by Sabina Leonelli
ESRC Centre for Genomics in Society (Egenis)
University of Exeter
Web interactions and ICTs enable the production and dissemination of huge masses of data of different types. On the one hand, this leads to the proliferation of new data types, which were not available to social scientists before the advent of Internet and information technology. On the other hand, this also involves the creation of new tools to collect, disseminate and model these data, which are bound to embody specific conceptual framings of the phenomena under investigation (e.g. social entities such as individuals, groups, behaviours).
Within natural science, this situation has long been subject to debate and huge investments have already been made in constructing giant cyberinfrastructures to support e-science research. The use of these resources is hugely affecting how research is carried out as well as how research objects are conceptualised. One example of this is research on model organisms such as fruit-flies or mice. These organisms are now largely conceptualised with the help of ‘community’ databases and so-called ‘virtual organisms’, which enable the analysis of data on those organisms in silico. Thus, virtual representations are shaping experimental practices in vivo (Leonelli and Ankeny, under review). A second example of how e-science is affecting the natural sciences are the emerging classification tools used to collect and organise experimental data. Arguably, the most important among such tools are the so-called ‘ontologies’, which provide classificatory categories for data and phenomena at the same time, and thus shape the ways in which data are interpreted towards the discovery of new phenomena (Leonelli 2010).
The social sciences have been slower than the natural sciences in exploiting the potential of e-science, and therefore the ways in which e-social science is affecting our understanding of social phenomena are not yet as clear. I here wish to highlight four sets of issues that I think should be addressed by anyone engaging in e-social science, and where the traditional approaches proposed by social science methodologies need to be re-examined and revised in order to retain their relevance. As highlighted by Annamaria Carusi and Anne Beaulieu in their introduction to this panel, these issues are at once ethical, epistemic and methodological: they are key concerns for the conceptualisation, pursuit and future use of e-science as a tools for social analysis, and thus deserve to become part of a manifesto of e-social science.
1. A reflexive, expert approach to information technology.
Now more than ever, social scientists need to be extremely aware of how the technology that they use to conduct their analysis works. For instance, consider the software used to classify and disseminate huge masses of data; or the modelling tools used to visualise statistical patterns. Programs used to analyse data online make use of specific assumptions about what constitutes an individual, an intention, a group – and one imperative for e-social science is to make such assumptions as clear and questionable as possible. A consequence of this is the necessity for social scientists to become as familiar and well-informed as possible not only on how e-science tools can be used, but also on how they were/are developed in the first place, and how flexible they are to further manipulation. The question that then arises concerns the expertise that social scientists need to possess. In order to be reflexive about the assumptions embedded in the technologies that they study and use, and their social implications, do social scientists need to reach the same level of technical expertise as the producers of, for instance, specific algorhythms? And/or do they need to devise new modes of collaborative work, involving team-work with IT experts? In other words, does the entry of computer science in social science imply a shift in the type of training, and division of labour, characterising social scientific work? A related question concerns the entry points available to social scientists for investigating the power and use of specific technologies. The early stages of design of a technology are an excellent moment for intervention and collaboration between IT engineers and social scientists, and yet often this is not the stage where the social effects of using a specific technology on a large scale can be observed and studied. How do social scientists study the assumptions embedded in technologies that are already widely implemented and used? Yet another conundrum arising in relation to this issue is the global dimension of e-science. How do local practices of technology development and use intersect with the distributed nature of the social networks involved in these practices? And what does that mean in terms of choosing loci for fieldwork?
2. A renewed attention in documenting the dynamics of social reality across time and space.
Online representations of society can be updated with unprecedented speed, and connected to historical data documenting previous periods. Now more than ever, e-social science has the capacity to study the development of social life by bringing together data in real time with historical data accumulated over the last few decades.
For instance, the possibility to track changes, updates and hyperlinks in online texts constitutes an excellent new resource for e-social science, whose methodological value needs to be explored and whose very existence is tied to the properties of online communication technologies. The fluid nature of online text reflects the ways and temporality in which actors define their own identity and their relations to others. Far from being an obstacle to social science research, the revisable, volatile nature of online texts (whether they are featured in institutional webpages, personal blogs or online news sites) constitutes a powerful empirical source for the study of social change and identity-formation. At the same time, investigating the pace and content of changes to texts online challenges more traditional understandings of social science sources, particularly of qualitative analyses of texts as ‘static’ objects and of technologies as shaping, and at the same time expressing, individual and collective perceptions of the self.
3. The use of virtual reality to challenge the boundaries of social reality.
Virtual, online representations of reality are often used to probe our understanding of the material world as we know it. This is obviously the case with using virtual organisms to test biological and medical understandings of genetics and physiology; and similarly, with using ‘artificial markets’ as a playground for testing ideas about how markets develop and how individuals and groups interact within them. Still, just like in the case of virtual organisms, the main problem arising from these research methods is the problem of inference: what precisely do we infer about reality from these virtual representations? And on what grounds?
4. A revived interest in critically probing alternative conceptions of society as a ‘system’.
When using e-social science to analyse the masses of survey data / consumption data available from companies / government, it is tempting to restrict one’s notion of society to the idea of aggregates of individuals, which can be described through statistical analysis and classification into groups, or to some form of methodological individualism (such as that implicit in many forms of agent-based modelling). It is important, however, to consider how other conceptions of society as an emergent whole, or a ‘system’, can inform e-social science. Just like the life sciences are exploiting the new online resources to explore the idea of organisms as biological systems, and question what should be seen as part of such systems, social scientists could use online data and tools for analysis to question what constitutes the ‘social’: is it an emergent, supra-individual property? Could we meaningfully appeal to a hermeneutic, dialectic understanding of social groups, and what would that imply in terms of data production and collection? One reason why it is worth thinking about retaining a strong notion of what constitutes the ‘social’ is to legitimate the very role of e-social science as separate from and complementary to e-natural science (e.g. cognitive or evolutionary explanations of human behaviour). Some provocative aspects of this legitimisation issue are explored in Steve Fuller’s book on the need of re-imagining the social (Fuller 2006).
Fuller, S (2006) The New Sociological Imagination. Sage.
Leonelli, S. (2010) Packaging Data for Re-Use: Databases in Model Organism Biology. In Howlett, P and Morgan, MS (eds) How Well Do ‘Facts’ Travel. Cambridge University Press.
Leonelli, S. and Ankeny. R.A. (under review) Re-Thinking Organisms: The Epistemic Impact of Databases on Model Organism Biology. Studies in the History and the Philosophy of the Biological and Biomedical Sciences: Part C.