YOU have to admire the authors鈥� ambition: Small Wars, Big Data has the laudable aim of reducing the toll of misery and waste in war by building a new explanation of conflict.
As the technological divide widens between the world鈥檚 richest and poorest combatants, imbalances in the capacity of warring parties are now common.
But do we really need new thinking to explain what follows? Yes, say the authors: most previous ideas have not recognised important differences between asymmetric and symmetric warfare. They assume that both forms of conflict follow the same rules. This glaring inadequacy is caused by people looking 鈥渇or one big answer 鈥� the kind of unifying theory physicists have been after鈥�.
Advertisement
This is a stirring call to arms (as it were). But the underpinning claim simply isn鈥檛 true. The foundational text on asymmetric conflict is Small Wars: Their principles and practice, written by C.E. Callwell, an Anglo-Irish officer who later directed operations and intelligence for the British Army in the first world war. Published in 1896, Small Wars proceeds from the premise that 鈥渟mall鈥� (asymmetric) wars and 鈥渞egular鈥� wars have distinct differences, which he explicates with a fluency and cogency that has yet to be matched.
I can鈥檛 believe the authors of Small Wars, Big Data haven鈥檛 heard of Callwell, so why would they have disregarded him?
As a professor of war studies, not informatics, I am guessing here, but I think the authors may have been dazzled by the 鈥渂ig data鈥� shibboleth without ever getting properly to grips with it.
Take this passage, in which the authors discuss the attitudes of civilians in a conflict: 鈥淣aturally, citizens鈥� attitudes will also be tempered by politics (co-ethnicity, sympathy for the goals of one side or the other, etc.), but we believe that their shifts in attitudes and their resulting actions can be measured and understood in terms of our model of conflict as a three-sided game [of government, rebels and civilians].鈥�
Of course model-building requires simplification 鈥� but this is over the top. It probably goes without saying that insurgency is predominantly a political act, rather than a military one, yet here we have all the difficult bits of the insurgency concept packed with an 鈥渆t cetera鈥� into a pair of parentheses!
Standard statistical and computational methods were sufficient for the data analysed here. Although they did have to hand-code some information from paper records, their data sets seemed neither too large nor comprised of streams arriving too fast to process readily 鈥� the conditions that normally call for big data techniques.
The authors claim that 鈥渂ig data allows us to identify cause-and-effect relationships in ways we never could before鈥�. To date, however, these techniques seem powerful but dumb: excellent for spotting correlations, lousy at explaining them. By all means apply maths to strategy; it may do good. But don鈥檛 skimp on the history, without which it won鈥檛.
Book details
Princeton University Press
This article appeared in print under the headline 鈥淏ig data goes over the top鈥�
