In the past decade, a number of European Union (EU) Member States, with Hungary and Poland most notable among them, have been chipping away at liberal democratic institutions upon which EU membership is premised (Bakke and Sitter Reference Bakke and Sitter2020; Kelemen Reference Kelemen2017; Vachudova Reference Vachudova2020). Despite the serious, and arguably existential (Kelemen Reference Kelemen2019), nature of the challenge posed to the EU’s normative and functional foundations, the response of European institutions has been slow (Kochenov Reference Kochenov2017; Kochenov et al. Reference Kochenov, Magen and Pech2016; Pech and Scheppele Reference Pech and Scheppele2017) and often criticised as insufficient (Hernández and Closa Reference Hernández and Closa2024; Kelemen Reference Kelemen2024). The situation is typically referred to as the Rule of Law Crisis (RoLC), a catchall term to cover a range of national government-led challenges to EU fundamental values but chosen to highlight the EU’s emphasis on violations directly or indirectly related to the rule of law (RoL) as the primary institutional check on autocratic power.
The European Parliament (EP) is a key site where consequential, public debate relating to the RoLC has been carried out, with deep divides between a growing pro-intervention camp of left-wing and liberal MEPs and a vocal opposition to intervention from more right-wing and Eurosceptic representatives (Coman Reference Coman2022: ch. 5; Wunsch and Chiru Reference Wunsch and Chiru2024). EP discourse so far has not been defined by acrimonious forms of political opposition. This is in part because of the nature of the main body of work carried out by the EP – to amend and pass technical legislation proposed by the Commission, which elevates a form of problem-solving discourse (Garssen Reference Garssen2016). But compromise-building logics have also resulted from the historical dominance of centrist and pro-EU parties in the EP that did not contest fundamental values (Bartolini Reference Bartolini2006). These conditions, however, may be shifting. Over the last three election cycles, the ideological composition of the EP has changed dramatically with significant increases in the representation of Eurosceptical and far-right actors (Brack Reference Brack2020; Chiru and Wunsch Reference Chiru and Wunsch2023). A marked increase in the number of debates on the RoLC has also likely exacerbated tensions between mainstream and fringe political groups. These debates are essentially accusatory contexts where some Member States – and perhaps others by implication – are presumed to be in the wrong by certain MEPs, not on matters of policy, but on fundamental values that directly problematise existential questions of national sovereignty and European integration.
This article seeks to explore the nature and determinants of political conflict in EP debates within the unique context of the RoLC. While previous research on this topic has mainly focused on the ideological dimension of this conflict, we shift focus to the affective dimension, analysing the roles of negative sentiment and motive-based criticism in these debates. The latter concept is a more extreme form of negativity involving disrespect, where MEPs attack the motivation of opponents rather than engaging with their arguments directly. Specifically, we ask: To what extent does negativity in general and motive-based criticism in particular characterise debates on RoLC-related issues in the EP? And what are the factors driving different forms of political opposition on these issues? The first research question is an inductive one, relying on descriptive statistics concerning the weight of negative sentiments and motive-based criticism relative to more positive and principle-based arguments. The second research question is an explanatory one, where our independent variables are drawn deductively from cognate studies on factors affecting MEP behaviour in RoL debates.
While our primary goal is to deepen existing research on political opposition in RoL debates, we submit that exploring these features of political contestation in the EP has wider significance for understanding discursive developments in the EP more broadly. Most significantly, precisely because it concerns questions of fundamental values, how MEPs debate with one another about the RoLC is among the strongest available indications of the extent to which affective polarisation may be present or emerging in the EP. Indeed, any patterns of animus that manifest in these debates are unlikely to remain isolated to RoL issues indefinitely.
To achieve our aims, we use qualitative content analysis to study 53 EP debates on the RoLC between 2012 and 2022, coding for the prevalent position of MEPs’ speeches in these debates, the sentiment of their intervention and whether they resort to motive-based criticism. Descriptively, we find that the proportion of pro-intervention statements far outweigh anti-intervention statements across the 53 debates, while negativity dominates and there are concerning levels of motive-based cynicism On the explanatory dimension of our study, findings from our regression analysis suggest that while MEPs from more right-wing national parties represented in the EP are more likely to make anti-intervention statements and to use motive-based rhetoric, they are not more likely to use a negative tone. A similar pattern emerges across all three outcomes for MEPs from more Eurosceptic parties. MEPs representing parties from Central and Eastern Europe (CEE) show similar patterns when it comes to their anti-intervention positioning and the negativity of their statements, but there is limited evidence that their discourse is more likely to be motive-based. Finally, MEPs from parties that form part of the government at the national level and are thus represented in the Council seem to be somewhat more likely to use more neutral language and to make anti-intervention statements.
From a normative perspective, two main concerns arise for the democratic quality of discourse within the EP. Firstly, the intensity of motive-based discourse appears to be high in RoL debates when compared to the limited historic record on levels of disrespectful discourse in EP debates more generally (Lord and Tamvaki Reference Lord and Tamvaki2012). This suggests that RoL debates may be fertile ground for the development of tendencies that both express and foster affective polarisation in the EP. Secondly, the large share of this motive-based discourse is concentrated on the right flanks of the EP. Should this prove persistent and increase over time with future election results favouring right-wing parties or new waves of democratic backsliding, the EP will be increasingly susceptible to the normative jeopardies of attack politics.
The article proceeds as follows. The first section explains our central concepts in the article – negativity and motive-based criticism – and details the significance of our research for understanding political opposition in the EP concerning the RoLC. In the next section, we provide an overview of our hypotheses concerning the factors affecting different forms of contestation in these debates. We then outline our coding strategy and methods before moving on to present our findings. Our conclusion puts our findings into a broader perspective, first by comparing them with limited existing studies on sentiment and motive-based criticism in the EP and then considering their implications for political opposition in the EP more broadly.
