- What is Risk? A Historical-Narrative Approach (new)

(Paper presented in 2013 in The Society for the Humanities. Cornell University)

( Published in Spanish in Mairal Buil, G.- 2013 La Década del riesgo. Situaciones y narrativas de riesgo en España a comienzo de siglo. Madrid. Ediciones Catarata.)

Risk is a broad concept, and its meaning may change drastically depending on the situation. Through risk we can discern the probability of evil though also of good. The word risk has undergone a dual semantic development, pointing on the one hand to the recognition of losses under the influence of maritime insurance and trading in the Mediterranean of the 13th and 14th centuries, and on the other to the possibility of gain, whether in games of chance, where players know that the probability of losing is greater than that of winning, or in business, where the relationship between gains and losses should usually tip in favor of profits. After the Spanish conquest and colonization of the Americas, risk also developed as narrative, above all in the new genre of journalism, which recounted natural disasters like volcanic eruptions, earthquakes and floods, all phenomena that quickly came to be identified with the new lands beyond the Ocean. Meanwhile, in Europe, the scourges of plague and epidemic provided fertile ground for new narratives of risk. As the virtual inventor of such novel narratives of risk and of the new genre of journalism in The Storm (1704) and above all in A Journal of the Plague Year (1722), the case of Daniel Defoe is extraordinary. Not long afterwards, the great Lisbon earthquake of 1755 would echo throughout Europe, producing multiple accounts that ranged from apocalyptic sermons and poetry, through letters and descriptions in personal diaries. The earthquake aroused a great debate in Enlightenment Europe between the fire and brimstone party, who saw in it only divine punishment, and enlightened thinkers like Voltaire and the Marquis of Pombal, who advocated prevention and public management of disasters. The Industrial Revolution and the rapid growth of the great European cities would give rise to the characterization of the "mean streets" as places fraught with peril. This context set the scene for a novel journalistic account of reality which made use of a new narrative to induce fear and dread. This is best illustrated historically in the case of Jack the Ripper, the murderer who prowled the streets of London in 1888. Crime and insecurity fostered sensationalist literary and journalistic genres in the 19th and early 20th centuries, resulting in the emergence of narrative styles like the horror story and the crime novel. Finally, in the 20th century the probabilistic calculation of risk spread to all manner of activities from economics to construction, epidemiology and medicine in general, politics, environmental management and insurance, to name but a few. Following the indiscriminate use of new military technologies in the World War II, culminating in the atom bombs of Hiroshima and Nagasaki, narratives of risk related with the massive spread of technology began to proliferate in a trend that only gathered strength in the 1970s with the appearance of the green movement, as intensifying criticism of the damage being done to the natural world became a repeated theme of risk narrative. In recent decades, risk has emerged as a key issue in the human and social sciences, becoming ever more global thanks to the extraordinary development of the mass media, which are today able to report almost any violent or catastrophic event worldwide in a matter of minutes, however remote the location. Two events that occurred in the 1980s set the pattern for the proliferation of risk narratives in today's world. These were the explosion of the nuclear reactor at Chernobyl in 1984 and the AIDS epidemic, which first broke out around 1981.

In light of this brief review of historical and contemporary examples, we may ask whether it is possible to frame a coherent conceptualization of risk capable of spanning the whole range of possibilities. This is my intention. After years of research, I believe the historical approach to be the best, because risk is an expert concept and its genesis can be traced, though with difficulty. In doing this, we shall see that the concept emerged in singular historical contexts, gradually taking on new meanings over time in response to new situations and needs. However, it conserves some basic properties which I propose to identify and examine, and to this end I would begin by asking a question that arises logically from my proposition that risk is a concept. If so, what does it conceptualize?

Risk is time

Any concept is an idea that refers to something, and it exists only to inform our understanding of that "something". I propose, then, that risk conceptualizes time in the form of a future that does not yet exist. For human beings, time is not a single, universal dimension but a mental construct that is both plural and diverse, and it is a cultural creation. Looking back over the whole of our existence, we may observe that the ability to foresee the future has been one of mankind's great challenges. What comes after death? What will happen in the course the voyage on which I am about to embark? What will be the outcome of a war? Will I succeed in my business ventures? People have repeated these and other such questions throughout human history, finding answers of all kinds in the form of concepts about events that have not yet happened. By providing an account of the future that assures life after death, religion offers a future, as do magic, divination and fate or destiny. In all of these cases, something that does not exist, because it has not yet happened, is formulated, foretold, predicted, shown or anticipated. When I set out I wonder whether I will have a safe journey, and if I am going to make an investment, I want to know whether it will be successful. Obviously, the future exists only in our imagination and it is therefore nothing but a representation of a reality waiting to exist.

The root of risk as a concept lies in language and in the grammatical structures that allow us to create and describe the future. Aristotle alludes to this dichotomy between the actual and the potential with reference not to risk or anything like it but to the difference between history and poetry:

"It is, moreover, evident from what has been said, that it is not the function of the poet to relate what has happened, but what may happen- what is possible according to the law of probability or necessity. The poet and the historian differ not by writing in verse or in prose. The work of Herodotus might be put into verse, and it would still be a species of history, with meter no less than without it. The true difference is that one relates what has happened, the other what may happen."[1]

Reading this paragraph carefully, we may observe even in this ancient text a suggestion of narrative about what could happen, which Aristotle refers to as "poetry", at the same time pointing towards a conceptualization of risk by capturing the same temporal dimension of what has not yet existed but could exist. The first conclusion we may draw from this example is that language constructs time through narrative. This was a key principle in the philosophy of Paul Ricoeur (1987), and it is the basic inspiration for the whole narrative argument concerning risk that I shall set out here.

