Monday, 23 March 2015


Pacifying Babel's Tower: Lessons for comparative urbanism


Michiel van Meeteren  (Ghent University)

Ate Poorthuis (University of Kentucky)

Ben Derudder (Ghent University)

Frank Witlox (Ghent University)

In our paper 'Pacifying Babel's Tower: A scientometric analysis of polycentricity in urban research', we analyze the origins of the fuzziness of the concept of polycentricity in urban studies. While seemingly little more than an adjective to describe the multi-cored character of settlements, the term 'polycentricity' disguises the much thornier and fundamental question of defining the notion of ‘city’ itself. This masking of a different debate suggests that, as a community of urban scholars, we have been talking past each other for a long time – hence our reference to Babel's Tower. We assume that we are discussing the same object, but are in fact often not. The paper extensively draws on the work of political scientist Giovanni Sartori (and his interlocutors) from which the notions of 'conceptual stretching', and 'Babel's Tower' are derived[1]. While conducting this research, we increasingly became convinced that Sartori's oeuvre might be able to provide important resolutions to some of the methodological issues plaguing (comparative) urban studies and human geography more widely (an insight that has proven productive elsewhere). In this post, as an extension of our paper, we explore the broader implications of Babylonian confusions caused by conceptual stretching for comparative research in urban studies.



The Tower of Babel (1563) by Pieter Breugel the Elder, Illustration used for the cover of Collier and Gerring's (2009) appreciation of the oeuvre of Giovanni Sartori. (Source: Wikimedia Commons, Accessed 23/03/2015)


Polycentricity is a concept that fits comfortably in the methodological toolbox of comparative urbanism. It provides a seemingly coherent yardstick to compare urban regions. Of course, in a strict ontological sense, every city in the world is unique. However, finding similarities and differences between places allows us to abstract from these unique and concrete realities. Based on these abstractions, we then define wider 'urban processes' that serve to structure the debates in our field of research. The crucial point in these abstractions is to avoid what Sartori calls 'miscomparing': drawing undue conclusions regarding similarity or difference.


Miscomparing can either take the form of a wrong 'case definition' (the answer to the question 'what is this a case of?', which leads to the proverbial comparing of apples and oranges, or a wrong 'case scoring' (assessing how this case compares to others, a kind of measurement error). When a specific miscomparison is routinely made (as has happened in the debate on urban polycentricity), it is solidified in the literature and may consequently become 'paradigmatic'. In this way, reminiscent of Kuhn's description of scientific revolutions, communities of researchers may collectively and unknowingly hold on to a theoretical proposition that ought to be revised given the available empirical evidence.


By way of example, let us consider an 'institutionalized miscomparison' identified in our paper that can potentially influence comparative urban studies in a detrimental way. In the polycentricity debate, differences between American and European contexts have commonly been overstated, despite the fact that the processes behind polycentricity are very similar across contexts. Because the processes behind polycentricity overlay distinct historical trajectories, incorrect assessments of difference between Europe and the United States have been made when studying the concrete situation 'on the ground'. At the correct level of abstraction (in this instance this involves comparing changing agglomeration-economy regimes across contexts), these cases should have been labeled as the 'same', whereas an empiricist assessment has led to a conclusion of 'difference'.


In general terms, this example highlights once again how the adopted theoretical and methodological perspective determines the results of comparative research. Miscomparing can result in accepting 'wrong' theory as 'valid', similar to type II errors in statistics, a concern that is at the heart of postcolonial critiques in urban studies, for example in the interventions of Jennifer Robinson or Ananya Roy. But it can similarly lead to type I errors: rejecting 'right' theory as 'invalid'. The polycentricity debate is one such example; another possibly arises if one takes too literally Roy’s sweeping suggestion that 'EuroAmerican theories' of the city cannot be used to say something useful about cities in Asia.


Therefore the most important conclusion of our study, which transcends the more narrow confines of the polycentricity debate, is that (as Sartori would phrase it) the proper application of the ‘ladder of abstraction’ is of paramount importance. We should always ask: are we comparing the same thing across contexts? If not, then we end up comparing apples to oranges, leading to erroneous conclusions about similarity and difference. This is remedied by abstraction, which results in the equally fascinating activity of comparing fruits.

[1] A wonderful introduction to Sartori's work on comparitivism is Collier & Gerring's 2009 edited volume Concepts and Method in Social Science.