The Conflict/War and Social Media: Politics of Information and Communication during Euromaidan and War in Donbas

October 22, 2015
Вважаєте відгук корисним?

The Conflict/War and Social Media: Politics of Information and Communication during Euromaidan and War in Donbas

Live streaming from the Danyliw Research Seminar on Contemporary Ukraine is available at

The third panel of the Danyliw Research Seminar on Contemporary Ukraine opened with Jennifer Carroll’s presentation on visual imagery of the protest movement and its mimicry during the Antimaidan protests and the events in the Crimea and war in Donbas. Joshua Tucker and Megan Metzger presented results from their study on the use of social media in relation to questions of identity as the events in Ukraine developed over time.

Jennifer Carroll emphasized that the visual imagery was quite consistent over time of protest: large crowds, flags, and even food. The images from Antimaidan were quite different – mainly mimicking what was happening “down the street,” to match the management decisions of the Euromaidan movement.

A prominent case in this regard are the self-defense brigades that were formed after the violence broke out. Instances of mimicry occurred soon after: men dressed in self-defense garb who vandalized shop windows in Kyiv. A brief analysis makes it obvious that these men were impostors – not least of all because they were “too clean,” evidently receiving their garb fresh from laundry.

Other instances of mimicry occurred during the Crimea takeover: “little green men” were presented as self-defense brigades in the media; they also used the props used previously by the self-defense brigades in Kyiv. For instance, the use of tires as a deescalation tactic was heavily used in Kyiv, and they were adopted also in Donbas at the barricades. 

Carroll emphasized that there is a subtle semiotic shift at play here. From Ferdinand de Saussure’s semiotic model, we speak of the signified and the signifier. Relationships that exist between them are those of resemblance (icon), interaction (index), and arbitrary pairing (symbol). All these relationships are evoked as we look at powerful images from Maidan.

When imitation happens, the signifier and the signified are subtly replaced. The imitation evokes a pre-existing sign and the knowledge about it, but substitutes the content (political message). This strategy can be labeled as “evoke and inflect”: the same signs are evoked but their meaning is changed (e.g. masculinity and nationalism in the case of imitation of self-defense brigades, but with a negative connotation of vandalism).

This kind of manipulation has to do with the fact that powerful images can be effectively used to change the message they send. We know that both in Ukraine and Russia people rely heavily on televised images as the main source of information. This is also rooted in the general affinity of humans for visual perception: Roland Barthes explains this with “the myth of photographic objectivity” which points to the fact that we are predisposed to inherently trust the image as an objective representation of reality.

Joshua Tucker and Megan Metzger presented their research on communication in social media, on Twitter in particular, in Ukrainian and Russian as the Euromaidan events and the conflict with Russia developed. The team of researchers at New York University sought to test the popular argument in the ethnic politics literature on identity that ethnic identity is constructed, fluid, and can change over time, but that the dominant identity becomes activated at the time of political upheaval. Availability of unprecedented data from social media allows to effectively test this assumption.

Tucker pointed out that social media data is more than just text; the language use matters as it represents the “real world,” not a lab environment, and, because it is a digital record, its extent is unprecedented.

Significant additional data – besides the tweet itself – can be collected here: the actual language of the interaction on Twitter (Ukrainian or Russian, for instance – posts in English were excluded from analysis), geographic location, city of origin, occupation of the person tweeting, age, gender, etc. This data can be combined in different ways to come up with a representative sample.

Tucker shared that one of their original assumptions was that during political actions the dominant language of Ukrainian users will be Ukrainian, and the language of Russian users would be Russian. However, they did allow for the possibility that evidence may falsify theory, especially as language choice online may not equal ethnic activation, or that the choice of the particular social media (Twitter) and the data from language use are not the most effective factors to raise these questions.

