A striking case of coincidence in the history of twentieth-century psychology, the concept of accident proneness, suggests that historians of science may have unwisely marginalized and dismissed the idea of independent simultaneous discovery – or as many scholars would now say “construction” – of a scientific idea. The discovery or construction of the idea of accident proneness or Unfallneigung is so remarkably close to an ideal model of simultaneous discovery as to resurrect the question and perhaps to contribute to demystifying multiple discovery. I approach the question in the spirit of the call of Peter Galison (e.g., Reference Galison1987, ix, 270–278) and others who recommend that historians focus less on theory and more on what happens in the laboratory.

Simultaneity in this case is particularly striking. The concept of accident proneness, or the tendency of a particular person to have more accidents than most people, originated independently in Britain and Germany during the World War I period. In 1926, both the English term, accident proneness, and the German, Unfallneigung (inclination to accident), were introduced into the technical and general literature – again, simultaneously – signaling a final and definitive formulation of the idea. In Germany, psychologist Karl Marbe became an advocate of the formulation. In Britain, another psychologist, Eric Farmer, emerged as the chief advocate among a group following the earlier work of epidemiologist Major Greenwood. Both Marbe and Farmer, however, had already given the idea a technical name in obscure local publications in 1925 (Marbe Reference Marbe1925; Farmer Reference Farmer1925). Altogether the picture is one of persistent, recurring coincidence.Footnote ¹

The General Problem of Multiple Discovery

The idea of simultaneous or multiple discovery has long excited historical curiosity (Merton Reference Merton and Norman1973, 353–355). Any time a new idea appears spontaneously in two or more places, general historians and historians of all the sciences (as well as technology) still take special notice. Simultaneous discovery suggests that there was something in the intellectual community or the surrounding culture, either a general “Zeitgeist” or a point at which knowledge in some way had reached a state that precipitated an inevitable next step, as in the celebrated case of Newton and Leibniz who independently devised calculus (or something like it) in the seventeenth century.

Interest in multiple discovery grew out of the internal histories of ideas. Historians portrayed the progress of science through a logical sequence of ideas, each one produced by a genius, or at least by a contributor to science. The question naturally arose, how could two geniuses independently come up with the same idea at the same time? Moreover, scholars believed that any scientific discovery was on some level unique or irrational; as Aharon Kantorovich puts it, “Scientific discovery seems to be the most impressive and mysterious feature exhibited by modern natural science” (Kantorovich Reference Kantorovich1993, 1; see also Magnani Reference Magnani2001, 15). Multiple discovery therefore particularly acquired an implicitly mystical overtone.Footnote ²

One could naturalize multiple discovery, as E. B. Wilson, Jr., reported from the laboratory: “It happens almost universally that the next logical step in the field will occur to several people at the same time, if they have access to the same background” (Wilson Reference Bright1952, 7). E. G. Boring in his classic history of experimental psychology invoked the idea of a Zeitgeist that brought forth discoveries, but in a naturalistic, not a Hegelian sense. He did, however, argue against the personal agency of the discoverer. That is, like many other thinkers, Boring held that circumstances give rise to intellectual developments (Boring Reference Boring1950, 3–4, 22–23). By 1959, Thomas Kuhn could declare that the ideas that “a scientific discovery must fit the times, or the time must be ripe” were “totally unexpressive truisms.” He himself repeatedly used the possibly Freudian metaphor of an idea that lay “close to the surface of scientific consciousness” (Kuhn [Reference Kuhn and Marshall1959] 1969, 321, 340, 341).

In a systematic examination of multiple discoveries, philosophers David Lamb and Susan M. Easton concluded that “scientific development has its own autonomy.” They went on to utilize Edward Sapir's idea of cultural drift – to which all members of a scientific community are subject – as the factor that can set off simultaneous, independent discoveries. They thus combined external forces and culturally conditioned individual agency as explanatory factors for the apparent fact that “multiple discovery is a normal feature of science” (Lamb and Easton Reference Lamb and Easton1984, 173, 70). Dean Keith Simonton points out that a Zeitgeist is both disciplinary and cultural. He argues that although a Zeitgeist is necessary to cause a creative next step in science, it is not sufficient. In the end, he would credit “chance” rather than any determinism for the appearance of multiples (Simonton Reference Simonton2004).

