Keywords

The invitation and motivation for this workshop contained the following statement:

… our understanding of safety as a construct daily enacted by a multitude of artefacts, actors and institutions has never really been conceptualised from the angle of these drawings, pictures, visualisations, images, but also videos or movies.

Showing or explaining something visually, by pictures or graphics, rather than by words, is generally assumed to be a more effective means of communication. It is indeed a common saying that a picture is worth a thousand words.Footnote 1 Safety is, however, only one word, so is it also the case that a picture of safety is a thousand times more valuable than the word or concept itself? Leaving this bogus question aside, the underlying issue is what, if anything is gained by substituting a word, a construct, or a concept by a picture.

In this commentary I propose to consider whether it in fact is possible to visualise safety as a word, a construct, or a concept. I will do that by considering two possible interpretations of what the invitation refers to, namely visualising for safety and visualising of safety. Visualisation for safety refers to the use of visual techniques or visual communication to assist in making systems and system performance safe. (This immediately begs the question of what “safe” actually means, as pointed out by the epigraph and as discussed later in this commentary.) In other words, visualising or visualisation with the purpose of promoting whatever processes or behaviours that are believed to support whatever safety is, to influence the behaviour of people (so that they behave “safely”) and to present an understanding of how something has happened or could happen (as in accident and risk analyses). It is the issue of visualising what should be done—although in practise it more often is what should not be done—to ensure that safety, or a state of safety, is maintained. This obviously also includes visualising what happens when safety is absent, such as copious pictures that show the consequences of accidents, etc.

Conversely, visualisation of safety refers to the use of visual techniques to show what safety is or means, assuming that there is something called safety and that that something can be shown. With the risk of being pretentious I might also suggest that visualising for safety addresses instrumental problems while visualising of safety addresses metaphysical, or perhaps even ontological, problems. To relieve any suspense the reader may have I can already now reveal that the second interpretation in my view is impossible.

6.1 Visualising for Safety

In the case of visualising for safety, the examples are many and varied as shown by the invitation. Although this may appear perplexing, it is nevertheless possible to assign the examples of visualising to a relatively small number of categories. The proposal here is to distinguish between visualising the outcomes of safety, visualising the mechanisms of safety or how such outcomes can come about, and finally visualising the safety shaping factors—what can or should be done to ensure that the specified outcomes occur. Other categorisations may, of course, also be possible.

6.1.1 Visualisation of Safety Outcomes

The visualisation of outcomes are pictures, naturalistic or symbolic, of the consequences of incidents and accidents ranging from the benign or even humorous to the gruesome. From a Safety-I perspective, their purpose is presumably to make the viewer aware of what could happen if care is not taken or if rules and regulations are breached or neglected. Ideally, this should then trigger some kind of avoidance behaviour, in the sense that people will try to avoid these outcomes. Visualisation of outcomes can also be used as an inducement to change behaviour generally, for instance as in the gory pictures shown on cigarette packages in many countries.

In addition to showing possible outcomes, visualisation has also typically been used to show the distribution of various outcomes. The most famous rendering of that is undoubtedly Heinrich’s accident pyramid and the iceberg model. The accident pyramid illustrates the advantages as well as the disadvantages of visualising a possible relationship between different types of outcomes. On the one hand, it is easy to understand and use as a reference, specifically when precise ratios of outcome types are assigned, but on the other hand it also suggests a causal relationship that neither exists—or at least has never been proven—and furthermore never was intended (cf., [2]).

6.1.2 Visualisation of Safety Mechanisms

The visualisation of safety mechanisms or of how outcomes happen is best represented by the graphical renderings of accident models, ranging from Heinrich’s Domino model and Ishikawa diagrams to Leveson’s STAMP, with the Swiss cheese model and the Bow-tie in between. In these cases, the graphical models of how accidents happen are really worth a thousand words, or even more, since a corresponding text would be quite lengthy. (To be fair, the description of the Domino model provided by [3] used only 388 words. But later and more complicated models most likely exceed the 1000 word limit).