"........between the action of narrating a story and the temporal character of human existence there is a correlation which is not purely accidental, but it represents the shape of a cross-cultural need. In other words: time becomes human time if it is articulated in a narrative way, and the narration gets its full meaning when it turns into a condition of temporal experience."[2]

If risk is time, and especially future time, as I have argued here, then narrative is a condition for its very existence, which leads me to the general proposition that risk is, and was created as, a narrative act. In this light, it seems to me that a theory of risk must meet two requirements: it must address narrative and it must be historical. Let us now turn to the second of these conditions.

The notion of risk is inspired by narratives about the future, and it arose from the material needs of mankind in situations where nature posed a challenge in which the human response involved the use of language to frame a narrative. I believe it was the desolate waste that posed this great challenge. Historically, the deserts, the sea, the frozen wilderness of the North and the sky were the great wastes, which mankind has conquered in the course of his wanderings and journeys of exploration, yet these feats could only be undertaken with the development of new ideas, tools, devices, technologies, skills and knowledge. Among these developments is the invention of risk.

The open sea was one of the great challenges facing ancient peoples from the Greeks to the Polynesians. The ancient Greeks dominated the eastern Mediterranean, though they ventured beyond its western end and even down the Atlantic coast of north Africa, according to some more or less legendary accounts. Greek ships would normally keep safely within coastal waters, but they also braved the deep in the Black Sea or Pontus Euxinus, creating the myth of the Argonauts in which Jason and the crew of the Argo sail into the unknown across a mysterious sea. The Greeks' account of this wide waste of water is bound up with the destiny of these heroes.

The invention of risk was a moment of fundamental change in human history. Until then, men believed they could see the future using all kinds of divination practices from oracles to studying the entrails of sacrificial animals. As Cicero laments in his De Divinatione:

Now what nation or what state disregards the prophecies of soothsayers, or of interpreters of prodigies and lightnings, or of augurs, or of astrologers, or of oracles, or the forewarnings of dreams, or of frenzy?[3]

There was no doubt about the future, because it was known (to the Gods) and, as the Greeks believed, foreshadowed in portents and omens of all kinds. There was no room for chance in the mind of the ancient world, and it was possible to learn what might befall before embarking upon any important undertaking. Chance was unimaginable because men were bound by their fate, an attitude that represents neither more nor less than a belief in a future that has already been seen. In The Greeks and Greek Civilization, Jakob Burckhardt compares the ancient world of the Greeks with the modern world which the Arabs began to forge with the expansion of Islam, as we shall see below.

"Overall, it could be said that one of the main differences between the ancient world and ours is that the ancients sought to foretell the future, and we do not . Nobody then doubted the omens and the auguries, and (...) they were used in daily life as a means to save on the need for forethought and decision. This way of thinking is separated from ours as by a chasm." [4]

The early Moslems did away with the belief that the future could be foretold, which they saw as a heresy since the future was in the hands of God alone. Instead, they applied a concept that was upheld by God but allowed intervention in the future. The difference is crucial. The ancients could see the future but could not influence it in any way, while the moderns cannot see the future, which is in the hands of God or is pure chance, but they can act to prevent it. With great clear-sightedness, Burckhardt himself alludes to what sets the ancient world apart from ours, and this is neither more nor less than "forethought and decision", knowledge and action. Knowledge is, then, essential to the construction of the future that is risk, as we shall see in this account of the concept's emergence, for it is knowledge that we project forwards, allowing us to act in, or on, the future. There is something incipiently scientific about this idea, of course, and the religious context in which it first appeared, the Koran and the expansion of Islam, shows how the early Moslems contributed a new rationalism and a set of principles that allowed mankind to control nature.

My thesis is that rizq[5] was conceived in the expanding Islamic world of the early 7th century as a first challenge to the uncertainty of a future that was in the hands of God alone. Riskwas thus born out of Islamic theology, which denied that mankind could in any way foresee the future and instead defined the future as inherently uncertain, even if this uncertainty was limited by "rizq", Allah's providence for the faithful and their possessions in any long journey or venture. The future cannot, then, be foretold or foreseen, but man can intervene to reduce uncertainty. In this initial period, then, risk is God-given. Let us now consider the historical and geographical context in which the concept of risk was invented in the deserts of Arabia and North Africa, in the Indian Ocean and in the Mediterranean.

Any journey across an empty space and into the unknown produces enormous pressure in future time. As travelers we must consider both the material dimension, because we will be crossing an uncharted space, and the immaterial, because we will be travelling in imagined time, which must be invented. Where will we get to and what will happen after we leave our point of departure? We cannot know, and this is exactly what makes the future a great enigma.

The Arabs conquered the deserts, one of the great empty spaces, as the great historian Fernand Braudel vividly describes:

"Islam is the desert", as the essayist Essad Bey has affirmed. It is its emptiness, its ascetic harshness, its inherent mystique, its devotion for the implacable sun as the unifying principle behind the myths, and the thousand consequences of the absence of men. Meanwhile, the greatness of the Mediterranean civilizations was determined by the empty sea. Where the salt waste was populated with shipping, the desert was populated by caravans and the eternal nomadic peoples. Like the sea, the desert is movement and so, therefore, is Islam."[6] (Braudel, F. 1986, Pp. Vol. II, 246)

There was, then, a dual conquest in space and time over the emptiness of the desert and the sea, and over the uncertainty of the future. These conquests were the route and risk. Those who ventured out into the waste could only defeat it if they had the knowledge to find their way to their destination. In the historical period in question, the time of the expansion of Islam, new discoveries and devices contributed decisively to making this possible. Though their knowledge was still rudimentary, the Arabs had already learned how to calculate latitude, and the compass rose and northward-pointing magnetic needle had just been invented. In general, these new techniques had arrived from the Far East, and the Arabs adopted them to dominate navigation first in the Indian Ocean and later in the Mediterranean. Suddenly, it had become possible to follow a known route across the desert or the sea to reach a given destination.