One additional factor attracts scholars to the idea: Simultaneous discovery in science is particularly persuasive to people who know that, in evolution, parallel environments can independently produce strikingly similar organisms in widely distant geographical areas (Kantorovich Reference Kantorovich1993, 185–188).Footnote ³ And obviously in Big Science research groups, simultaneous discovery within the group could be expected.

In general terms, multiple discovery is of historical interest because examining roughly contemporary and comparable instances helps show how science worked at any time, as Robert Merton contended years ago (Merton Reference Merton and Norman1973, 343–412). But since Merton wrote, philosophers, sociologists, and historians have tended to pursue reductionistic inquiries into the nature of discovery. Of course questioning the very idea of discovery undermines any idea of multiple discovery.Footnote ⁴

Scholars pursuing reductionistic analyses of discovery had legitimate and important goals. They were at least in part responding to social demands of the late twentieth century to explain how to foster creativity and scientific innovation. Philosophers were additionally inspired by artificial intelligence studies to investigate the logic and structure of innovative thinking. Sociologists were examining scientific priority claims and disputes, which could become legal questions of intellectual property. But as Alan Gross explains, no “moment of scientific discovery” can exist, because a “discovery” is created only by retrospective judgment (Gross Reference Gross1998, 167). Following up on all of this work, even historians avoiding presentist viewpoints were dissecting specific discoveries of the past. They, too, found, as did Ilana Löwy in Reference Löwy1990, that there are great difficulties with the theoretical idea of discovery. Terms may change in meaning. Retrospective accounts, particularly, may not be accurate. Again, if the idea of discovery is called into question, multiple discovery is also called into question.

Many scholars therefore have raised points of doubt about all of the elements of simultaneous discovery/construction. Augustine Brannigan indeed held that because the idea of simultaneous discovery is itself a social construction, the phenomenon in some sense is not real (Brannigan Reference Brannigan1981). Much of the scholarship on multiple discovery was in fact secondary to interest in priority disputes in science. In particular, the idea of multiple discovery has not gained support from scholars who are preoccupied with epistemology – to the point that Crosbie Smith has poked fun at the image of “several ‘independent’ discoverers all sleepwalking toward the same physical truth” (Smith Reference Smith1998, 12).

Yet such a scenario is exactly what historian of science Buhm Soon Park recently found when he examined the details of the discovery/construction of the idea of “hybridization” in physics and chemistry by John C. Slater and Linus Pauling in the interwar period of the twentieth century: “their different routes to a common destination – hybridization.”Footnote ⁵ Indeed, Park explores a discovery that came out of two dissimilar disciplinary viewpoints. As a result, a “diversity of aims and approaches . . . happened to produce the same result.” Under the stimulus of quantum mechanics, the independent discoverers were each venturing into what turned out to be new territory for both disciplines. With disciplinary differences and different goals, and yet later construction of convergence of ideas, it is no wonder that in the case that Park interrogates, the definition of what constitutes multiple discovery could be called into question – and yet intensify interest in the phenomenon (Park Reference Park2000; see also Logan Reference Logan2002). The fact therefore remains that historians can and do identify innovations and shifts in ideas and outlooks in scientific communities and the societies that surround them. It is in this perspective and this scholarship that the history of the idea of accident proneness has a place. And, after all, in another context, coincidence is one form of validation in some scientific investigations.

Characteristics of Multiple Discovery

The instance I am describing, accident proneness, can be shown in concrete terms to have emerged directly from widespread social and cultural circumstances. No mystical Zeitgeist or vague cultural determinism was sufficient. It is, in fact, possible to reconstruct in a plausible way the concrete context, in effect a commonsense de-mystification and naturalization of coincidence.

Beyond simultaneity, other characteristics of multiple discovery emerge from the history of science literature. One is that the discoveries were independent. In the case of accident proneness, there is evidence that neither originator had heard of the work of the other prior to 1926 or, probably, 1927 (Anonymous 1927; Farmer Reference Farmer1927, 436n). And another characteristic is that the idea is new. Both proponents and critics of the idea of accident proneness believed it to be distinctive.Footnote ⁶

Still another requirement is that the independent discoveries are indeed discoveries of the same thing, that is, that they were commensurate. In the case of accident proneness, there are two kinds of evidence. One is the fact that people at that time explicitly stated that the English and the German versions were equivalent. As one British reviewer wrote: “That individuals vary in liability to accident has been frequently asserted; but Dr. Marbe's [1926] book, like Dr. Greenwood's work in this country, raises what was once merely a ‘view’ to the status of an established fact” (Anonymous 1927, 278). The second type of evidence is that safety experts utilized the two versions as equivalent.

