Defining the precautionary principle: an empirical analysis of elite discourse

Abstract

The precautionary principle (PP) has gained influence in environmental politics as a ‘policy principle’ – an idea that can spur policy change. Yet, exact definitions of the PP remain elusive, making evaluation of its actual political influence difficult. Given the controversy over the PP's meaning and policy utility, broader empirical analysis of its public formulations is overdue. Elite discourse on the PP is analysed in the search for a dominant formulation among 238 articles in a variety of disciplines. The modal PP formulation is found to be a mix of stronger and weaker elements, broadly resembling Principle 15 of the 1992 Rio Declaration. The data suggest that the principle has become weaker over time, and that its critics formulate it more strongly than proponents. Contrary to some assertions, however, American and European authors do not differ significantly in their interpretations of the PP.

Keywords:

• precautionary principle (pp)
• Rio Declaration on Environment and Development
• wingspread statement
• risk analysis
• risk management
• discourse analysis

Acknowledgements

The authors are grateful for research funding for this project by National Science Foundation grant number CBET-0523393. The authors thank Inez Hua, Thomas Seager, Andrea Olive, Jay McCann, and the anonymous reviewers at Environmental Politics for their helpful suggestions and comments. Any errors or omissions that remain are, of course, our own.

Notes

1. Both authors coded the initial set of approximately 20 articles to pilot test and refine the typology and to assure consistent interpretation of each category. When we confirmed that the typology required no additional interpretive categories and that our inter-coder agreement was at a rate of 80% or higher, one author independently coded the remaining articles over several months, consulting with the other author as needed.

2. We weigh each category equally in calculating HRS because no other weighting method is theoretically justified. In addition, we weigh each article equally in calculating average HRS scores because there is no means of comparing the political influence of each author's ideas – potential weights like journal impact factor or citation counts are ineffective at assessing an article's overall ‘impact’, especially with respect to journals from a wide range of disciplines including law reviews, where impact factors are not widely used. Nor would academic citations or impact factors necessarily represent an author's influence in the policy realm.

3. Shifting the division between ‘hazard’ and ‘risk’ backwards (counting 1999 as part of ‘risk’ time, for instance) only strengthens the difference between the independent samples, as would be expected. Moving the break to 2001, by contrast, renders the difference statistically insignificant, which is consistent with our supposition that the year 2000 appears to be the critical tipping point in our data.