Negativity and motive-based cynicism in the RoLC
Research on how the RoLC is being contested in political discourse is predominantly qualitative in nature. Crucial advances have been made in elucidating how backsliding dynamics are justified and opposed in domestic politics, specifically focusing on the domestic discourses of national politicians orchestrating backsliding in the EU’s Member States (Csehi and Zgut Reference Csehi and Zgut2020; Dawson and Hanley Reference Dawson and Hanley2019; Fabbrini and Zgaga Reference Fabbrini and Zgaga2024; Schlipphak and Treib Reference Schlipphak and Treib2017). Single and comparative case-study research has also been conducted on domestic opposition responses to backsliding (Ayan Musil and Yardımcı-Geyikçi Reference Ayan Musil and Yardımcı-Geyikçi2023; Gamboa Reference Gamboa2022; Stanley Reference Stanley2016; Til and Krasztev Reference Til and Krasztev2015).
A smaller literature analyses how these discursive dynamics play out within EU institutions. Existing qualitative studies tend to focus on how offender states justify their position within EU arenas for domestic audiences (Mos Reference Mos2020), and how in turn the discourse of backsliding governments is received and countered by EU institutions (Emmons and Pavone Reference Emmons and Pavone2021). Natasha Wunsch and Mihail Chiru (Reference Wunsch and Chiru2024) explore norm contestation as a crucial dimension of the oppositional dynamics in EP RoL debates, while Ramona Coman (Reference Coman2022: 137–147) discusses dissensus between different European Parliamentary Groups (EPGs) concerning how their members understand the RoL. What these authors respectively refer to as polarisation is the distance in perspectives on RoL issues that emerges between opposing MEPs in these debates.
In this article, we focus on complementary aspects of contestation, namely negativity and motive-based criticism. Negativity is the extent to which MEPs formulate arguments with a negative or critical tone, for example, criticising opponents or pointing to the negative consequences of intervention or lack of intervention. On the other hand, MEPs may also defend arguments in essentially positive terms, where MEPs provide positive reasons or examples for why intervention should be supported or opposed. Limited existing sentiment analyses on EP speeches – specifically automated dictionary approaches across all policy areas (Schumacher et al. Reference Schumacher, Schoonvelde, Traber, Dahiya and de Vries2016) and specific topics during crisis periods (Salgado Reference Salgado2021) – indicate that positive sentiments tend to dominate EP discourse. To the extent that RoLC debates are accusatory, significant negativity beyond the norm is expected. The interesting question is precisely the extent of this negativity (where some actor or proposal is targeted) relative to more positive assertions (where some value or proposal is advocated). This will provide an indication of both the high salience of the issue and the extent of political divides on fundamental values.
Negativity can be principle-based, aimed primarily at the practices of opponents and demonstrating how said practices contradict or undermine key values. Motive-based criticism, instead, is directed primarily at the intentions of opponents and thus questions the integrity of the reasons they may have to do or say something. This may include, for example, accusations of malign partisanship and corruption, or moral failings like double standards or ignorance driven by prejudice, laziness or bandwagoning. Motive-based criticism is therefore a specific kind of negativity with a particular normative valence.
Adopting motive-based forms of discourse can be particularly problematic for democratic politics at both the elite and population level, producing a tit-for-tat cycle where actors seek to exploit the inherent ambiguity of political motivations by putting the worst possible spin on the actions of their opponents (Herman and Muirhead Reference Herman and Muirhead2021). Indeed, even when there is good reason to attribute negative motivations to specific actors, doing so publicly tends to involve at least some degree of speculation by challenging the integrity of agents. In fact, all political actors are exposed to such attacks to the extent that motivations in politics are typically mixed: even the best intentions for the common good will also tend to involve some degree of self-interested and partisan motivations (Bagg and Tranvik Reference Bagg and Tranvik2019; Gutmann and Thompson Reference Gutmann and Thompson2010).
While negativity on its own may be an indication of ideological or discursive polarisation in the context of norm contestation (Wunsch and Chiru Reference Wunsch and Chiru2024), motive-based reasoning as we understand it can be seen as a set of discursive patterns that both express and contribute to the moralising dimension of affective polarisation. Rather than merely seeing out-partisans as fundamentally different (othering) or as unlikeable and untrustworthy (aversion), the moralising dimension of affective polarisation involves thinking of the other side as morally corrupt in some way (Finkel et al. Reference Finkel, Bail, Cikara, Ditto, Iyengar, Klar and Druckman2020). While moralising affective polarisation can be expressed in different ways, attributing cynical or otherwise ill-founded motivations to opponents captures a significant aspect of this phenomenon.
When elites and mass publics suspect duplicity from opponents and charge them with ill-intentions, affective polarisation and related breakdowns in dialogue become increasingly likely (Enyedi Reference Enyedi2016; Luttig Reference Luttig2017; McCoy et al. Reference McCoy, Rahman and Somer2018; Milstein Reference Milstein2020). This in turn produces risks for the long-term resilience of democratic institutions. At the elite level, moralising polarisation shapes enmity among competing political actors (Banda and Cluverius Reference Banda and Cluverius2018), making cooperation more difficult, producing ‘an unwillingness to defer to or accord legitimacy to institutions that produce outcomes inconsistent with one’s convictions’ (Campos and Federico Reference Campos and Federico2025). It also perpetuates a cycle of limited discursive quality focused on undermining the position of others and eliciting fear over the potential consequences of allowing the other side to have its way. To the extent that cooperation is fundamental to maintaining the daily problem-solving capacity of the EP and its function as the main forum for public justification in the EU, the development of a more pernicious partisanship in the EP could jeopardise both roles. At the population level, the negativity of motive-cynicism can alienate citizens from the political process and erode trust in institutions (Lau et al. Reference Lau, Sigelman and Rovner2007), while encouraging the spread of affective polarisation among the citizenry (Finkel et al. Reference Finkel, Bail, Cikara, Ditto, Iyengar, Klar and Druckman2020). Intriguingly, in the EU context, this may stoke antagonistic attitudes towards the EU itself and not merely pro-EU partisans, allowing for Eurosceptics to construe it as a morally corrupt organisation bent on eroding national sovereignty (Oleart Reference Oleart2021: 37–39).
