Project Description

Initial Situation

It can be argued that the accident at the Japanese nuclear plant Fukushima Daiichi in March 2011 turned into a global watershed for domestic atomic energy policies and public attitudes towards the use of nuclear power and the use of renewable energy. While in Germany and Switzerland the so-­called “Fukushima effect” has led to a phase-­out in the year of the incident, in Japan itself nuclear energy still remains a core element of the domestic energy mix.

Various commentators have argued that it was the largely pro-­nuclear mass media in Japan, in combination with a change of government in 2012, who are to blame for this course. Nevertheless, at the same time, Japan has also witnessed some of the largest protest movements against nuclear power since the 1960s. It is generally accepted that social media, by providing platforms and infrastructure, have played a key role in these recent forms of contentious politics. Not only has the media system itself globalized, many of the most recent protest movements have evolved into transnational protest movements as well.

Both developments are key elements of the emergence of a phenomenon we call the “algorithmic” and “transnational” public sphere. However, despite the growing prevalence of this new form of mediated publicness, particularly the inner architecture and mediating mechanics of social media remain mostly unknown. Research is hindered not only by the inaccessibility of the algorithms underlying these platforms, but also by the sheer amount of primary data, which is hard to analyze with traditional methods such as quantitative content analysis. Moreover, the fact that one is dealing with a transnational phenomenon, i.e. one crossing borders of culture and language, complicates this situation even further.

Thus, classical content analysis is faced with a range of methodological problems (language barrier, cultural differences between annotated categories, text duplication, etc.) and cannot achieve a valid comparative analysis. Existing studies into single aspects of the transnational algorithmic public sphere either analyze the impact of “Fukushima” on the attitudes and public opinion across different media in a single national context, or from a transnational perspective but only with regard to a single medium (usually newspapers), applying manual or computer-­assisted methods to data sets from a relatively short time period (usually from 4 weeks to at most 12 months after March 2011).

Objectives of our Project

Accordingly, this project pursues both thematic and methodological goals. On a theoretico­-methodological meta-­level, we aim to develop a general method that mediates between effective but often shallow and over-­simplifying big-­data analysis and visualization, and insightful manual studies that are highly work-intensive and thus limited to small-­scale data sets. We thus develop a toolkit and use it to perform a comparative analysis of the discourse on nuclear and renewable energy in Japan and Germany in the mass and social media in the years 2011–2013.

Our innovative methodology strives for a deep integration of hermeneutic, computational and corpus-­linguistic approaches. On a thematic level, we exploit the new methods to analyze and visualize the intra-media process of dissemination of discourses regarding nuclear energy in social and mass media, and the inter-­media reciprocities between social media and the mass media; both on a temporal (i.e. Mar 2011–Mar 2013) and a transnational axis (i.e. Germany and Japan) – thereby shedding light on the emergence of the transnational algorithmic public sphere.

Against the background of a lack of cross-­media (i.e. social media and mass media) research on the transnational public sphere and the lack of effective methods for studying this phenomenon, we aim at developing an integrated intercultural-­analytic and computer-­aided approach, taking the shift in public opinion and attitudes towards nuclear and renewable energy as a case study. We consider the formation of a global discourse regarding nuclear energy in the aftermath of the Fukushima Daiichi disaster to be a prototypical example of a transnational algorithmic public sphere, because it emerged from intra-­ and inter-­media connections between national “manifest” public spheres as represented in the mass media and the “latent” semi-­public sphere of social media on multiple levels.

Methodological Vision

The methodological vision of the EFE project is to close the hermeneutic circle in the digital humanities and computational social sciences by integrating qualitative and quantitative approaches. Achieving this goal depends on three major innovations:

• to replace the high-­level abstractions and appealing, but non-­transparent visualizations of big data approaches with quantitative analyses that retain a clear link to the primary data and are directly amenable to human interpretation;
• to replace the unidirectional flow of information from quantitative analysis to qualitative-­hermeneutic interpretation to a final visualization with an iterative process in which all three aspects are genuinely integrated and on equal footing
• to achieve a precise understanding and awareness of the limitations of semi-­automatic analyses by comparing their results with an in-­depth manual exploration of targeted samples, and to develop new computer-­assisted techniques that achieve a better approximation of fine-­grained manual annotation.

Thus, we aim at studying the mechanisms leading to the emergence of a transnationally connected algorithmic public sphere, namely how the topics and corresponding attitudes and opinions constituting a discourse spread via slogans/memes and other linguistic realizations across different media and national borders.