4. Context-Sensitive Mapping of Rule of Law Narratives. Sources and Methods

To explore parliamentarians’ narratives, we gathered speeches from plenary debates² stored in digital parliamentary archives. In contrast to many studies of the rule of law, which often focus on a narrow range of sources, mostly government statements, our final collection comprises written verbatim records of hundreds of plenary speeches. Our approach also differs from typical policy analysis using the Narrative Policy Framework (Smith-Walter and Jones 2020). These studies often focus on a particular policy field and debates on issues originating in that field. The rule of law, in contrast, is a cross-sectoral issue. It can be addressed in debates on diverse topics and policy fields. As a result, statements related to the rule of law do not necessarily follow the logic of actors addressing a certain speech setting, plot, characters and favourite policy solutions, a pattern observed in the Narrative Policy Framework, but focus on various issues. To understand how actors in parliaments address and narrate the rule of law in practice, we need to take into account this reality by covering a broader set of debates.

When choosing the relevant sources, we have sought to provide a valid, that is case-sensitive, basis for cross-country comparisons. This necessitated case-specific strategies. Paying due attention to the context of our cases meant, for example, adjusting the periods covered by our analysis in two cases. In general, the period of study starts in the year following the beginning of the transition to democracy and the rule of law, after the Soviet Union had relinquished its hegemonic power in the macro-region, allowing for independent political rhetoric and action. It was only after 1989 that the new democracies in Central and Eastern European countries removed “communist ‘pseudoparliamentary’ constraints” (Ilie 2015, p. 6). Accordingly, our body of documents covers the period from 1990 to the end of 2021 for Poland, Hungary and Romania. As Czechia and Slovakia emerged from the split of Czechoslovakia into two independent states in 1993, it was appropriate to analyse each republic’s parliamentary discourse from the moment of its newly defined statehood while also considering the constitution-making processes that prepared its founding in late 1992.

We also had to be aware of the different structures of the parliaments, which are unicameral in the case of Hungary and bicameral in the other four cases. For Poland and Czechia, we did not include speeches from the second chambers because bicameralism in both countries is asymmetrical. Essential legislative and oversight powers are vested in the first chamber (Andrews 2014).³ In Romania, in contrast, the Camera Deputaților and the Senate fulfil almost the same tasks (although this has changed slightly since 2003 (Szabó and Küpper 2021, p. 90)) and are elected similarly, with members of the lower chamber later often elected to the upper chamber. In consequence, debates in both chambers fulfil an equivalent role in the legislative process, and the speech contexts of the deputies and the senators are rather similar. Therefore, we included speeches from both parliamentary chambers.

All the parliamentary debates examined deal with the rule of law in one way or another. However, political debates usually focus on a particular legal or political issue and less on theoretical problems or concepts. Moreover, “parliamentary discourse is audience-oriented”, with MPs speaking “in front of a wide (present and virtual) audience” (Ilie 2015, p. 13). In their interventions, speakers may refer to a specific occasion or audience, which does not necessarily require the use of elaborated terms for a detailed rule of law discussion. Nevertheless, they can still refer to established analytical concepts of the rule of law, such as judicial independence or the principles of legality and non-retroactivity. In doing so, individuals may associate different meanings with the rule of law, depending on their “standpoints and representations of reality” and their belonging to different parties or camps (Ilie 2015, p. 7). We aimed to include such implicit references and different understandings when analysing narratives on the rule of law. For this reason, we paid particular attention to identifying speech acts that were relevant to our analysis.

We used two strategies to select the material for our analysis (Fig. 4.1). First, we collected debates in which MPs explicitly referred to the rule of law. Second, considering that political actors may talk about the rule of law and related issues without explicitly referring to the term, we collected debates on key constitutional provisions and legislative proposals regulating the rule of law. These laws institutionalise the rule of law or its elements; thus, they represent key legislation related to the subject of our study (even if the speakers do not mention the rule of law explicitly).

Bild vergrößern

Debates with direct mentions of the rule of law. To identify parliamentary speeches with explicit references, we used the search engines available on official parliamentary websites and the official digitalised full-text transcripts of parliamentary debates. We collected all debates on a specific item on the day’s agenda in which speakers directly used the term ‘rule of law’ or its grammatical variations (e.g. all declinations) at least three times. For Slovakia, for example, this meant searching for nine variations of the term ‘právny štát’. We additionally considered relevant derived terms such as ‘materiálnoprávny štát’, ‘formálnoprávny štát’, ‘neprávny štát’, ‘právny charakter štátu’ and ‘právny a demokratický štát’. In the following, we also refer to such grammatical and semantical variations when we write about ‘mentions of the rule of law’.