Limited research on the relationship between ideological and affective polarisation indicates that the degree of ideological polarisation tends to increase the extent of affective polarisation (Riera and Madariaga Reference Riera and Madariaga2023; Webster and Abramowitz Reference Webster and Abramowitz2017). In this light, one may expect the structure of RoL debates to make moralising forms of affective polarisation especially likely to find their expression and development, precisely because they centre around fundamental differences in values where actors are strongly motivated to advance and defend their own positions.
The EP has long been recognised as a consensus-driven organisation where predominantly technical debates take on a problem-solving structure (Garssen Reference Garssen2016). Yet there are good reasons to think debates on the RoLC may take on a more clearly adversarial logic. For pro-intervention MEPs, the RoLC might raise existential questions on the continued integrity of the EU’s legal order (Kelemen Reference Kelemen2019) and its ability to maintain ‘an area of freedom, security and justice’ (Article 3.2 Treaty on European Union, henceforth TEU) under conditions of ‘sincere cooperation’ (Article 4.3 TEU) (Bellamy and Kröger Reference Bellamy and Kröger2021; Bellamy et al. Reference Bellamy, Kröger and Lorimer2022). For those countries that are potential targets of EU intervention, in turn, considerable financial and political interests are in the balance, given that, in principle, they risk heavy fines, loss of access to EU funds and, eventually, possible deprivation of their voting rights within EU institutions. Changes in the composition of the EP since 2014 have only compounded this growing tension, with increased ideological diversity and greater representation of Eurosceptic forces resulting in the emergence of radically different understandings of liberal democratic values and European integration (Chiru and Wunsch Reference Chiru and Wunsch2023; Fabbrini and Zgaga Reference Fabbrini and Zgaga2024; Treib Reference Treib2021). At the same time, the expression or development of intense forms of affective polarisation is not a foregone conclusion in debates on fundamental values, as they can still be conducted at the more principled level of norm contestation. What is most important, however, is not the presence of motive-based cynicism itself, but rather its proportion relative to principled forms of speech and how its expression is distributed among different political actors.
Our article seeks to make two main contributions to the literature on the RoLC. Firstly, by moving beyond norm contestation and ideological polarisation to study the sentiment of speeches and markers of affective polarisation, we will be able to assess the degree to which the EP is susceptible to forms of pernicious partisanship. While there have been a range of studies on affective polarisation among European citizens at the national level (Reiljan Reference Reiljan2020) and among populations during EP elections (Hahm et al. Reference Hahm, Hilpert and König2023), affective polarisation in the EP (and among elites generally) has been understudied. We know, however, that a higher degree of ideological polarisation among elites, as well as sustained and coordinated negativity towards out-partisans (Kollberg and Bantel Reference Kollberg and Bantel2025), intensifies affective polarisation among citizens (Gidron et al. Reference Gidron, Adams and Horne2020; Hahm et al. Reference Hahm, Hilpert and König2023). There are few studies on either negative sentiment or motive-based criticism in EP discourse, thus providing us with few real points of comparison for our study (though for some partial exceptions see Salgado Reference Salgado2021; Schumacher et al. Reference Schumacher, Schoonvelde, Traber, Dahiya and de Vries2016; Lord and Tamvaki Reference Lord and Tamvaki2012 as per above).
By exploring these phenomena, we will contribute to building a picture of little-explored dimensions of political contestation in the EP. To the extent that fundamental values are constitutive of the EU and designed to serve as a framework within which disagreement ought to take place, high levels of motive-based criticism on these issues would be of particular concern. They would indicate emergent lines of enmity that go far deeper than (and may eventually seep into) contestation over more standard forms of policy disagreement. By contrast, low levels of motive-based critique on such a normatively charged set of debates would indicate a degree of EP resilience to the emergence of more pernicious forms of partisanship.
Our second contribution is to quantitatively explore the key factors influencing the sentiment and motive-based criticism of MEPs who make speeches in our dataset. Quantitative methods in the context of the RoLC politics have been primarily used to understand how MEPs vote on these issues rather than how they discuss them, specifically with the aim of understanding the puzzle of EU ‘inaction’ on fundamental values in the early 2010s (Chiru and Wunsch Reference Chiru and Wunsch2023; Meijers and Van der Veer Reference Meijers and Van der Veer2019; Sedelmeier Reference Sedelmeier2014, Reference Sedelmeier2017).
Qualitative studies have demonstrated the value of studying how the RoLC is being politicised in the EP, and by whom (Coman Reference Coman2022: ch. 5; Wunsch and Chiru Reference Wunsch and Chiru2024). These studies combine descriptive statistics and qualitative analysis to study the EP’s growing activism on the RoLC since the early 2010s. They find clear lines of division remaining among MEPs according to ideological and geographical lines, with discursive opposition to EU intervention mostly driven by right-wing Eurosceptics and MEPs from CEE. Pro-intervention discourse, in turn, is over-represented among liberal, left-wing MEPs from Western Europe who appeal to supranational unity and principles.