German Origins

Marbe himself traced his thinking on the subject of accident proneness, as such, back to his school days. The only child of a German merchant initially operating out of Paris, Marbe attended school in Freiburg i. B. in Germany after his father became ill and moved back to Germany.Footnote ⁷ Marbe later recalled that as a boy attending elementary school,

Marbe went on to note that there are capitalists and bankers who never fail to incur losses, while others produce good outcomes from their ventures. And in general, some people have bad luck all of their lives, and others, Glückskinder [lucky people, “darlings of fortune”], enjoy success.Footnote ⁸

Marbe, who was one of the key figures in the history of the Würzburg School in psychology, was not a likely person to originate an idea of accident proneness as part of applied psychology. He was a defender of scientific and rigorously experimental psychology. But he was drawn into applied psychology through some pioneer work that he did in forensic psychology. In 1911, he gave expert opinion in a case of sexual abuse, and in 1912, he got involved in two more court cases (Mülberger Rogele 1995, esp. 305–369).Footnote ⁹ The next year, Marbe published a report of his opinions concerning the psychology of employees involved in a terrible train wreck near the Swiss border. The engineer had a history of unreliable behavior, and the guard who might have stopped the train did not have the mental capacity to make an appropriate judgment (Marbe Reference Marbe1913; Marbe Reference Marbe1935, 102–103). These court cases aroused Marbe's interest in psychological factors in accidents.Footnote ¹⁰ After his experience as a court expert, he and his students were fairly launched into applied psychology, including now the psychology of accidents (Mülberger Rogele 1995, esp. 311–317; Marbe Reference Marbe1924).Footnote ¹¹

In 1916, in a major philosophical work interpreting natural phenomena, Marbe briefly revealed in print that he had decided that someone who had suffered one accident must expect to suffer further accidents. Or, as he finally put it, the probability of an accident was greater for the person, A, who had had an accident, than for the person, B, who had not. Marbe meant this as conventional statistical probability. He indicated that he had validated his idea using accident statistics that were available to him and anyone else from the German social insurance establishments (Marbe Reference Marbe and Emil1916, 384–387).Footnote ¹²

In 1916, Marbe viewed statistical grouping in terms of his philosophic belief that similarities occur in nature and tend to group together. With regard to human beings, Marbe held that they would behave consistently in similar circumstances – as they did in word association tests. He even referred to the law of repetition by which an individual would show consistent behavior. Marbe believed that the basis of the consistent behavior was a combination of heredity and acquired habit, a combination that produced a person's personality, a term that was becoming commonplace in psychology at that time (Marbe Reference Marbe1926).

In a short paper published in the proceedings of a local society, Marbe referred to the probability of a person's having accidents as Unfallneigung, although he also used the term “disposition” (Marbe Reference Marbe1925, especially 172–173). This paper came out the year before his more publicized (and cited) use of Unfallneigung in 1926. Marbe in this way first introduced a new name for the phenomenon in an obscure publication, and, as I have noted, in exactly the same year as did Farmer, the English psychologist who named the concept.

Marbe in his earlier formulations had used other familiar terms. As late as 1923, he did not use the term Unfallneigung when he reported on industrial accident cases analyzed by his student, Nico Greiveldinger. There was, wrote Marbe at that time, a personal factor that made some people disposed (disponiert) to accidents. Accident-burdened individuals had different personalities (“die verschiedenen Persönlichkeiten der Unfälle erleidenden Individuen”). He relied most on the medical term, “disposition.” By the last part of this 1923 article, Marbe was referring casually to Unfalldisposition and Unfällen disponierender Faktor, and he ended up with Disposition zu Unfällen, a term that he used repeatedly.

Marbe's definitive publication on accident proneness, Unfallneigung, was his book of 1926 that summarized his thinking. He cited four articles published by others from 1924 alone on the subject of people who had repeated accidents. For the most part, however, his book was a synthesis of his own previous publications. He explained in detail how consistencies in behavior were based on inherited constitution, upbringing, mental habits, and training. Those consistencies inclined an individual to have accidents, “individuelle, persönliche Unfalldisposition oder Unfallneigung.” Moreover, he pointed out again, the inclination was proportional to the number of accidents previously suffered. Evidence from those who consistently had no accidents was even more convincing than that of those who had a record of accidents (Marbe Reference Marbe1926, esp. 14, 54). He thus had grouped people into two categories: those who did not suffer accidents, and those, Unfäller, who had accidents in proportion to the number of their earlier accidents.

