Andrew Brown recently published a piece about war and civilization on his Guardian blog, an interpretation of a recent multi-authored paper in Proceedings of the National Academy of Sciences, which mathematically models the relationship between war, space and the evolution of complex societies.
Brown concludes that this paper demonstrates a relationship between war and civilization that renders a leftist politics untenable in an age of (European) peace. In summary, he concludes that the findings of this paper should worry the left, as it suggests social solidarity is dependent upon war fighting and that in the absence of “the Spirit of 45” the left will struggle to articulate a better way.
Superficially this may be an attractive argument. In reality, it is indicative of the sorry state of the relationship between history and politics in the present. As such it should act as a wakeup call to historians to start rethinking the role of history in contemporary politics and ideology. There are two problems that Brown’s piece reveals. The first is related to the original paper itself, and the problem of positivist storytelling and historical representation. The second is Brown’s own re framing of history to support the notion that a leftist politics is only possible as a consequence of war. Together, these say much about the ideological context in which historians work today, and against which they must struggle.
The problem of positivist storytelling
To be fair to the authors of the original article, Brown’s interpretation of their article will no doubt come as a surprise. The original paper sets itself rather more modest objectives than Brown’s conclusions suggest, seeking to demonstrate a causal relation between the evolution of ‘complex’ societies and the diffusion of military technologies and warfare. The context for the study is the Afro-Eurasian landmass between 1500 BCE and 1500 CE and the environmental and ecological conditions of that space. The model that the authors deploy incorporates environmental and geographical variations and was relatively successful in reproducing the historically observed patterns of the spread of what the authors term ‘large-scale societies’. Nowhere does the suggestion appear to suggest that this model predicts or explains the potential for a leftist politics in the present, which, after all, is not the question that these researchers have set themselves. Nor is it at all clear that the authors’ would support the claim made by Brown that “states are costly for their inhabitants”.
Given the careful claims to scientificity in the original study, arguably Brown has done a significant disservice to the paper’s findings. It is clear that his reframing of the paper does not accord with the original intent of the authors themselves. However, it is arguably no accident that Brown draws the political conclusions that he does from it. In part the paper itself opens itself up to appropriation through the nature of its claims to historical veracity, which are consistently counter-pointed to mainstream ‘narrative historiography. Throughout, the paper contains a number of remarks about the superiority of mathematical modelling over other forms of historical representation, which are dropped into the text in an uncritical manner, but which suggest that the authors see their mode of representation as distinct from and even superior to) history as such. For instance, paper draws a distinction between their approach and the “traditional method of inquiry that historians use”; a claim that presumes a unity to historical methodology that is empirically dubious, but nonetheless suggestive. Against this “traditional method” the authors pose the idea that mathematical models could be “unambiguously tested against data”, and also used to test “whether alternative hypotheses are equally good at explaining the observed data”.
This claim to be able to unambiguously test data is both the most problematic part of the paper, and arguably exactly what Brown finds useful in his reframing. Here we have a claim to an understanding of a past historical process that offers absolute positivity of knowledge and the prospect of future experimentation. Of course, this positivist enterprise is largely dependent upon the veracity of ‘observed data’, and of the data selected for eventual inclusion in the process of modelling. Interestingly, in the supporting information to the paper, the authors are rather more careful about the claims they make. They point out that they use one-hundred year time slices in their model which misses the “peak of some substantial polities”. Moreover, the data used also leaves out small politics, such as the Greek city states, something the authors’ acknowledge, but surely an intriguing omission in a model claiming to tell us about the evolution of complex societies.
Perhaps more troubling than these, however, is the manner in which the simulation model is tested against historical ‘data’, which they authors tell us come from ‘historical atlases’. Remarkably, given the care taken elsewhere in the paper these atlases are apparently taken as givens. There is no discussion of the problems that may arise from the tendency of historians, archaeologists and geographers to disagree vehemently over the interpretation of past evidence. There is no discussion of the epistemological status of the historical atlas, the process of its composition, how disputes over the geography of past polities are resolved in their composition, at all. In short, this form of historical representation, whilst visual, is subject to all the same narrative ambiguities that the authors are precisely attempt to insure against.
The paper only partially acknowledges this hermeneutic circle, and far too quickly casts aside the problems of relying on ‘data’ which is the result of the precisely the kinds of historical analysis that is elsewhere suggested to be unscientific. This is a serious problem for a process which the authors claim, in a statement which says much about their rather dismissive view of historical method, avoids “cherry-picking examples that support our ideas”. Ultimately, insights along the lines that history is not “just one damned thing after another”, and that, “there are general mechanisms at play in shaping broad patterns of history”, will likely come as no surprise to most historians, especially those responsible for the composition of historical atlases. It never strikes the authors to ask whether they are actually straining to catch a gnat.