This article bridges these qualitative and quantitative literatures by identifying the factors driving MEP political opposition on the RoLC through quantitative content analysis. In other words, we depart from existing qualitative studies on EP RoL discourse and existing quantitative studies on EP RoL roll-call votes by adopting a quantitative focus on EP RoL discourse. Our study complements roll-call research by testing the extent to which ideological variables, geographical variables and the strategic interest of national governments in pro- and anti-intervention positions on these issues that have been shown to play a role in informing MEPs’ voting behaviour on the RoLC also track pro-intervention and anti-intervention discursive positioning – what we call intervention-based disagreement. To the extent that patterns in voting and discourse can diverge (Lauderdale and Herzog Reference Lauderdale and Herzog2016) and given the volatility of RoL voting patterns measured by the evolution of high levels of intra-party disagreement (Coman Reference Coman2022: 134), we think there is value in verifying that EP discursive patterns are reliably driven by the same factors as voting patterns. Furthermore, the role of national government status has been so far under-explored as a predictor of voting behaviour across parties on these issues. By exploring this variable in the present context, it will be the first time that it has been considered for the EP as a whole.
But the focus on political discourse over voting behaviour also allows us to add nuance to current understandings of opposition over the RoLC in the EU. While roll-call studies provide an understanding of the factors that drive MEPs to vote for or against intervention, they do not capture the salience and significance of these issues for MEPs, and indeed the nature of their political disagreements. We therefore aim to capture the adversarial nature of these debates not only by deepening existing research on issue-based disagreement, but also by exploring the extent to which the factors driving this form of disagreement influence our core variables of interest: negative sentiment and motive-based criticism.
Hypotheses for the determinants of political opposition in RoLC EP debates
Given the limited research on the role of negative sentiment and motive-based criticism in the EP in general, and especially with respect to the RoLC, the descriptive aspect of our article is purely exploratory and without explicit hypotheses. For the explanatory dimension of our article, there is a rich literature on the RoLC and beyond that provides a solid basis for several hypotheses concerning the characteristics of MEPs’ national parties that may increase their likelihood of a) defending anti-intervention positions, b) engaging in negative argumentation and c) using motive-based language. Specifically, we explore the impact of ideological, strategic and geographical positioning of MEPs’ national parties on each of our three dependent variables.
Ideologically, we analyse the role of left-right positioning and Euroscepticism as key factors likely to drive discursive contestation on the RoLC in the EP. Precisely because the radical right is more likely to be tolerant of democratic subversions (Albertazzi and Mueller Reference Albertazzi and Mueller2013; Blokker Reference Blokker2019; Crum et al. Reference Crum, Alvaro and Overeem2023; Müller Reference Müller, Kaltwasser, Taggart, Espejo and Ostiguy2017) and Eurosceptics tend to be resistant to EU interventions in national affairs (Halikiopoulou et al. Reference Halikiopoulou, Nanou and Vasilopoulou2012), existing research has highlighted these variables as potentially important drivers that can explain MEP stances on issues relating to the RoLC (Herman et al. Reference Herman, Hoerner and Lacey2021; Meijers and Van der Veer Reference Meijers and Van der Veer2019). However, we also expect both the left-right dichotomy and Eurosceptical attitudes to affect negativity and motive-based cynicism in the EP.
Following well-established research that associates the type of populist anti-pluralism that drives polarising rhetoric with the right-wing of the political spectrum (Crum et al. Reference Crum, Alvaro and Overeem2023; Medzihorsky and Lindberg Reference Medzihorsky and Lindberg2023), qualitative evidence points to the currency of such motive-cynicism among anti-intervention proponents in the EU, especially by offender executives on the right (Csehi and Zgut Reference Csehi and Zgut2020; Emmons and Pavone Reference Emmons and Pavone2021; Mos Reference Mos2020; Schlipphak and Treib Reference Schlipphak and Treib2017; Wunsch and Chiru Reference Wunsch and Chiru2024). Motive-based reasons of this kind have also been shown to be prevalent among far-right parties in the EP when it comes to issues of gender equality, an issue on which the EP also appears to be especially polarised (Kantola and Lombardo Reference Kantola and Lombardo2020). Furthermore, because of the typically radical and populist profile of Eurosceptical national parties, their expected negativity in resisting EU intervention may spill over into motive-based discourse. Taking these considerations together, we propose the following two sets of hypotheses concerning the relationship between ideology and forms of discursive contestation.
Hypothesis 1: MEPs from national parties with more right-wing ideological preferences are more likely to defend a) anti-intervention positions; b) their positions in negative terms; and c) their positions using motive-based arguments compared to MEPs with more left-wing ideological preferences.
Hypothesis 2: MEPs from Eurosceptical national parties will be more likely to defend a) anti-intervention positions; b) their positions in negative terms; and c) their positions using motive-based arguments compared to MEPs with more pro-European ideological preferences.
Geographical factors may also impact discursive contestation over the RoLC in the EP. Not only are target countries in RoL debates – most notably Hungary and Poland – in CEE, but democratic quality is on average lower in CEE (Bakke and Sitter Reference Bakke and Sitter2020; Schimmelfennig and Sedelmeier Reference Schimmelfennig and Sedelmeier2020). CEE governments that fall short of democratic standards would have good reasons to stand in solidarity with targeted executives in the hope of future reciprocity. Regardless of the state of the RoL in their country, CEE MEPs might also be inclined to oppose intervention or remain silent on the issue in a bid to preserve good neighbourly relations with countries such as Poland or Hungary. What we know of Member States positioning on the RoLC points in this direction (Sedelmeier Reference Sedelmeier2024: 838–841; Wunsch and Chiru Reference Wunsch and Chiru2024: 18). Furthermore, in defending their actions at the EU-level, the Polish Law and Justice Party (PiS) and Hungarian Fidesz are also known for questioning the motivations of their opponents, and to present the targeting of CEE states as evidence of ideological and geographical bias (Csehi and Zgut Reference Csehi and Zgut2020; Emmons and Pavone Reference Emmons and Pavone2021; Mos Reference Mos2020; Schlipphak and Treib Reference Schlipphak and Treib2017; Wunsch and Chiru Reference Wunsch and Chiru2024). This provides grounds for believing that negative and motive-based discourse may be more likely to proceed from members of national parties with the greatest stake in resisting EU action.
