Political Narratives of War: A Comparative Analysis of MEP and U.S. Presidential X Posts on Ukraine

Download

Authors

Marius Dragomir & Benjamin Marks

About the project

The Political Narratives of War: A Comparative Analysis of MEP and U.S. Presidential X Posts on Ukraine is part of MJRC’s Media Content Analysis Series, which focuses on systematically examining media output to uncover patterns in coverage, bias, framing, and editorial choices. This series includes both thematic studies, such as crisis coverage, disinformation, and political framing, and cross-national comparisons of media narratives. In recent years, MJRC has integrated AI-driven tools and machine learning models into its methodology, enabling large-scale analysis of news texts, sentiment, and visibility trends across multiple languages and platforms. The work as part of this series combines computational analysis with media research to expose trends in global media attention.

Cite this study

Dragomir, M., & Marks, B. (2026). Political Narratives of War: A Comparative Analysis of MEP and U.S. Presidential X Posts on Ukraine. Media and Journalism Research Center. https://doi.org/10.5281/zenodo.18374727

Communications and outreach

Mihaela Groza, Paul Theobald

Data access

The datasets generated through this project are available on the Media and Journalism Exchange platform. To access them, please register here.

Executive summary

Over the past decade, social media has accelerated an epistemic fragmentation of public debate: competing narratives, selectively amplified, weaken a shared base of facts. This study examines how that fragmentation plays out in narratives on the war in Ukraine.

We analysed posts on X (formerly Twitter) by Members of the European Parliament (MEPs) across the 9th and 10th parliamentary terms and a defined set of U.S. administration accounts (Biden and Trump eras), covering 24 February 2022 to 1 May 2025. From 919,933 collected posts by 1,303 accounts, an AI-assisted relevance filter, validated against expert human coding, retained 80,716 Ukraine-related items. Each post was coded for narrative (pro-Ukrainian, pro-Russian, neutral/other), theme (e.g., Western support & sanctions; sovereignty & defence; geopolitics & power; humanitarian impact), and framing (e.g., moral, security, geopolitical/economic, demonisation/othering, victory/resilience, crisis & suffering). This research design enabled a like-for-like comparison across time, actor type, political family, country, and even language of communication.

Three main findings stand out. First, the study identified a dominant, but not monolithic, pro-Ukrainian consensus. Across all actors combined, roughly three quarters of posts endorse Ukrainian sovereignty and condemn Russian aggression; neutral content is secondary, and explicitly pro-Russian content remains a minority. Yet that overall picture conceals sharp segmentation inside the European Parliament. The centrist core, represented by parties such as EPP, S&D, Renew and Greens/EFA, communicates with striking discipline, with about 90% of their messaging pro-Ukrainian and pro-Russian content near negligible. By contrast, Patriots for Europe and the Europe of Sovereign Nations cluster significant pro-Russian discourse (the latter overwhelmingly so), while ECR sits in an ambiguous middle: majority pro-Ukraine overall but with a noticeably higher pro-Russian share than the centre. The landscape of independent MEPs is fragmented but taken together leans markedly towards pro-Russian narratives. The study thus shows narrative unity at the centre, and fissures at the sovereigntist and unaffiliated flanks.

Secondly, the research has identified a clear transatlantic asymmetry in consistency. U.S. administration accounts under Biden were almost uniformly pro-Ukrainian across 2022–2024, with pro-Russian content virtually absent. The Trump-era accounts, by contrast, exhibit greater fragmentation, with a larger share of neutral language and a small but present pro-Russian component. Measured across the full period, the U.S. therefore contributes both extremes in our dataset: the most consistently pro-Ukraine communication under Biden and a more variegated, neutrality-tolerant profile under Trump. Europe’s profile remains pro-Ukraine overall, but the internal spread between groups, as well as within some groups, drives the diversity that adversaries can exploit.