In some cases, depending on how parliamentary debates are archived, the search engines allowed us to identify debates on a specific item on the day’s agenda with at least three mentions of specific terms. If such a procedure was impossible because the transcribed debates were stored for an entire sitting day or even an entire multi-day plenary session without a machine-readable internal structure, we first retrieved all potentially relevant documents and then extracted relevant passages from them by individual searches.

In Poland, the digital parliamentary archives did not allow the identifying of debates on a specific item of the agenda with at least three mentions of specific terms. Here, we relied on a list from the team of researchers from the Polish Academy of Sciences, who prepared a corpus of all Sejm debates for quantitative text analysis, the Korpus Dyskursu Parlamentarnego.⁴ As it revealed that there were considerably more direct mentions than in other countries, we decided to include a selection in our corpus of sources. For Romania, the digital archive of the Camera Deputaţilor only covers the period from 1996 to 2021. Therefore, it was beyond our capabilities to systematically include speeches with three references to the ‘stat de drept’ for the period of 1990–1995⁵ (for further information, see Anders et al. 2024).

Debates on key legislation. To collect debates with implicit references to the rule of law or elements of it in a context-sensitive way, we started by reviewing laws and plenary debates, identified debates on rule of law issues, familiarised ourselves with these debates, and derived a list of the typical rule-of-law-related topics in the five countries. We then supplemented this list based on information from various background interviews⁶ and a review of indicators from common rule of law indices (Freedom House, V-Dem, Democracy Barometer and the WJP Rule of Law Index).⁷ Guided by this list, we identified key constitutional provisions, constitutional amendments, laws and amendments related to the (institutionalisation of the) rule of law or its elements and identified the respective bills and parliamentary debates (i.e. all readings) for our period of analysis.⁸

After reviewing and comparing the lists of debates on key legislation in our five countries, we selected the debates that touched upon several key aspects of the rule of law. These were debates on constitutions, constitutional laws and laws related to fundamental rights (including anti-discrimination and minority rights), constitutional laws and laws on the appointment of judges, the competences and procedures of the courts of the different branches of the judiciary (constitutional court, administrative court, ordinary courts), the public prosecutor’s office and judicial councils, and constitutional laws and laws on corruption as well as on lustration. We also included debates on laws on restitution, lay judges and the ombudsperson. In this way, we sought to capture the broadest possible thematic perception of the rule of law. In order to be able to trace narratives over the whole period of analysis, we aimed to include laws from each legislative term, with an optimal number of one law per year on average. We analysed between 25 and 33 debates on key legislation in each country⁹ and usually focused on the first reading (see Anders et al. 2024 for an overview).

The process described resulted in a collection of plenary debates selected using both inductive and deductive logic. The next step was to identify the relevant passages/speech acts within these debates (see Fig. 4.1). We did this by examining the explicit use of the term ‘rule of law’ and other terms that might be used synonymously. At the same time, we searched for implicit references by looking at the context of the statements (we examined, for instance, whether the preceding paragraphs mentioned the rule of law) and by checking for references to elements of the rule of law. To prevent individual researchers from relying on their personal understanding of the elements of the rule of law, we jointly developed and inductively compiled a non-exhaustive list of examples of elements or aspects of the rule of law. In addition, our research team regularly exchanged information on the issues related to the rule of law and its elements debated in the five countries. The idea was to gain a common understanding of the multifaceted debates in the five parliaments and to ensure that we all searched for relevant passages in a similar way.

Having identified the relevant paragraphs, we systematised all themes related to the rule of law within these paragraphs. For this purpose, we developed thematic categories to mark the issues addressed in a statement. Drawing on conceptual works on the rule of law (Chap. 2), but also on literature about our region with its distinctive features of transition to democracy (Chap. 3), we first developed provisional categories, systemising all the aspects of or related to the rule of law found in the relevant paragraphs. For example, we created the category ‘Judiciary’ for all rule-of-law-related statements on the judiciary or the broad category ‘RoL’ for all statements containing an explicit or implicit definition of the rule of law as well as statements concerning its elements, function and purpose. More context-related categories were ‘Lustration’ or ‘Transition’ for statements referring to rule of law issues with post-1989 personnel checks or problems of democratisation and economic liberalisation.

In a pilot phase, we tested whether the provisional categories captured all the rule-of-law-related themes in the sources and supplemented the categories inductively where necessary. This led us, for example, to include the additional category ‘value’ to capture statements about the rule of law as a value and a precious good worth protecting.¹⁰ We also created technical categories to capture aspects such as the occasion of the speech acts, and speaker categories to capture the party affiliation of speakers and whether they were members of the executive.