Hypothesis 3: MEPs from CEE countries are more likely to defend a) anti-intervention positions; b) their positions in negative terms; and c) their positions using motive-based arguments compared to MEPs from Western European countries.
Finally, we examine how the strategic interests of parties in government affect contestation in RoLC debates. The Council and the European Council are known to cultivate an atmosphere of collegial bargaining, whereby states build coalitions of mutual interest to facilitate decision-making (Puetter Reference Puetter2012; Warntjen Reference Warntjen2010). National parties in power, therefore, have incentives to shy away from open criticism to maintain good intergovernmental relationships with backsliding states. At the same time, MEPs tend to be more responsive to instructions by their national party when the latter is in government and therefore a member of the European Council and Council (Hix et al. Reference Hix, Noury and Roland2006; Willumsen Reference Willumsen2018). What we know of the politics of the RoLC supports these expectations (Blauberger and Kelemen Reference Blauberger and Kelemen2017: 332; Emmons and Pavone Reference Emmons and Pavone2021: 1619; 1622–1623; Herman et al. Reference Herman, Hoerner and Lacey2021). This leads us to propose the following:
Hypothesis 4: MEPs from parties in government are more likely to defend a) anti-intervention positions; b) their positions in neutral terms; and c) their positions using principle-based arguments compared to MEPs from parties in opposition.
Methods and models
Qualitative coding strategy
To construct our dataset, we selected 53 EP plenary debates on the RoL from the 7th, 8th and 9th EP terms (up to 2022), starting when the Parliament cast its first vote on the Hungarian situation on 16 February 2011. We identified relevant debates by first conducting a search for ‘rule of law’ within our date range on the EP plenary website and then selecting debates that meet the following criteria, based first on their title and then on a review of their contents.Footnote 1 Our selected debates were a) consequential by virtue of having a resolution attached to them, b) about developments within EU Member States rather than EU institutions and c) about general state-led violations of liberal democratic institutions (rather than specific individuals or social events). Topics in the final list of debates span a wide range, including threats to media freedom, attacks on LGBTQ+ rights and gender equality in the EU. Some of the selected debates are explicitly discussing RoL breaches in a specific country (for example Hungary or Slovakia), but others discuss the RoL in the context of broader debates – for example, regular EP debates on the situation of fundamental rights in the EU (e.g. Debates 18 or 25 in our dataset), or debates concerning the implementation of a RoL conditionality mechanisms (e.g. Debates 49 and 51 in our dataset). Overall, our selection allows for a far broader analysis of debates around the RoL in the EP than existing studies, while remaining focused on EU internal affairs. A full list of debates can be found in the Supplementary Material, Section A.
The debates were hand-coded by the authors as well as two research assistants using NVivo software.Footnote 2 We developed the codebook in an iterative manner, starting from our theoretical expectations as outlined above and making minor amendments after an initial round of coding. Importantly, our unit of analysis is the individual MEP speech, to which coders applied one code of each of the following categories. Coding was facilitated by the relatively short format of each MEP speech, limited by the Rules of Procedure of the European Parliament.Footnote 3 In line with the hypotheses section above, our dependent variables are the following categories of our coding scheme. Concerning the position of MEPs on EU intervention, we coded as anti-intervention those speeches that primarily opposed intervention against backsliding states but also those that doubted the existence of the problem or minimised the RoLC. The code pro-intervention was applied to those speeches that primarily supported intervention against backsliding states but also those emphasising the negative implications of the RoLC or criticising backsliding states. Finally, intervention neutral was used for those speeches that did not formulate a position on whether the EU should intervene in backsliding states or on the consequences of the RoLC. While we use the terms pro- and anti-intervention positions to define the different poles of issue-based disagreement, the coding therefore covers a wider range of arguments than those strictly formulated in favour of or against the use of a specific type of sanction against backsliding states in the EU.
For our second dependent variable, operationalising negativity in the EP, we coded for the dominant tone or sentiment conveyed by MEPs in their speeches.Footnote 4 MEPs may defend their position in negative terms, harnessing negative examples or verbalising positions against the opposing camp; in positive terms, providing positive reasons to support their argument against or in favour of intervention; or in neutral terms, where MEPs do not attach clear positive or negative sentiments to their positions. While coding the dominant sentiment of a speech can sometimes result in difficult coding choices when a speech is well-balanced in terms of sentiment, these instances do not appear so frequently as to risk substantially altering results – especially when considering that coding positive sentiments in borderline cases will cloak negative sentiments and vice versa. Considering the short length of MEP speeches, typically just a few hundred words, the dominant sentiment is usually clear.
Finally, we operationalised motive-based criticism by identifying the type of arguments made, distinguishing between principled and motive-based justifications for MEP positioning. Here we draw on the framework presented by Lise Esther Herman (Reference Herman2017) to operationalise pluralist values in partisan discourse. Arguments were coded as principled when speakers justified their argument based on the merits of one or several specific principles, whether that principle was positively defended or rendered as being under threat. For instance, an anti-intervention argument might defend the democratic merits of specific institutional reforms under scrutiny or focus on the negative impact of EU intervention on national sovereignty. A pro-intervention argument, in turn, might insist on the importance of the EU protecting the RoL, or emphasise the negative consequences of its foundational values being under threat. In turn, arguments were coded as motive-based when they were justified based on a criticism of the motives held by opponents. Such arguments would also always be coded as negative. For instance, an anti-intervention argument may be formulated as an accusation of double standards: the pro-intervention camp being suspected of predominantly focusing on backsliding initiated by right-wing government for partisan reasons. A pro-intervention argument, in turn, may describe backsliding governments as acting in their own corrupt interests and seeking to advance their personal power. Details on our coding strategy and codebook can be found in the Supplementary Material, Section B.