Thirdly, the informational landscape has been shifting over time toward greater polarisation at the margins. Early 2022 was defined by near-unanimity: MEPs’ pro-Ukrainian share peaked around the Russian invasion. However, by early 2025, pro-Russian content had multiplied several-fold from its 2022 baseline, while neutral messaging rose and fell in waves, briefly surging late-2023. The centre held in volume terms, but dissenting narratives became more visible and more engaging. This is mirrored in the framing trajectory: moral appeals, strong at the outset, have been declining steadily whereas security/defence and demonisation/othering frames gained ground and geopolitical/economic frames remained the backbone throughout. The humanitarian lens, both as theme and frame, stays marginal, a consistent finding across actor types and years.

Two other insights refine this picture. First, the study found that discourse varies sometimes by language. English, the European Parliament’s working lingua franca, functions as a channel for disciplined, coalition-compatible messaging: pro-Ukrainian content is higher and neutral content lower in English than in native-language posts. When MEPs switch to their country of origin language, messaging becomes measurably more cautious and diffuse. Mainstream groups (Renew, Greens/EFA, EPP, S&D) maintain near-perfect cross-language alignment, but right-of-centre and sovereigntist formations show more recalibration while independent members diverge the most. For citizens following only national-language feeds, the emphasis they encounter may thus differ meaningfully from the line projected in English at the European level.

Finally, engagement dynamics underscore the strategic costs and incentives driving this evolution. On a per-post basis, U.S. administration content attracts far more interactions than MEP posts, though the latter generate larger cumulative engagement due to volume. Within MEP communications, pro-Russian posts over-perform substantially on average engagement, outpacing both pro-Ukrainian and neutral content. Frame-wise, demonisation/othering and security/defence travel furthest per post, while conversations anchored in geopolitics/economics and sanctions attract the largest total attention by scale. Late-2024 to early-2025 saw simultaneous spikes across narratives, with pro-Russian and neutral content registering especially steep spikes, evidence of a more contested and attention-sensitive environment.

Taken together, the evidence depicts a communication ecosystem that remains broadly united in principle but diverse in rhetoric, by group, region, language, and actor type. The mainstream European centre projects a consistent pro-Ukraine stance; sovereigntist, nationalist and unaffiliated actors account for the bulk of dissent; U.S. governmental messaging ranges from uniformly aligned (Biden) to more fragmented (Trump). Over time, moral appeals have faded, security and demonisation frames have hardened, and humanitarian language has receded. Because audience-specific tailoring is stronger outside the mainstream (and because oppositional narratives outperform on intensity) a persistent structural risk of narrative drift is evident. Addressing epistemic fragmentation therefore depends not only on reinforcing the dominant pro-Ukrainian narrative, but also on managing intra-European variation, language-specific messaging, and the engagement bias that rewards polarising content.

Introduction

Political Narratives of War: A Comparative Analysis of MEP and U.S. Presidential X Posts on Ukraine is grounded in the understanding that narratives and framing play a crucial role in shaping public perception and international discourse. The project aims to shed light on how various actors communicate about the conflict, highlighting how competing narratives influence the global understanding of Ukraine’s ongoing struggles.

Methodology

We constructed a dataset of X (formerly Twitter) posts authored by Members of the European Parliament (MEPs) and selected U.S. administration officials, covering the period from 24 February 2022 (the start of the war in Ukraine) until 1 May 2025.

Data collection

For the MEPs, accounts were identified through a two-step procedure. First, we parsed the XML provided on the European Parliament’s website to compile the official list of representatives for both the 9th and 10th parliamentary terms. We then scraped each profile to extract the Twitter handle where available. For cases where no handle was listed, accounts were identified through manual verification. This ensured coverage of both parliamentary terms, with data collected from the first day of the war until the end of the mandate for 9th-term MEPs, and until May 2025 for 10th-term MEPs. For the U.S. administrations, we relied on a predefined set of official and cabinet-level accounts from both the Trump and Biden administrations. Tweets were collected using the twitterapi.io API, resulting in a dataset(raw) of approximately 919,933 from 1,303 accounts.

Relevance filtering

To ensure topical focus, we applied an automated relevance classifier built with OpenAI’s API, guided by expert instructions. Random validation against human coding showed strong performance: the classifier matched human judgment on 93% of tweets, with a precision of 95% (tweets labeled as relevant were almost always correct) and a recall of 85% (most, though not all, relevant tweets identified by humans were captured). We considered this level of agreement sufficient to scale the approach to the full dataset. Only tweets deemed relevant were retained for further analysis. After filtering, the dataset contained 80,716 tweets.

Content coding

The content of the tweets was coded along three dimensions defined by media experts: narratives, themes, and framing strategies as follows:

Narratives

• Pro-Ukrainian (support for Ukrainian sovereignty, critique of Russian actions, highlighting resilience)
• Pro-Russian (support for Russia, critique of Ukraine/NATO, or justification of aggression)
• Neutral/Other (no clear stance, factual, or procedural updates)

Themes

• Sovereignty & Defense
• Western Support & Sanctions
• Russian Aggression & Justification
• Humanitarian & Civilian Impact
• Geopolitical Strategy & Power

Framing Strategies

• Moral Framing
• Security/Defense Framing
• Geopolitical/Economic Framing
• Demonisation/Othering
• Victory/Resilience
• Crisis & Suffering

Automated classification using OpenAI’s reasoning model was refined over several iterations to align with expert expectations. Actor type was coded as either MEP or U.S. administration.

Limitations

Our dataset is restricted to publicly available posts on X. Deleted or private tweets are not captured. Moreover, activity levels vary substantially across actors: some MEPs and U.S. officials were highly active, while others posted rarely or not at all. As a result, the dataset reflects the communicative practices of those more active on the platform rather than providing uniform coverage across all representatives. The automated classification was validated by human experts with strong agreement, and any residual discrepancies are negligible for the overall patterns or insights.

Key findings

Common narratives

At the aggregate level where both MEPs and US administration officials are counted, pro-Ukrainian messaging dominates the discourse. Across all sources, three quarters of the material (75.8%) expressed explicit support for Ukrainian sovereignty, condemned Russian aggression, and highlighted Ukraine’s resilience. Neutral or procedural messages accounted for 15.3%, while explicitly pro-Russian narratives remained marginal.

When disaggregated by actor type, we see a striking asymmetry. U.S. administrations, under both Biden and Trump, overwhelmingly deployed pro-Ukrainian narratives. Biden officials expressed this line almost uniformly (95.8% of statements), while Trump officials, though less consistent, still leaned strongly pro-Ukrainian (80.3%).

By contrast, MEPs show more variation: although still dominated by pro-Ukrainian narratives (75.4%), there was a significantly larger share of both pro-Russian narratives (9.1%) and neutral statements (15.6%). This demonstrates that U.S. discourse was more consistently pro-Ukrainian, while the European Parliament displayed greater discursive diversity, including a notable minority of pro-Russian voices.

Narratives by European political groups

Looking across the European Parliament, the variation in narrative alignment is striking. The mainstream centre-right and centre-left families (the European People’s Party (EPP), the Progressive Alliance of Socialists and Democrats (S&D), Renew Europe, and the Greens/EFA) are consistently and overwhelmingly pro-Ukrainian. In these groups, between 88% and 91% of discourse falls within the pro-Ukrainian category, while pro-Russian narratives rarely exceed 1–2%. The remaining share is largely neutral or procedural, underscoring how firmly these groups converge around the defence of Ukrainian sovereignty and condemnation of Russian aggression.

On the sovereigntist and nationalist flank, the picture changes substantially. The Europe of Sovereign Nations (ESN) group emerges as the most pro-Russian of all groups, with 75.7% of its narratives taking a pro-Russian stance and only 4% supporting Ukraine, a near-complete reversal of the pattern seen among the mainstream families. Similarly, the newly formed Patriots for Europe group displays high levels of pro-Russian discourse (39.6%), just 21.3% pro-Ukrainian, and nearly a third neutral (31.8%).

The European Conservatives and Reformists (ECR) group present a more ambiguous picture: while a strong majority (74.1%) is pro-Ukrainian, the group still shows 9.2% pro-Russian content, far above the levels found in EPP, S&D, or Renew, suggesting internal divisions and space for sovereigntist positioning.

At the other end of the spectrum, The Left (GUE/NGL) shows a majority pro-Ukrainian stance (52%), paired with a large neutral share (41.2%) and 6.8% pro-Russian. This suggests a more nuanced and sometimes critical approach compared with the mainstream centrist families.

