ANALYTICAL STRATEGY FOR CHAPTER 7-9 11

whereby I compare the daily responsibilities and the way the scientists go about their work with the highly idealised ideas about proper scientific conduct.

‘truth’, the ‘business’ and the ‘public’ (see chapter 6,7,8). The modes are inspired by John Law’s ‘modes of orderings’ (1994) and each ‘mode’ entails a specific (patterned) way of understanding and acting in the world: What is seen as right and wrong, the role of the scientist, the role of science in society and a range of other issues that render meaning to the scientists in a given situation.

The construction of the modes was done by assembling descriptions of actions from the field notes that were similar to each other and described common ways of behaving in specific situations, also called ‘conduct’.¹² Similarly, I assembled justifications from field notes and interviews that were alike. I also (especially from the STIR inventions) assembled statements that compared a way of doing things with a way of justifying them. I did so in order to establish connections between conducts and justifications. I do realise that the connections between

‘conducts’ and ‘justifications’ are a construction. This is because, firstly, actors make sense and thus justifications for their actions in hindsight as, among others, Karl Weick has described (Weick, Sutcliffe, and Obstfeld 2005). In that way, the actors’ connections are a meaningful reconstruction of their ways of behaving and the accompanying motives, but not necessarily identical with what took place in the specific situation – which was probably much more ‘messy’, as John Law would say (Law 2004). Secondly, I, the researcher, also construct the connections between conducts and justifications on the basis of the informants’

descriptions and my theoretical understandings. Thereby, I also influence how these connections are constructed. In a similar way to the informants, I also create meaningful stories about responsibility, which makes sense in relation to the goal of making a dissertation. Given that I maintain (following Law 1994:

109) that both the actors and I draw on well-known general stories about science’s role in society in order to construct a sense of meaning and

responsibility in specific situations, I think it is valuable to describe the connections between conducts and justifications. Not as a ‘correspondence’

description of ‘reality’, but rather as a way of suggesting that how the scientists make sense of their work and how they conceive of the relation between how they act and why they do so is an important clue as to how the scientists adapt to current demands for ‘responsibility’ in science. As John Law writes about his own research aim:

‘[…] one of the points that I tell is that how the Laboratory members tell stories, how they formulate their past, is an important clue to a much more general issue: how it is that they would like to order the organization in a much wider range of circumstances; and how it is that the organization is being performed and embodied in a wide range of circumstances.’ (Law 1994: 19, original emphasis)

In relation to actually constructing and distinguishing the different modes, it was necessary – besides using ‘conduct’ and ‘justification’ – to look at some of the other descriptions that Law gives of the modes in order to get inspiration and guidance. John Law (1994) also describes ‘shared stories’ in the labs, that is, shared ways of thinking about the laboratory’s past, present and future, its boundaries, internal content and external environment, which help the lab to maintain itself (Law 1994: 54). I have not explicitly worked with the concept in the presentation of the analysis, however, during the construction of the modes, I became aware of working with both conducts and justifications that seemed to be ‘shared stories’ – either in both labs or within one of them. I want to describe modes as shared, as something most people in the laboratories will recognise in contrast to more private ideas and positions.

Furthermore, Law also describes ‘ordering effects’, which is in his words actually the only observable parts of the modes themselves, as they are literally the sum of the ‘recursive patterned effects of the networks of the social’ (Law 1994: 15). Law describes that the modes may generate ‘patterns of deletion’;

that is, the modes tend to put some aspects of the social in the background, while others are empowered and therefore very visible (Law 1994: 111). He also states that modes generate and perform

‘…distributions, defining or embodying a characteristic approach to what might, does or should pass from whom to what under what circumstances’ (Law 1994: 111).

In my work on constructing the modes, I also looked out for these kinds of patterned effects in my material. I sorted them according to how some entities, for instance, ‘the scientific profession’, were extremely visible and active in some situations, while being utterly absent in others. And I distinguished between the different ways of constructing knowledge that the scientists told me about; at times, they spoke of public science as a stable institution that disseminates knowledge to a ‘society’ made up of huge institutions such as the

‘state’, ‘industry’ and ‘the public’. Other times, they spoke about how knowledge is sold on a market where they compete with other laboratories.

Again, the task was to group these different stories that I saw in my material, since I treated them as effects of a specific way of looking at the responsibility the scientists have.

These groups of shared stories, ways of doing their work, justifications for their actions and the ordering effects were then combined and distinguished from each other in order to make up each of the modes. All of these elements did not

just fit together nicely like pieces of a puzzle. Therefore, I chose justifications as the most important part of the mode in order to be able to distinguish them from each other. If the justifications for different forms of behaviour or more general effects were the same, then they belonged in the same mode, even though the effects looked very different. If it was not possible for me to construct all elements, that is, conduct, justification, shared stories and ordering effects, then I did not consider the data material consistent enough to support the construction and description of a mode. In that way, the different components also worked as a way of triangulating my analysis: if all the elements were not in place, I refrained from further description and these ‘half-baked’ modes are thus not part of the analysis. In the end, I ended up with three ‘modes of responsibility’, whereby each generated some specific ways of doing the daily work, each argued in different ways for the reasons for this specific responsibility and each had some general consequences for how the labs were organised.