To venture out into the desert or across the sea required at least some knowledge of what might happen, as the denial of divination and other ways of foretelling the future left only uncertainty in the minds of travelers[7]. It is not difficult to imagine them wondering what would happen. Where would they find land? How long would their voyage take? Let us remember here that at the time of the great Arab-Islamic expansion, these questions were couched in eminently religious terms. Meanwhile, those who set out to cross the desert or the sea took with them both their own personal possessions and goods of all kinds to trade, and the cloud of uncertainty hung over both. This fact is essential to understand the subsequent development of the concept of risk. The Arab term rizq offered the faithful who put their trust in Allah a protective mantle of divine providence both for themselves and for their goods, like a divine insurance policy that makes no distinction between a person and his chattels[8]. As the centuries passed, this distinction became more clearly drawn, and the concept of risk came to focus on goods rather than on persons. As this secularization of the concept gradually gained ground, risk took on a new legal form in maritime contracts like the Moslem qirad and the commenda, a type of consignment agreement that evolved after the appropriation of Arab partnership arrangements by the Christian cities of the Mediterranean like Genoa, Venice, Marseille and Barcelona in the 13th century.

My central argument is that risk is time, and as such it is a concept about the future. In his Sociology of Risk, Luhmann refers to this conceptualization, linking the notion of the future with that of probability:

"The result seems to be that the future can only be perceived through probability, which means that its features can only be perceived as more or less likely, or more or less unlikely. (Luhmann, N. 2006, 94)

The key thing about risk is its nature as a cultural artifact representing what does not yet exist but could exist, or commonly what has already happened and could happen again. In dealing with risk, we link together the different times that emerge from diverse cultural constructions, and in this way memory and historical accounts, the narrative of past events, are projected forwards as a narrative of what could recur, becoming a narrative of risk in our imagination of the future.

Risk is knowledge

As a concept, risk embodies the knowledge that is gleaned from experience. Once again, this characteristic is revealed by the historical conditions in which the concept of risk arose. As we have seen, risk first emerged in a context of new astronomical, geographical and navigational discoveries achieved thanks to the use of new techniques and inventions. Moreover, the notion of risk was in itself a novel form of shared knowledge, an idea developed in the world of expert knowledge. As explained above, knowledge of risk was essential, together with knowledge of the route, to navigate the open sea. Meanwhile, the concept of risk has come down to us in writing, which means we can place it in the enlightened spheres of the Islamic world among the people who read the Koran and the early works of Arabic literature, who organized the great caravans and who sailed the Indian Ocean and the Mediterranean to trade. In later medieval documents, the work "risk" becomes a specialist legal term, and from there it gradually worked its way into general parlance, by the 16th century in Spanish and French, and by the 17th century in English. In this way, risk found its place in the modern languages as an ordinary term in popular speech.

Risk is not a perception, and we cannot use our senses to deal with it. Rather, we use it as an idea that is capable of cementing our knowledge of something and, as may be observed in numerous cases[9], the notion of risk spreads through multiple channels, the most important of which today are the mass media. If the community in an area with a large number of cell phone antennas or overhead power lines[10] believes the infrastructure poses a risk, their belief will arise not so much from their actual perceptions of the offending installations but from information received about the supposed adverse effects on health produced by electro-magnetic radiation. This information is both believable and received in many different ways. For better or worse it contains knowledge, but in any case it provides a sufficient foundation to construct a basic account of risk: if a community lives in close proximity with installations of this kind, people's health could suffer. This probability can only be constructed on the basis of the knowledge contained in the information received. This is how the notion of risk works today, but in the 13th century when a ship and its cargo were insured, the terms of the arrangement, expressed in the words "ad meum risicum" in the medieval documents, were based on the whole cumulative body of knowledge about routes, vessels, currents, winds and pirates. All of this knowledge allowed the creation of an account which in its simplest terms meant, for example, that a cargo of wine transported by ship from the port of Barcelona to Mallorca sailed at its owner's own risk, as the vessel could run aground, sink or fall prey to pirates.

Risk is not a reaction to any existing person or event, but proposes the probability that something could happen in relation to a person or event. This means that risk cannot be the result of a perception. Nevertheless, theories of perception dominated the study of risk for decades under the influence of psychology, the first of the human sciences to address risk as a subject of academic research. This eventually led to the spread of the expression "perception of risk". On reflection, however, we can see clearly that perception has nothing to do with it. Risk is not a phenomenon of perception but a knowledge-based conceptualization of something. Here we may follow the German philosopher, Hans Georg Gadamer, who proposes a particularly apt definition of perception that goes beyond the realm of the senses:

"To perceive is not merely to gather sense impressions, but rather, as the German word wahrnehmen beautifully says, to perceive means to "hold (nehmen) something to be real (wahr)". This means that what is offered to the senses is perceived and taken as being something."[11] (Gadamer, H.G. 1991, 78)

However, the notion that something is held to be real or is perceived and taken as being something cannot be applied to the concept of risk, because risk perceives nothing. On the contrary, its nature consists precisely of the inability to perceive anything or to hold anything to be real, because it is oriented towards the future, as I have argued. Allow me to illustrate my explanation with an example. Let us imagine a situation in which a powerful earthquake far out to sea causes a tsunami to roll towards the coast. The quake is spotted thanks to early warning systems and alert networks, allowing the authorities to alert the population to the existence of the still distant tidal wave, so that they can react and flee the areas at risk in time. Obviously, this would be a model situation, in which the something (i.e. the tsunami) is still far off and cannot be seenas something real. Following Gadamer's definition, then, the tidal wave cannot yet be directly perceived, but its potential victims already know about it and they can therefore make it into an "object of risk" which could endanger their very lives, so that they themselves become the "objects at risk". This relationship is at the core of the concept of risk, but it can only be established through prior knowledge. As this example clearly shows, the great virtue of the concept of risk is that it allows us to anticipate events based on available knowledge, granting the population precious hours in which to escape.