Finally, the non-attached members (independent MEPs) stand out for their fragmentation but, in aggregate, skew pro-Russian (57.4%), with 8.5% pro-Ukrainian and 34.1% neutral, the strongest pro-Russian profile outside the ESN group.

Taken together, the data demonstrate that the European Parliament is sharply segmented. The centrist core strongly backs Ukraine; Patriots for Europe, ESN, and many non-attached members concentrate pro-Russian discourse; ECR sits between these poles; The Left is majority pro-Ukraine but unusually neutral compared with the centre. These divisions underline how foreign policy narratives in the EU are increasingly structured along the same ideological cleavages that define the broader political landscape.

Focusing on message consistency, two blocs emerge. The pro-EU mainstream is highly disciplined: Renew is the most consistent (91.2%), with EPP close behind (89.6%) and Greens/EFA (86.6%) and Socialist & Democrats (85.9%) also maintaining tight alignment. Notably, EPP pairs high consistency with the largest output (22,812 tweets), suggesting coordination scales well in the centre.

By contrast, the right-of-centre European Conservatives and Reformists shows markedly looser discipline (66.8%) despite heavy activity (11,181 tweets). On the Eurosceptic/populist flank, coherence fragments: Identity and Democracy (52.3%), Patriots for Europe (46.9%) and non-attached Members (49.5%) all hover around or below the halfway mark. The standout exception is Europe of Sovereign Nations, which, despite smaller volume, keeps a relatively high internal coherence (75.7%) around a pro-Russian line. Finally, The Left records the lowest consistency of any group (42.0%), indicating substantial internal divergence even while the group’s dominant narrative remains pro-Ukrainian.

Overall, consistency is a hallmark of the pro-European centre, while fragmentation characterises most of the Eurosceptic/populist and unaffiliated space.

Across the corpus, five thematic categories structured discourse. The most common was Western Support & Sanctions at 30.7%, followed by Geopolitical Strategy & Power at 22.8% and Sovereignty & Defence at 21.5%. References to Russian Aggression & Justification and Humanitarian & Civilian Impact were less frequent, together accounting for around a quarter of discourse. This thematic distribution highlights that discourse is heavily oriented towards power, strategy, and support mechanisms, rather than humanitarian or civilian-centred perspectives.

In terms of framing, Geopolitical/Economic framing and Demonisation/Othering were the most common, together constituting nearly half of all frames. Moral arguments and Security/Defence frames followed, while Victory/Resilience and Crisis & Suffering were less frequent. Interestingly, the least used frame (Crisis & Suffering) corresponds with the least discussed theme (Humanitarian & Civilian Impact), suggesting that the humanitarian dimension is consistently marginalised in elite political discourse.

Narrative trends over time

The dataset spans February 2022 to April 2025, a period in which the communications landscape shifted from near-unanimity to a more fragmented and contested space.

The first year of the war was marked by overwhelming consensus. Across all actors combined (both MEPs and US administration officials), pro-Ukrainian discourse averaged nearly 78% of monthly mentions, with neutral/other and pro-Russian narratives accounting for the remainder. The European Parliament mirrored this almost exactly, and in some months even amplified it, reaching almost 86% in pro-Ukrainian support in February 2022.

The shift in balance is gradual but perceptible. In the first three months covered in 2022 (February to April), the MEPs share was nearly 81% pro-Ukrainian and barely 3% pro-Russian. Compared with the same months in 2025 (February to April), those figures had changed to roughly 71% and 16%, respectively. Neutral coverage, meanwhile, rose to prominence in late 2023, peaking at a quarter of all mentions, before retreating again as polarity reasserted itself.

The US administrations tell a different story. The Biden administration’s discourse is the most disciplined in the dataset: pro-Ukrainian tweets near 95% in 2022 and exceed 97% in both 2023 and 2024. Pro-Russian content is practically absent. The Trump administration’s communication, by contrast, fractures quickly. In 2024 it was still 93% pro-Ukrainian, but by 2025, only around two-thirds of mentions were in favour of Ukraine while almost 28% were neutral and 6% pro-Russian.