There was one element that was new in 1926: Marbe explicitly extended the idea of accidents from personal injury to production errors, picking up on previous work that he had done on psychomotor errors. Errors made by children sorting peas showed the same incidence pattern found among workers who suffered injuries. Finally, Marbe emphasized accident proneness as an endogenous factor, and, despite earlier words about training, he wrote pessimistically. These were not ordinary people. The statistics for criminal convictions, Marbe pointed out, showed the same pattern as for accidents: the more convictions, the more chance of recidivism. And accident statistics, showing the law of repetition of behavior, therefore suggested action, Marbe believed. Insurance premiums for accident prone people ought to increase. Personnel policies should likewise take accident records into account. An accident prone chauffeur, Marbe contended, needed to find another kind of occupation (ibid., esp. 15–25, 56–59, 78–83, 106–109).Footnote ¹³

By 1926, however, Marbe could point out that some industrial executives were already implicitly recognizing the problem of accident proneness. Most conspicuous were the personnel practices of several German railways, which included some psychological testing as early as 1917 ([Glasel] 1926).Footnote ¹⁴ Indeed, Marbe used evidence from some of these industrial practices to vindicate his theory. The managers of transport systems were well aware of the potential for injury and property damage from employees who might in some way tend to have accidents. Some managers therefore did employ psychological tests, at first for reaction time and alertness, and then for ability to estimate distance and speed and for quick decision-making. At some point, the obvious tests for color blindness and night blindness, in addition to optical and auditory acuity, were added. Finally, special talents were tested, such as physical strength for a switchman to throw a switch and motor and manipulative abilities for various tasks. And of course a certain level of intelligence was also required. All of these traits, physical and mental, were implicitly, if not explicitly, designed to insure safe operation of transportation systems.Footnote ¹⁵ This same kind of testing could be applied to other occupations, and already was, sometimes – such as woodworking, in which workers often sawed off a finger. Could the worker see? Could the worker react? (Heller Reference Heller1924).

All of the explications in such discussions tended to be stated in terms of what was desirable to make a good and safe worker. Moreover, the dangers were specific: a signal not detected, a sudden danger not reacted to, a switch not thrown. What Marbe, by contrast, was describing was a person who was not disqualified by some specific disability such as color blindness or indecision, but who, on the contrary, could be identified simply because he or she had suffered one or more accidents.Footnote ¹⁶

British Origins

Where Marbe presented the concept of accident proneness as his own theory, in Britain the concept of accident proneness was initially presented by a group of researchers. Although they were describing the same phenomenon, the German and British principals followed different paths. Indeed, until Eric Farmer became the chief advocate, British investigators considered the idea of an accident prone employee an awkward, isolated empirical finding.

The Industrial Fatigue Research Board was set up during World War I to facilitate war production in England. In 1919, in a publication of the Board, Major Greenwood and Hilda M. Woods reported individual differences as one of the findings of the British investigators. Greenwood was an eminent epidemiologist (his university title was medical statistician);Footnote ¹⁷ Woods was a staff assistant. Greenwood and Woods noted that their work was based on extensive records held by the Ministry of Munitions, but they did not say who first noticed that a small group of workers produced “the bulk of the accidents.” Moreover, Greenwood and Woods added to the munitions workers' records those of another large factory. By Reference Greenwood and Woods1919, their published discussion of alternative explanations and additional hypotheses suggests strongly that the profile of workers who had accidents had stimulated much thought among investigators associated with the work.Footnote ¹⁸ They summarized their findings: “the bulk of the accidents occur to a limited number of individuals who have a special susceptibility to accidents, and suggests that the explanation of this susceptibility is to be found in the personality of the individual” (Greenwood and Woods Reference Greenwood and Woods1919, 2).

Greenwood and Woods saw other implications in their new way of viewing the data. Twice they mentioned practical consequences: “both initial selection of recruits and also a rapid elimination of those sustaining multiple accidents should have a great effect in reducing the casualty rate of the hospital,” and “by weeding out susceptibles the accident rate would necessarily decline.” Also they referred to “unsafe people” and “a more dangerous person.” Accidents, they concluded, did not occur just by chance, which they portrayed as the explanation that was alternative to their finding. They concluded instead that “varying personal liability” was an important factor in accidents (ibid., especially 5, 7, 9, 11). Altogether, their determination that one worker was someone who had accidents while another was free of accidents was remarkably parallel to Marbe's categories of Unfäller and Nichtunfäller.