This example and discussion of risk as perception lead me to certain other considerations which have formed part of the numerous errors and wrong turnings found in studies of risk. In particular, I refer to the widespread confusion between risk and danger. Let us now imagine the same tsunami, a giant wave rolling towards the coast. If there are still any people who have not yet fled the area at risk, and tragically this has happened on countless occasions, they will see things that they will take to be only too real. As we know, the wind will pick up, the sea will change and the great wave may be made out in the distance as it approaches. This is indeed perception and what is "offered to the senses will be perceived and taken as being something". The key point, however, is that what is perceived is no longer a risk; it is not something that might happen but something that will. It is neither more nor less than a danger. A danger can be perceived in a way that a risk cannot be. Animals react to their perceptions of warning signs indicating the presence of a danger, such as a predator. Obviously, such reactions can be futile or come too late, or it may be that the signs go unheeded with fatal results. In the case of a tsunami, the chances of escape are slim unless precautionary measures are taken in advance in view of the risk identified before the wave approaches the coast. This means that danger forms part of the thing in itself but risk does not, because risk constitutes a relationship as I shall go on to explain in more detail. Neither in English nor in Spanish could we describe a tsunami as risky or arriesgado, but we could certainly say that the phenomenon is dangerous or peligroso. This shows that danger is a property inherent in the object, something we can recognize in a person or a thing, while risk would in any event be a relationship established with something dangerous. By managing their situation in terms of a notion of risk, the victims of our hypothetical tsunami will flee the path of the oncoming wave, and by doing so they will be managing their relationship with the advancing threat. Those who do not, or cannot flee, will change their situation from one of risk to one of peril, as the danger inherent in the imminent arrival of the tidal wave translates into the actual perception of this new, critical situation. In short, danger is the capacity to cause harm that we can recognize in things, while risk is the relationship we maintain with this property in a given object. The difference between risk and danger is, then, substantial.

Niklas Luhmann was the theoretician of risk who addressed this issue most fruitfully, noting by the way how little interest scholars, most of them writing in English, had shown in the difference between danger and risk. In the first place, we may recall that Luhmann wrote in German, a fact that spared him the confusion brought about in English by the use of the ambiguous term hazard[12], which dictionaries tend to define equally as danger and risk. It may be that many authors writing in English, and they are the majority, have not paid sufficient attention to this matter because it is obscured by the language itself.

In The Sociology of Risk, Luhmann refers to this difference as follows:

"Two possibilities may be observed, then. It might be supposed that the possible harm is a consequence of the decision, and then we would talk of risk, or more precisely decisional risk. Or it might be supposed that the possible harm was caused externally, attributing it to the environment. In this case, we would talk of danger." (Luhmann, 2006, 67)

It is worth examining[13] both definitions, which in my view reveal a distinction that is partly correct, yet needs some qualification. The expression "possible harm" attributed to both risk and danger is problematic, since risk as such cannot be related with a possibility but only with a probability. This is because the function of risk is to convert a possibility into a probability through knowledge. To say something is probable makes it a little more certain than if it were merely possible. Furthermore, risk is not a decision, for although we can talk of risky decisions, risk management and risky behavior, risk per se is none of these things but the relationship established in terms of probability between an object of risk and an object at risk.

If I am driving down a country road and I pass a sign indicating rock slides ahead, I can say there is a danger of falling rocks, and this statement concurs with Luhmann's definition of danger. The risk, however, cannot be anything other than the probabilistic relationship I establish based on the information or knowledge contained in the road sign between the object of risk identified (i.e. rock slides) and the object at risk (my own physical integrity). By establishing this probabilistic link between the two, I am applying a notion of risk, and on this basis I can take a decision. Nevertheless, the risk is not in the decision, which could be either to turn around or to press ahead regardless. In the first case, I would neutralize any danger by avoiding the possibility of harm thanks to a notion of risk. In the second, after ignoring, underrating or consciously adopting this notion of risk, I decide to drive on, entering the sphere in which there is a danger rather than a risk. As in the example of the tsunami described above, what this shows is the temporal and spatial continuity existing in situations where we may have begun with a notion of risk but in which danger starts to dominate. There is thus a certain progression marked by transit points between risk and danger. To continue with the same example, the sign put up to alert drivers to the danger of rock slides must be positioned so as to give sufficient advance warning and far enough away from the place where the danger is imminent, allowing drivers time construct an account of the risk that would go something like: if I go beyond this point I could get caught in a rock slide. If I drive on regardless, any concept of risk I might have will lose its utility as the distance to the danger zone grows shorter and arrival becomes imminent, until at some point the play of probabilities loses its predictive effectiveness. Once again in this case, we see a possible distinction between risk and danger: risk is a future construction, and it exists in present only to conceive that future which is always imaginary, though it may be more or less believable, while danger is something we can perceive and describe as acting on our reality. A rock fall is dangerous because it is capable of causing actual harm. In light of these considerations, it is clear that the old debate over the objectivity and subjectivity of risk is pointless.

Luhmann does not quite manage to define risk, leaving the distinction he draws a little confused. However, his definition of danger is close to what I believe would be a possible definition. Danger is the capacity to cause harm that we recognize in people and things, but it is never a relational concept linking different things. Rather, danger is inherent in things or in people, and it becomes manifest in their relations with the world. This is key to the differentiation of danger from risk, which is a relational concept and not a thing or a property. To place risk in the decision-making sphere, as Luhmann does, means confining the concept to only a part of the phenomenology of risk, namely to risk-taking, an area in which the proximity between the formation of the concept of risk and the eventual decision can create the illusion that the risk lies in taking decisions. In reality, however, risk is an idea on the basis of which decisions are sometimes taken and sometimes not.