Overall, the pro-Ukrainian narrative decreased by roughly 15 percentage points from the Biden administration to the Trump administration. Among U.S. administration officials specifically, combined anti-Ukraine sentiment (pro-Russia + neutral) increased by 368% from Biden to the Trump administration.

But measured from the start of the analysis in February 2022 to the last month analyzed (April 2025), the picture is quite dramatic. Across the combined dataset of both MEPs and U.S. administration officials, pro-Russia narratives grew more than eightfold, surging from just 1.92% at the outset to 15.96% by April 2025. Neutral narratives remained comparatively stable, increasing only slightly, from 12.08% to 12.78%.

Within the MEP dataset, a nearly identical pattern emerged: pro-Russia content expanded by over eightfold, jumping from 1.96% to 16.00%, while neutral content showed only a modest rise, from 12.17% to 12.65%.

The most striking development was seen in the U.S. administration dataset. In February 2022, there were no Pro-Russia narratives recorded, yet by April 2025, the share of neutral narratives had skyrocketed tenfold, leaping from 6.67% to 66.67%. This indicates a dramatic realignment in messaging priorities within the U.S. administration, moving away from a clear pro-Ukrainian stance toward a more neutral, and potentially ambiguous, position.

Taken together, these trajectories outline a coherent arc. The opening phase of 2022 was defined by consensus; 2023 diluted that unanimity through the rise of neutral narratives; 2024 re-polarised around pro-Ukrainian and pro-Russian poles; and early 2025 marks the high-water mark so far for pro-Russian messaging, even as overall volumes remain lower than in the first year of the war.

Throughout, the Parliament has mirrored the field rather than diverged from it. The United States, however, contributes both extremes: an administration under Biden that is the most uniformly pro-Ukrainian actor in the dataset, and under Trump one of the most fragmented, with space for neutrality and even a modest expansion of pro-Russian content.

At the onset of the war, moral and geopolitical frames were almost equally strong. Over time, however, moral framing steadily declines from almost 20% in 2022 to less than 14% by 2025. In parallel, security/defence and demonisation frames grew sharply, each surpassing 20% in the latest data. Geopolitical/economic perspectives remain consistently central, hovering around a quarter of all narratives. Victory/resilience is a secondary but persistent theme (9–12%), while crisis and suffering declined markedly, from nearly 11% in 2022 to below 5% in 2025.

MEPs closely track the overall trends. By 2024–25, their profile converges with the all-actor averages: moral and crisis/suffering frames diminish, security/defense and demonisation grow, and geopolitical/economic framing remains the backbone of the overall narrative body.

The framing trajectory contrasts with the more volatile shifts in narrative dominance (pro-Ukrainian vs pro-Russian). While narratives have moved from consensus (2022) to fragmentation (2025), the framing lenses have proven remarkably stable. The war is consistently narrated as a geopolitical and security crisis, with moral appeals present but gradually declining. Demonisation has grown in relative prominence as the conflict prolonged, while the humanitarian “crisis and suffering” frame has been pushed to the margins.

The MEPs narrative exemplifies this continuity: it began more moralistic than the overall field but gradually re-centred around the same trio of frames, geopolitics, security, and demonisation, by 2025. The persistence of victory/resilience frames (hovering above 9%) underlines the discursive effort to maintain a language of endurance despite war fatigue.

Engagement dynamics: who shapes the debate?

The engagement landscape is profoundly asymmetric across actor types. On a per-post basis, content from U.S. administrations attracts an average of 7,501 interactions, compared to just 346 for MEP posts. In other words, a typical U.S. administration post generates nearly 22 times more engagement than an MEP post over the same period.

In aggregate terms, however, tweets from MEPs generate much higher cumulative engagement, roughly double that of U.S. administrations (27.28 million vs. 13.41 million). This difference largely reflects the much larger sample of MEP accounts included in our dataset. When engagement is calculated on a per-account basis, the disparity is striking: U.S. officials average 685,502 total engagements per account, compared to just 21,264 per MEP account.

Within MEP communications, pro‑Russian content is an outlier for intensity: despite being the smallest narrative by volume, it delivers an average of 811 engagements per post, more than double pro‑Ukrainian and nearly triple neutral content. Among the US administrations, pro-Russian content engagement again spikes with an extremely high per‑post average, though the sample is tiny and should be treated with caution. The consolidated view is below. 