In Reference Greenwood1927, Greenwood in retrospect provided the only record that exists as to who, among their group, actually noticed the workers who had a markedly disproportionate number of accidents. When he and Woods first published, a critic noted that it was trivial work, for everyone knew that “liability to accident varies.” But Greenwood, working with another statistician during the war, saw a theoretical interest in accident statistics.

Air Force personnel were subject to “unequal exposure” and data about individuals would not be comparable; therefore, Greenwood and his statistical colleague, G. Udny Yule, turned to the records available to them, trivial accidents in munitions factories (Greenwood and Yule Reference Greenwood and Yule1920).Footnote ¹⁹

So it came about that “two statisticians” noticed the unequal distribution of accidents, as Greenwood recalled at the end of his life (Greenwood Reference Greenwood1950). The phrasing suggests that the statisticians were Greenwood and Yule. Clearly Greenwood was one of the two who saw the statistical distribution, even if, given the gender relations of the time, it might have been Woods and not Yule who was the other “statistician.” The fact remains that the idea was taken up by the whole research group.

In 1921 in a general book on The Health of the Industrial Worker, Greenwood repeated and expanded his conclusion that “in industrial life we may remove from risk those who show a tendency to accidents.” He reported on the ways in which varying conditions could change “the human factor” in accidents, but he then took up the findings he reported jointly with Woods, that multiple accidents happened to the same individuals. To the earlier findings, Greenwood added, from further research by the group, confirmation that some individual workers have a “personal predisposition to accidents” (Collis and Greenwood Reference Collis and Greenwood1921).Footnote ²⁰ While Greenwood did not pursue this line of investigation further, he had clearly launched the idea of personal predisposition to accidents in an important general book on industrial medicine.Footnote ²¹

The first annual report of the Industrial Fatigue Research Board included substantial recognition of this line of research: “The inferences drawn [by Greenwood and Woods], though new and of great interest, are admittedly based on slender data, and cannot yet be regarded as finally established. The Board, however, considers them of sufficient importance to justify a continuance of the investigation on wider lines, so as to include men and boys, and hope to extend it by psychological inquiry into the causation of accidents.” Then later on the same page, the tone of the report became more assertive, suggesting that “accidents are very largely due to a special susceptibility inherent in the personality of the individual, so that the bulk of accidents occur amongst a limited group of individuals,” although, the anonymous author added, a single accident has no significance (Medical Research Council and Department of Scientific and Industrial Research 1920, 12).

Eric Farmer joined the Board in 1919 working mostly on time and motion studies. In 1925, Farmer's first report, the one that incorporated the term, “accident proneness,” appeared as an appendix to the board's 1924 annual report (Farmer Reference Farmer1925; the report for the year 1924 was published in 1925).Footnote ²² Eric Farmer was born in Northampton in 1888 and was a 1914 graduate of Trinity College, Cambridge, where he read moral sciences. Charles Samuel Myers, who was a central figure in British industrial psychology, recalled Farmer as one of the research students who passed through his experimental psychology lab at Cambridge (Myers Reference Myers and Carl1936, 222). From 1935 to 1953, Farmer was reader in industrial psychology in the University of Cambridge, a special post financed by the Medical Research Council for him personally. Students remembered him as a clear lecturer but not as an innovative thinker; they recalled that he was formal in manner, reserved, and had military bearing and attitude. He was not an easy mixer.Footnote ²³ His influence in spreading the idea of accident proneness, then, was destined to be formal, rather than personal, much less charismatic.

Farmer's short article, “The Method of Grouping by Differential Tests in Relation to Accident Proneness,” was a progress report on the research sponsored by the Board to find psychological tests to identify accident prone people. “We already know from work the Board has published,” wrote Farmer, “that people differ in their degree of proneness to accidents. This does not mean that there is a group who have accidents and a group who do not, but that people exposed to equal risk will tend to sustain accidents according to their degree of proneness” (Farmer Reference Farmer1925, 43–45).Footnote ²⁴ This idea of accident proneness, which Farmer, along with Chambers, described in the Reference Farmer and Chambers1926 report, entered the scientific literature at the same time Marbe's monograph did. Before that time, they were using a medical concept, individual susceptibility. Thus a person could suffer attacks of an illness under the same or similar circumstances while the disease might pass other people by. And since the investigators had originally been working with labor in the munitions industry, their viewpoint was clear: “Trivial accidents are indicators of unsafe people whom the record of the ambulance room can be employed to discover” (Greenwood and Woods 1919, esp. 11).