In light of the above, we may affirm that danger is a property or attribute of things, while risk is a relationship which we establish between things. In the first place, we can say that they are different by nature, and it is this that makes the confusion of the two so problematic. As we have seen in the preceding pages, risk is a notion of the future, and it cannot be identified with any particular thing without reification. Danger, on the other hand, is an attribute of things, and therefore it is a property that things and people can have. We can say that a person is dangerous, but we cannot say that a person is risky. Danger is not time, but risk is. Likewise, danger cannot be described as a relationship, and nor can any specific knowledge or probability be ascribed to it. The condition of danger has nothing to do with that of risk, and yet the two notions are related so frequently as to become confused. Danger is a property that is identified causally by the harm produced. Risk is not a cause of harm or anything else for that matter, but rather a concept that warns us of the likelihood of an event, which will normally, but not necessarily, be one that causes harm. We may, then, say that the notion of risk often serves to forestall dangers which might otherwise cause harm.

On August 7, 1996 a great thunder storm broke over the ravine called the Barranco de Arás in the Aragonese Pyrenees near the small town of Biescas. Heavy rain over a period of some three hours produced total precipitation of more 250 mm with a peak of over 150 mm between 6:00 and 7:00 p.m., causing a flash flood which crashed down the ravine, washing away the Las Nieves campsite and killing 86 holidaymakers. Responsibility for the tragedy was investigated in two judicial inquiries. The first case ended in 1999 with the exoneration of the authorities from any criminal liability. Meanwhile, the regional government of Aragon and the Spanish Ministry of Environment were ordered to pay compensation of €11.2 million to the victims in tort proceedings, which ended only in 2005. In its judgment the court cited the testimony of an expert witness, who declared, "Danger cannot be prevented, but risk certainly can be." The context in which this assertion was made is important.

One of the features of the ravine is an alluvial fan down which torrents of water and debris pour after the summer thunderstorms that are so common in the Pyrenees. Two major flash floods had in fact caused devastation in 1929 and 1959, although in both cases it was only the road that was washed away, as there were no buildings or other infrastructure in the ravine at the time. After the 1929 flood, major civil engineering works were undertaken to control torrents by building channels and plunges along the last stretch of the alluvial fan at the bottom of the ravine. Various dams were also built upstream to control flow. In 1986, however, a campsite was built inside the area covered by the alluvial fan at the bottom of the ravine. The facility had all of the requisite licenses and permits, which were granted despite a technical report describing the planned location as a "high risk" area. In 1996 a wall of water and debris crashed down the ravine, sweeping away the campsite and campers alike. In answer to media questioning, the authorities claimed the flash flood had been entirely unforeseeable, and that only fate could explain what had happened, as if this terrible disaster were a Greek tragedy. This argument remained the official cause for years afterwards. However, subsequent investigation brought to light an official report drawn up in connection with the license to build the campsite, which warned against the siting the facility within the area of the alluvial fan at the bottom of the ravine, where it would be in the path of any flash floods that might form. The existence of this report played a key role in the final verdict of the National High Court in the tort case. The Barranco de Arás was in danger of flash floods in spite of the preventive infrastructure that had been built. We can say, then, that danger was (and still is) an attribute of flash floods in the ravine, but the probability of this danger materializing and causing harm to people was (and is) a risk. Had the authorities heeded the warning not to allow a campsite to be built at the bottom of the ravine, there would have been no risk to people, and the flood would not have caused a tragedy. The Barranco de Arás was dangerous, and the location of the campsite was risky. The danger was an attribute of the ravine itself, but the risk was a relationship between the campsite and the ravine, which eventually had fatal consequences. From this standpoint, we may affirm that the not-guilty verdict handed down in the criminal proceedings reflects an assessment of the danger, on the basis of which it was found that the authorities were not responsible, but the verdict in the tort case addressed the risk and on this basis the authorities were indeed found to be liable.

In all of the above arguments, I have sought to stress the basis of knowledge that underlies risk. The historical facts prove this to be true, and in this light we may place the concept within the sphere of expert thinking from its inception. Risk is a sophisticated concept that has spread over time into every human activity worldwide, but always bearing the stamp of expertise and knowledge.

Risk is a relationship

In 2000 and 2001 a group of sociologists and anthropologists from five European countries formed a research network[14] to share the experience and insights gained from our applied investigation of risk. One of the key issues that arose in our debates on the definition of risk was the problem of reification. There was a tendency to make risk into something material and capable of acquiring permanence, so it seemed to us essential to stress its "relational" nature, and we placed this notion at the center of our own definition:

" Risk can be considered as a frame that creates contexts which bring together an "object of risk" (a hazard or source of potential harm), an "object at risk" ( a target of potential harm) and an evaluation (implicit or explicit), of human consequences. Risk is therefore a relational order through which connections between people and "things" are constituted and which guides its interpretation".[15]

We must avoid the reification of risk, because the quality of objectivity is not a part of the risk itself, but of the objects which contextualize[16] it through a relationship. The description of objects of risk and objects at risk is methodologically essential, as the materialization of risk arises under the terms in which the subjects relate the two. There is thus no contradiction between the real and the ideal, the object represented and the representation, that would oblige us to choose between two options, because risk is always a way of relating objects that actually exist. It is not these objects that characterize risk, but the nature of the relationship we establish between them, which is probabilistic.

A nuclear power station, for example, is not a risk as such[17], but its existence as a plant using nuclear fission to generate electricity, the shortcomings that may be attributed to it and the possibility of malfunctioning and the escape of radioactivity may make it into an "object of risk" for the neighboring community. This occurs when another object is brought into a relationship with the object of risk. In this example, such an object would logically be the health or life of the community, which would become the object at risk. Thus, risk springs into existence at the moment when a group of people identifies a relationship between an object of risk (e.g. a nuclear power plant) and an object at risk (the health of the community living in the vicinity of the plant), and recognizes the probability of harmful consequences.

