Isaac Asimov did not only deal with robot laws. One of his most famous works is the Foundation Triology. The hero is a mathematician named Hari Seldon who develops psychohistory as a science. Seldon assumes that due to the mass of mankind – it populates an entire galaxy – empirical-statistical methods can be used to calculate laws of behavior. They can be used to predict the future, because some patterns do not change. Seldon calculates that the existing empire will collapse in a few centuries. This is followed by 30,000 years of chaos and barbarity before a second empire restores peace and prosperity. Seldon ensures that his successors intervene at decisive turning points. In this way he wants to shorten the barbaric time – with success.
In between, however, there was a point at which Seldon’s plan threatened to fail – a mutant who calls himself a “mole” – develops abilities that Seldon did not foresee. However, Seldon foresaw the possibility of such a deviation and developed his Plan B – a second association of psychohistorians who, apart from the galactic center, strengthen their mental abilities and oppose the mole, a fanatical seducer of mankind.
In 1988, the American author and statistician Michael Flynn wrote an introduction to psychohistorics on Isaac Asimov’s Foundation Triology, in which he examines the question of whether the future can actually be predicted by means of statistical methods – tracking down patterns in the history of mankind. He shows different ways of identifying such patterns. I asked Michael Flynn a few questions on this subject.
Today, scientists are working on computer models with which they can analyze the behavior of masses in the past or present and thus also calculate the future. So is your text more valid today than ever?
Today I am less convinced that specific predictions are possible. The theory of evolution, for example, cannot predict which descendants will develop out of a species. However, a mathematical ecology is possible: we can establish equations that, for example, calculate the predator-prey ratio, population growth and similar things. Something of this kind could also be possible for historical processes – for example, how ideas spread in a culture. There is an interesting book on this subject, “Looking at History through Mathematics” by Nicolas Rashevsky.
But what cannot be predicted, for example, are regime changes. In the process, a system changes in such a way that mathematical patterns that may have been set up previously no longer function. It’s like a rubber band with a weight. When the weight becomes heavier, the band stretches. There is a positive correlation between length and weight. But at a certain point the rubber band no longer becomes longer, but tears. The rubber/weight system has reached a point where the equation is no longer valid. The mole in Foundation Triology was in some sense such a moment – unpredictable because there is no variable in the equation for it.
Do you think that the Internet and especially the social media could help to develop a new statistical method to at least predict social behavior?
No, if at all, these media are part of the social behaviour to be predicted. They certainly facilitate the collection of certain information. But when I look at the Internet like that, I also gather a lot of fictitious information. Some things may go faster, but the Internet does nothing to ensure that data is accurate, relevant and complete.
Future visionary Ray Kurzweil at least believes that a technological singularity can be foreseen – a time in our future when artificial intelligence will replace biological-human intelligence.
Singularities are mathematical abstractions. In history, things usually do not continue exponentially. Singularity is quite simply a Christian belief that has been rewritten into a secular scientific belief. There is no guarantee at all that technological progress will continue indefinitely – there have been dark periods of time in the past – and perhaps artificial intelligences are even impossible in principle. In addition, technological availability does not mean that it will be used. A culture encompasses much more than just its toolbox.
So predictions based on previous developments cannot work?
The future will always be unpredictable if the prediction is to be so specific that it benefits individuals. That is why the climate is easier to predict than the weather. The more accurate the prediction, the more likely it is to be wrong. And the less accurate it is, the more likely it is to be useless.
In addition, there is the human uncertainty factor. Unlike inanimate bodies, people do not always react the same way to a stimulus. Sometimes a sabre rattling convinces another country to retreat. Sometimes the same behaviour leads to war. That is irreducible. No science is a science of the individual. Physics does not study a single atom, chemistry not a single molecule – but how these things behave in the masses. But history does not behave in a general way. Even if sabre rattling prevents a war in 80 percent of all previous cases, the next case cannot be predicted.
Nevertheless, some scientists believe that comprehensive socio-economic simulations are helpful in assessing the impact of political decisions – they can intervene in time and prevent negative developments.
Governments have always tried to manipulate the development of society. By subsidizing A and taxing B, one hopes to strengthen A and weaken B. However, the law of unintended consequences says that while you actually strengthen A and weaken B, you also influence C, D and E – and these influences are unexpected, unplanned and sometimes unpleasant. A culture is like a mobile. Pull in one place and all the other elements wiggle with you. And then C, D and E have influence on A and B again. When the USA banned alcohol, they wanted to prevent people from getting drunk. This led to an increase in organized crime, more corruption in the police and courts, and the illegal bar continued to drink diligently.
Isn’t it also a contradiction in Hari Seldon’s theory that the development of society is predictable and influenceable at the same time?
A truly psychohistorical theory would have to include all efforts of influence. Then the efforts are part of the effect. And part of the cause too. In other words, a historical process does not develop despite people’s intentions and desires, but because of them.