In later reports, the same group continued to use the formulation of individual susceptibility, or, at most, “tendency” to having accidents, until Farmer and Chambers introduced accident proneness and accident prone people to a more general audience.Footnote ²⁵

An Instance of Simultaneous Discovery/Construction

From this basic narrative, it is possible to use the discovery/construction of accident proneness to explore the meaning of simultaneous discovery. First, was this a discovery/construction of something new? The naturalistic explanations that people used to explain accidents were carelessness – which incurred moral condemnation – and clumsiness. Clumsiness was a lack of the ability to function normally so as to avoid injuries and damages. Moreover, clumsiness, as part of one's character, was generally excusable behavior, as, for example, physical weakness would be. Accident proneness was a personality characteristic or trait that was an added element – a part of one's self, but a behavior that one did not will. In psychology, the model of slowed reaction time was already well known, but reaction time was psychophysiological, while accident proneness was a behavioral artifact of statistics. Because proneness or inclination to have accidents occurred consistently over time, like a habit, people at the time understood that it functioned as a personality trait independent of psychophysiological causes. This was a trait that was added to the person, not a deficit (as were carelessness and clumsiness). It was therefore something new. But the idea in both Germany and Britain came out of statistics. In Marbe's case, it was simply a matter of examining accident records collected for insurance purposes. In the British case, investigators again were reviewing statistical records of accidents, this time in a factory setting.

What did Marbe and the British investigators see? They saw that over time, some people had more accidents than others to the point that one could identify people whose profiles were notably different from the records of other people. In the eyes of the investigators, these people fell into a classification of people who regularly experience accidents, the Unfäller, as opposed to the much larger group of people who have few or no accidents. Part of the discovery of the idea lay in devising the classification.Footnote ²⁶

The naming of the personal characteristic of accident proneness therefore signified more than devising a terminology for bad luck or the traditional weeding out of unproductive industrial employees by managers. Both Marbe and Greenwood and his collaborators cited standard work in statistics and emphasized that they were describing an incidence of accidents that varied markedly and distinctively from chance distribution.

And, further, by naming accident proneness and Unfallneigung, both Marbe and Farmer unmistakably moved away from the concept and name of “disposition.” Disposition implied physical weakness, possibly hereditary, that invited illness.Footnote ²⁷ With the new terms, the names moved to an independent personality or character trait – for which both British and German originators immediately began to devise psychological tests. They thus with a new term moved the idea out of the realm of medicine (disposition) and into the realm of psychology (accident proneness or Unfallneigung).

Evidence of the Cultural Setting

What conditions and ways of thinking in European countries called the idea into being? Many scholars have commented that industrialization and mechanization were standard parts of modernization. In addition, gathering and analyzing statistics, as has been pointed out, was a standard part of modern, bureaucratized societies. And the cultural impact of technology and machinery has attracted comment from literary people as well as historians. Indeed, Bernhard Rieger has drawn a dramatic parallel between the ways in which the British and the Germans welcomed technology in the decades between 1890 and 1945 (Rieger Reference Rieger2005). And in the Europeanized nations in general, there evolved a generalized “risk society,” in which people devised ways to deal with modern dangers (Beck Reference Beck and Mark1992).

Before and especially after World War I, it was observed casually or clinically that some people had much more than their share of accidents. Just as Marbe as a boy had drawn conclusions from observing his schoolmates, so physicians and other professionals who worked with groups of children reported over and over the common experience of seeing individual children who had all the accidents. In a pediatric journal an author commented: “Years ago, before I knew the term ‘accident proneness,’ I recognized in my summer camp [for girls] the tendency among certain children to have more than their share of accidents. ‘There goes Jane again,’ or ‘Who do you suppose fell out of the boat? Mary, of course'” (Ross Reference Ross1953).

Even Greenwood and Woods in the first British publication describing accident prone people, noted that there was a common explanation for individual susceptibility to having accidents: “carefulness or carelessness; as one says, there are people whose fingers are all thumbs and there are others who are neat fingered; or again some people are scatter-witted and others circumspect . . .. Do our results amount to more than an arithmetical verification of this?” (Greenwood and Woods 1919, 9).

