Study title

Parliaments are key institutions of democracy. The documents that parliaments produce - including plenary protocols, legislative bills, and ultimately adopted laws - thus contain invaluable information for understanding how democratic governance works while advances in natural language processing offer an ever increasing potential to extract and analyse this information systematically. Yet, despite this significance of analysing parliamentary documents systematically, accessing them often involves very notable hurdles in applied projects (Sebők et al., 2025). To conduct computational text analysis effectively and to compare across different legislatures, researchers require comprehensive, consistent, and machine-readable text corpora. ParlLawSpeech (PLS) contributes to and extends recent efforts - e.g. the CLARIN project (De Jong et al., 2022), the Comparative Agendas Project (Baumgartner et al., 2019), ParlSpeech (Rauh & Schwalbach, 2020), or Parl-EE (Sylvester, et al., 2022) - to make such corpora available to the research community. Besides extending the coverage of readily available parliamentary text corpora, PLS innovates in two respects. First, it is not constrained to only one type of parliamentary document but rather provides full texts for speeches, bills, and laws. Second, it offers data linkage across these different types of parliamentary texts in a way that is typically not even offered by institutional repositories and data providers (Kiss & Sebők, 2022). These two innovations allow users to map and to combine different parliamentary outputs, thereby enabling them to address research questions spanning the full process of parliamentary decision-making from draft documents, over plenary debates, to the contents of the finally adopted rules. In sum, PLS offers machine-readable full-text vectors of parliamentary speeches, bills, and laws, as well as relevant metadata for seven European countries (Austria, Czech Republic, Croatia, Denmark, Germany,...