Disaggregating Biden and Trump administrations clarifies the intensity gap even further. Over the whole period, Biden‑era posts average around 5,289 interactions, while Trump‑era posts average 32,886, a multiple driven by far fewer, but much higher‑impact, posts (notably pro-Russian and neutral). Again, sample sizes for some narratives are small and, hence, volatile.

Across all actors, the late‑2024 to early‑2025 window shows pronounced spikes. Pro-Ukrainian support peaks in December 2024, with two additional high points in October 2024 and October 2023. Pro-Russian narrative surged in September–October 2024 and again in February 2025. Neutral narratives climaxed in March 2025, with strong levels in February 2025 and September 2024. These crests suggest that late‑cycle developments intensified audience reactions across all narrative types.

Demonisation/Othering has the highest per‑post engagement at 651 and also commands the largest share of total engagement (25%). Security/Defence is next on intensity (571), while Geopolitical/Economic generates the largest volume and the second‑largest share (23%), but with the lowest per‑post average (483). In short, Demonisation/Othering frames seem to travel far and fast; Geopolitics/Economic drives scale more than intensity.

Russian Aggression & Justification has the highest per‑post average engagement (636), followed by Sovereignty & Defence (539) and Geopolitical Strategy & Power (532). Western Support & Sanctions has the largest share of total engagement (29.6%) thanks to its scale, even though its per‑post average (515) sits below the top trio.

Put together, the data point to a consistent logic of engagement. Intensity is fuelled by polarising rhetoric and attack–defence frames, especially when paired with pro-Russian content, which, even when rare, overperforms dramatically. Scale, however, accrues to mainstream geopolitical and sanction‑policy conversations, accounting for the largest slices of total attention despite lower per‑post traction. The late‑2024/early‑2025 surge window amplified all narratives, with pro-Russian and neutral content registering the steepest crests.

In terms of communications strategy, that suggests a two‑track reality. If the goal is depth of reaction per item, moralised othering and hard‑security frames dominate. If the goal is breadth of total attention, sustained coverage in geopolitical/economic and sanction‑policy registers remains the main narrative. Finally, the actor effect is important: US‑level messages operate on a different intensity curve than MEP communications, even when addressing similar narratives. 

Narratives across countries and administrations

The debate on Ukraine within both the United States and Europe is not a single, uniform conversation but rather a blend of narratives, themes, and framings shaped by distinct political, geographic, and institutional contexts.

At the broadest level, one striking feature is the consistency of the U.S. narrative environment compared to the more fragmented European picture. In Washington, both administrations overwhelmingly positioned themselves in support of Ukraine’s sovereignty, producing a communicative environment with little room for ambiguity. The data shows that over 94% of all U.S. statements fell under the pro-Ukrainian category, leaving barely half a percentage point for pro-Russian voices.

By contrast, Europe is more variegated. Northern Europe emerges as the most U.S.-aligned, with nearly four in five statements unequivocally pro-Ukrainian. Western Europe posts the highest pro-Russian share (over 14%), exceeding Southern Europe, whose discourse is nearly one-quarter neutral, and Eastern Europe, where the pro-Russian presence reflects domestic heterogeneity and the proximity of Russian influence (e.g., Hungary, Slovakia, Bulgaria).

Beyond the binary of pro- and anti-Ukrainian narratives, political communication reveals significant differences in thematic focus. For some states, the war is first and foremost a question of sovereignty and territorial defense; for others, the dominant lens is Western sanctions and external support; and in still others, the war is integrated into a wider geopolitical chessboard. Clear divergences emerge: MEPs in Malta stress sovereignty (38.5%), Hungarian ones frame the war almost entirely as geopolitics and sanctions (39% each), while representatives of Denmark and Sweden anchor their narratives in Western support and sanctions (nearly 40%). Luxembourg, uniquely, balances sovereignty and humanitarian dimensions (26%).

If themes tell us what policymakers are talking about, frames reveal how they package their arguments. The contrast here is especially striking: some countries stress the moral imperative of supporting Ukraine, others highlight security considerations, while still others rely heavily on demonisation or othering.