This is precisely what happened at the Almaraz nuclear power plant in the western Spanish region of Extremadura. In 1996 a group of activists founded the Platform for the Victims of the Almaraz Nuclear Power Plant to protest against the high rates of cancer and leukemia in the vicinity of the plant according to data provided by local doctors. Four years later the campaigning efforts of the community living in the surrounding area succeeded in gaining official recognition of the need for epidemiological studies to assess the reality of the health concerns harbored by local people. On November 30, 2004, meanwhile, the television station Tele 5 broadcast a prime time report entitled Diary of a threat: nuclear power plants presented by Mercedes Milá, one of Spain's leading news anchors, who dedicated her best efforts to the issue. Ms. Milá appeared repeatedly in the documentary interviewing local people, most of whom described the health problems of family members, neighbors and friends, though some also discussed their own cases. These accounts were in general very similar, referring mostly to somebody who had fallen sick or died after living near the Almaraz plant, and occasionally to children born with congenital malformations and the problems experienced by expectant mothers. This highly emotive description of the link between living in the neighborhood of the nuclear plant and sickness suggested a nuclear causality in which the plant was plainly at fault. No data were provided or figures shown, however. All the activists did was to tell a story. Among other matters, this reveals how the notion of risk is expressed through narrative by the supposed victims. A number of other people were also interviewed, including a worker at the power plant, who spoke in a distorted voice without showing his face, a doctor working in the area, municipal authorities and residents of nearby villages, and a senior manager of the facility, who answered the journalist's questions in his office. With the exception of the latter, all of the interviewees concurred with one key point of the narrative, and this was that something was happening. While some accounts made no bones about blaming the nuclear plant, other activists were more cautious, calling for proof at the same time as recognizing the likelihood of a nuclear cause, so that suspicion still fell roundly on the Almaraz facility. Though there was a consensus demand for an epidemiological study, this had not yet been performed at the time the documentary was broadcast even after more than ten years of protests. It was not until 2009 that the Spanish Nuclear Safety Board and the Carlos III Health Institute finally undertook the survey, which covered all of the nuclear power plants in Spain. The results did not confirm the complaints contained in the discourse of the alleged cancer victims, failing to detect any statistically significant increases in mortality due to the disease. Like many others, this case exemplifies how the narrative version of risk assumed by a whole community may not agree at all with the experts' version obtained by probabilistic calculation. Social conflicts over the existence or future siting of major infrastructures or construction projects are common, because the notions of risk applied by the communities affected and by the experts differ markedly, being based on conflicting epistemologies, and it is often impossible to reconcile the two. This example not only illustrates the model of risk represented in Chart 1, it also reveals the essentially narrative character of the version of risk found in the Almaraz case, which was built up out of accounts of life and death, health and sickness. In the end, the events and protests affecting the nuclear plant produced a narrative of risk that had a national impact and was followed by millions of Spanish viewers on television.

If two or more objects, things, events or behaviors that do not necessarily have anything in common, like nuclear power and people's health can create a new entity, risk, this is because of the narrative that relates the two. Hence, the conceptual core of risk (i.e. the relationship between an "object of risk" and an "object at risk") is narrative.[18] Any narrative[19] must have a temporal structure, because time is a part of its essential nature, and narratives of risk always have the same time structure, which is none other than the future as probability[20]. Following the model of risk outlined above, a narrative of this type tells us that there is a probability of harm or loss (C) when a relationship is formed between (A) and (B). Before going on to discuss probability, we may conclude by proposing that risk should be considered as a relationship between "objects", of which we may predicate the ability to cause harm or to be harmed. This relationship is established in narrative terms.

Risk expresses probability

According to P. Ricoeur (1987), narrative is based, in the first place, on a temporal structure, and we can therefore distinguish between different kinds of narrative depending on their time frames. If, as we have seen, risk is a relationship constructed in narrative form, the time structure of this narrative is none other than that of probability. If we are to conceptualize risk as fully as possible, however, we must understand the nature of probability not so much as a mathematical ratio but as plausible information about the future.

Basically, mathematical probability links or relates the number of actual cases of an event with the number of possible cases, thereby structuring the possible outcomes. This would be an arithmetical definition of probability, and it was proposed in rudimentary form as early as the 16th century by the Italian mathematician and gambler Gerolamo Cardano in his Liber de ludo aleae, ("On Casting the Die"), a work that was published only posthumously in 1620. The first formulation of probability as a mathematical concept, however, dates from 1654 and is found in the correspondence between Fermat and Pascal. It appears again in 1657 in another treatise on gambling, De ratiociniis in ludo aleae penned by the Dutch mathematician, Christiaan Huygens. As a word, "probability" is much older and is derived from the Latin probabilitas from the verb probare, which means to "prove" and the suffixes bilis, denoting capability, and tat, denoting a quality, so that the meaning of probabilitas would be something like "the quality of being capable of proof". Many dictionary definitions of "probability" are either circular, like "the extent to which something is probable" in the Oxford Dictionary, or fall back on the mathematical definition. This is what the Diccionario de la Real Academia Española de la Lengua does where it defines "probability" in the following terms: En un proceso aleatorio, razón entre el número de casos favorables y el número de casos posibles ("In a random process, the ratio between the number of actual outcomes and the number of possible outcomes"). However, this dictionary also includes another definition: Verosimilitud o fundada apariencia de verdad ("Verisimilitude or convincing appearance of truth"). Both the Latin etymology of the word and the second definition given in the Spanish Royal Academy's Dictionary suggest that the semantic core of "probability" is related with access to a truth about something. I believe we can identify this something with the future, so that probability would mean having convincing or plausible knowledge about the future. In trying to express the meaning of the ratio resulting from the calculation of risk in words, we find that narrative and arithmetical probability converge, as each offers a kind of knowledge about what may occur. I believe, therefore, that we can properly say that narrative probability existed before mathematical probability was conceived. Let me provide some revealing historical facts in this regard.