To ensure a common understanding of which passages of the collected speeches were relevant and how they should be categorised, we regularly tested the consistency of the categorisation across all five cases and team members. We also discussed the meaning and further specified our definitions of the categories to guarantee that they were applied as consistently as possible across our five countries (see Anders et al. 2024 for an overview of the categories).

Again, we had to react to country specifics. For Romania and Czechia, we went through all selected parliamentary debates with at least three direct mentions of the rule of law. For Hungary, Poland and Slovakia, where the term was used much more frequently, this was neither possible nor desirable for the sake of comparability. Thus, in these cases, we categorised only parts of the selected debates.¹¹

The result of this process was a selection of parliamentary debates (chosen for their direct mentions of or implicit references to the rule of law) with all the relevant passages categorised by thematic categories. The final step was to decide which of these categorised passages to include in our qualitative content analysis. For the debates with direct mentions of the rule of law, we chose a more technical criterion: for all countries, we decided to code speeches by MPs in which the term ‘rule of law’ was explicitly mentioned at least once. Within these speeches, we categorised each paragraph explicitly or implicitly related to the rule of law.

For the debates on key legislation, this approach made little sense because these debates by definition revolved around rule-of-law-related constitutional provisions and their amendments, as well as laws and amendments related to rule of law institutions. Using the approach used for the debates with direct mentions of the rule of law would technically have meant that we would have had to consider every paragraph as relevant to our analysis. Therefore, we focused on the most relevant parts of the speeches. We considered parts from speeches in debates on key legislation as relevant only if two conditions were met: (1) if they contained statements on the definition, functions or properties/elements of the rule of law and thus could be assigned to the category ‘RoL’, and (2) if at least two additional content categories could be assigned (see Anders et al. 2024).

Overall, the process served two goals. First, it provided a selection of relevant sources and prepared them for further analysis by identifying the relevant passages. For Slovakia, for instance, we identified 1573 segments. Second, this procedure allowed the team to develop a shared understanding of the various issues related to the parliamentary discourses around the rule of law in each country under study and thus laid the foundation for our qualitative content analysis.

Because of our time-consuming identification and preparation of the relevant documents and passages, the qualitative analysis of the narratives in the five countries took less time. Nevertheless, this stage of the analysis was methodologically demanding. As outlined above, narratives are neither used systematically nor presented in a single speech act but taken up by various speakers in different contexts. From a methodological point of view, therefore, the crucial question was how “to unravel the form and content of a narrative, especially in circumstances where the narrative might be contained in multiple (…) documents” (Prior 2020, p. 550).

To explore and map the rule of law narratives, we analysed all the segments we had identified as relevant. We looked systematically for recurring patterns (typical definitions, justifications, argumentative patterns) that cut across the individual speech acts. To do so in a context-sensitive way, we conducted a qualitative content analysis, a text-based method described as a “qualitative data reduction and sense-making effort that (…) attempts to identify core consistencies and meanings” (Patton 2002, p. 453). It is particularly useful for the qualitative analysis of a large number of parliamentary debates because it combines two features. On the one hand, it allows for the systematic analysis of selected aspects of large amounts of text. On the other hand, it leaves room for the qualitative interpretation of both the explicit and the deeper meanings of the text, i.e. the (often context-dependent) implicit content (Devi Prasad 2019; Schreier 2012).

Our analysis is rooted in the interpretive strand of qualitative content analysis, which sees the inductive and contextual identification and interpretation of implicit meanings, themes and narratives as the main aim of the method. In this logic, the credibility of findings is guaranteed through ‘thick description’ rather than intersubjective replicability, as would be the case with quantitatively oriented content analyses (Devi Prasad 2019; Tracy 2010, p. 843).

The central device for structuring the empirical material in every content analysis is the coding scheme, which provides the basis for the rule-guided analysis of textual sources. It contains the main codes that represent the aspects one wants to focus the analysis on and the subcodes that help to specify what is said about the main codes (Schreier 2012, p. 60).¹² As laid out in Chap. 1, we were particularly interested in (1) the meaning and purpose that speakers ascribe to the rule of law and (2) what they say about its constitutive elements, (3) how speakers relate the rule of law and elements of it to democracy, (4) what they see as the central source of legitimacy of the rule of law, (5) how they talk about rule-of-law-related rights and (6) what challenges to the rule of law they identify. Accordingly, our coding scheme contained six main codes (‘Purpose’, ‘Elements’, ‘Relation to democracy’, ‘Legitimation’, ‘Rights’ and ‘Challenges’) and several subcodes. They served to systematise the narratives.

