Whether it's a shape, a structure or a work of art, once you see a meaning in something, there's no going back. The reasons why run deep.
The concept of chance as we understand it today emerged long after the birth of philosophy. The ancient Greeks distinguished between events which were inevitable and those which were seen as the will of a god. The idea that there could be events which were neither inevitable nor intentional is a difficult and abstract one which, with contributions from Cicero, the 1st century BC Roman philosopher, poet and statesman, has progressively gained strength as modern science has developed.
Even today, the assertion that a particular phenomenon is "random" is neither trivial nor obvious and, except in textbook cases, is the subject of debate. For example, after French biologist Jacques Monod wrote in his 1971 book Chance and Necessity that genetic mutations were due to chance, several biologists voiced objections to his use of language, accusing him of confusing his ignorance with a "roulette wheel". A simpler example can be found in the image you will see by clicking here, which we will come back to later.
We tend not to accept the possibility that something meaningful could actually be random: it has to be down to chance or it has a meaning. So as this image has no meaning, we accept that it is random.
Artists, of course, have long realised this incompatibility, which asserts that chance and meaning are mutually exclusive categories. In the early 20th century, when modernist architects such as Frank Lloyd Wright, Le Corbusier and Jacobus Oud wanted to escape from historical styles and classical symmetries, they used randomness in their composition, something which was a revolutionary feature of Art Nouveau and the Dutch De Stijl movement. When the Greek composer Iannis Xenakis wanted to abandon tonal music and the habits created by classical harmony, he used random choices governed by Markov chains.
Erasing meaning can also reveal new sensations. The pioneer of abstract art, Wassily Kandinsky, said that he turned to that mode of art because of a revelation he had one evening. At dusk in his studio he saw, lying against the wall, a canvas that made a huge impression on him. Moments later he realised that it was one of his own paintings upside down. Abstract painting was, for him, a quest for such precious experiences.
Conversely, when we do not understand something, we tend to think that this is the result of chance. In front of an abstract painting that leaves us unmoved, for example, we tend to dismiss it and say: "This is just random squiggles of paint!"
The question of meaning is at the heart of scientific knowledge. Both philosopher John Stuart Mill and mathematician, economist and philosopher Antoine Cournot stressed the importance of the human ability to identify patterns, that is, to guess a physical law and expect it to hold in future.
For example, suppose we plot the results of an experiment. If they lie on a circle, then we take this to suggest a law, not an accident, even if the points are not exactly on the circle. Cournot remarked that the points could just as well lie on a parabola, an ellipse or a spiral as a circle. Our propensity to believe that there is likely to be an underlying law somewhere is linked to the simplicity of the law, which in this case is the simplicity of the geometric shape. The simpler the law, the stronger our belief in it. But there is no obvious classification system for showing the relative simplicity of laws.
For example, a spiral may, in some ways, be better than an algebraic curve at representing nature but it is not clear which is simpler. Cournot deduced that our belief in a law cannot be quantified and therefore lies in the field of interpretation.
Let's go back to our image. You see it as random because you do not perceive any meaning in it. But if you click on the next link, you will suddenly see a meaning and the image on this page will no longer seem random. But before you do so, be warned: after making the new interpretation, you will never be able to regain your current ingenuousness - the interpretation will remain and will henceforth appear as reality.
This is the irreversibility of interpretation. It holds quite generally: once you perceive something in the world - a shape, structure or a meaning - you can't go back. You can access other new interpretations, certainly, but you can never go back to the state of mind you had before that first interpretation.
This irreversibility gives society a feeling of progress in the field of interpretation. It is clearly seen in politics, where it turns political doctrines into beliefs among their adherents. For example, those who have read Karl Marx and adhere to his theory that human consciousness is conditioned by social class start to see this influence behind the words of politicians, journalists and businesspeople - and they cannot stop seeing things in this way. Meanwhile, others see the invisible hand of Adam Smith regulating the economy and they also cannot escape from this interpretation of social behaviour.
All this is crucial to truly understanding risk. The belief some people have that risks can be objectively measured means expunging their interpretative aspect, even though that aspect is an essential part of understanding risk. From the epistemic point of view, it is the meaning of the event that determines the risk. The probabilistic representation of risk as a pair of mathematical quantities - a probability law that governs the states that can arise, and a random variable that assigns to each state a measure of the financial implications of that state - is too simplistic.
Usually there isn't enough information for such a model: we do not know the probabilities of rare events occurring since there will never be enough data, we do not have a full description of what can happen, and we do not know how to calculate the cost of that event occurring.
Above all, that kind of representation ignores the reasons why we are interested in these events because it acts as if translating them into financial terms is - or could ever be - automatic and objective. True risk analysis necessarily involves understanding interpretations. That's much more difficult, not least because it is sensitive to the information that is available to concerned social groups, and to their imagination, not in the sense of dreaming or delirium, but in their ability to perceive the field of possibilities. And those concerns are part of social reality.
For example, the Japanese nuclear crisis in March this year changed many people's perception of nuclear energy. As a result, many different interpretations of the situation are possible. For example, one interpretation might be that the management responsible for the Fukushima plant was negligent but that nuclear power stations of other advanced countries are well managed and safe.
Another view might hold that there are still risks but that technological advances will eliminate them. And yet another could argue that while security may be effective in well-organised countries, in countries where there are higher levels of bombings and corruption, security will be impossible and nuclear power should be abandoned even where it is otherwise safe.
The bottom line - quite literally, sometimes - is that to really understand risk, we have no choice but to take account of the way people interpret events.
Nicolas Bouleau is a mathematician and professor at École des Ponts, part of the Paris Institute of Technology, France. He heads one of the first French research groups to work with banks. His new book, Risk and Meaning: Adversaries in art, science and philosophy is published by Springer.