Megan Metzger outlined that Twitter data shows changes in use at key moments such as the Crimean secession or the downing of the MH17 airplane (significant increase of posts). There was an enormous change in the amount of tweets published by Russian users (in Russian) after Yanukovych flees Ukraine. Interestingly, at the very same time with the Russian users, Ukrainian users increase their use of the Russian language for tweeting.  The reasons for this may be that language choice and ethnic activation indeed do not overlap.

The other possible reason, Metzger suggested, could be that retweets were included in the first round of analysis – but even after removing them the results looked similar in terms of proportions. Yet another possible explanation could be the employment of bots and trolls by the Russian state. These can be identified by analyzing the number of followers that a possible bot has (low) vs. the number of people that a possible bot follows (high). Another possible explanation could be that different users talk about different topics: audiences that talk about Euromaidan are not the same that talk about the Crimea.

Metzger pointed out that the number of tweets is also interesting: after the spike in February 2014, the number of tweets goes back to almost the same levels, although in the Russian users’ case the level remains overall higher. A possible explanation here could be the fact that Russian is a more universal language than Ukrainian, so one can assume that in crisis users revert to the more universally used language in order to spread relevant information. The question is also whether the audience is predominantly domestic or international, as it influences the use of language.

The next steps of the project will consist in studying ethnicity salience (focus on Ukrainian users, shifting analysis from the aggregate to the individual level), adding English tweets, connecting language and political sentiment, networks, and strategic concerns.

The lively discussion revolved around the applicability of the criterion of language on social media for conclusions about ethnic identity of the social media users. Mayhill Fowler emphasized that the combination of language choice, ethnicity, and politics is highly problematic, as one would first need to define the term “Ukrainian.” Fowler underscored that all these identities are very fluid and complex – perhaps studying similar cases of bilingual and multi-ethnic societies could be helpful in explaining the link between the choice of language and the users’ identity.

In response, Tucker pointed out that all that was measured was the language in which people tweeted, largely because the researchers were interested in testing a theory that argues that in times of crisis people revert to their dominant identity. The findings are thus very simple: the number of people choosing to interact in Russian went up, while the number of those who chose to interact in Ukrainian went down. This might be a futile way to test the issue of identity, but it is an interesting way to approach this set of questions.

Discussion also included questions of language proficiency that influence the language choice, but also the arguments (e.g. by Volodymyr Kulyk) that users may identify as Ukrainian but not speak the language. Similarly, the effect of the age of Twitter users (generally lower), the use for strategic purposes rather than identity formation, and the performative nature of online publication were discussed as important factors in shaping the possibility of a link between language choice and identity. Cases such as the massive influx of citizens from Western Ukraine (mostly Ukrainian speakers) and singular actions such as Edinaia strana / Iedyna kraïna [One Country, in Russian and Ukrainian] that asked Twitter users to switch their language for a short time could also complicate such analysis.

In discussion of the visual imagery of the popular protest and its mimicry, questions focused on the difference between satire and mimicry. Carroll pointed to the fact that satire can be read as resistance against power, but it is also playful, while mimicry signifies an effort to obscure its imitative character. In response to the question of the top-down character of some popular slogans (such as “Yanukovych – pid aresht” [Yanukovych – under arrest; the slogan involves a pejorative slang term for homosexual] and “NeZlyi Maidan” [a word play of “Not an Evil Maidan” and “Do Not Anger Maidan”]), Carroll – while agreeing on the top-down nature of the slogans – pointed to the very small scale of such phenomena, as compared to Russian propaganda in particular. 

About the Author

We do critically important work in a world that increasingly neglects rational discussion and dialogue in favor of emotional proclamations. We need and deserve your full support. Please subscribe or become a sustainer by donating today.

Thank you!

We are a not-for-profit organization and all earnings go entirely towards production of high-quality analysis on Ukraine and the region.

Join the conversation!

Guest's picture
Please login or register in order to post your comments or questions.

Please consult our Terms of Use and Privacy Policy. We welcome a rational, respectful and matter-of-fact debate of all issues that this publication raises.