The oldest written instance of the term risicum, which is to say the first Europeanized version of the Arabic rizq, extant today is a maritime contract made in Genoa in 1248, which reads,

"debet dicta navis in mari varari ad meum risicum et fortunam de omni casu"[21] (Villain-Gandossi,C., 1990,80).

It is worth considering this brief sentence, for it certifies the birth of narrative risk in Europe. The contexts in which the concept first appears are, in the first place, navigation and, in the second, maritime contracts. Thus, the term risk was contextually loaded from the outset. Its narrative properties are evident in this case, as the text tells a story, despite its brevity: if the ship sets sail, the charterer shall be liable for any loss resulting from shipwreck or any other adverse events that may occur. The contract seeks to single out risk, but it is a risk that is placed in context. Turning to the possessive "meum", we find that both risk and fortune belong to someone, and they are therefore not things that exist in and of themselves but relationships established by a person with something, whether it be an object, an idea or a situation. What produces the risk, the "object of risk" is the sea, and what is affected by that risk is the "object at risk" (i.e. the ship). Risk is in any event the relationship between these two "objects", which in this case takes on narrative form, "Debet dicta navis in mare varari." ("The said ship shall sail the sea.") Consequently, risk appears for the first time in Europe as both contextual and relational, structuring a narrative that contains plausible knowledge about something in the future, a future in which an event like a shipwreck or pirate attack could occur, and the risk of loss in these cases lies with a given party to the commenda contractunder which the vessel was chartered. Risk as probability is plausible knowledge about the future.

I have referred above to the enormous pressure which uncertainty has always exerted on mankind, and how divination was and continues to be used to look into the future, if only through a peephole. In its origin, risk was an alternative to soothsayers inspired by Islamic theology through an appeal to Allah to make the shift from the possible to the probable. Thus, the concept of rizq as employed by the Arabs in the 7th century was a discovery of enormous importance, making it possible to exchange absolute for relative uncertainty, as the rizq of Allah meant that an expedition across the desert appeared to the believer as a journey that would "probably" see the traveler and his goods arrive safely at their destination. Not long afterwards, both the Arabic rizq and then the Latin risicum were secularized, focusing increasingly on trade goods as the core probability shifted to the consideration of potential losses. Risk thus came to define the probability of shipwreck or pirate attack, and the consequences for a given party in the event of an adverse outcome. For these purposes, maritime contracts described all the possible circumstances of loss in narrative terms, employing the expression ad meum risicum to determine who ran the associated risks. Uncertainty was bridled in the first place through plausible contextual knowledge of the possible losses involved. This knowledge could in turn be used to foresee and compensate for losses within the realm of the probable. With the emergence of maritime insurance, risk and probability were honed to ever greater levels of precision. This was especially true of probability, resulting in an ever more accurate quantification of value, though not yet in the transformation into an arithmetical ratio.

Various scholars[22] place the first maritime insurance policy in Genoa in 1347. The text of this contract, which refers to the vessel Santa Clara bound for Mallorca, has been preserved. Georgius Lecavellum affirms that he has received 107 livres of silver from Bartholomeus Bassus by way of a free and friendly loan, which he undertakes to return. He further undertakes in consideration of that amount to assume the risks inherent in the voyage from Genoa to Naples by the direct route under the following terms:

"I personally assume all the risk, and responsibility for said amount of money until said boat shall have arrived at Majorca, being navigated by direct route as above. (...) In said manner and under said conditions I promise to make said settlement, otherwise I promise to you to pay and incur the penalty of double the stipulated amount of said money together with restitution of damages and expenses which may arise on that account or be sustained in litigation, the aforesaid remaining secure under the pledge and security of my property, goods and possessions."[23] (Golding, Ll, 1927, appendix I)

Here we find, then, an insurer, who receives from a ship owner or charterer a certain sum of money, which will fall to him if the insured vessel and cargo arrive at their destination without loss within a period of six months. Otherwise, the insurer will pay double the sum received together with compensation for damages, securing this pledge against all of his assets. The arrangement represents the beginning of insurance as a commercial activity in which the object of exchange is neither more nor less than risk understood as the probability of loss. We could, then, say that risk is a good which has a given, stipulated monetary value. This objectifies risk, which is no longer merely a "hazard" but also the amount concerned. In this way, a rationally religious notion of the future gradually became an intangible good that could be valued in terms of probability.

The concept of risk was born out of narrative, which described facts in the context of a future scenario and only subsequently took on the form of a numerical calculation of future events when an adverb of time like "when" was replaced by a mathematical ratio. Thus, the proposition "When A, a vessel, sails by B, the open sea, C, a shipwreck, may occur," can be substituted by "The ratio between A, the number of shipwrecks, and B, the number of voyages made, gives us C, the probability that the outcome will be a shipwreck." These are the narrative and mathematical versions of probability. Meanwhile, the historical facts show that narratives of risk already existed long before the discovery of mathematical probability. These narratives were at first theological, gradually evolving a legal and financial nature, and they expressed the probability of some harmful event.

