In the Spring of 2020, Alyssa Goodman will begin teaching a "GenEd" (General Education) course on Prediction at Harvard. Human beings are the only creatures in the animal kingdom properly defined as worriers. We are the only ones who expend tremendous amounts of time, energy, and resources trying (sometimes obsessively) to understand our futures before they happen. While the innate ability of individual people to predict has not changed much in the past few millennia, developments in mathematical and conceptual models have inordinately improved predictive systems. These systems have integrated comparisons to past results and quantified how “certain” we can be about various aspects of the future -- processes that were, in many cases, inconceivable at one point in the past. This course is a coordinated investigation of the history and future of prediction, beginning with Ancient Mesopotamians reading signs in sheep entrails and ending with modern computer simulations for climate, health, wealth, and the fate of our Universe. In this class, you will design your own predictive systems to critically engage with assumptions about how the world works and situate your explorations in a study of how motivations and techniques for divining the future have changed–and not changed–throughout human history.
The Framework for Predictive Systems shown here captures much of the fundamental content of our curriculum. By the end of GenEd 1112, all students should be able to apply this Framework to the analysis of any predictive system.
Students will become familiar with various PREDICTIVE SYSTEMS (shown as a grey box in the Framework). Over the course of the term, each student will choose to analyze particular extant systems, as well as invent their own, so not all students will be expert on the same systems. Instead, the goal is to gain and appreciate expertise on just a few systems, or types of systems, while using the Framework to understand how many systems fall into general groupings, appreciating the general features of those groupings. By way of a simple example: predictive systems with a great deal of human input can all be very subjective.
The bubble in the Framework labeled EVALUATE ACCURACY is also key to GenEd 1112, as it encompasses the concept of uncertainty. There are vast differences in how uncertainty is and has been evaluated (if at all) and valued in predictive systems throughout history. By the end of GenEd 1112, students will have a deep appreciation for how, for example, Ancient Greeks may have thought about the Oracle of Delphi’s accuracy, while also appreciating how to understand the uncertainty associated with weather forecasts, or Bayesian analyses of jury trials’ likely outcomes.