Identifying stance in legislative discourse: a corpus-driven study of data protection laws

The increasing importance of data in the digital world has heightened concerns about data protection and privacy rights. Lawmakers globally have responded by drafting and refining data protection laws to mitigate risks associated with data misuse and breaches. This study focuses on the stance expressed in data protection laws of the United States, the European Union, and China – three jurisdictions with significant global influence. The EU's early adoption of data protection rights, the US's extensive experience with data collection and regulation, and China's rapid development of its legal framework make a comparative analysis particularly insightful. The study aims to address three main research questions: (1) What are the divergences and convergences of stance expressions between the data protection laws of these three jurisdictions? (2) What is the overall stance orientation of data protection laws across the three jurisdictions? (3) What are the underlying rationales for the choices and representations of stance among the three jurisdictions? The study aims to contribute theoretically by filling a gap in the literature on stance in legislative discourse and proposing a specialized research model. Practically, it seeks to guide individuals in understanding their data protection rights and help legislators convey their intentions more effectively.

Existing research on stance primarily focuses on media and political discourse, where explicit stance expressions are more easily identified. While stance analysis in legal discourse is less abundant, previous studies have explored aspects like stance construction in judicial opinions, courtroom discourse, and specific legal areas. However, research on stance within legislative discourse, particularly concerning data protection, remains limited. This study expands upon previous research by extending the scope to include China, focusing on stance rather than terminological variation or ideology, and providing a more comprehensive linguistic exploration of data protection laws.

The study employs a corpus-driven approach, using three self-compiled corpora of data protection laws from the US, EU, and China. The corpora contain a total of 169,167 tokens, with the US corpus being the largest (92,812 tokens), followed by the EU (58,902 tokens) and China (17,453 tokens). The US corpus includes federal and state-level legislation, while the EU and China corpora contain key national-level laws. Hyland's (2005, 2018) stance and engagement framework is used for analysis, focusing on attitude markers, hedges, boosters, and self-mentions. The analysis is conducted in three steps: (1) automated search and manual verification of stance markers using Lancsbox; (2) analysis of stance dimensions (evidentiality, affect, presence) based on Hyland's framework; (3) concordance analysis to examine contextual information. A mixed-methods approach, combining quantitative frequency data and qualitative concordance analysis, is used.

The analysis reveals significant differences in the proportion and types of stance markers across the three corpora. The EU corpus shows the highest proportion of stance items (28%), followed by the US (17%) and China (14%). Hedging devices are the most prevalent across all corpora, reflecting the careful and nuanced nature of data protection legislation. Self-mentions are remarkably low, particularly absent in the EU and China corpora, indicating a focus on objectivity. The analysis of hedging devices reveals that 'may' and 'should' are the most frequent, with variations in their usage reflecting differences in the emphasis on permissions and obligations across jurisdictions. Boosters, expressing conviction and assertion, appear with slightly higher frequencies in the US and China corpora compared to the EU, contrasting with the higher frequency of hedges in the EU. The analysis of attitude markers highlights the prevalence of 'appropriate' across all corpora, reflecting a cautious and prudent stance. Important and essential, conveying positive meanings, show corpus-specific collocations reflecting the prioritization of data security (China) and consumer rights (US) and public interest (EU). Self-mentions, though rare, are present in the US corpus, mainly in the form of references to 'our society' or 'our personal information', functioning to foster a sense of shared values and concerns.

The findings indicate two overarching stance orientations in data protection laws: modesty (prevalence of hedging and scarcity of self-mentions) and sufficient discursive space for judicial interpretation. The differences in stance practices across jurisdictions reflect distinct legislative values and public ideologies. The EU's high frequency of hedges reflects its emphasis on precision and legal reasoning, driven by its pioneering role in data protection. The US corpus, while not as high in overall stance items, presents a balance between national security concerns and data privacy rights, reflecting its market-economy focus. China's lower proportion of stance items might reflect a strategic approach of neutrality, balancing individual rights with the development of a data-driven economy. The study suggests that the overall representation of stance aligns with the performance of data protection legislation, showcasing a preference for covert stance expressions to achieve an overtly neutral appearance.

This study contributes to the understanding of stance in legislative discourse by revealing the unique patterns of stance construction in data protection laws. It highlights the importance of legal constructiveness in emerging legal fields, the prevalence of covert stance markers in legislative discourse, and proposes a refined model of stance in law. Future research could explore the judicial domain to further elucidate the interplay between legislative and judicial stance in data protection.

The study's limitations include the focus on a limited number of jurisdictions and the potential for bias in the selection of legislative texts. The size of the Chinese corpus, relatively smaller than the US and EU corpora, could also influence the findings. Future studies could address these limitations by expanding the corpus to include more jurisdictions and legislative texts.