Datasets and models for the quantitative analysis
To analyse the impact of the various ideological and strategic factors highlighted in our hypotheses section, we aggregated the dataset resulting from our hand coding by national party and debate. Overall, our dataset contains 8,401 party observations containing a total of 1,558 speeches words nested in the 53 debates. A list of all interventions by country, EP group and national party (and thus a full list of parties) can be found in Supplementary Material, Section F. As highlighted above, our dependent variables are the various categories of our coding scheme: pro/anti-intervention, positive/negative sentiment and principled/motive-based arguments. This takes the form of counts of statements coded into a particular category for each party per debate. Each individual outcome (i.e. pro-intervention, anti-intervention or neutral) is modelled separately.
We merge the resulting dataset at the level of national parties and the month during which a given debate took place with our various independent variables of interest. We make use of the following data sources for our independent variables in the first set of models (Figure 1). Firstly, to estimate the impact of a party’s general left-right position and its support for European integration, we rely on data from the Chapel Hill Expert Survey (CHES) (Jolly et al. Reference Jolly, Bakker, Hooghe, Marks, Polk, Rovny and Vachudova2022). From this dataset, we also include the cultural left-right dimension (Green/Alternative/Libertarian to Traditional/Authoritarian/Nationalist, henceforth GAL-TAN) as an alternative measure (see Figures E1 and E2 as well as Table E1 in the Supplementary Material). We use linear interpolation between the different CHES waves at the monthly level and match the respective debates from our dataset to the month in which they took place. If a party is not covered in the subsequent wave and for all parties and months after January 2019, we carry forward the last available value. For the EU position variable, we create the inverse so that higher values indicate more Euroscepticism.
Share of the Different Coding Categories for Type of Intervention, Sentiment and Type of Argument
Secondly, for information regarding the government status of a party at the national level and its presence in the EP, we rely on ParlGov data (Döring et al. Reference Döring, Huber and Manow2022). These variables are merged with our dependent variables at the start date of a cabinet and the date of the last EP election, respectively. Thirdly, we include a dummy variable for whether a party is from a CEE country or not. We also include a variable for the absolute seat share of a party in the EP as a control variable. All models also include fixed effects for EP parliamentary terms. Descriptive statistics can be found in the Supplementary Material, Section C.
Given that the dependent variables in all our models are counts, we fit negative binomial regression models with fixed effects for the EP term and standard errors clustered at the level of national parties. A negative binomial model was chosen given the larger number of ‘zeros’ in our dataset, i.e. instances in which given parties do not make any statements in a debate. A negative binomial rather than a Poisson model was therefore chosen. Zero-inflated models are alternatives in this context, but the negative binomial model has the advantage of easier interpretability (Allison Reference Allison2012). Moreover, the zeros in our dataset are ‘true zeros’, i.e. only those parties are included which were represented in the EP at the time of the debate and could thus have made a statement, further highlighting that a zero-inflated model is not required.
Findings
Before we run our models, we first examine the relative distribution of our key variables of interest, displaying the number of interventions in a particular category divided by the sum of all interventions for each (Figure 1). Descriptively, we find that there are almost three times more pro-intervention statements than anti-intervention statements; negative statements outweigh positive statements by a similar margin, while motive-based statements represent a total of around 28% of the total discourse.
There is a relatively consistent share of both negativity and motive-based statements over the three terms, with only a slight increase in the share of negative tone and a slight decrease in the share of motive-based statements over time (Figure 2).Footnote 5 If we disaggregate these patterns by EP group, some interesting patterns emerge. As Figure 3 shows, the highest share of motive-based statements can be found among members of the European Conservatives and Reformists party (ECR) and the radical right groups as well as among the non-inscrits (NI). Moreover, these groups also show high levels of negativity in their tone. However, here we also observe high values for the radical left, whose negativity has also increased over the last three EP terms. While trends over time are difficult to interpret, we note a decrease in the share of motive-based statements among European People’s Party (EPP), ECR and NI MEPs, and an increase in these statements among radical right MEPs.
Share of Negative Tone and Motive-Based Statements over Time
Share of Negative Tone and Motive-Based Statements over Time by EP Group
Figure 3 Long description
The bar graph shows the share of negativity and motive-based statements across EP groups for the years 2009, 2014 and 2019. The groups on the x-axis are GUE/NGL, Greens and EFA, S&D, ALDE/RENEW, EPP, ECR, Radical Right and NI. The y-axis represents the proportion, ranging from 0 to 1. Each group has vertical bars for each year, with two bars per year: one for negativity and one for motive-based statements. Negativity is represented by darker bars, while motive-based statements are shown with lighter bars. Notably, the Radical Right group shows consistently high negativity across all years, reaching close to 0.9 in 2019. The ECR group also displays high negativity, particularly in 2019. In contrast, the Greens and EFA group maintains lower negativity levels. The motive-based statements are highest for the Radical Right and ECR groups in 2019. The graph highlights trends over time, with some groups showing increased negativity and motive-based statements, while others remain stable or decrease.
We now turn to our substantive regression models. In Figure 4, we show the coefficients for our main variables of interest as specified in our hypotheses. We first discuss the effects of a national party’s left-right position. In line with Hypothesis 1a, we find that MEPs from national parties with more right-wing ideological preferences are more likely to defend anti-intervention positions than MEPs with more left-wing ideological preferences. By contrast, we do not find any evidence for Hypothesis 2a relating to sentiment – it does not seem to be the case that more right-wing positions are associated with more negative statements by MEPs. However, we do find that the statements of these MEPs are more likely to be motive-based (Hypothesis 3a). Table 1 shows the coefficients in numerical form.