References:

Braudel, F.- 1987 El Mediterráneo y el mundo mediterráneo en la época de Felipe II. México: Fondo de Cultura Económica.
Burckardt, J.- 2006 Historia de la cultura griega. Vol II. Barcelona: Editorial Iberia
Epalza, Mikel de.- 1990 "Origines du concept de risque: de l' Islam a l'Occident". En Le risque et la crise. European Coordination Centre for Research and Documentation in Social Sciences. Pp. 63-70
Defoe, D.- 2005 The Storm. London: Penguin Books
Defoe, D.- 2001 A Journal of the Plague Year. New York: the Modern Library.
Escohotado, A.- Los enemigos del comercio. Historia de las ideas sobre la propiedad privada I. Madrid: Espasa.
Faugeres, L., Vasarhelyi, P. and Villain-Gandossi, C. (eds.) Le risque et la crise. Malta: Foundation for International Studies.
Gadamer, H.G. 1991 La actualidad de los bello. Barcelona: Paidós
Golding, LL.- 1927 A History of Reinsurance with Sidelights on Insurance. Londres:Waterlow & Sons.
Luhmann, N.- 2006 Sociología del riesgo. México: Universidad Iberoamericana.
Ricoeur, Paul.- Tiempo y Narración I. Configuración del tiempo en el relato histórico.
Thomson P, and Dean, W. 1996 Competing Conceptions of Risk. https://www.piercelaw.edu/risk/vol17/fall/thompson.htm
Villain-Gandossi, C.- 1990 "Origines du concept de risque en Occident. Les risques maritimes ou fortune de mer et leur compensation: les débuts de l' assurance maritime". En Faugeres, L., Vasarhelyi, P. and Villain-Gandossi, C. (eds.) Le risque et la crise. Malta: Foundation for International Studies. Pp.71-85

[1] (Translated by S.H. Butcher, Internet Classics Archive, classics.mit.edu/Aristotle/poetics.html)

[2] Ricoeur, P. 1990 Time and Narration. Vol I. Chicago: the University of Chicago Press

[3] Loeb Classical Library, Harvard University Press, vol. XX, 1923; Latin text with facing English translation by W.A. Falconer, available online at;

https://penelope.uchicago.edu/Thayer/E/Roman/Texts/Cicero/de_Divinatione/home.html)

[4] Burkhardt, J.- 2006 The Greeks and the Greek Civilization. Vol II, 339

[5] It seems that the Arab root rizq is the etymological origin of the English word "risk", Spanish "riesgo", French "risque" and so on. It appears for the first time in the Koran, meaning the divine providence which protects and guides the faithful who are obedient to the will of Allah. Sometime in the 13th century, the term found its way into the maritime contracts made in the Christian ports of the Mediterranean (Venice, Genoa, Marseille, Barcelona, Ragusa, etc.) in the expression ad meum risicum. The earliest documentary evidence for its use is found in a contract made in Genoa in 1248. (See De Epalza, M. in Faugeres, L, Vasarhelyi, P. and Villain-Gandossi, C. (eds.), 1990, pp. 63-70)

[6] Braudel, F. 1976 El Mediterráneo y el mundo mediterráneo en la época de Felipe II. México: F.C.E. p. 234

[7] This is very like the situation of Christopher Columbus before he embarked on his voyage of discovery to America. However, Columbus made his voyage centuries later, and by his time sailors had developed a much more precise concept of risk, as revealed in the contracts signed by the Admiral, which include the Spanish expression "a riesgo de...". Historical research tells us that Columbus already had considerable knowledge of what he might find, as well as information which some believe may have come from sailors who had already made landfall in the New World in the course of voyages from the Gulf of Guinea to Portugal following the route known as the "volta da Mina", which passed close to what is now the coast of Brazil. He was also well versed in astronomy and was deeply influenced by Ptolomy's geography.

[8] Christianity had no similar theology because the Fathers of the Church had never looked kindly on trade and some, like Saint Ambrose, had even condemned merchants for going to sea, accusing them of greed.

[9] I have researched a number of cases ranging from opposition to new dams in the Pyrenees to the sinking of the oil tanker Prestige and the oil slick that was washed ashore on the northwest coast of Spain in 2003. In these studies, I have investigated how the basic knowledge that supports the shared notion of risk among a community is created and spread. (Mairal, 1997, 2000, 2008)

[10] I have mentioned these two examples because considerable evidence exists from actual situations of this kind in Spain.

[11] Gadamer, H.G. 1991 La actualidad de lo bello. Barcelona: Paidós. P. 78

[12] The English word "hazard" is a borrowing of the French "hasard", which is in turn etymologically derived from Spanish "azar", which according to Joan Corominas would lead us to the Arabic term "az-zahr", which means "flower". Both in Arabic and in Spanish and French, the term means "chance". It appears to owe its origin to the game of dice, which was introduced to Spain from Syria, as the faces of the die were marked with flowers. The semantic shift in the English word "hazard" is certainly curious.

[13] Unfortunately, I have no German and am obliged to rely on the Spanish edition of this work.

[14] This was the Network for Research into the Construction of Environmental Risk (2000-2001) (https://www.unizar.es/risk/) financed by the European Science Foundation. The Network comprised the Universities of Bordeaux II, East Anglia, Gothenburg, Naples and Zaragoza, involving the researchers François Lafaye, Peter Simmons, Asa Boholm, Amalia Signorelli and Gaspar Mairal, as coordinator.

[15] The materials developed by the research network can be found at https://www.unizar.es/risk/. The paper Representing Risk is especially significant.

[16] Thompson and Dean (1996) distinguish between probabilistic and contextualized conceptions of risk. In my opinion, all risk is contextual, and if we wish to draw a distinction like that this, it would be between calculated and narrated risk. The difference between these two types of risk is historical, as it was only after the invention of arithmetical probability that risk, a concept born out of narrative, came to include numerical calculation.

[17] The risk inherent in a nuclear power plant can of course be calculated by experts in probabilistic terms to establish safety levels. From the standpoint of the social sciences, however, this remains a mere algorithm, and it can never form part of the nuclear plant itself.

[18] The narrative nature of risk is supported and explained by the historical facts I have outlined above.

[19] In this case, I would define narrative as a series of structured facts or events communicated in a temporal order around a theme and expressed using a given style.

[20] I refer to this matter above where I explain that risk is time.

[21] "The said ship shall sail the sea at my own risk and fortune in all events."

[22] Villain-Gandossi, C. (1990), Golding, LL. (1927)

[23] English translation of the original document in Latin.