We might predict events – not impacts (part I)

(I’m fully indebted for this post to Duncan J. WATTS (2011): Everything is obvious (once you know the answer) – How common sense fails. Atlantic Books, London)

I cannot repeat it often enough – small events may have great consequences. One could also say: great impacts. If we are smart enough we may predict “events”, but we are hardly ever able to predict “impact”. Let me explore this assertion. In the period 1629-44 AD there were 1.8 recorded insurrections per hour in China[1]. One could make the plausible guess that throughout China’s history insurrections were a daily occurrence. All imperial dynasties in China were overthrown by insurrection – except for two, which succumbed to foreign invasion. Insurrections then were not “uncommon events” – Black Swan in TALEB’s parlance[2]. What made a few insurrections decisive were the circumstances that transformed a local affray into a nationwide convulsion. No wonder Yellow emperors took even the smallest signs of insurrection seriously[3]. This is what happens when we make a prediction. Out of the infinite circumstances of today we pick one item (we hope the right one); we project it into the future, free of co-evolving circumstances; and we predict the ensuing “event”; as it plunges anew into the real world of evolving circumstances we lose control on impact. Circumstances will instantly transform any “event” into an unforeseable “path dependent outcome”. The projection is like a flying fish rising out of the water for a short while – it flutters, plunges, and is gone. Here a supporting example. On 6^th August 1945 the US dropped an A-bomb on Hiroshima – a signal “event” for sure. What was its political impact? Historians[4] still debate whether Japan’s decision to surrender preceded, followed, or was accelerated by the bomb, or whether other factors were determinant – like the Soviet Union’s entry into war; or even the evolving attitude of Hirohito, who was looking to save his own skin and the position of Tenno. Once the “event” plunged into the circumstances of the day we are no longer able to follow its effects precisely. I’ve put “event” is quotation marks, for in human affairs events do not exist. It is a fig-leaf behind which we hide a complex reality. What we say is: “for the purposes of analysis we may disregard the complex reality hidden behind the label and assume this label to represent an unchanging reality” – the world reduced to billiard balls. We delude ourselves that, by hiding a complex reality behind a label, it will stay put and not bother us. Yet remember: “atom” means indivisible; well we split it into nucleus and electrons, and then protons into quarks, and quarks have flavors and charms…and depending on circumstances all of these elements may play a major role. Was the storming of the Bastille an “event”? The historian may think so – from his perspective – or from the level of generality he argues. Let’s stop for a moment and reflect. A crowd has assembled that day: where, precisely? how many? Who were they? with what ambitions, hates, and despairs? Could we understand the dynamics of a crowd in a predictable way? In order to do so, we would have to understand the behavior of each individual. But could we reasonably predict the behavior of an individual, and the innumerable factors, which moved him in the event? Every time we chose a “level of generality” – event – crowd – individual: deeper abysses of critical detail open up. We may not make a prediction about the “upstairs” without having understood “downstairs”… all the way down – except that there is no “bottoming out”. For whenever we think we imagine having found a causa causans, a prime mover, it dissolves upon closer inspection into myriads of other “movers”. In this sense history has no beginning. Only tales have a beginning, and they are fairy tales. Still unconvinced? I went to buy a couple of bottles of wine, this morning. I asked the caviste how business was. “Slow”, he answered “it is election time. People never buy before elections.” The Italian “paidina” vendor is nervous: business is slow for him too. The pharmacist confirms the slow trend, but rejects the explanation: “people just don’t have money”. Whatever the cause for people’s behavior, it seems to peter out into the intangible, rather than “free will” or a precise cause.

Let’s move on. We may predict simple systems, like the interaction between the sun and the planets. Even the interaction between three suns, however, would be beyond our mathematical capacities (said Poincaré at the beginning of the XX^th century – and he has not been disproven). Imagine complex systems, where infinite small interactions occur all the time. I’ve just come across a splendid example of the “three body problem”. George FRIEDMAN[5] is a professional strategic analyst and a scholar of war. From prediction to prediction he has gone to establishing STRATFOR[6] as one of the prime sites for “strategic analysis”. FRIEDMAN’s views might be considered “influential” among decision-shapers worldwide. In 1994 he co-authored: Coming war with Japan[7]. Five years after Tiananmen he had come to the conclusion that China would “re-fragment”. As China implodes – FRIEDMAN argues in his book – the US and Japan will fight each other over the spoils awaiting the winner. Even had the “re-fragmenting event” taken place – the interplay between China, US, and Japan could not have been predicted. Enough for today. Many reasons prevent us from predicting “impact”. One is left for tomorrow – creativity.

Link of original post