Main Models
Numerical Results of the Regression Analysis
Table 1 Long description
Regression results link party characteristics to five rhetoric outcomes: anti-intervention, negative sentiment, motive-based, neutral sentiment, and principled argument. Anti-intervention has the strongest and most consistent positive associations: higher left-right scores, more euroscepticism, being based in the East, being in government, and having more seats all predict higher anti-intervention counts, with especially large effects for East and government. Motive-based rhetoric also increases with left-right position, euroscepticism, East, and seat share, but the government effect is not clearly different from zero. Negative sentiment is mainly associated with seat share; left-right, euroscepticism, East, and government show no clear relationship. Neutral sentiment increases with government status and seat share, while other predictors are not clearly related. Principled argument shows little relationship to most predictors, aside from a positive association with seat share. All five models use the same sample size of 8,401 observations, and results should be read with the note that uncertainty is reported via clustered standard errors and that term indicators are included but not listed.
Notes: Negative binomial regression models with standard errors clustered at the party level. Includes dummies for parliamentary term (omitted). * = significant at the 0.1 level, ** = significant at the 0.05 level and *** = significant at the 0.01 level.
Figure 5 shows the substantive size of the effects. Substantively, the most right-wing parties make around 10 times more anti-intervention statements than a left-wing party, with a value of 2 on the CHES left-right scales, as shown in the predicted values in the upper left-hand panel. Moreover, the lower left-hand panel shows that the most right-wing parties issue more than twice as many motive-based statements as the most left-wing parties in our dataset, holding all other variables at their mean.Footnote 6 The absence of an effect for negative sentiment is again illustrated in the upper right-hand panel of Figure 5. We find similar results for the effects of Euroscepticism across intervention-based positioning, negativity and antagonism, thereby supporting Hypothesis 2a and c (see Figure 6).
Predicted Values at Different Levels of the Left-Right Position
Predicted Values at Different Levels of Euroscepticism
Moving to the effect of the East/West variable, MEPs belonging to CEE Member States issue significantly more anti-intervention statements, providing support for Hypothesis 3a. We do not find evidence for Hypothesis 3b, which predicts greater negativity for CEE MEPs. In line with Hypothesis 3c, however, these MEPs seem to issue more motive-based statements, but the coefficient is only significant at the 0.05 level, so our evidence for this hypothesis is somewhat weaker than the effect of ideology on the same dependent variable (see Figure 4 and Table 1). The predicted values can be found in Figure 7. MEPs from parties from CEE countries issue around three times more anti-intervention statements than MEPs from parties from non-CEE countries (upper left-hand panel). Note that the overlapping confidence intervals in the lower left-hand panel seem to confirm that our evidence for Hypothesis 3c (motive-based statements) is more limited.
Predicted Values for CEE and Non-CEE Countries
Lastly, we focus on the effect of government participation. Figure 4 shows that the coefficient for anti-intervention statements reaches significance at the 0.01 level. There is thus evidence for Hypothesis 4a, which predicts that MEPs belonging to national parties represented in the Council are more likely to defend anti-intervention statements. The coefficient for neutral tone only reaches the 0.05 level of significance. The evidence for Hypothesis 4b is thus somewhat less clear. By contrast, we find clear evidence against Hypothesis 4c, according to which MEPs from parties represented in the Council are more likely than those belonging to domestic opposition parties to use more principled arguments. Figure 8, showing the predicted values, confirms that there is some evidence for an effect of government participation on anti-intervention statements and neutral tone (Hypothesis 4a and b) and substantial evidence against a stronger use of principled statements (Hypothesis 4c).
Predicted Values for Government Participation
We have also explored the impact of alternative specifications and measures, running three further sets of models. Firstly, we focus on the impact of the cultural GAL/TAN dimension on our independent variables. Secondly, we distinguish by the effect of whether a debate is focused on a specific country, a distinction made by Coman (Reference Coman2022). Thirdly, we focus on the impact of a debate being targeted at a far-right party (e.g. the Fidesz government in Hungary). Details on all three models and associated results can be found in the Supplementary Material, Section E.
Discussion and conclusion
This article aimed to determine the extent to which debates over fundamental values are fertile ground for negative and motive-based forms of political opposition in the EP and to identify the main drivers of these different forms of opposition. Descriptively, we found these debates to have a high level of negativity and that a significant part of this negativity involves motive-based discourse. Given the limited research on these dimensions of EP discursive dynamics, we have few direct points of comparison for our study. In this sense, the descriptive aspect of this article serves as one of the few baseline reference points against which future studies on sentiment or motive-based criticism can compare their findings across policy areas or over time. There are, however, some points of comparison we can draw upon to help put our findings into perspective.
Compared to existing sentiment analyses of EP discourse, which suggest that positive sentiments dominate over negative ones (Salgado Reference Salgado2021; Schumacher et al. Reference Schumacher, Schoonvelde, Traber, Dahiya and de Vries2016), the extent of negativity in RoL debates is striking. The focus on targeting specific actors with negativity, as opposed to the norm of advocating for something more positively, is a powerful indication of the high salience associated with these debates and a reflection of the ideological contestation that characterises them (Coman Reference Coman2022; Wunsch and Chiru Reference Wunsch and Chiru2024). As a rare point of comparison for our findings concerning motive-based reasoning, Christopher Lord and Dionysia Tamvaki (Reference Lord and Tamvaki2012: 41) find that 16% of statements are disrespectful, where they use a similar set of measures to evaluate 44 EP debates on different areas of policy debates, as well as some relating to Treaty change and European integration in the period 2004–2009.
