I talked about it before and some, but the concept is so interesting to me that I need to elaborate.

I believe that it is intrinsic in human nature to possess a kind of confirmation bias, extrapolating severely limited set of data points (say, the time span of one generation) to make statements about timeless truths, concepts lasting for very long time frames. In other words, we all have a kind of local view of history which is causing us to make incorrect assumptions about the past, or is preventing us from questioning things enough to make good predictions about the future.

There are numerous examples to support this idea. Until about 2006, the general populace was convinced that housing prices will always go up. We fear of terrorists taking over planes but forget that in the seventies, plane hijackings were rather common. We love good food but just fifty years ago Americans considered food to be a rather utilitarian exercise.

It is commonly thought (and by “It is thought” I mean “at some point we all thought that, even if now we may not admit it because of a certain pressure from those who tend to expose common conceptions as myths who may make us appear stupid”) that the Middle Ages were by and large a waste — several centuries of backwardness. However, I believe that we think that only because we live in a technophiliac age where one assigns value to a very specific kind of progress that was, admittedly, absent in the Middle Ages. This is most likely coupled with a phenomenon in which we conveniently forget that progress has an exponential nature — sure, the Renaissance seemed like a huge step forward, but it very likely needed to be bootstrapped by a much slower progress that was brewing in the centuries before it (plus, following the Durants in the Lessons of History, “Sometimes we feel that the Middle Ages and the Renaissance, which stressed mythology and art rather than science and power, may have been wiser than we, who repeatedly enlarge our instrumentalities without improving our purposes”). But, even more fundamentally, our belief in progress is an even more deeply rooted assumption about the human race. Can we extrapolate it into the future? Will we always aim towards progress? What if we deplete our natural resources?

And how about political systems? We think of monarchy as a less developed, more primitive form of government that suffers from a lot of problems that we have since managed to mitigate with other systems. But again, this is just confirmation bias: we were raised in a culture that relishes in the importance of democracy. We seem to think that democracy will reign forever — or at least, that monarchy will never come back. But that doesn’t have to be the case (Frank Herbert was onto something). It doesn’t have to be the case because the path between concept and execution is chaotic — small variations in the original assumptions can lead to massive differences in execution, so what seems like a massive problem with monarchy in the context of eighteenth-century world citizen’s worldview may easily be solvable with twenty-first-century’s tools and mindset. For example yes, there is potential for abuse, but what if we use technology to limit it just as technology today ensures we don’t abuse some of the privileges given to us. Once we solve these problems (which centuries of oppression have brought to the foreground), the underlying framework may provide a much more efficient way to stabilize the society (in fact, I’ve always thought that democracy is ostentatiously very inefficient — just think about the years of training, campaigning, and money spent by the candidate who loses the elections).

That’s exactly the problem with extrapolation: it assumes that some phenomenon can be approximated well with a linear model. So long as the extrapolation is minimal, most phenomena do behave pretty nicely: if today we’re burning up coal like crazy, tomorrow we are also likely to use up a lot of it. But it does not necessarily follow that in fifty years we will. Extrapolation over a long time window is unable to take into account the interactions of the thousands of variables that affect the outcomes we’re trying to predict.

History, or social sciences in general, suffer from this problem particularly much, precisely because there are so many variables at play (i.e. the systems they try to explain are so chaotic). In a way, one of my biggest realizations was that social sciences are enormously more complex than exact sciences because the latter have the liberty of operating on kiddie worlds where everything can be controlled and measured and there are relatively few degrees of freedom. It’s only when we move on to sciences that rely on inaccurate, nondeterministic and complex objects, such as economics (struggling with the complexity of human nature when applied to incentive-driven behavior), politics (struggling with the complexity of understanding the impact of policies on e.g. economics) and history (struggling with the complexity stemming from the interplay of a number of economic, political and social factors with individuals and their decision-making process), that the number of variables increases beyond our comprehension. Those models are nowhere near linear, so let’s not extrapolate.

There are some positive consequences of the abandonment of superfluous extrapolation. It is not necessarily the case that food will look less and less like food — in fact, we are slowly starting to see people go back to natural sources of food, from organic food to CSAs to growing their own vegetables. We may run out of rare earth metals in twenty years, but that doesn’t necessarily imply armageddon.