—to eat spaghetti at Luigi’s. Having made this decision, it only remains to specify and assemble the action sequences that will take you to Luigi’s by the most convenient route.

The plan at this stage consists of a stated aim and a rough outline of the actions necessary to achieve it. Notice that the planner does not need to fill in the fine details of each operation. These action steps are largely automatic subroutines that are implicit in the jottings already made on the mental scratchpad. They are manipulated in thinking by a series of verbal tags and mental images. The more we engage in these habitual action sequences, the fewer are the number of ‘tags’ required to specify them in our planning. Repetition reduces the number of low-level control statements necessary to guide our behaviour.

In the very broadest sense we can say that our actions are in error when they fail to achieve the objectives of our current plan. But here we run into a problem, as the previous example of planning indicates. Let’s assume that our self-indulgent planner arrives at Luigi’s without a hitch and consumes a large plate of spaghetti carbonara. In the sense that his actions fulfilled his intentions exactly, they could hardly be said to be in error. But what about his longer-term dieting plan? Viewed in this light, the successfully executed short-term plan to consume a calorie-laden dish was clearly a mistake.

The point being made here is that our lives are governed by many plans. Sometimes they nest together in close harmony, but at other times they conflict, as shown by our example. The existence of co-existing and conflicting plans would make even a working definition of error beyond our reach, were it not for two built-in limitations to human performance. First, our physical capacity to turn personal plans into action is limited. We can only be in one place at any time. In addition, we possess only a limited mental capacity for carrying out plans. Although there may be many stored plans competing for our attention, usually only one of them is maximally active at any one time.

Error, as we have seen, is not an easy notion to pin down. Dictionaries send us on a semantic circular tour through other like terms such as mistake, fault defect and back to error again. That dictionaries yield synonyms rather than definitions suggests that the notion of error is something fundamental and irreducible. But we need to probe more deeply into error’s psychological meaning.

Error is intimately bound up with notions like plan, intention, action and its outcome. The success or failure of our actions can only be judged by the extent to which they achieve, or are on the way to achieving, their planned consequences.

For our present purposes, therefore, we can define error as follows:

The term error will be applied to all those occasions in which a planned sequence of mental or physical activities fails to achieve its desired goal without the intervention of some chance agency.

Two qualifications are important here. Firstly, there is the inclusion of the notion of intention, and secondly, the absence of any chance intervention.

Should you inadvertently shoot a red light as the result of being stung by a wasp, you would have suffered an automatism rather than having committed an error, because the actions responsible for the undesired outcome were not what you intended or could reasonably have avoided. Unfortunately, these excuses will not wash with the police—shooting a red traffic light is a crime of absolute liability.

Similarly, if you were struck down on the street by a piece of returning space debris, you would not achieve your immediate goal, but neither would you be in error, since this unhappy intervention was outside your control. By the same token, achieving your goal through the influence solely of happy chance—as when you slice a golf ball that bounces off a tree and onto the green—could hardly be called correct performance.

The logic of this definition yields two distinct ways in which you can fail to achieve your desired objective:

•  In the first case, the plan of action may be entirely appropriate, but the actions do not go as planned. These are slips and lapses (absent-mindedness) or trips and fumbles (clumsy or maladroit actions). Such failures occur at the level of execution rather than in the formulation of intentions or planning.

•  The second category of failure can arise when your actions follow the plan exactly, but the plan itself is inadequate to achieve its desired goal. These are termed mistakes and involve more complex, higher-level processes such as judging, reasoning and decision-making. Mistakes, being more subtle and complex, are much harder to detect than slips, lapses, trips and fumbles. In the case of actions-not-as-planned we have a conscious record of what was intended and so the discrepancy is easily discovered. But it is not always obvious what kind of a plan would be ideal for attaining a particular objective. Thus mistakes can pass unnoticed for long periods—and even when detected they can be a matter of debate.

In the next chapter we will consider three levels of performance: skill-based, rule-based and knowledge-based. These ideas have proved extremely useful in the categorization of errors and violations and were a major step onwards.

Chapter 3

Three Performance Levels

It was argued in the previous chapter that slips and mistakes arise from quite different psychological mechanisms. Slips were said to stem from failures at the level of execution, often arising from the unintended activation of largely automatic procedural routines—also known as action schemas. Mistakes, on the other hand, derive from higher level mental processes involved in formulating plans, setting objectives, decision-making and judging the available information.

If that were really so, we would expect slips and mistakes to take quite different forms. But that is not the case. Both slips and certain kinds of mistakes can take very similar ‘strong-but-wrong’ forms, where the erroneous actions are more in kee...