ABSTRACT
This chapter tackles the issue of how software packages are selected. As with many topics in the social sciences, the debate surrounding the choice and purchase of technologies is polarised across a number of incommensurable positions. A major line of argument has been between, on the one hand, technocratic analyses advanced, for example, by Economics, Management and Engineering accounts, where the assumption is that sufficient information is available about the properties of artefacts to enable rational choice to be made, and, on the other, more sociological and critical approaches which emphasise the profound uncertainties surrounding procurement, the consequent contestability of claims about the properties of technology, and the ‘negotiability’ of the criteria used to assess objects. In this latter view, choice of one technology over another is seen to reside not in the objective properties of the artefact as revealed by a formal technical or economic assessment but to be, of necessity, refracted through or, in some accounts, driven by the micro-politics of the organisation, the commitments of the various actors, prevalent rhetorics, fads, etc. (Grint and Woolgar 1997; Neyland and Woolgar 2002). While sympathetic to the critical account, we are dissatisfied with the
relativist outcome and portrayal of technology selection as divorced from formal decision-making criteria, particularly when it seems to us that procurement is subject to powerful, albeit complex, rationales. In contrast, we suggest that the formal assessment criteria we have adopted will guide and transform the technology selection process. Our thinking is influenced by the work of those scholars who have become sensitive to the interrelationship between these two positions, in particular the work of Callon (1998, 1999; Callon and Muniesa 2005) and MacKenzie (1992, 2003, 2006a, 2006b) which has examined the contrasting explanations that economists and sociologists offer for the functioning of economic and financial markets. They both suggest that the gulf that exists between these two viewpoints is unhelpful. MacKenzie (1992) raised the idea that these disciplines offer tools with different kinds of explanatory power. Economic tools, he suggests, are
well honed for assessing the aggregate outcomes of highly regularised behaviour, where there are more or less formal criteria in play. Sociological tools (and here he refers to contemporary actor-centred accounts) are best honed for exploring the particularities of behaviour in their more or less unique historical, geographical and social settings. Callon too focuses on the efficacy of the tools that both disciplines bring to bear, but he takes the discussion further by investigating why one set of explanatory mechanisms appears to be more successful. His conclusions, which have sparked an intense debate in the field of Economic Sociology and beyond, are that certain theories and tools not only describe but also help to create the settings in which they are applied. In other words, certain theoretical constructs appear to more accurately reflect their setting because they are performed in (and in so doing help constitute) that setting. It is the notion of ‘performance’ (and ‘performativity’) which is crucial
here. The term suggests that certain phenomena are, to a substantial degree, brought into existence and sustained through the actual ‘doing of them’ (Callon 1998, 1999; MacKenzie 2006a, 2006b). Callon and MacKenzie have applied this concept in Economic Sociology to explain how markets may be brought into being. Callon suggests that if people are to trade and purchase goods in a ‘market’ (as opposed to any of the other ways the exchange of goods might occur) then the market has to be continually performed. The strength of his work lies in the conceptual framework he introduces to show this performance. For instance, he discusses and confirms the existence of homo economicus, suggesting that economic man does indeed exist but only because ‘he’ is brought into being through a process of ‘disentanglement’ and ‘framing’.1 It is because of disentanglement/framing, says Callon, that actors can make decisions which appear ‘calculated’ and ‘rational’. Callon also introduces a third term, ‘overflowing’, to emphasise the constant need to reframe as new information relevant to the decision comes forward. Petter Holm (2007) describes how, in Callon’s approach, actors and
objects are so thoroughly entangled in ‘sticky cultural contexts’ that these processes of framing and disentanglement are crucial if market actors are to evaluate and calculate the likely results of their decisions: buyer and seller must be constructed as ‘autonomous agencies’ and the object to be sold must be constructed as stable and commodity-like. Importantly, the mechanisms that enable this decontextualisation are not part of ‘human nature’ but have to be actively constructed (ibid.). To put it more in the terms we want to develop in this chapter, various assessment and comparative measures must be defined, constructed and put in place if framing and disentanglement are to take place. Paraphrasing Holm, the more taken for granted, embedded and ‘thing-like’ these measures become, the more effective they will be in untangling objects from their social, cultural and technological contexts, and thus enabling actors to make the calculated decision they desire (ibid.). These are important ideas, which, if they can be applied to the study of decision-making during economic transactions, can also be useful for understanding the choice of one software package over another.
