Psychohistory: using Big Data to predict the Future
Psychohistory is a fictional science invented by Isaac Asimov and used in the Foundation trilogy. It combines history, sociology and disciplines such as statistics or mathematics to make general predictions about future events and behaviors of the human race. The use of this science has its rules: it only works with large populations of individuals and for long periods of time, it can only handle a limited number of independent variables, it works best when freedom of action is strongly restricted and it only works when its findings are kept in secret.
Asimov used the analogy of a gas to explain it: an observer has great difficulty predicting the movement of a single molecule in a gas, but can predict the action of the mass of the gas with a high level of precision (kinetic theory). Obviously this “science” is part of the science fiction relm but there is a lot of debate around whether it could be a reality using Big Data and Artificial Intelligence. In Factor Magazine they dedicate a really interesting article to psychohistory and the initiatives that could make it possible, such as the Global Database of Events, Language, and Tone (GDELT).
GDELT is an Open Data project designed to answer the question: what if we could take all the data around us and use massive computing power, algorithms and mindsets to try to understand the world? It currently translates fully automatically all the content that arrives in 65 different languages, and another 35 translated by hand, with a “sporadic” frequency by experts worldwide. It’s accessible on Google Cloud for anyone who wants to try to bring psychohistory to reality.
Another interesting reflection is to mix psychohistory with cryptoeconomics, a new discipline that mixes philosophy, sociology, psychology and economy based on technology. There is a very interesting video of Vitalik Buterin, Russian programmer and writer, known mainly for being the co-founder of Ethereum and the co-founder of Bitcoin Magazine, which explains it in great detail.
Obviously, there are many people who think that psychohistory should remain in the field of science fiction, that there is no way to bring it to reality, as is the case of Dan Hirschman who talks about it in Scatterplot.
My primary argument against the possibility of psychohistory is that the human condition — who we are and what we will do — is not a solvable problem. Emergence is the idea that systemic characteristics arise organically from the interaction of the system’s parts, and these emergent outcomes cannot necessarily be determined in advance. More detailed discussions of emergence can be found in organization theory (Padgett and Powell 2012) and critical realism (Gorski 2016), but I think it is fair to say that emergence is a central concept in sociology precisely because we are not just studying collections of individuals. A very similar set of ideas could be expressed in language more familiar to the Santa Fe Institute: Given that human societies are complex non-linear systems with multiple independent actors who co-evolve with their environment, these are not systems which lend themselves to being ‘solved.’ A more fun treatment (I’m assuming you’re here because you like to read science fiction!) can be found in Liu Cixin’s The Three-Body Problem, whose fascinating story uses the non-repeating motion of a planet orbiting three suns as a central plot point. The critique of quantitative and predictive social science I offer here is related to George Box’s idea that ‘all models are wrong, but some are useful.’ Unlike the ambitions of psychohistory, the good model — to my mind — functions as a paradigm whose value lies in eschewing nuance (to paraphrase Healey 2017).
My second main argument against the possibility of psychohistory is based on the meaning-making activity which is fundamental to human life. We imbue our lives and actions with meanings that we construct, but these constructions are subject to change over time. I have in mind a version of Giddens’ (1984) structuration, but there are other theoretical formulations which will work. Humans think back at you. Religion, kingship, family, the good life: these are not constant over time, and human behavior reshapes itself as the basic categories of how we make sense of our own lives change over time. The vector space of social life keeps changing because the categories we use to make sense of it are impermanent, and because our interventions shape future possibilities.
And you, what do you think? Will it be possible to create a large-scale system that anticipates big milestones for humanity?
#365daysof #futurism #scifi #bigdata #AI #day202