To develop subcodes, we proceeded inductively. Based on our sources, we selected all the categorised paragraphs from which we expected statements related to the six main codes. Starting with one country, we then read all these paragraphs and wrote down keywords to capture the main themes or arguments within these paragraphs as well as the aspects that surprised us, disturbed us or intruded (Sunstein and Chiser-Strater 2012, p. 115, cited in Saldaña 2016, p. 23). Next, we jointly grouped the keywords thematically to generate provisional subcodes that captured the different narratives within the empirical material. With these ideas in mind, we then read the paragraphs for the other countries, again captured their content through keywords and, based on that, revised and refined the provisional subcodes.

For two of the main codes (‘Purpose’, ‘Legitimation’), our inductive procedure yielded a handful of subcodes that directly captured the narratives’ essence. With regard to the main code ‘Relationship to democracy’, the material could be covered by two subcodes—one reflecting a liberal understanding of democracy and the other expressing a majoritarian understanding of democracy. For the main codes ‘Rights’ and ‘Challenges’, the statements made were far more multifaceted and diverse. Here, we used the subcodes to systematise the themes addressed by speakers, and the narratives were identified later through qualitative interpretation of the text passages that fell into a particular subcode. We read the passages, looked for typical definitions, justifications and argumentative patterns and then grouped the empirical material accordingly. A similar procedure was employed for the last main code ‘Elements’, where we first identified all the elements of the rule of law as mentioned by the speakers in the analysed debates and then grouped these elements into a few thematically distinct groups that served to interpret the statements and identify narratives.

In our qualitative content analysis of selected parliamentary debates, all text passages identified as relevant to the rule of law narratives formed our sources. The units of coding were paragraphs. They could be assigned to multiple codes if their content referred to different issues, e.g. the purpose of the rule of law or its legitimation. To ensure reliable coding, we produced a codebook describing all codes and providing detailed coding instructions. During the coding process, all codes and subcodes were discussed repeatedly within the team to check that they were sufficiently abstract to reflect the statements made in all five countries. In addition, we regularly compared the content of the text segments coded with the residual subcodes to see if we needed new subcodes to capture all the topics mentioned in the empirical material (Schreier 2012, p. 77). Since identifying the narratives required scholarly interpretation, it was important for us to have country experts perform a context-sensitive reading of the speech acts instead of relying on more technical coder-independent coding.

This stage of the investigation resulted in a refined selection of text segments with attributed codes and subcodes.¹³ These systematically categorised text passages were then interpreted by individual experts to identify significant narratives in a country under study. For every main code, the country experts elaborated a report providing information on identified narratives and their contextual features.

Firstly, we were interested in the prevalence or contentiousness of certain narratives among political parties. Therefore, we analysed whether particular narratives were overlapping, diverging or one-sided. Overlapping narratives refer to those used by representatives of different political parties, regardless of their competitive or cooperative relationships and ideological or programmatic backgrounds. Although there may be slight differences in their wording, the core messages of overlapping narratives are identical. Therefore, there is no substantive conflict between the speakers, even though, depending on the circumstances of the debate, particularly the relations between the government and opposition, these narratives can serve ad hoc political struggles. Diverging narratives represent narratives that are not linked to the current political issue but rather to party ideology, a party’s agenda or to long-term inter-party competition. Counter-narratives used by other actors can mirror this conflict. One-sided narratives are used by only one party or a group of like-minded parties without their counterparts taking an explicit position on them in favour of focusing on other issues.

Secondly, we also sought to capture the evolution of the use of the identified narratives over time. Within the framework of a temporal comparison within each case (country), we focused on whether changes in the intensity of the use of the narratives in question can be observed at specific periods (legislative terms, periods of cohabitation, waves of legislation etc.). The temporal aspect of our analysis intersected with the dimension mentioned above, where we also looked at whether the type of a particular narrative changed over time, e.g. from diverging to overlapping or vice versa.

The 30 reports produced for the six thematic categories (main codes) for each of the five examined countries served as the basis for a cross-country comparison. Figure 4.2 illustrates the framework of the comparative analysis presented in Chap. 5.

Bild vergrößern

To ensure the trustworthiness and plausibility of our interpretative work, the empirical chapters rely on ample direct quotations to provide evidence for our findings. In doing so, we enable readers to gain insights into the various discourses and narratives we have identified and allow them “to evaluate the evidence without relying entirely on the author’s own authority” (Gerring 1998, p. 298).