Compared to our limited knowledge of historical trends, therefore, our findings cautiously suggest that RoL debates are particularly prone to negativity, as well as expressions of moralistic affective polarisation. Further research is required to determine whether RoL debates are unique in this respect or if the growing presence of right-wing and Eurosceptical forces in the EP has also led to greater affective polarisation overall. A third possibility is discursive spillover, where elevated levels of motive-based cynicism in RoL debates on key issues of national sovereignty and European integration will have caused a souring of relations between political groups, thereby leading to more corrosive forms of discourse in EP debates generally. To the extent that a supranational cleavage may be increasingly structuring partisan politics in the EP (Otjes and van der Veer Reference Otjes and van der Veer2016), this is a distinct possibility as RoL debates are a key site where this cleavage emerges most nakedly.
Concerning the drivers of intervention-based discursive positioning of MEPs in RoLC debates, we have largely confirmed that similar factors are at play as with voting behaviour. Right-wing, Eurosceptic and CEE MEPs are all more likely to take up an anti-intervention position. We also find substantial evidence that the interests of national parties in government domestically are predictive of anti-intervention discourse among affiliated MEPs. This supports the findings of a paper focused exclusively on the voting behaviour of the European People’s Party (Herman et al. Reference Herman, Hoerner and Lacey2021). MEPs thus show signs of feeling constrained by their national parties in government, both when speaking and when voting.
However, our article goes beyond intervention-based positioning to investigate the factors that shape the sentiment and content of oppositional discourse on RoL debate in the EP. Firstly, we find that MEPs from more right-wing parties are more likely to make motive-based arguments. However, the left-right position of an MEP’s party does not seem to influence the dominant sentiment of their speech. The result for the anti-EU position of a MEP’s national party follows a very similar pattern. Secondly, concerning the effect of geography, we find that MEPs from parties from CEE countries are more likely to adopt motive-based discourse, albeit at a lower level of significance, but again, no effect for negative sentiment. Finally, when focusing on the effects of government participation at the national level, we find some evidence at the 0.05 level for a more frequent use of neutral language by MEPs of such parties. But this coefficient does not quite reach the same levels of significance compared to other variables we have tested.
Contrary to our expectations, none of the independent variables we have considered make it more likely for MEPs from a particular party to issue a negative statement than another. This is a particularly intriguing finding when we consider the structure of RoLC debates, which typically involve resolutions to be adopted or rejected. One would thus expect the pro-intervention side to have a preponderance of positive statements since they will tend to argue for the resolution, while the anti-intervention side would be more likely to issue negative statements because they are arguing against the resolution. From our reading of the content of these debates, the negativity of pro-intervention MEPs is likely due to their critique of those accused of falling afoul of EU fundamental values, including EU institutions for failing to act or other EPGs for tolerating breaches of fundamental values among their national party members. It thus appears that pro-intervention MEPs are very keen to express negative sentiments against the accused, if their negativity is sufficient to crowd out the tendency towards positive statements for those supporting the resolution embedded in the very structure of these debates. This aligns with qualitative findings from our previous research, according to which far-left parties in EP debates on the RoLC are particularly likely to criticise both backsliding states and the relative inaction of EU institutions, results that might well extend to other parliamentary groups (Herman et al. Reference Herman, Hoerner and Lacey2026).
Negativity on its own, however, does not amount to the kind of motive-based statements that can express and fuel moralistic affective polarisation. From the large body of research on right-wing discourse, it is expected that discourse of this kind will most likely be perpetrated by the right. What is more counterintuitive is the extent to which left-wing and pro-intervention parties are less likely to use such discourse. Indeed, it is not hard to find such MEPs throughout the debates attributing ill-motivations to national governments accused of backsliding, including accusations of corruption, power-grabbing and self-interest. However, such discourse has clearly been constrained on the pro-intervention side by a significant focus on more civil and principled statements, even if these are often couched in negative terms.
Not all negativity is bad for democracy, and some argue that forms of motive-based arguments may be justifiable when they point to genuine duplicity and are coupled with more principled forms of argumentation (Schedler Reference Schedler2021). Notwithstanding these points, the level of negativity and antagonism in RoLC debates is striking and a concern when considering the elite- and population-level democratic jeopardies associated with this kind of politics. Debates on the RoLC bring ideologically opposed actors into foundational and high-stakes conflict, setting the stage for a politics of negativity and affective polarisation that could affect the wider discourse of the EP. Considering existing research on the problematic consequences of moralising forms of affective polarisation, the rise of these trends in the EP might make elite cooperation more difficult. Antagonistic speech may feed the tendency of Eurosceptical elites and citizens to view the EU as illegitimate when it is perceived as encroaching upon national sovereignty, while also starving higher quality discourses focused on public justification of oxygen. The recent increase of radical right and Eurosceptic MEPs, and the growing willingness of the EPP group to cooperate with them, suggests that we should not expect to see an improvement in the extent of negative and motive-based discourse in the EP, at the very least when it comes to debates relating to the RoLC.
Supplementary material
The supplementary material for this article can be found at https://doi.org/10.1017/gov.2026.10053.
Data availability
The data and code to reproduce all tables and figures in the manuscript and in the supplementary materials are stored in the Harvard Dataverse: https://doi.org/10.7910/DVN/NWUG7G.
Acknowledgements
We are grateful for the constructive feedback we received on a first draft of this article from attendees at the Standing Group on the European Union (SGEU) ECPR Conference in June 2024, specifically comments from Professor Michael Blauberger, Vlagyiszlav Makszimov, Professor Arndt Wonka and Professor Natasha Wunsch. Many thanks also to Jihed Ncib and Sofia Quaglia, who assisted us in parts of the data coding and analysis.
Disclosure statement
The authors are entirely responsible for the scientific content of the article and declare that the article adheres to the journal’s authorship policy. The authors declare no competing interests that could be perceived to exert an undue influence on the content or publication of this work.
Open access
