The Commodification Gap. Matthias Bernt

the kind of relationship that a particular element has to the concrete totality of which it is a part. In this respect, it is crucial to distinguish between external or contingent relations and internal or necessary relations (Sayer 1992, pp. 89–90). External or contingent relations can be found when two objects can exist without each other. Sayer (ibid.) uses the example of a person and a piece of the earth for this; though their interference can have significant effects, each of the two can exist without the other. By contrast, the relationship between a master and a slave or a tenant and a landlord is internal: both entities form a social relation that ceases to exist without both elements being part of it. Third, it is possible to identify asymmetrical internal relations. These appear when ‘one object in a relation can exist without the other, but not vice versa’ (Sayer 1992, p. 90). Sayer provides the example of the state and council housing; states can exist without council housing, but no council housing can exist without a state.

      Clarifying the type of relationships that link the different elements of an object to each other is the first step towards conceptualising, as it forces the researcher to explain what it is about an element that makes it a necessary part of a relationship. With regard to gentrification, the questions would thus be: What is it about phenomenon A (renovations of dilapidated houses) that enables it to cause phenomenon B (displacement of poor households)? Why and how are renovations done and how does this relate to the likelihood of poor households being displaced? Is there a phenomenon C (say an immigration of middle‐class households) that necessarily goes together with phenomenon A to achieve phenomenon B?

      In this sense, causation is the identification of the actual mechanisms producing the event (as different from identifying patterns of regularity). This, however, leads to new problems as the ‘open systems’ character of the social world leads to situations in which the same causal power can produce different outcomes in different contexts, or different causal mechanisms can work at the same time. As multiple causation is fundamental to social phenomena, it follows that a particular event (e.g. the displacement of low‐income households) can involve several causal mechanisms that may or may not internally be related to each other. Depending on the conditions, the same mechanism can produce different results, or different mechanisms can lead to the same outcome. If one was to relate this approach to the example of gentrification, the research design would become even more complex: not only would a causal relation between A (renovations of dilapidated houses) and B (displacement of poor households) need to be established and the impact of the intervening variable C (immigration of middle‐class households) be explained, but research would need to expand to other contexts in which the impact of C (immigration of middle‐class households) would possibly be neutralised by variable D (say strong welfare support) or intensified by E (say ethnic discrimination). From this, theorising would move backwards and re‐examine the causal relationship between renovation activities and displacement established at an earlier stage of research and reformulate the theory on this basis.

The realist approach towards theory building is, thus, built upon empirical observation and abstraction. It starts with a theory, but moves on to empirical data capable of inducing an alternative conceptualisation, and then moves back to revising the original theory, thus producing a need for new empirical observation and so on and so forth. The actual processing of research is, therefore, highly iterative. It implies a back and forth between inductive and deductive cycles of data collection and explanation in which deduction helps to identify the object of interest and suggests explanations about the mechanisms at play, whereas induction provides data that is to be explained and tested. Both the use of data as an illustration of theoretical arguments and the use of empirical data to falsify a theory without giving any specificities that would allow for alternative conceptualisations are seen as preliminary stages in the development of new knowledge that need to be overcome by the development of new theories.

Design of this Study

      How can these ideas inform a comparative study on gentrification? Without further ado, I draw four basic orientations from the preceding discussion.

      First, I conceptualise gentrification not as an object, but as an abstraction. Advancing its study is hardly possible on the basis of a holistic approach that tries to integrate a broad variety of empirically observable phenomena all at once and indiscriminately apply the whole range of available, and to some part contradictory, explanations for gentrification. Learning from realist approaches, it is necessary to disaggregate the various explanations and start with one established theory whose validity can be examined in different contexts. Only when this is done, does it become possible to carefully examine the limits of this theory of gentrification and to study how other factors impact on the causal relationship between different phenomena. Meaningfully studying gentrification, thus, demands a clear point of reference towards an existing theory, as well as a focus on the phenomena to be observed.

      Second, there is a close relationship between studying how an established concept works in different contexts and theoretical advancement. The task of an empirical study is, thereby, not to verify or falsify an established theory, but to work out alternative causations. In this context, it becomes crucial to empirically distinguish between necessary and contingent relations. Do different structures (say welfare states) influence a mechanism in a way that changes the causal relationship between two phenomena (say renovations and displacement)? Do contextual factors lead to differences in kind or in degree? Can the same outcomes (e.g. displacement) be explained by different mechanisms in different environments?

Third, neither universalising nor individualising modes of comparisons suffice to promote this process of iterative theory development. The problem, as I have pointed out with reference to Tilly, is not that particularising or universalising comparisons were misleading, as such, but that they are good in doing different things. While universalising comparisons work well to support a theory, individualising comparisons do exactly the opposite, demonstrating where a theory has reached its limits. The benefits and drawbacks of the two types of comparison are laterally inverted. What is needed is not a decision for the Scylla of universally valid categorisations or the Charybdis of thick descriptions of singular cases, but a way between the two that engages established theories with new empirical material that aims to build better explanations. In this sense, comparisons can be generative for a new round of theorising – but only when they are built upon a reference to existing theories and a commitment towards abstraction. Working with concepts is of key importance here.

Fourth, the relationship of theorising and empirical research in the course of investigation can be described as ‘ascending from the abstract to the concrete’, as Karl Marx famously put it in Grundrisse (1971 [1894). The abstract defines the object of empirical research that then leads to understanding the points where the theory stops explaining its particular objects and new theories must be integrated or found. This leads to an iterative back and forth movement between the empirical observation, abstraction and conceptualisation that is illustrated in Figure 1.2.

      The study starts with a simple, abstract concept that seems to apply to all sorts of gentrifications at all times; moves on to the examination of actually existing forms of this abstraction; examines the interrelation of this abstraction with different concrete sociospatial formations; and then moves back to the abstract and examines the dialectic developed at a systemic level so far.

      How can these recommendations be taken on board? In essence, restriction is demanded both with regard to the issues studied and the theories examined. In this respect, the following points form critical junctures for this book.

       Definition of gentrification: I follow Clark’s (2005) definition of gentrification and count all processes as gentrification that ‘involv[es] a change in the population of land users such that the new users are of higher socio‐economic status than the previous users, together with an associated change in the built environment through a reinvestment of fixed capital’ (Clark 2005, p. 258). With this, I position myself within close range to what has become known as ‘supply‐side theories’ in the world of gentrification researchers.³