One of the main motivations for developing Seshat is to use our store of coded information to cut through the tangled undergrowth of competing theories that has grown up around our understanding of human history. An article written by Laura Spinney just published in the latest issue of New Scientist does a nice job of bringing this key feature of our approach front-and-centre. The article provides an accessible overview of the aims of our project, the current state of play and some of our plans for the future. Popular coverage of science often likes to have a simple narrative of scientists making rare or unexpected discoveries, or finding support for a certain hypothesis. However, progress in science is actually as much about working out what theories aren’t supported. In academic disciplines such as history, which have traditionally been non-scientific in their approach, there has not been an agreed upon way of deciding which ideas are better supported. This has led to a proliferation of ideas, especially as all these theories have proponents who have been able to muster some evidence consistent with their thoughts. A main aim of our analyses is therefore to create what is referred to in the article as a “cemetery of theories” that don’t explain the data as well as others.
As an example, one of the aspects of human history that we are interested in is the origin and evolution of large-scale, complex societies. This area of inquiry is home to a vast number of ideas, often proposing different factors that acted as drivers of social complexity. Some of these main ideas were set out in an article published in the journal Cliodynamics written in the early phases of project. To test these hypotheses we have been collecting data on social complexity, warfare, rituals and also developing models of agricultural productivity for past societies (as mentioned in the New Scientist article and discussed in more depth in a 2015 article). Analyses involving these data allow us to assess how these factors relate to one another and to assess which ideas are better than others at explaining the emergence and spread of the kind of complex societies we live in today.
The project itself continues to evolve. We have recently come to the end of the first wave of data collection, but have already moved on to the next phase. Although written extremely recently, the quoted figure of more than 130,000 data points is already out-of-date with our present total standing at 146,763. We are currently conducting a number of analyses, and the article contains a teaser of some analyses where we are quantifying social complexity and trace its evolutionary trajectories in our 30 world regions. We aim to submit a paper on this in the near future, so more will follow once we have charted the choppy waters of the academic peer-reviewed publication process.
An important angle that the article does not cover, yet is one of the key features of our approach to data collection, is the fact that we work with a range of historical and archaeological experts at all stages of the data collection process. This enables us to find the best way of getting information from the existing literature, and also helps ensure that what we code best reflects this information. A potential criticism of projects like ours, that is given voice in the article, is that the reporting and coding of information is coloured by the people involved in collating and collecting that information. It is worth pointing out that it is for exactly these reasons that we are so keen to reflect different points of view about whether certain features of past societies were present of absent, and to provide referenced information to justify coding decisions. We are not trying to make final decisions about the veracity of any particular piece of historical information, but instead reflect what is the current state-of-the-field (which may change as new information or interpretations come to light). One of the things we would like to stress is that we want to work with humanities scholars, rather than trying to replace them. This is shown by our general approach and the fact that several Seshat members are based in the Cliodynamics Lab in the Digital History Research Centre at the University of Hertfordshire under the directorship of Seshat editor, Pieter Francois.
The article ends by highlighting some of our future directions, including our continued move towards incorporating exciting new developments in computer science. Indeed, the seeds of the article were sown when another New Scientist journalist, Kate Douglas, visited our Dublin meeting and the associated 2nd international workshop on computational history and data-driven humanities. As the article mentions, we are exploring ways in which we can use Seshat and the analyses we conduct to better inform public policy. This is hopefully part of a growing trend as envisaged by the recent call by Graham Allison and Niall Ferguson for historians to be in a closer dialogue with politicians. However, we feel that in order to truly learn the lessons from history we need to study it in a systematic way by assessing alternative explanations, weighing the evidence and putting to one side the ideas that don’t work, and that is where Seshat can help.
P.S. The article includes some funky, bespoke artwork by Beppe Giaobbe. While there is a picture of Egyptian deity, it is not Seshat – so here she is: