Conclusions and discussion

Discussion

The salience of environmental policy displays rather different trajectories in Terni and Perugia. Its distribution over time follows social movements’ and interest groups’ success rather closely in Terni, while it seems rather independent from other policy areas in Perugia. This emerges from the analysis of both the YouTube and UmbriaPress corpora: Environmental policy in Perugia holds a stable place in public discourse, while its salience seems more issue-oriented in Terni.

Topic co-occurrence adds a significant piece to the puzzle: a larger portion of public discourse around environmental policy in Terni is related to the noxiousness of industrial activity, both on YouTube and in online news articles. The opposite, however, is not true. While the proportions of what we could call topic overlap is somewhat stable in Perugia, they are almost completely one directional in Terni: though articles and videos about the environment trace a connection with industrial activity, articles and videos about heavy industry do not mention environmental policy whatsoever. This asymmetrical relationship is further substantiated by a semantic analysis of the corpus. The overlap is not only identified in terms of term occurrence in the corpora, rather it shows even as a coexistence of terms connected to air quality and industry in the same semantic space.

The emerging picture is that of environmental policy as a highly politicised topic in Terni. This makes its salience fluctuate together with the popularity and success of social movements and interest groups who centre their political activity around climate change or environmental policy. This was especially the case between 2017 - the year in which the relevance of anti-incinerator mobilisations peaked - and 2019, when Fridays for Future came under the spotlight.

These actors’ tendency to connect it to industrial noxiousness is reflected in the media, but industrial activity seems to be discussed without ever taking environmental concerns into account. This feeds back to Vona’s claim that job security is of utmost priority even for those who suffer the most from the adverse effects of toxic pollution (Vona 2019).

Differences in tone between between Terni and Perugia are subtle and nuanced. They appear mostly negligible when analysing news articles as a whole, but become rather interesting if one restricts the analysis to the keywords’ immediate context. This kind of analysis unearths surprisingly high sentiment in Terni around ecology-related keywords, and generally low values in every other one. This might hint at an idealisation of nature as opposed to a noxious urban environment.

Limitations

This study has a series of limitations. First and foremost, the retrieval techniques employed here employed are rudimentary. While quick and computationally cheap, keyword-based techniques are inherently prone to error. Future developments of the present research might benefit from a more “sophisticated” approach to text retrieval, i.e. a classification model trained to classify news articles and press releases based on their content. Second, the data collection process directly influences the results of text analysis. Each different language variety included in a sample of local political communication has a series of idiosyncrasies, especially in terms of tone. What this means is that in order to accurately assign a sentiment value to a chunk of text, it would be necessary to take into account the political alignment and stylistic habits of at least every news outlet. Finally, and most importantly, this study does not advance any explanatory claims. Its role is paving the road for future contributions to define the causal chain linking patterns of loss to the support for environmental policy.

Conclusion

Empirical results highlight an interesting connection between salience and exposure. When faced with a trade-off between job security and environmental concerns, the public’s priorities seem to lie with job security. This is certainly consistent with an interpretation of public discourse around environmental policy centred off patterns of loss, although the way in which it manifests in media is different than expected. Salience (or rather, its absence) appears to be more telling than tone. The most distinctive trait of Terni’s public discourse is its utter disregard of environmental concerns when discussing industrial planning.YouTube videos concerning industrial policies were, in fact, completely void of information on environmental policy. In the same corpus, however, a connection between toxicity and industrial activity can be identified. Additionally, “nature” as an idealised concept has an extremely positive connotation in a (formerly) industrial town like Terni in comparison to the “baseline” case of Perugia. In the context of this study’s proposed framework, this could be interpreted as a way to avoid what is perceived as a zero-sum game. What is left for future research to ascertain is whether it is, in fact, a fear of loss that causes these differences in the connotation of public discourse, and to investigate the subsequent connection between public discourse and electoral results, as well as public opinion formation.

Vona, Francesco. 2019. “Job Losses and Political Acceptability of Climate Policies: Why the ‘Job-Killing’ Argument Is so Persistent and How to Overturn It.” Climate Policy 19 (4): 524–32. https://doi.org/10.1080/14693062.2018.1532871.