This expanded view highlights several notable contrasts. First, MEPs hailing from Austria and Poland rely strongly on demonisation frames, with over a quarter of communications portraying Russia as an existential enemy. Hungary’s representatives in Brussels diverge dramatically, devoting nearly half (48.5%) of their communication to geopolitical frames. This reflects Budapest’s transactional diplomacy and strategic ambiguity within the EU. Malta is exceptional as more than a third of its MEPs’ discourse is moralistic, with nearly a quarter relying on victory framings, an unusual pattern for a small state with limited military weight. Ireland, Italy, Luxembourg, and Slovenia lean more heavily on moral framings than their peers, often paired with humanitarian appeals. MEPs from Finland, Latvia, and Sweden, located at the EU’s northeastern edge, unsurprisingly stress security frames, reflecting their frontline anxieties. Finally, EP representatives from Cyprus and Greece rely heavily on geopolitical framings, aligning with their Mediterranean positioning and exposure to wider regional tensions.

Aggregating these results reveals a series of regional communication logics. Northern Europe (Nordics and Baltics) stresses security and deterrence, with Finland and Latvia among the most security-heavy communicators. Western Europe blends moral and geopolitical frames, with France, Germany, and Belgium balancing ethical appeals with strategic policy pragmatism. Southern Europe is the most fragmented, oscillating between humanitarian appeals (Cyprus, Greece) and symbolic moralism (Malta, Italy). Finally, Eastern Europe mixes hard demonisation (Poland, Slovakia) with occasional geopolitical ambivalence (Hungary).

To conclude, Europe speaks in many accents, unified in support but differentiated in rhetorical packaging, variations that adversaries can exploit in targeted disinformation.

When the focus shifts to Washington, the picture is one of continuity in direction but divergence in emphasis. Both administrations prioritised Western support and sanctions, but their secondary emphases differed. The Biden administration relied on moral and humanitarian framings, presenting the war as a struggle between right and wrong. The Trump administration, in contrast, emphasises security and victory, projecting resilience and deterrence rather than empathy.

Bringing these analyses together reveals two dynamics. First, pro-Ukrainian narratives dominate across the board, but their articulation differs sharply: Europe reveals diverse framings rooted in regional contexts, while the U.S. maintains a consistent overarching line. Second, differences between Biden and Trump are less about substance than style: Biden stressed morality and humanitarianism, Trump stressed deterrence and power.

At the European level, regional logics matter: Northern Europe’s security lens, Western Europe’s moral-geopolitical balance, Southern Europe’s fragmented humanitarianism and symbolism, and Eastern Europe’s paradox of hard demonisation vs. Hungarian realpolitik form a patchwork of discourses.

This complexity underscores that narrative unity does not equal communicative uniformity. While Ukraine’s allies remain united in principle, the diversity of frames, moral, security, geopolitical, or humanitarian, provides multiple points where adversaries can exploit differences, sowing discord or diluting urgency. Sustaining solidarity thus requires not only maintaining a dominant pro-Ukrainian narrative but also actively managing intra-European variations to avoid narrative drift in the long run.

Lost in translation? Cross-language political narratives in the European Parliament

The cross-language comparison confirms two communication logics. English, the Parliament’s working lingua franca, remains the channel for disciplined, coalition-compatible messaging, while native languages carry more domestically tailored, constituency-facing content. Aggregate narratives shift accordingly: in English, posts are 81.3% pro-Ukrainian, 6.2% pro-Russian and 12.5% neutral; in native languages, the balance moves nearly 4 percentage points less pro-Ukrainian and 4 percentage points more neutral, with the pro-Russian share essentially unchanged. In short, native-language communication is measurably more cautious and diffuse as MEPs are very conscious about how to frame their messaging for the local voters.

On consistency, measured as both the share of MEPs keeping the same top narrative cluster across languages and the average dominance of that top cluster within each language, the pro-EU mainstream is exceptionally disciplined. Renew and Greens/EFA show perfect cross-language alignment (100%), with very high concentration (Renew: 94.7% English and 88.3% country of origin language; Greens/EFA: 92.5% / 87.1%). EPP follows closely and S&D remains highly coherent. The Left also posts strong alignment (90.0%), albeit with lower concentration, indicating message unity with more internal variety.