The main problem with accident models as a visualisation of safety mechanisms is that they do not visualise safety at all—quite apart from the uncertain epistemic status of what a causal mechanism is [4]. As the epigraph states, safety is defined by its absence rather than its presence—or as a dynamic non-event [5]. An accident is a consequence of the absence of safety, in part or in whole, and a visualisation of an accident either of the outcome or of the way it happens can therefore not be considered a visualisation of safety.

6.1.3 Visualisation of Safety Shaping Factors

If safety is defined by the absence of accidents, then the ways in which accidents can be prevented or avoided must clearly be accepted as visualisations of safety. Examples of this can easily be found ranging from simple matters such as holding on to handrails when using a staircase [6] to a flowchart for maintenance of power transformers (accompanied by about 100 pages of text!).

More generally accidents can be avoided by putting in some barriers which therefore are assumed to serve as safety shaping factors. The purpose of a barrier is to hinder access or passage either in a direct physical sense or in a more metaphorical sense. Barriers, or barrier systems, can be characterised as either physical or material, functional (active or dynamic), symbolic, or incorporeal. (An incorporeal barrier lacks material form or substance in the situations where it is applied and instead depends on the knowledge of the user in order to achieve its purpose, cf., [7].) The three first types of barrier systems rely on visualisation in the sense that they are required to be seen. (It may be argued that a physical barrier system such as a wall will achieve its purpose even if it cannot be seen, but in practice that rarely happens. In fact, when someone bumps into a wall it will most likely be classified as an accident itself).

6.2 Visualising of Safety

The visualisation of safety obviously requires a definition of the essence of safety or of what safety actually is. There is little help to find in the common definitions of safety, which generally equate safety with the relative—or even absolute—freedom from danger, risk, or threat of harm, injury, or loss to personnel and/or property [1, 8]. This is also the essence of the epigraph. (All such definitions reflect the etymology of the English word “safe”, which comes from the French word sauf that means both “without” and “unharmed.”) By using definitions such as these as a starting point, the problem in effect becomes how to visualise nothing. Even if it somehow was possible to visualise nothing, there would be nothing to see, hence no help to understand what safety is.

The concern for the meaning of safety was part of the discussions that led to the formulation of resilience engineering [9]. This later developed into the proposal that it was possible to consider two different ways of interpreting safety which were called Safety-I and Safety-II. According to a Safety-I perspective, safety is defined as a condition where as little as possible goes wrong hence as being without unacceptable outcomes. According to a Safety-II perspective, safety is defined as a condition where as much as possible goes well hence as being with acceptable outcomes. There is therefore something to visualise but since there is little tradition for doing that, the problem is what should be shown. Perhaps it is what affords safety, rather than what safety is?

6.3 Conclusions

The challenges are obviously not the same for the visualisation of Safety-I and the visualisation of Safety-II, cf. Table 6.1. For the visualisation of Safety-I some of the answers have been given above (at least according to my interpretation). Visualising for Safety-I is clearly possible but visualising of Safety-I is not. Visualising for Safety-II is also possible, although we then need to reconsider how we best show what the outcomes are, how they come about, and how they can be furthered or facilitated—rather than prevented. Visualising of Safety-II remains contentious. Even though Safety-II can be associated with the presence of something, it is still an elusive concept. In practice, the differences between the two perspectives do not matter much, since the overriding concern is how we best manage or control the processes that lead to specific outcomes rather than how we manage or control safety as such.

Table 6.1 Differences in visualisation for Safety-I and Safety-II
Full size table

6.3.1 Visualisation as a Means to an End

Since there are two significantly different interpretations of what safety is, any analysis of or suggestion for how to visualise safety should recognise the plurality of interpretations and clearly refer to either one or the other rather than to “safety” in general. Visualising safety, whether as Safety-I or Safety-II, should furthermore not be a purpose in itself, but a means to achieve a purpose. Visualisations of safety fall into the category of artefacts in Schein’s [10] model of organisational culture, hence must be seen in conjunction with the espoused values and shared basic assumption that also determine performance. While it is beyond doubt that various forms of visualisation can be useful to ensure that work and working environments function as intended—whether in relation to safety, quality, or something else—the pros and cons of visualisation should always be considered relative to the specific purpose. Visualising something is not a magical way to make a diffuse idea intelligible.