Another question that remains unasked is whether a model demonstrating a link between the evolution of complex social development, techniques of war and the environment, is not actually just reproducing the the analytical assumptions of historians. Why do we believe that there are such things as ‘complex’ societies, for example? Why do we think first of war, rather than peace, as a motive factor in their evolution? What determines the legitimacy of the historical question asked in the first place?
Along with Brown’s easy appropriation of their arguments, these unasked questions are indicative of the real problems with this paper. This is not its use of mathematical models, or even the search for causative explanations of history, but rather the paper’s representational naivety. The key question, especially for a paper that omits Greek city-states in a study of the evolution of complex polities, is precisely that of whether mathematical models, stripped as they must be of much of the complexity of social existence or the ambiguities of hermeneutic practice, can ever offer an adequate mode of representation of historical change.
Unfortunately beyond a few chiding remarks at the inadequacies of traditional historiography, it is apparent that the authors have little interest in the detailed problems of representation which have so underpinned historical debate. They do not even acknowledge their positivist progenitors, or the controversies their work brought into focus. There are no references to the problems of past projects that sought to produce causative explanations of the evolution of human civilization based upon natural laws and environmental conditions such as Buckle’s History of Civilisation in England. But, as Brown’s attempt to appropriate their project for a political argument demonstrates, one cannot insulate history from the problem of representation simply through claims to scientificity.
Media framings of history
Brown’s (ab)use of the findings of this paper, however, is quite extraordinary, drawing conclusions in no way justified by either the object of analysis or the results. For instance, the authors of the original paper offer absolutely no 'explanation' of “why Western Europe has welfare states and the US has not”, an historical difference that Brown makes no attempt to empirically verify. Why does Brown make this connection? Similarly, the argument that Bismarck established a welfare state for “war preparation” not only ignores the significant empirical fact of the presence of a strongly organised working-class movement on the development of the German state before 1914, it goes far beyond any of the conclusions of the PNAS paper. Nor is it at all clear, as Brown suggests, that the levels of national ‘solidarity’ experienced during the world wars have any connection to the evolution of complex modern states, which is not a question posed by the paper. Indeed, revisionist histories of both world wars have suggested quite the opposite.
It is also unclear whether Brown’s lesson – that a strong left is impossible without war - doesn’t apply equally, or even more strongly, to a conservative politics for which, of course, the preservation of existing social connectivity and hierarchy is paramount. Similarly, the most rudimentary knowledge of the political left would surely leave one very well aware that internationalism and peace movement have long been core to its historical mission, such that it is hardly ignorant of the great obstacles these pose to the achievement of socialism. Overcoming such obstacles is precisely the heart of its political project. The fact that many reformists have abandoned this left project in no way changes these facts.
Overall, one is left asking why Brown wants to make this paper suggest that war fighting and ‘civilization’ are inherently linked; to the permanent detriment of socialist objectives. This says much about the ideological state of our times, and the ideological abuse of history. Under the influence of neo-liberal ideology we have come to fear the details of history, which constantly expose complexity, unpredictability and the essential openness to possibility of the human experience. The appeal to scientific objectivity, through socio-biology, behaviourism, and the like, has always been partly bound up with the reproduction of the idea that humans are rugged individualists by nature. The idea of reducing solidarity to a few variables in a cost-benefit analysis has arguably become an almost reflexive ideological response in an intellectual culture whose understanding of history and historicity is arguably increasingly impoverished. This political-ideological context does not simply disappear by claiming access to an explanatory model. Arguably such representations simply play into the predominant ideological presuppositions about how we represent the past.
Brown’s article is surely partly influenced by this anti-historical ideological yearning for scientific certainly which the authors of the PNAS paper have set up so neatly in their argumentation. To this he yokes a series of more or less unsupported observations, unrelated to the objectives of the original paper, in the expectation that they will thus acquire some kind of objective historical status. It is an object lesson in how not to use history to think about the present.
The politics of narrative complexity
There is a good reason why historical representation is founded on the recognition of narrative complexity and the irreducibility of complex historical phenomena to straightforward attempts at causative explanation. Radical openness, and dialectical sensitivity, actually helps to resist the political capture of the past. It is a mistake to think that the historian’s ultimate role is to explain the past. Such an attitude is driven by a desire to shut down awkward questions. Rather, the historian works to ensure history remains forever open to the complexity and possibility of being human, to ask questions with history about the present. Both Brown and the authors of the PNAS paper miss the inherent political and ideological character of historical representation. Taking the notion of an experimentally verifiable modelling of past historical processes, they quite ignore the politics of what questions are asked, what data is produced, and how argument is represented. Brown’s article stands as a warning to those seeking to pursue a positivist historical project that they still need to pay attention to the problem of historical representation. Like it or not, if you are doing history, what you have to say and how you say it will ultimately be read politically.
Timothy Cooper is senior lecturer in history at the University of Exeter.