Further right, ECR keeps a moderate level of coherence (85.2%) but sees a notable fall-off in native-language concentration (85.4% English versus 73.5% native), consistent with domestically adapted rhetoric. ESN is an outlier: despite a small cohort, it combines 80.0% cross-language consistency with higher concentration in native languages, suggesting a tighter national-language line. By contrast, Patriots for Europe and especially the Non-attached display marked fragmentation. For Non-attached MEPs, the dominance of the top cluster plunges from 93.5% in English to 62.0% in native languages, a strong indicator of message recalibration between European and domestic audiences.

Taken together, the data point to a structural duality: mainstream pro-European families pair near-perfect cross-language consistency with high message concentration, while Eurosceptic/populist and unaffiliated actors are more heterogeneous and prone to audience-specific tailoring. The practical implication for transparency is that citizens following only national-language feeds may encounter a meaningfully different emphasis than the message projected in English at the European level.

Conclusions

This study maps a communication battlefield in which political elites shape public meaning about the war in Ukraine within a fractured online sphere. The point of departure is epistemic fragmentation: the erosion of a shared base of facts as different audiences encounter different blends of narrative, theme, and frame. The study’s findings show a pro-Ukrainian centre that still holds, but it does so amid organised pockets of dissent, systematic cross-language reframing, and platform (in this case, X) incentives that reward the sharpest edges of debate. How these forces play out matters for citizens, for journalism, and for democracy.

For citizens, there is no single conversation to follow. English, the working language of European institutions, carries disciplined, coalition-friendly messaging, while national-language feeds often present more cautious or domestically tailored versions of the same story. This segmentation nurtures parallel publics: people consuming only local-language content may come away with a meaningfully different emphasis than the one projected in English. At the same time, the humanitarian lens remains marginal.

As security and othering frames grow and moral language recedes in the communication of politicians, empathy is crowded out by strategic and punitive registers. That imbalance risks normalising the human costs of war and deepening fatigue, with real consequences for support to people displaced or under fire. A citizenry equipped with attention literacy, able to see virality as a property of distribution rather than a proxy for truth, would be better placed to resist the pull of manipulative frames and the manufactured impression of consensus.

For journalism, the relationship with polarisation runs both ways. Platform logics prize engagement over accuracy, and that bias does not stop at the newsroom door. The frames that travel furthest per post are those that dramatise or harden conflict while the ones that complete the human picture travel least. In this environment it is easy for editorial judgment to be bent toward polarisation, dividing headlines, thinner context and fewer human-impact angles, not necessarily because newsrooms wish to inflame, but because the metrics that govern visibility reward it.

The country patterns in our data also illustrate the domination of local political logic. In Hungary, for example, the pro-government, often pro-Russian tilt visible in most of the country’s media ecosystem finds echoes in the X feeds of affiliated political actors. The X feeds and the news content of the national media reinforce one another, and each then appears to confirm the other’s narrative “common sense.”

Finally, for democracy, the fragmentation that this study unveils can look like pluralism from a distance. Citizens may appreciate a cacophony of views as a sign of a diverse and vibrant public sphere. But epistemic fragmentation is not mere diversity; it is rather diversity without a stable centre of shared facts. When messages are readjusted by language and audience, oppositional frames reap disproportionate attention, and party-aligned ecosystems echo and amplify those frames, the result is not only a livelier debate but a thinner epistemic commons, which carries a democratic price. The European project depends on a minimum of unity around fundamental values and on collective resolve in the face of aggression and war. A public sphere that cannot hold those baselines steadily in view becomes easier to pry apart by both foreign powers that seek to exploit divisions as well as domestic actors who use these contradictions for their own purposes, most of the time political.

In sum, our analysis of MEP and U.S. officials’ tweets yields a conclusion neither alarmist nor complacent: the current communication environment, rewarding provocation and shifting tone across languages, sustains epistemic fragmentation. The centre will still hold, but only if citizens understand how attention is engineered, journalists stay inclusive and critical without becoming conduits for polarisation, and institutions protect a shared stock of facts.

Appendix