Not least of earlier observers of the phenomenon were managers who transferred or discharged workers who had a record of damaging goods or causing injuries. We have many detailed records from transport companies before 1926, for example, on both sides of the Atlantic. The general efforts of managers to reduce accidents, however, do not translate into economic motives, remote or immediate, that would lead them to an abstract idea that individual workers’ personalities appeared consistently in accident statistics. Employers had always known that some employees were more desirable than others, and that knowledge made many of them ready for the articulated insight that accident records foretold more accidents.

The phenomenon was also observed by clinicians who had to treat injured people. A Swiss military surgeon, Chs. Widmer, provides an example from Reference Widmer1919. He wrote specifically about Unfalldisposition [accident disposition]. Any physician, he observed, has treated trauma, and in each case the standard anamnesis, finding out “how did this happen?” revealed that the patient failed to react automatically to avoid the accident. Moreover, with extensive experience, such as he had with the same group of military personnel, one would notice that a small number of people did not react appropriately and so repeatedly became injured. Anyone with extensive clinical experience such as he had, Widmer wrote, would identify people who were likely to suffer injuries. One could, he concluded, see an “accident neurotic” long before the accident came to pass (Widmer 1919).Footnote ²⁸

In 1920, a member of the State Safety Board in the state of Washington who had experience in investigating industrial accidents had concluded that there were individual differences in the realm of accidents. He observed “many unfortunate human beings engaged in very hazardous work who, through intemperance, hereditary mental disturbances, domestic troubles and generally speaking ‘all around hard luck victims,’ have grown disgruntled, careless and indifferent, thus jeopardizing their own lives and the lives of those surrounding them” (Flyzik Reference Flyzik, Martin1920). Managers and safety experts had noticed that some individuals were a constant source of errors and injuries. William H. Tolman and Leonard B. Kendall, early leaders in the American safety movement, all but gave accident proneness a name, and they anticipated later discussions around the idea: “It is always a good plan to watch out for men who are hurt frequently. It is hard to discharge them, but this is better than running the constant risk that their clumsiness will endanger their fellows” (Tolman and Kendall Reference Tolman and Kendall1913, 29).

All of this common experience, lay and professional, in all Western countries, set the stage for the idea of accident proneness to appear explicitly in Germany and Britain. As Thomas Kuhn commented in his classic case of simultaneous discovery, energy conservation, accounts of other people who had come close to the basic idea “could be increased, but not fruitfully” (Kuhn [Reference Kuhn and Marshall1959] 1969). No one before Marbe and Greenwood and associates, however, named the phenomenon, pointed out the statistical basis for it, and made it a part of an emerging psychological idea, personality, that is, an unwilled trait or habit. Moreover, they immediately set to work to perfect psychological tests that could indicate the presence of the trait of accident proneness in people. In moving accident proneness into psychology with psychological testing (a model newly operative only in the World War I era), Marbe and Farmer were not just asserting disciplinary hegemony. They were making conceptualization of accident proneness a means for helping to alleviate a terrible social problem.

Readiness for Receiving the Idea

Once the idea of accident proneness was formulated in modern terms, people recognized the phenomenon. In 1924, a German railway administrator attending a safety meeting heard Marbe report on his idea of accident proneness. The administrator had been trying to reduce accidents by employee selection and recognized at once that Marbe described exactly the concept that fitted his work. He was inspired to embrace the idea, to which he referred, following one of Marbe's early formulations, as “accident affinity” (Schmitt Reference Schmitt1926).

From across the Atlantic, the medical director of the Selby Shoe Company in the United States told how he personally came to embrace the idea as soon as he heard about it:

By the early 1930s, then, the formal concept of accident proneness was widely familiar, and people immediately recognized the phenomenon. Moreover, they were able to reconceptualize their own experience with humans who encountered machines, experience that antedated even the naming and categorizing of accident proneness. Clearly the encounter of humans with accidents of any kind had set off observations that some people get hurt consistently more than others. Common circumstances therefore preceded and accompanied the formulation and use of the idea of accident proneness or Unfallneigung. The specific problems that the human-machine interface created were the same all over: pieces of bodies cut off or crushed, people smashed by moving objects and vehicles. And one can add: the collection of statistics and the existence of psychologists who could think of different patterns of human behavior and devise tests to detect them.