Can we predict the future with math? So far, social science has had less predictive success in real life than in Asimov’s Foundation series.
In his Foundation series of books, Isaac Asimov foresaw mathematics predicting and shaping the future. 65 years on from the release of Foundation, the social sciences use computers to test increasingly complex models of human behavior.
Is there a chance that Asimov’s Foundation is just around the bend?
Asimov’s Foundation Series
First published in the early 1950s, the Foundation sci-fi series was set against the backdrop of the dying days of a galactic empire. Seeing millennia of death and darkness ahead, a mathematician named Hari Seldon develops psychohistory, a mathematical tool for predicting human behavior. Using it to analyze the different outcomes, he sees one in which the coming dark age is much shorter. And so he sets up the Foundation, an organization whose purpose is to shepherd humanity to its least painful future.
Asimov’s Foundation series was inspired by the growing field of social science. In turn, they inspired social scientists and others who would shape the world. Nobel-winning economist Paul Krugman, politician Newt Gingrich, and entrepreneur Elon Musk are among those who have cited the books as an important influence on them.
In the same way that Star Trek inspired a generation of technologists and aerospace engineers, the influence of Asimov’s Foundation series lay in the minds it inspired, men and women who set out to emulate the mathematical predictions of psychohistory.
The Tough Reality
Social science has had less predictive success in real life than in Asimov books. Economics, sociology, and social psychology have all provided deep insights into the broad trends of how society behaves. But their ability to predict the actions of individual humans has been far more limited.
In the later Foundation books, covert agents are able to identify the times to act and the people to act on to push history back on track. In reality, pundits can seldom predict the outcomes of elections, never mind the decisions of individual voters.
The Nature of Social Science
This difference comes down to the nature of social science.
Isaac Asimov looked at social science and foresaw something that would become like the physical sciences, in which direct patterns of cause and effect can be seen. This atom splits that atom, leading to a nuclear explosion. This person meets that person, leading to a child who will be super smart.
But this isn’t how social science has developed. Instead, it deals in probabilities and big patterns. Mark Duggan and Steven D. Levitt can point at sumo wrestling data and confidently say that cheating has taken place, but they can’t prove who, when, or how. Opinion polls can give a candidate a 58% chance of taking the White House, but that isn’t the same as saying “it will be them”.
Economic Models and Big Data
Social science continues to make big advances. An important part of this has been the spread of mathematical models used in economics and game theory. In this sense, the social sciences are coming closer to psychohistory. Equations such as the Nash equilibrium can be used, however imperfectly, to predict human behaviour. The building blocks of the complex equations written by the Foundation may be emerging.
Some of this comes from academia. Economist Steven D. Levitt has reached a broad audience applying his skills to different aspects of society. On a smaller scale are projects such as physicist Taha Yasseri’s prediction of box office figures based on Wikipedia use.
Big business is also driving us toward more predictions through math. The past five years have seen the widespread adoption of predictive analytics, with businesses trying to understand everything from consumer trends to when they should hire new staff. Like the Foundation, they are using the numbers not just to predict but to shape the future. The popularity of predictive analytics in important journals such as the Harvard Business Review indicates that it is working.
Predictions are Still Failing
With so much effort going into these techniques, you would expect that we would have accurate predictions for the biggest events. But recent news has proved that wrong. In Britain, the electorate were expected to vote to remain in the European Union, with canny politicians such as David Cameron and Boris Johnson gambling their future on the outcome. Then came the results – hours after a remain victory was still predicted, Britain had chosen to leave the EU.
The same happened with the American election. Hilary Clinton’s persistent lead over Donald Trump evaporated and he won the electoral college by a comfortable margin.
Look back a decade and the same problem can be seen in the economic crash. In retrospect, economists have easily modelled the sub-prime bubble to see why it collapsed. But they didn’t see it coming.
Wouldn’t the Foundation have saved us from this?
E.T.A. Hoffmann’s short story, “Der Sandmann” is a compelling Gothic sci-fi thriller that inspires more theories and speculation than any modern mystery. Published in 1817, “The Sandman” (Der Sandmann in the original German) became one of the most influential and unique early uses of automatons, and using epistolary-style narratives (stories told from a series of letters or journal entries) to create anticipation. Though generally regarded as a pillar of the German Romantic movement, Hoffmann has been referenced as a powerful influence for fiction and science fiction writers such as [Click here to read more…]
Nudging and Specifics
It may be that our predictive models just aren’t sophisticated enough yet. After all, it took centuries for physics and chemistry to consistently get their results right. Perhaps sociology needs more time.
But the challenge for the social sciences isn’t just accuracy, its specificity. Accurately predicting the outcome of an election is useful. But if that prediction is general, it doesn’t tell us who to influence to change the future. Who is the Stalin substitute we could take out to prevent a new dictatorship? Which angry young man do you calm down to turn back a mob?
Then there are the conflicts that come up once we can predict the future. This is one of the ways in which life is like the Asimov’s Foundation series. More than one faction is learning to predict the future. Their use of that information changes what will happen, making other predictions less reliable. When looking at what will happen changes it, even accurate predictions become wrong.
Forward the Foundation?
So are we any closer to psychohistory than we were 65 years ago? Undoubtedly, yes. We have more complicated mathematics, better computers to run it on, and better data with which to make predictions. Asimov’s visionary future isn’t with us here and now. It may never be possible. But data analysis can identify a teenage pregnancy before the parents know; it can dig economies out of depression; it can even tell who will be top of the cinema box office. Not the grand dreams of saving a galactic empire, but a start.