Overall, our sources provide a novel basis for analysing rule of law narratives and perceptions. The principal added value of our collection of documents is that it allows for a context-sensitive exploration of the wide variety of meanings and issues related to the rule of law, covering both the explicit and the context-specific implicit references to the rule of law. Based on this, we can systematically analyse narratives on the rule of law in the parliaments of East Central Europe and, for the first time, make the relevant debates accessible to a broader public.

At the same time, the particularities of our collection of primary documents need to be taken into account when interpreting the results, as they limit the possibility of generalising from our analysis. They arise from (1) the strategy of compiling the relevant sources, (2) the particular institutional contexts, (3) the specific party organisation and (4) the peculiarities of politicians’ rhetoric.

For this reason, we recommend that readers, firstly, read our analysis as a description of narratives, not of policies or political activities surrounding the rule of law. Our strategy for compiling the sources was not geared to explaining policy decisions, but to observing the use of terms and related arguments. This can entail that particularly polarised debates (which might have been relevant for decisions on certain rule-of-law-related issues) are not included in our corpus. Also, since we first identified the most important constitutional provisions, laws and amendments related to the rule of law or its elements and then collected the corresponding parliamentary debates, our corpus does not contain debates on rule-of-law-related bills that were not passed because MPs could not reach a consensus.¹⁴ This means that the sampled material does not represent the full range of opposition activities, which generally did not succeed in getting their own bills passed.

Secondly, one needs to be aware that the relevance of certain narratives cannot be assessed by calculating how often certain statements are made. Instead, we classified statements as relevant based on case interpretation. This is because the speeches analysed are given by speakers embedded in very different institutional contexts. These have an impact on who has the right to speak in the plenum, for instance. The rules of procedure regulate the order of speakers and the time allocated to speeches differently. Under standard procedures, the speaking time of MPs in Czechia is not subject to any time limitations, unlike the other four cases. Furthermore, the importance of the plenary and the committees differs across countries. In two of our five cases (Chechia and Slovakia), committees only consider bills after the basic principles of a bill have been agreed in the plenary first reading (Kopecký 2004, p. 147), which influences what is discussed in plenary and in what detail.

Other factors also affect the issues that are raised and the emphasis placed on them. In Poland, for example, the ombudsperson (officially called Commissioner for Citizens’ Rights) is constitutionally obliged to “annually inform the Sejm and the Senate about his activities and report on the degree of respect accorded to the freedoms and rights of persons and citizens” (Constitution of the Republic of Poland, Article 212). These reports are regularly followed by debates in the plenary. As a result, the Polish case is likely to have more debates related to citizens’ freedoms and rights than the other cases in our sample, which does not necessarily imply that these issues are more critical for Polish MPs.¹⁵ Furthermore, as amply discussed in Chap. 2, specific events such as the upcoming EU accession or the logic of party competition may also affect the emphasis on the rule-of-law-related issues and the ways in which they are discussed. A greater number of references to matters relating to the rule of law does not necessarily mean that the level of attention paid to the rule of law was higher.

Thirdly, it is important to bear in mind that the narratives described in this book do not represent every individual MP and were sometimes used in a context of fluid party memberships. Our study intends to describe everyday speaking in parliament and accordingly our selection of quotes/speakers aims to reflect the discourse on the rule of law in parliaments. This everyday speaking is influenced by party organisation, such as internal functions and hierarchies. Therefore, quotes were not evenly distributed among MPs. We found more quotes from certain people simply because they were specialists in their party on issues around the rule of law or were particularly high-ranking in their party. For the governing parties, it was primarily the responsible ministers who made the relevant speeches. Although we tried to provide a relatively broad picture of statements, it would be inaccurate to cite more backbenchers if the few high-ranking politicians were the more relevant ones. We consider their statements to represent social narratives because their parties supported them as holders of the key functions. However, we cannot rule out the possibility that people who did not speak might have described the purpose, elements and other features of the rule of law differently. These few people per country have sometimes belonged to different parties over the course of time, which means that inferring party positions from quotes in this book might be somewhat misleading.

Fourthly, one needs to remember what we have already mentioned in the introduction—the narratives that we reconstruct do not necessarily reflect the individual speakers’ perceptions. Political actors, reacting to their particular environments, can communicate strategically (Chaps. 2 and 3). As politicians they can be expected to debate controversial issues rather than the uncontroversial ones or the issues everyone takes for granted. If, for example, the independence of the courts is not discussed at length in parliaments, this does not mean that it is irrelevant to the speakers but might indicate that the actors agree that this independence is a necessary element of the rule of law. Again, this underlines that the frequency of specific statements does not tell us anything about the importance that speakers in the five countries attach to certain topics. In our study we do not pay too much attention to the frequency, but more to the content of statements, basing our interpretation on in-depth knowledge of the cases.