by Kevin Hong
Throughout human history, philosophers have been wondering about the future of our species. At the societal level, such intellectual inquiries have been mostly prescriptive (what should be) rather than descriptive (what is/what will be). For example, Plato famously proposed his ideal state in the Republic where individuals with different capacities occupy different societal roles to achieve harmony, efficiency, and justice; more recently social Darwinists attempted to use evolutionary theory to improve the human race genetically. At the individual level, there has been a mixture of prescriptive and descriptive theories regarding what will happen during someone’s lifetime. Some of the descriptive theories generates genuine predictions such as physiognomy where a person’s character and fate can be revealed by examining their facial and bodily features, an extremely common practice in early civilizations from Mesopotamia to China.
Modern people are interested in the same questions, but are approaching these questions with more powerful tools. Peter Turchin, a professor at University of Connecticut, has developed a series of models on human societal evolution based on extensive analysis of historical data (http://peterturchin.com/cliodynamica/quantitative-prediction-political-v...). Turchin specifically predicts that the US will experience a period of heightened social and political instability during the 2020s (he made the prediction in 2010).
Other predictive models focus more on the mechanisms, and a variety of approaches have been utilized to that end. For example, agent based simulation has been employed to model political decision making (https://pdfs.semanticscholar.org/7628/740832bff9a7d5d23e1ea5ddfbeb5e063f...). These computational models do not usually make grand predictions about the future evolutionary trajectory of human societies; rather, they offer insights regarding how different variables affect some outcomes of interest such as voting behavior.
At the individual level, as physiognomy has generally been discarded by the scientific community as pseudo-science, genetics and epidemiology emerged as powerful tools to generate health-related predictions based on personal genetic background and environmental factors. Today, scientists have acquired a significant amount of knowledge of the human genome, and tremendous amount of efforts have been devoted to understanding how specific genes influence phenotype, particularly in the medical context. We now know, for example, that Huntingtin’s disease is 100% genetic, meaning every child of a parent with HD has a 50/50 chance of inheriting the faulty gene, and the symptoms will develop between the ages of 30 to 50 (additional read: why hasn’t natural selection got rid of the HD gene? https://evolution.berkeley.edu/evolibrary/article/side_0_0/medicine_05). Knowing this information can be very valuable because it not only “reveals one’s fate”, but also may influence important life choices (such as reproduction).
But most diseases are the result of complex interactions between genes and environment. Epidemiologists use complex statistical methods and machine learning algorithms to evaluate the likelihood that a given person may develop a particular disease. The methods and algorithms usually identifies variables that associate with an increased risk of disease of infection (called risk factors), which often offer intuitive predictive power. For example, smoking doesn’t guarantee that they will develop lung cancer, but it certainly increases the chance of developing lung cancer compared to someone who doesn’t smoke. You can probably see how insurance companies would use this type of data.
The advance in medical technology has drastically reduced disease occurrence and increased life expectancy. This, however, may have an unintended effect. Prominent quantitative geneticist Michael Lynch has recently suggested that due to a lack of selective pressure against mildly deleterious mutations. The long-term consequence is an “expected genetic deterioration in the baseline human condition, potentially measureable on the timescale of a few generations in westernized societies”. (https://www.genetics.org/content/202/3/869). This prediction is based on an understanding of the human genetic architecture and the logic of natural selection. The paper, published in the journal Genetics, generated a bit of controversy (https://www.ncbi.nlm.nih.gov/pmc/articles/PMC5068865/).
There are other phenotypic traits that are not directly health-related that could nonetheless influence the future of humanity in significant ways. Our cognitive capacity, for example, has been shown to have a substantial genetic component. Two recent paper published in Proceedings of the National Academy of Sciences have shown a steady decline in the genetic component of educational attainment in the US and Iceland over the past century despite the undeniable increase in the phenotype (https://www.pnas.org/content/114/5/E727.short, https://www.pnas.org/content/113/28/7774). The implication is that the increase in phenotype (IQ score, number of years of schooling) has been primarily driven by environmental factors whereas the genetic component has been deteriorating. I recently constructed a simulation showing that the increase in phenotype is unlikely to increase indefinitely and may start to decline once the environmental contribution “saturates” (https://www.biorxiv.org/content/10.1101/605311v2).
In an important sense, predicting what will happen to X or how X will change depends on our knowledge of the system that consists of X as well as X as a system consisting of other factors. Therefore, as our knowledge increases, we should be able to make more accurate predictions about X. These predictions often appear in the form of if-then conditionals, and is perhaps best exemplified in economic policy evaluations: What will happen to the overall economy if we raise the tax for the top one percent? What will happen to the average consumer if the US starts a trade war with China? Is universal basic income a good idea? These are vitally important questions that have been taken up by scholars and politicians alike (https://www.hup.harvard.edu/catalog.php?isbn=9780674052284).
Obviously, such predictions depend not only on a good grasp of the economic system but also an understanding of fundamental human nature, in particular, what shapes preferences and how humans behave in environments with different incentive structures. By specifying these parameters, economists are able to identify conditions that produce important economic and financial phenomena, such as booms and busts in asset prices (https://papers.ssrn.com/sol3/papers.cfm?abstract_id=2252) and perhaps even financial crises (https://www.nature.com/articles/460685a).
There is yet another group of factors that could influence the future of humanity in a rather brute-force way. Since humans’ primary habitat is the earth, any change in the habitability of earth is going to significantly affect humanity’s future prospect. We know that dinosaurs went through mass extinction due to geological and climate changes (https://www.nytimes.com/2019/10/21/science/chicxulub-asteroid-ocean-acid...), and the current anthropogenic climate change may lead to undesirable consequences such as the rising of sea level and the increasing occurrence of extreme weathers. In the very long term, earth climate will change no matter what we do because the sun will die (https://www.sciencealert.com/what-will-happen-after-the-sun-dies-planeta...). Thanks to our increased understanding of astrophysics and unprecedented computational power, astronomers have figured out when this will happen: in another 10 billion years. There is not much man can do to prevent that from happening at this point, but this piece of information may contribute to the on-going discussion of space colonization (https://phys.org/news/2017-03-future-space-colonization-terraforming-hab...). On a smaller scale, asteroid hitting earth could cause mass destruction as well (the 1908 Tunguska event which flattened hundreds of square kilometres of forests in Siberia, released energy equivalent to several hundred Hiroshima bombs). Thankfully, these events can also be forecasted by astronomers now (Rees, On the future).